Attempting to use an external library to handle bitstreams. Isn't going the grreatest.

include/bitstream/LICENSE

@@ -0,0 +1,28 @@
+BSD 3-Clause License
+
+Copyright (c) 2023, Krede
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+1. Redistributions of source code must retain the above copyright notice, this
+   list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright notice,
+   this list of conditions and the following disclaimer in the documentation
+   and/or other materials provided with the distribution.
+
+3. Neither the name of the copyright holder nor the names of its
+   contributors may be used to endorse or promote products derived from
+   this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

include/bitstream/NETSTACKLICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2018 Stanislav Denisov (nxrighthere@gmail.com)
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

include/bitstream/VERSION.txt

@@ -0,0 +1 @@
+0.1.4

include/bitstream/bitstream.h

@@ -0,0 +1,22 @@
+#pragma once
+
+// Quantization
+#include "quantization/bounded_range.h"
+#include "quantization/half_precision.h"
+#include "quantization/smallest_three.h"
+
+// Stream
+#include "stream/bit_measure.h"
+#include "stream/bit_reader.h"
+#include "stream/bit_writer.h"
+#include "stream/byte_buffer.h"
+#include "stream/serialize_traits.h"
+
+// Traits
+#include "traits/array_traits.h"
+#include "traits/bool_trait.h"
+#include "traits/enum_trait.h"
+#include "traits/float_trait.h"
+#include "traits/integral_traits.h"
+#include "traits/quantization_traits.h"
+#include "traits/string_traits.h"

include/bitstream/quantization/bounded_range.h

@@ -0,0 +1,104 @@
+#pragma once
+
+/*
+ *  Copyright (c) 2018 Stanislav Denisov
+ *
+ *  Permission is hereby granted, free of charge, to any person obtaining a copy
+ *  of this software and associated documentation files (the "Software"), to deal
+ *  in the Software without restriction, including without limitation the rights
+ *  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ *  copies of the Software, and to permit persons to whom the Software is
+ *  furnished to do so, subject to the following conditions:
+ *
+ *  The above copyright notice and this permission notice shall be included in all
+ *  copies or substantial portions of the Software.
+ *
+ *  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ *  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ *  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ *  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ *  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ *  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ *  SOFTWARE.
+ */
+
+#include <cstdint>
+
+namespace bitstream
+{
+	/**
+	 * @brief Class for quantizing single-precision floats into a range and precision
+	*/
+	class bounded_range
+	{
+	public:
+		constexpr bounded_range() noexcept :
+            m_Min(0),
+            m_Max(0),
+            m_Precision(0),
+            m_BitsRequired(0),
+            m_Mask(0) {}
+
+        constexpr bounded_range(float min, float max, float precision) noexcept :
+		    m_Min(min),
+            m_Max(max),
+            m_Precision(precision),
+            m_BitsRequired(log2(static_cast<uint32_t>((m_Max - m_Min) * (1.0f / precision) + 0.5f)) + 1),
+            m_Mask((1U << m_BitsRequired) - 1U) {}
+        
+        constexpr inline float get_min() const noexcept { return m_Min; }
+        constexpr inline float get_max() const noexcept { return m_Max; }
+        constexpr inline float get_precision() const noexcept { return m_Precision; }
+        constexpr inline uint32_t get_bits_required() const noexcept { return m_BitsRequired; }
+
+        constexpr inline uint32_t quantize(float value) const noexcept
+        {
+            if (value < m_Min)
+                value = m_Min;
+            else if (value > m_Max)
+                value = m_Max;
+
+            return static_cast<uint32_t>(static_cast<float>((value - m_Min) * (1.0f / m_Precision)) + 0.5f) & m_Mask;
+        }
+        
+        constexpr inline float dequantize(uint32_t data) const noexcept
+        {
+            float adjusted = (static_cast<float>(data) * m_Precision) + m_Min;
+
+            if (adjusted < m_Min)
+                adjusted = m_Min;
+            else if (adjusted > m_Max)
+                adjusted = m_Max;
+
+            return adjusted;
+        }
+
+	private:
+        constexpr inline static uint32_t log2(uint32_t value) noexcept
+        {
+            value |= value >> 1;
+            value |= value >> 2;
+            value |= value >> 4;
+            value |= value >> 8;
+            value |= value >> 16;
+
+            return DE_BRUIJN[(value * 0x07C4ACDDU) >> 27];
+        }
+        
+	private:
+        float m_Min;
+		float m_Max;
+		float m_Precision;
+
+		uint32_t m_BitsRequired;
+		uint32_t m_Mask;
+        
+        constexpr inline static uint32_t DE_BRUIJN[32]
+        {
+             0,  9,  1, 10, 13, 21,  2, 29,
+            11, 14, 16, 18, 22, 25,  3, 30,
+             8, 12, 20, 28, 15, 17, 24,  7,
+            19, 27, 23,  6, 26,  5,  4, 31
+        };
+	};
+}

include/bitstream/quantization/half_precision.h

@@ -0,0 +1,114 @@
+#pragma once
+
+/*
+ *  Copyright (c) 2018 Stanislav Denisov
+ *
+ *  Permission is hereby granted, free of charge, to any person obtaining a copy
+ *  of this software and associated documentation files (the "Software"), to deal
+ *  in the Software without restriction, including without limitation the rights
+ *  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ *  copies of the Software, and to permit persons to whom the Software is
+ *  furnished to do so, subject to the following conditions:
+ *
+ *  The above copyright notice and this permission notice shall be included in all
+ *  copies or substantial portions of the Software.
+ *
+ *  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ *  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ *  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ *  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ *  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ *  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ *  SOFTWARE.
+ */
+
+#include <cstdint>
+#include <cstring>
+
+namespace bitstream
+{
+	/**
+	 * @brief Class for quantizing single-precision floats into half-precision
+	*/
+	class half_precision
+	{
+	public:
+		inline static uint16_t quantize(float value) noexcept
+		{
+			int32_t tmp;
+			std::memcpy(&tmp, &value, sizeof(float));
+
+			int32_t s = (tmp >> 16) & 0x00008000;
+			int32_t e = ((tmp >> 23) & 0X000000FF) - (127 - 15);
+			int32_t m = tmp & 0X007FFFFF;
+
+			if (e <= 0) {
+				if (e < -10)
+					return static_cast<uint16_t>(s);
+
+				m |= 0x00800000;
+
+				int32_t t = 14 - e;
+				int32_t a = (1 << (t - 1)) - 1;
+				int32_t b = (m >> t) & 1;
+
+				m = (m + a + b) >> t;
+
+				return static_cast<uint16_t>(s | m);
+			}
+
+			if (e == 0XFF - (127 - 15)) {
+				if (m == 0)
+					return static_cast<uint16_t>(s | 0X7C00);
+
+				m >>= 13;
+
+				return static_cast<uint16_t>(s | 0X7C00 | m | ((m == 0) ? 1 : 0));
+			}
+
+			m = m + 0X00000FFF + ((m >> 13) & 1);
+
+			if ((m & 0x00800000) != 0) {
+				m = 0;
+				e++;
+			}
+
+			if (e > 30)
+				return static_cast<uint16_t>(s | 0X7C00);
+
+			return static_cast<uint16_t>(s | (e << 10) | (m >> 13));
+		}
+
+		inline static float dequantize(uint16_t value) noexcept
+		{
+			uint32_t tmp;
+			uint32_t mantissa = static_cast<uint32_t>(value & 1023);
+			uint32_t exponent = 0XFFFFFFF2;
+
+			if ((value & -33792) == 0) {
+				if (mantissa != 0) {
+					while ((mantissa & 1024) == 0) {
+						exponent--;
+						mantissa <<= 1;
+					}
+
+					mantissa &= 0XFFFFFBFF;
+					tmp = ((static_cast<uint32_t>(value) & 0x8000) << 16) | ((exponent + 127) << 23) | (mantissa << 13);
+				}
+				else
+				{
+					tmp = static_cast<uint32_t>((value & 0x8000) << 16);
+				}
+			}
+			else
+			{
+				tmp = ((static_cast<uint32_t>(value) & 0x8000) << 16) | (((((static_cast<uint32_t>(value) >> 10) & 0X1F) - 15) + 127) << 23) | (mantissa << 13);
+			}
+
+			float result;
+			std::memcpy(&result, &tmp, sizeof(float));
+
+			return result;
+		}
+	};
+}

include/bitstream/quantization/smallest_three.h

@@ -0,0 +1,156 @@
+#pragma once
+
+/*
+ *  Copyright (c) 2020 Stanislav Denisov, Maxim Munning, Davin Carten
+ *
+ *  Permission is hereby granted, free of charge, to any person obtaining a copy
+ *  of this software and associated documentation files (the "Software"), to deal
+ *  in the Software without restriction, including without limitation the rights
+ *  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ *  copies of the Software, and to permit persons to whom the Software is
+ *  furnished to do so, subject to the following conditions:
+ *
+ *  The above copyright notice and this permission notice shall be included in all
+ *  copies or substantial portions of the Software.
+ *
+ *  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ *  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ *  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ *  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ *  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ *  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ *  SOFTWARE.
+ */
+
+#include <cstdint>
+#include <cmath>
+
+namespace bitstream
+{
+	/**
+	 * @brief A quantized representation of a quaternion
+	*/
+	struct quantized_quaternion
+	{
+		uint32_t m;
+		uint32_t a;
+		uint32_t b;
+		uint32_t c;
+
+		constexpr quantized_quaternion() noexcept :
+			m(0),
+			a(0),
+			b(0),
+			c(0) {}
+
+		constexpr quantized_quaternion(uint32_t w, uint32_t x, uint32_t y, uint32_t z) noexcept :
+			m(w), a(x), b(y), c(z) {}
+	};
+
+	/**
+	 * @brief Class for quantizing a user-specified quaternion into fewer bits using the smallest-three algorithm
+	 * @tparam T The quaternion-type to quantize
+	*/
+	template<typename T, size_t BitsPerElement = 12>
+	class smallest_three
+	{
+	private:
+		static constexpr float SMALLEST_THREE_UNPACK = 0.70710678118654752440084436210485f + 0.0000001f;
+		static constexpr float SMALLEST_THREE_PACK = 1.0f / SMALLEST_THREE_UNPACK;
+
+	public:
+		inline static quantized_quaternion quantize(const T& quaternion) noexcept
+		{
+			constexpr float half_range = static_cast<float>(1 << (BitsPerElement - 1));
+			constexpr float packer = SMALLEST_THREE_PACK * half_range;
+            
+			float max_value = -1.0f;
+			bool sign_minus = false;
+			uint32_t m = 0;
+			uint32_t a = 0;
+			uint32_t b = 0;
+			uint32_t c = 0;
+
+			for (uint32_t i = 0; i < 4; i++)
+            {
+				float element = quaternion[i];
+
+				float abs = element > 0.0f ? element : -element;
+
+				if (abs > max_value)
+				{
+					sign_minus = element < 0.0f;
+					m = i;
+					max_value = abs;
+				}
+			}
+
+			float af = 0.0f;
+			float bf = 0.0f;
+			float cf = 0.0f;
+
+			switch (m)
+			{
+			case 0:
+				af = quaternion[1];
+				bf = quaternion[2];
+				cf = quaternion[3];
+				break;
+			case 1:
+				af = quaternion[0];
+				bf = quaternion[2];
+				cf = quaternion[3];
+				break;
+			case 2:
+				af = quaternion[0];
+				bf = quaternion[1];
+				cf = quaternion[3];
+				break;
+			default: // case 3
+				af = quaternion[0];
+				bf = quaternion[1];
+				cf = quaternion[2];
+				break;
+			}
+
+			if (sign_minus)
+			{
+				a = static_cast<uint32_t>((-af * packer) + half_range);
+				b = static_cast<uint32_t>((-bf * packer) + half_range);
+				c = static_cast<uint32_t>((-cf * packer) + half_range);
+			}
+			else
+			{
+				a = static_cast<uint32_t>((af * packer) + half_range);
+				b = static_cast<uint32_t>((bf * packer) + half_range);
+				c = static_cast<uint32_t>((cf * packer) + half_range);
+			}
+
+			return { m, a, b, c };
+		}
+
+		inline static T dequantize(const quantized_quaternion& data) noexcept
+		{
+			constexpr uint32_t half_range = (1 << (BitsPerElement - 1));
+			constexpr float unpacker = SMALLEST_THREE_UNPACK * (1.0f / half_range);
+            
+			float a = static_cast<float>(data.a * unpacker - half_range * unpacker);
+			float b = static_cast<float>(data.b * unpacker - half_range * unpacker);
+			float c = static_cast<float>(data.c * unpacker - half_range * unpacker);
+
+			float d = std::sqrt(1.0f - ((a * a) + (b * b) + (c * c)));
+
+			switch (data.m)
+			{
+			case 0:
+				return T{ d, a, b, c };
+			case 1:
+				return T{ a, d, b, c };
+			case 2:
+				return T{ a, b, d, c };
+			default: // case 3
+				return T{ a, b, c, d };
+			}
+		}
+	};
+}

include/bitstream/stream/bit_measure.h

@@ -0,0 +1,235 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/crc.h"
+#include "../utility/endian.h"
+#include "../utility/meta.h"
+
+#include "byte_buffer.h"
+#include "serialize_traits.h"
+
+#include <cstdint>
+#include <cstring>
+#include <limits>
+#include <memory>
+#include <type_traits>
+
+namespace bitstream
+{
+	/**
+	 * @brief A stream for writing objects tightly into a buffer
+	 * @note Does not take ownership of the buffer
+	*/
+	class bit_measure
+	{
+	public:
+		static constexpr bool writing = true;
+		static constexpr bool reading = false;
+
+		/**
+		 * @brief Default construct a writer pointing to a null buffer
+		*/
+		bit_measure() noexcept :
+			m_NumBitsWritten(0),
+			m_TotalBits((std::numeric_limits<uint32_t>::max)()) {}
+
+		/**
+		 * @brief Construct a writer pointing to the given byte array with @p num_bytes size
+		 * @param num_bytes The number of bytes in the array
+		*/
+		bit_measure(uint32_t num_bytes) noexcept :
+			m_NumBitsWritten(0),
+			m_TotalBits(num_bytes * 8) {}
+
+		bit_measure(const bit_measure&) = delete;
+
+		bit_measure(bit_measure&& other) noexcept :
+			m_NumBitsWritten(other.m_NumBitsWritten),
+			m_TotalBits(other.m_TotalBits)
+		{
+			other.m_NumBitsWritten = 0;
+			other.m_TotalBits = 0;
+		}
+
+		bit_measure& operator=(const bit_measure&) = delete;
+
+		bit_measure& operator=(bit_measure&& rhs) noexcept
+		{
+			m_NumBitsWritten = rhs.m_NumBitsWritten;
+			m_TotalBits = rhs.m_TotalBits;
+
+			rhs.m_NumBitsWritten = 0;
+			rhs.m_TotalBits = 0;
+
+			return *this;
+		}
+
+		/**
+		 * @brief Returns the buffer that this writer is currently serializing into
+		 * @return The buffer
+		*/
+		[[nodiscard]] uint8_t* get_buffer() const noexcept { return nullptr; }
+
+		/**
+		 * @brief Returns the number of bits which have been written to the buffer
+		 * @return The number of bits which have been written
+		*/
+		[[nodiscard]] uint32_t get_num_bits_serialized() const noexcept { return m_NumBitsWritten; }
+
+		/**
+		 * @brief Returns the number of bytes which have been written to the buffer
+		 * @return The number of bytes which have been written
+		*/
+		[[nodiscard]] uint32_t get_num_bytes_serialized() const noexcept { return m_NumBitsWritten > 0U ? ((m_NumBitsWritten - 1U) / 8U + 1U) : 0U; }
+
+		/**
+		 * @brief Returns whether the @p num_bits can fit in the buffer
+		 * @param num_bits The number of bits to test
+		 * @return Whether the number of bits can fit in the buffer
+		*/
+		[[nodiscard]] bool can_serialize_bits(uint32_t num_bits) const noexcept { return m_NumBitsWritten + num_bits <= m_TotalBits; }
+
+		/**
+		 * @brief Returns the number of bits which have not been written yet
+		 * @note The same as get_total_bits() - get_num_bits_serialized()
+		 * @return The remaining space in the buffer
+		*/
+		[[nodiscard]] uint32_t get_remaining_bits() const noexcept { return m_TotalBits - m_NumBitsWritten; }
+
+		/**
+		 * @brief Returns the size of the buffer, in bits
+		 * @return The size of the buffer, in bits
+		*/
+		[[nodiscard]] uint32_t get_total_bits() const noexcept { return m_TotalBits; }
+
+		/**
+		 * @brief Instructs the writer that you intend to use `serialize_checksum()` later on, and to reserve the first 32 bits.
+		 * @return Returns false if anything has already been written to the buffer or if there's no space to write the checksum
+		*/
+		[[nodiscard]] bool prepend_checksum() noexcept
+		{
+			BS_ASSERT(m_NumBitsWritten == 0);
+
+			BS_ASSERT(can_serialize_bits(32U));
+
+			m_NumBitsWritten += 32U;
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes a checksum of the @p protocol_version and the rest of the buffer as the first 32 bits
+		 * @param protocol_version A unique version number
+		 * @return The number of bytes written to the buffer
+		*/
+		uint32_t serialize_checksum(uint32_t protocol_version) noexcept
+		{
+			return m_NumBitsWritten;
+		}
+
+		/**
+		 * @brief Pads the buffer up to the given number of bytes with zeros
+		 * @param num_bytes The byte number to pad to
+		 * @return Returns false if the current size of the buffer is bigger than @p num_bytes
+		*/
+		[[nodiscard]] bool pad_to_size(uint32_t num_bytes) noexcept
+		{
+			BS_ASSERT(num_bytes * 8U <= m_TotalBits);
+
+			BS_ASSERT(num_bytes * 8U >= m_NumBitsWritten);
+
+			m_NumBitsWritten = num_bytes * 8U;
+
+			return true;
+		}
+
+		/**
+		 * @brief Pads the buffer up with the given number of bytes
+		 * @param num_bytes The amount of bytes to pad
+		 * @return Returns false if the current size of the buffer is bigger than @p num_bytes or if the padded bits are not zeros.
+		*/
+		[[nodiscard]] bool pad(uint32_t num_bytes) noexcept
+		{
+			return pad_to_size(get_num_bytes_serialized() + num_bytes);
+		}
+
+		/**
+		 * @brief Pads the buffer with up to 8 zeros, so that the next write is byte-aligned
+		 * @return Success
+		*/
+		[[nodiscard]] bool align() noexcept
+		{
+			uint32_t remainder = m_NumBitsWritten % 8U;
+			if (remainder != 0U)
+				m_NumBitsWritten += 8U - remainder;
+			return true;
+		}
+
+		/**
+		 * @brief Writes the first @p num_bits bits of @p value into the buffer
+		 * @param value The value to serialize
+		 * @param num_bits The number of bits of the @p value to serialize
+		 * @return Returns false if @p num_bits is less than 1 or greater than 32 or if writing the given number of bits would overflow the buffer
+		*/
+		[[nodiscard]] bool serialize_bits(uint32_t value, uint32_t num_bits) noexcept
+		{
+			BS_ASSERT(num_bits > 0U && num_bits <= 32U);
+
+			BS_ASSERT(can_serialize_bits(num_bits));
+
+			m_NumBitsWritten += num_bits;
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes the first @p num_bits bits of the given byte array, 32 bits at a time
+		 * @param bytes The bytes to serialize
+		 * @param num_bits The number of bits of the @p bytes to serialize
+		 * @return Returns false if @p num_bits is less than 1 or if writing the given number of bits would overflow the buffer
+		*/
+		[[nodiscard]] bool serialize_bytes(const uint8_t* bytes, uint32_t num_bits) noexcept
+		{
+			BS_ASSERT(num_bits > 0U);
+
+			BS_ASSERT(can_serialize_bits(num_bits));
+
+			m_NumBitsWritten += num_bits;
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes to the buffer, using the given @p Trait.
+		 * @note The Trait type in this function must always be explicitly declared
+		 * @tparam Trait A template specialization of serialize_trait<>
+		 * @tparam ...Args The types of the arguments to pass to the serialize function
+		 * @param ...args The arguments to pass to the serialize function
+		 * @return Whether successful or not
+		*/
+		template<typename Trait, typename... Args, typename = utility::has_serialize_t<Trait, bit_measure, Args...>>
+		[[nodiscard]] bool serialize(Args&&... args) noexcept(utility::is_serialize_noexcept_v<Trait, bit_measure, Args...>)
+		{
+			return serialize_traits<Trait>::serialize(*this, std::forward<Args>(args)...);
+		}
+
+		/**
+		 * @brief Writes to the buffer, by trying to deduce the trait.
+		 * @note The Trait type in this function is always implicit and will be deduced from the first argument if possible.
+		 * If the trait cannot be deduced it will not compile.
+		 * @tparam Trait The type of the first argument, which will be used to deduce the trait specialization
+		 * @tparam ...Args The types of the arguments to pass to the serialize function
+		 * @param arg The first argument to pass to the serialize function
+		 * @param ...args The rest of the arguments to pass to the serialize function
+		 * @return Whether successful or not
+		*/
+		template<typename... Args, typename Trait, typename = utility::has_deduce_serialize_t<Trait, bit_measure, Args...>>
+		[[nodiscard]] bool serialize(Trait&& arg, Args&&... args) noexcept(utility::is_deduce_serialize_noexcept_v<Trait, bit_measure, Args...>)
+		{
+			return serialize_traits<utility::deduce_trait_t<Trait, bit_measure, Args...>>::serialize(*this, std::forward<Trait>(arg), std::forward<Args>(args)...);
+		}
+
+	private:
+		uint32_t m_NumBitsWritten;
+		uint32_t m_TotalBits;
+	};
+}

include/bitstream/stream/bit_reader.h

@@ -0,0 +1,343 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/crc.h"
+#include "../utility/endian.h"
+#include "../utility/meta.h"
+
+#include "byte_buffer.h"
+#include "serialize_traits.h"
+#include "stream_traits.h"
+
+#include <cstdint>
+#include <cstring>
+#include <string>
+#include <type_traits>
+
+namespace bitstream
+{
+	/**
+	 * @brief A stream for reading objects from a tightly packed buffer
+	 * @tparam Policy The underlying representation of the buffer
+	*/
+	template<typename Policy>
+	class bit_reader
+	{
+	public:
+		static constexpr bool writing = false;
+		static constexpr bool reading = true;
+
+		/**
+		 * @brief Construct a reader with the parameters passed to the underlying policy
+		 * @param ...args The arguments to pass to the policy
+		*/
+		template<typename... Ts,
+			typename = std::enable_if_t<std::is_constructible_v<Policy, Ts...>>>
+		bit_reader(Ts&&... args)
+			noexcept(std::is_nothrow_constructible_v<Policy, Ts...>) :
+			m_Policy(std::forward<Ts>(args) ...),
+			m_Scratch(0),
+			m_ScratchBits(0),
+			m_WordIndex(0) {}
+
+		bit_reader(const bit_reader&) = delete;
+        
+		bit_reader(bit_reader&& other) noexcept :
+			m_Policy(std::move(other.m_Policy)),
+            m_Scratch(other.m_Scratch),
+            m_ScratchBits(other.m_ScratchBits),
+            m_WordIndex(other.m_WordIndex)
+        {
+            other.m_Scratch = 0;
+            other.m_ScratchBits = 0;
+            other.m_WordIndex = 0;
+        }
+
+		bit_reader& operator=(const bit_reader&) = delete;
+
+		bit_reader& operator=(bit_reader&& rhs) noexcept
+		{
+			m_Policy = std::move(rhs.m_Policy);
+			m_Scratch = rhs.m_Scratch;
+			m_ScratchBits = rhs.m_ScratchBits;
+			m_WordIndex = rhs.m_WordIndex;
+
+			rhs.m_Scratch = 0;
+			rhs.m_ScratchBits = 0;
+			rhs.m_WordIndex = 0;
+            
+            return *this;
+		}
+
+		/**
+		 * @brief Returns the buffer that this reader is currently serializing from
+		 * @return The buffer
+		*/
+		[[nodiscard]] const uint8_t* get_buffer() const noexcept { return reinterpret_cast<const uint8_t*>(m_Policy.get_buffer()); }
+
+		/**
+		 * @brief Returns the number of bits which have been read from the buffer
+		 * @return The number of bits which have been read
+		*/
+		[[nodiscard]] uint32_t get_num_bits_serialized() const noexcept { return m_Policy.get_num_bits_serialized(); }
+
+		/**
+		 * @brief Returns the number of bytes which have been read from the buffer
+		 * @return The number of bytes which have been read
+		*/
+		[[nodiscard]] uint32_t get_num_bytes_serialized() const noexcept { return get_num_bits_serialized() > 0U ? ((get_num_bits_serialized() - 1U) / 8U + 1U) : 0U; }
+
+		/**
+		 * @brief Returns whether the @p num_bits be read from the buffer
+		 * @param num_bits The number of bits to test
+		 * @return Whether the number of bits can be read from the buffer
+		*/
+		[[nodiscard]] bool can_serialize_bits(uint32_t num_bits) const noexcept { return m_Policy.can_serialize_bits(num_bits); }
+
+		/**
+		 * @brief Returns the number of bits which have not been read yet
+		 * @note The same as get_total_bits() - get_num_bits_serialized()
+		 * @return The remaining space in the buffer
+		*/
+		[[nodiscard]] uint32_t get_remaining_bits() const noexcept { return get_total_bits() - get_num_bits_serialized(); }
+
+		/**
+		 * @brief Returns the size of the buffer, in bits
+		 * @return The size of the buffer, in bits
+		*/
+		[[nodiscard]] uint32_t get_total_bits() const noexcept { return m_Policy.get_total_bits(); }
+
+		/**
+		 * @brief Reads the first 32 bits of the buffer and compares it to a checksum of the @p protocol_version and the rest of the buffer
+		 * @param protocol_version A unique version number
+		 * @return Whether the checksum matches what was written
+		*/
+		[[nodiscard]] bool serialize_checksum(uint32_t protocol_version) noexcept
+		{
+			BS_ASSERT(get_num_bits_serialized() == 0);
+
+			BS_ASSERT(can_serialize_bits(32U));
+
+			uint32_t num_bytes = (get_total_bits() - 1U) / 8U + 1U;
+			const uint32_t* buffer = m_Policy.get_buffer();
+
+			// Generate checksum to compare against
+			uint32_t generated_checksum = utility::crc_uint32(reinterpret_cast<const uint8_t*>(&protocol_version), reinterpret_cast<const uint8_t*>(buffer + 1), num_bytes - 4);
+
+			// Advance the reader by the size of the checksum (32 bits / 1 word)
+			m_WordIndex++;
+
+			BS_ASSERT(m_Policy.extend(32U));
+
+			// Read the checksum
+			uint32_t checksum = *buffer;
+
+			// Compare the checksum
+			return generated_checksum == checksum;
+		}
+
+		/**
+		 * @brief Pads the buffer up to the given number of bytes
+		 * @param num_bytes The byte number to pad to
+		 * @return Returns false if the current size of the buffer is bigger than @p num_bytes or if the padded bits are not zeros.
+		*/
+		[[nodiscard]] bool pad_to_size(uint32_t num_bytes) noexcept
+		{
+			uint32_t num_bits_read = get_num_bits_serialized();
+
+			BS_ASSERT(num_bytes * 8U >= num_bits_read);
+
+			BS_ASSERT(can_serialize_bits(num_bytes * 8U - num_bits_read));
+
+			uint32_t remainder = (num_bytes * 8U - num_bits_read) % 32U;
+			uint32_t zero;
+
+			// Test the last word more carefully, as it may have data
+			if (remainder != 0U)
+			{
+				bool status = serialize_bits(zero, remainder);
+				BS_ASSERT(status && zero == 0);
+			}
+
+			uint32_t offset = get_num_bits_serialized() / 32;
+            uint32_t max = num_bytes / 4;
+
+			// Test for zeros in padding
+			for (uint32_t i = offset; i < max; i++)
+			{
+				bool status = serialize_bits(zero, 32);
+				BS_ASSERT(status && zero == 0);
+			}
+
+			return true;
+		}
+
+		/**
+		 * @brief Pads the buffer up with the given number of bytes
+		 * @param num_bytes The amount of bytes to pad
+		 * @return Returns false if the current size of the buffer is bigger than @p num_bytes or if the padded bits are not zeros.
+		*/
+		[[nodiscard]] bool pad(uint32_t num_bytes) noexcept
+		{
+			return pad_to_size(get_num_bytes_serialized() + num_bytes);
+		}
+
+		/**
+		 * @brief Pads the buffer with up to 8 zeros, so that the next read is byte-aligned
+		 * @notes Return false if the padded bits are not zeros
+		 * @return Returns false if the padded bits are not zeros
+		*/
+		[[nodiscard]] bool align() noexcept
+		{
+			uint32_t remainder = get_num_bits_serialized() % 8U;
+			if (remainder != 0U)
+			{
+				uint32_t zero;
+				bool status = serialize_bits(zero, 8U - remainder);
+
+                BS_ASSERT(status && zero == 0U && get_num_bits_serialized() % 8U == 0U);
+			}
+
+			return true;
+		}
+
+		/**
+		 * @brief Reads the first @p num_bits bits of @p value from the buffer
+		 * @param value The value to serialize
+		 * @param num_bits The number of bits of the @p value to serialize
+		 * @return Returns false if @p num_bits is less than 1 or greater than 32 or if reading the given number of bits would overflow the buffer
+		*/
+		[[nodiscard]] bool serialize_bits(uint32_t& value, uint32_t num_bits) noexcept
+		{
+			BS_ASSERT(num_bits > 0U && num_bits <= 32U);
+
+			BS_ASSERT(m_Policy.extend(num_bits));
+
+			// This is actually slower
+			// Possibly due to unlikely branching
+			/*if (num_bits == 32U && m_ScratchBits == 0U)
+			{
+				const uint32_t* ptr = m_Policy.get_buffer() + m_WordIndex;
+
+				value = utility::to_big_endian32(*ptr);
+
+				m_WordIndex++;
+
+				return true;
+			}*/
+
+			if (m_ScratchBits < num_bits)
+			{
+				const uint32_t* ptr = m_Policy.get_buffer() + m_WordIndex;
+
+				uint64_t ptr_value = static_cast<uint64_t>(utility::to_big_endian32(*ptr)) << (32U - m_ScratchBits);
+				m_Scratch |= ptr_value;
+				m_ScratchBits += 32U;
+				m_WordIndex++;
+			}
+
+			uint32_t offset = 64U - num_bits;
+			value = static_cast<uint32_t>(m_Scratch >> offset);
+
+			m_Scratch <<= num_bits;
+			m_ScratchBits -= num_bits;
+
+			return true;
+		}
+
+		/**
+		 * @brief Reads the first @p num_bits bits of the given byte array, 32 bits at a time
+		 * @param bytes The bytes to serialize
+		 * @param num_bits The number of bits of the @p bytes to serialize
+		 * @return Returns false if @p num_bits is less than 1 or if reading the given number of bits would overflow the buffer
+		*/
+		[[nodiscard]] bool serialize_bytes(uint8_t* bytes, uint32_t num_bits) noexcept
+		{
+			BS_ASSERT(num_bits > 0U);
+            
+			BS_ASSERT(can_serialize_bits(num_bits));
+            
+            // Read the byte array as words
+            uint32_t* word_buffer = reinterpret_cast<uint32_t*>(bytes);
+			uint32_t num_words = num_bits / 32U;
+            
+            if (m_ScratchBits % 32U == 0U && num_words > 0U)
+            {
+				BS_ASSERT(m_Policy.extend(num_words * 32U));
+
+                // If the read buffer is word-aligned, just memcpy it
+                std::memcpy(word_buffer, m_Policy.get_buffer() + m_WordIndex, num_words * 4U);
+                
+                m_WordIndex += num_words;
+            }
+            else
+            {
+                // If the buffer is not word-aligned, serialize a word at a time
+                for (uint32_t i = 0U; i < num_words; i++)
+                {
+                    uint32_t value;
+                    BS_ASSERT(serialize_bits(value, 32U));
+                    
+                    // Casting a byte-array to an int is wrong on little-endian systems
+                    // We have to swap the bytes around
+                    word_buffer[i] = utility::to_big_endian32(value);
+                }
+            }
+            
+            // Early exit if the word-count matches
+            if (num_bits % 32 == 0)
+                return true;
+            
+            uint32_t remaining_bits = num_bits - num_words * 32U;
+            
+            uint32_t num_bytes = (remaining_bits - 1U) / 8U + 1U;
+			for (uint32_t i = 0; i < num_bytes; i++)
+			{
+                uint32_t value;
+				BS_ASSERT(serialize_bits(value, (std::min)(remaining_bits - i * 8U, 8U)));
+                
+				bytes[num_words * 4 + i] = static_cast<uint8_t>(value);
+			}
+
+			return true;
+		}
+
+		/**
+		 * @brief Reads from the buffer, using the given @p Trait.
+		 * @note The Trait type in this function must always be explicitly declared
+		 * @tparam Trait A template specialization of serialize_trait<>
+		 * @tparam ...Args The types of the arguments to pass to the serialize function
+		 * @param ...args The arguments to pass to the serialize function
+		 * @return Whether successful or not
+		*/
+		template<typename Trait, typename... Args, typename = utility::has_serialize_t<Trait, bit_reader, Args...>>
+		[[nodiscard]] bool serialize(Args&&... args) noexcept(utility::is_serialize_noexcept_v<Trait, bit_reader, Args...>)
+		{
+			return serialize_traits<Trait>::serialize(*this, std::forward<Args>(args)...);
+		}
+
+		/**
+		 * @brief Reads from the buffer, by trying to deduce the trait.
+		 * @note The Trait type in this function is always implicit and will be deduced from the first argument if possible.
+		 * If the trait cannot be deduced it will not compile.
+		 * @tparam Trait The type of the first argument, which will be used to deduce the trait specialization
+		 * @tparam ...Args The types of the arguments to pass to the serialize function
+		 * @param arg The first argument to pass to the serialize function
+		 * @param ...args The rest of the arguments to pass to the serialize function
+		 * @return Whether successful or not
+		*/
+		template<typename... Args, typename Trait, typename = utility::has_deduce_serialize_t<Trait, bit_reader, Args...>>
+		[[nodiscard]] bool serialize(Trait&& arg, Args&&... args) noexcept(utility::is_deduce_serialize_noexcept_v<Trait, bit_reader, Args...>)
+		{
+			return serialize_traits<utility::deduce_trait_t<Trait, bit_reader, Args...>>::serialize(*this, std::forward<Trait>(arg), std::forward<Args>(args)...);
+		}
+
+	private:
+		Policy m_Policy;
+
+		uint64_t m_Scratch;
+		uint32_t m_ScratchBits;
+		uint32_t m_WordIndex;
+	};
+
+	using fixed_bit_reader = bit_reader<fixed_policy>;
+}

include/bitstream/stream/bit_writer.h

@@ -0,0 +1,400 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/crc.h"
+#include "../utility/endian.h"
+#include "../utility/meta.h"
+
+#include "byte_buffer.h"
+#include "serialize_traits.h"
+#include "stream_traits.h"
+
+#include <cstdint>
+#include <cstring>
+#include <memory>
+#include <type_traits>
+
+namespace bitstream
+{
+	/**
+	 * @brief A stream for writing objects tightly into a buffer
+	 * @tparam Policy The underlying representation of the buffer
+	*/
+	template<typename Policy>
+	class bit_writer
+	{
+	public:
+		static constexpr bool writing = true;
+		static constexpr bool reading = false;
+
+		/**
+		 * @brief Construct a writer with the parameters passed to the underlying policy
+		 * @param ...args The arguments to pass to the policy
+		*/
+		template<typename... Ts,
+			typename = std::enable_if_t<std::is_constructible_v<Policy, Ts...>>>
+		bit_writer(Ts&&... args)
+			noexcept(std::is_nothrow_constructible_v<Policy, Ts...>) :
+			m_Policy(std::forward<Ts>(args) ...),
+			m_Scratch(0),
+			m_ScratchBits(0),
+			m_WordIndex(0) {}
+
+		bit_writer(const bit_writer&) = delete;
+        
+		bit_writer(bit_writer&& other) noexcept :
+            m_Policy(std::move(other.m_Policy)),
+            m_Scratch(other.m_Scratch),
+            m_ScratchBits(other.m_ScratchBits),
+            m_WordIndex(other.m_WordIndex)
+        {
+            other.m_Scratch = 0;
+            other.m_ScratchBits = 0;
+            other.m_WordIndex = 0;
+        }
+
+		bit_writer& operator=(const bit_writer&) = delete;
+
+		bit_writer& operator=(bit_writer&& rhs) noexcept
+		{
+			m_Policy = std::move(rhs.m_Policy);
+			m_Scratch = rhs.m_Scratch;
+			m_ScratchBits = rhs.m_ScratchBits;
+			m_WordIndex = rhs.m_WordIndex;
+
+			rhs.m_Scratch = 0;
+			rhs.m_ScratchBits = 0;
+			rhs.m_WordIndex = 0;
+            
+            return *this;
+		}
+
+		/**
+		 * @brief Returns the buffer that this writer is currently serializing into
+		 * @return The buffer
+		*/
+		[[nodiscard]] uint8_t* get_buffer() const noexcept { return reinterpret_cast<uint8_t*>(m_Policy.get_buffer()); }
+
+		/**
+		 * @brief Returns the number of bits which have been written to the buffer
+		 * @return The number of bits which have been written
+		*/
+		[[nodiscard]] uint32_t get_num_bits_serialized() const noexcept { return m_Policy.get_num_bits_serialized(); }
+
+		/**
+		 * @brief Returns the number of bytes which have been written to the buffer
+		 * @return The number of bytes which have been written
+		*/
+		[[nodiscard]] uint32_t get_num_bytes_serialized() const noexcept { return get_num_bits_serialized() > 0U ? ((get_num_bits_serialized() - 1U) / 8U + 1U) : 0U; }
+
+		/**
+		 * @brief Returns whether the @p num_bits can fit in the buffer
+		 * @param num_bits The number of bits to test
+		 * @return Whether the number of bits can fit in the buffer
+		*/
+		[[nodiscard]] bool can_serialize_bits(uint32_t num_bits) const noexcept { return m_Policy.can_serialize_bits(num_bits); }
+
+		/**
+		 * @brief Returns the number of bits which have not been written yet
+		 * @note The same as get_total_bits() - get_num_bits_serialized()
+		 * @return The remaining space in the buffer
+		*/
+		[[nodiscard]] uint32_t get_remaining_bits() const noexcept { return get_total_bits() - get_num_bits_serialized(); }
+
+        /**
+         * @brief Returns the size of the buffer, in bits
+         * @return The size of the buffer, in bits
+        */
+		[[nodiscard]] uint32_t get_total_bits() const noexcept { return m_Policy.get_total_bits(); }
+        
+		/**
+		 * @brief Flushes any remaining bits into the buffer. Use this when you no longer intend to write anything to the buffer.
+		 * @return The number of bytes written to the buffer
+		*/
+		uint32_t flush() noexcept
+		{
+			if (m_ScratchBits > 0U)
+			{
+				uint32_t* ptr = m_Policy.get_buffer() + m_WordIndex;
+				uint32_t ptr_value = static_cast<uint32_t>(m_Scratch >> 32U);
+				*ptr = utility::to_big_endian32(ptr_value);
+
+				m_Scratch = 0U;
+				m_ScratchBits = 0U;
+				m_WordIndex++;
+			}
+
+			return get_num_bits_serialized();
+		}
+
+		/**
+		 * @brief Instructs the writer that you intend to use `serialize_checksum()` later on, and to reserve the first 32 bits.
+		 * @return Returns false if anything has already been written to the buffer or if there's no space to write the checksum
+		*/
+		[[nodiscard]] bool prepend_checksum() noexcept
+		{
+			BS_ASSERT(get_num_bits_serialized() == 0);
+
+            BS_ASSERT(m_Policy.extend(32U));
+            
+			// Advance the reader by the size of the checksum (32 bits / 1 word)
+			m_WordIndex++;
+            
+            return true;
+		}
+
+		/**
+		 * @brief Writes a checksum of the @p protocol_version and the rest of the buffer as the first 32 bits
+		 * @param protocol_version A unique version number
+		 * @return The number of bytes written to the buffer
+		*/
+		uint32_t serialize_checksum(uint32_t protocol_version) noexcept
+		{
+			uint32_t num_bits = flush();
+
+			BS_ASSERT(num_bits > 32U);
+
+			// Copy protocol version to buffer
+			uint32_t* buffer = m_Policy.get_buffer();
+			*buffer = protocol_version;
+
+			// Generate checksum of version + data
+			uint32_t checksum = utility::crc_uint32(reinterpret_cast<uint8_t*>(buffer), get_num_bytes_serialized());
+
+			// Put checksum at beginning
+			*buffer = checksum;
+
+			return num_bits;
+		}
+
+		/**
+		 * @brief Pads the buffer up to the given number of bytes with zeros
+		 * @param num_bytes The byte number to pad to
+		 * @return Returns false if the current size of the buffer is bigger than @p num_bytes
+		*/
+		[[nodiscard]] bool pad_to_size(uint32_t num_bytes) noexcept
+		{
+			uint32_t num_bits_written = get_num_bits_serialized();
+
+			BS_ASSERT(num_bytes * 8U >= num_bits_written);
+
+			BS_ASSERT(can_serialize_bits(num_bytes * 8U - num_bits_written));
+
+            if (num_bits_written == 0)
+            {
+				BS_ASSERT(m_Policy.extend(num_bytes * 8U - num_bits_written));
+
+                std::memset(m_Policy.get_buffer(), 0, num_bytes);
+                
+                m_Scratch = 0;
+                m_ScratchBits = 0;
+                m_WordIndex = num_bytes / 4;
+                
+                return true;
+            }
+            
+            uint32_t remainder = (num_bytes * 8U - num_bits_written) % 32U;
+			uint32_t zero = 0;
+
+			// Align to byte
+			if (remainder != 0U)
+				BS_ASSERT(serialize_bits(zero, remainder));
+
+			uint32_t offset = get_num_bits_serialized() / 32;
+            uint32_t max = num_bytes / 4;
+
+			// Serialize words
+			for (uint32_t i = offset; i < max; i++)
+				BS_ASSERT(serialize_bits(zero, 32));
+
+			return true;
+		}
+
+		/**
+		 * @brief Pads the buffer up with the given number of bytes
+		 * @param num_bytes The amount of bytes to pad
+		 * @return Returns false if the current size of the buffer is bigger than @p num_bytes or if the padded bits are not zeros.
+		*/
+		[[nodiscard]] bool pad(uint32_t num_bytes) noexcept
+		{
+			return pad_to_size(get_num_bytes_serialized() + num_bytes);
+		}
+
+		/**
+		 * @brief Pads the buffer with up to 8 zeros, so that the next write is byte-aligned
+		 * @return Success
+		*/
+		[[nodiscard]] bool align() noexcept
+		{
+			uint32_t remainder = m_ScratchBits % 8U;
+			if (remainder != 0U)
+			{
+				uint32_t zero = 0U;
+				bool status = serialize_bits(zero, 8U - remainder);
+
+				BS_ASSERT(status && get_num_bits_serialized() % 8U == 0U);
+			}
+			return true;
+		}
+
+		/**
+		 * @brief Writes the first @p num_bits bits of @p value into the buffer
+		 * @param value The value to serialize
+		 * @param num_bits The number of bits of the @p value to serialize
+		 * @return Returns false if @p num_bits is less than 1 or greater than 32 or if writing the given number of bits would overflow the buffer
+		*/
+		[[nodiscard]] bool serialize_bits(uint32_t value, uint32_t num_bits) noexcept
+		{
+			BS_ASSERT(num_bits > 0U && num_bits <= 32U);
+
+			BS_ASSERT(m_Policy.extend(num_bits));
+
+			// This is actually slower
+			// Possibly due to unlikely branching
+			/*if (num_bits == 32U && m_ScratchBits == 0U)
+			{
+				uint32_t* ptr = m_Policy.get_buffer() + m_WordIndex;
+
+				*ptr = utility::to_big_endian32(value);
+
+				m_WordIndex++;
+
+				return true;
+			}*/
+
+			uint32_t offset = 64U - num_bits - m_ScratchBits;
+			uint64_t ls_value = static_cast<uint64_t>(value) << offset;
+
+			m_Scratch |= ls_value;
+			m_ScratchBits += num_bits;
+
+			if (m_ScratchBits >= 32U)
+			{
+				uint32_t* ptr = m_Policy.get_buffer() + m_WordIndex;
+				uint32_t ptr_value = static_cast<uint32_t>(m_Scratch >> 32U);
+				*ptr = utility::to_big_endian32(ptr_value);
+				m_Scratch <<= 32ULL;
+				m_ScratchBits -= 32U;
+				m_WordIndex++;
+			}
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes the first @p num_bits bits of the given byte array, 32 bits at a time
+		 * @param bytes The bytes to serialize
+		 * @param num_bits The number of bits of the @p bytes to serialize
+		 * @return Returns false if @p num_bits is less than 1 or if writing the given number of bits would overflow the buffer
+		*/
+		[[nodiscard]] bool serialize_bytes(const uint8_t* bytes, uint32_t num_bits) noexcept
+		{
+			BS_ASSERT(num_bits > 0U);
+            
+			BS_ASSERT(can_serialize_bits(num_bits));
+            
+            // Write the byte array as words
+            const uint32_t* word_buffer = reinterpret_cast<const uint32_t*>(bytes);
+			uint32_t num_words = num_bits / 32U;
+            
+            if (m_ScratchBits % 32U == 0U && num_words > 0U)
+            {
+				BS_ASSERT(m_Policy.extend(num_words * 32U));
+
+                // If the written buffer is word-aligned, just memcpy it
+                std::memcpy(m_Policy.get_buffer() + m_WordIndex, word_buffer, num_words * 4U);
+                
+                m_WordIndex += num_words;
+            }
+            else
+            {
+                // If the buffer is not word-aligned, serialize a word at a time
+                for (uint32_t i = 0U; i < num_words; i++)
+                {
+                    // Casting a byte-array to an int is wrong on little-endian systems
+                    // We have to swap the bytes around
+                    uint32_t value = utility::to_big_endian32(word_buffer[i]);
+                    BS_ASSERT(serialize_bits(value, 32U));
+                }
+            }
+            
+            // Early exit if the word-count matches
+            if (num_bits % 32U == 0U)
+                return true;
+            
+            uint32_t remaining_bits = num_bits - num_words * 32U;
+            
+            uint32_t num_bytes = (remaining_bits - 1U) / 8U + 1U;
+			for (uint32_t i = 0U; i < num_bytes; i++)
+			{
+				uint32_t value = static_cast<uint32_t>(bytes[num_words * 4U + i]);
+				BS_ASSERT(serialize_bits(value, (std::min)(remaining_bits - i * 8U, 8U)));
+			}
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes the contents of the buffer into the given @p writer. Essentially copies the entire buffer without modifying it.
+		 * @param writer The writer to copy into
+		 * @return Returns false if writing would overflow the buffer
+		*/
+		[[nodiscard]] bool serialize_into(bit_writer& writer) const noexcept
+		{
+			uint8_t* buffer = reinterpret_cast<uint8_t*>(m_Policy.get_buffer());
+			uint32_t num_bits = get_num_bits_serialized();
+			uint32_t remainder_bits = num_bits % 8U;
+
+			BS_ASSERT(writer.serialize_bytes(buffer, num_bits - remainder_bits));
+
+			if (remainder_bits > 0U)
+			{
+				uint32_t byte_value = buffer[num_bits / 8U] >> (8U - remainder_bits);
+				BS_ASSERT(writer.serialize_bits(byte_value, remainder_bits));
+			}
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes to the buffer, using the given @p Trait.
+		 * @note The Trait type in this function must always be explicitly declared
+		 * @tparam Trait A template specialization of serialize_trait<>
+		 * @tparam ...Args The types of the arguments to pass to the serialize function
+		 * @param ...args The arguments to pass to the serialize function
+		 * @return Whether successful or not
+		*/
+		template<typename Trait, typename... Args, typename = utility::has_serialize_t<Trait, bit_writer, Args...>>
+		[[nodiscard]] bool serialize(Args&&... args) noexcept(utility::is_serialize_noexcept_v<Trait, bit_writer, Args...>)
+		{
+			return serialize_traits<Trait>::serialize(*this, std::forward<Args>(args)...);
+		}
+
+		/**
+		 * @brief Writes to the buffer, by trying to deduce the trait.
+		 * @note The Trait type in this function is always implicit and will be deduced from the first argument if possible.
+		 * If the trait cannot be deduced it will not compile.
+		 * @tparam Trait The type of the first argument, which will be used to deduce the trait specialization
+		 * @tparam ...Args The types of the arguments to pass to the serialize function
+		 * @param arg The first argument to pass to the serialize function
+		 * @param ...args The rest of the arguments to pass to the serialize function
+		 * @return Whether successful or not
+		*/
+		template<typename... Args, typename Trait, typename = utility::has_deduce_serialize_t<Trait, bit_writer, Args...>>
+		[[nodiscard]] bool serialize(Trait&& arg, Args&&... args) noexcept(utility::is_deduce_serialize_noexcept_v<Trait, bit_writer, Args...>)
+		{
+			return serialize_traits<utility::deduce_trait_t<Trait, bit_writer, Args...>>::serialize(*this, std::forward<Trait>(arg), std::forward<Args>(args)...);
+		}
+
+	private:
+		Policy m_Policy;
+
+		uint64_t m_Scratch;
+		int m_ScratchBits;
+		size_t m_WordIndex;
+	};
+
+	using fixed_bit_writer = bit_writer<fixed_policy>;
+
+	template<typename T>
+	using growing_bit_writer = bit_writer<growing_policy<T>>;
+}

include/bitstream/stream/byte_buffer.h

@@ -0,0 +1,22 @@
+#pragma once
+
+#include <cstddef>
+#include <cstdint>
+
+namespace bitstream
+{
+	/**
+	 * @brief A byte buffer aligned to 4 bytes.
+	 * Can be used with bit_reader and bit_writer.
+	 * @note Size must be a multiple of 4
+	*/
+	template<size_t Size>
+	struct byte_buffer
+	{
+		static_assert(Size % 4 == 0, "Buffer size must be a multiple of 4");
+
+		alignas(uint32_t) uint8_t Bytes[Size];
+
+		uint8_t& operator[](size_t i) noexcept { return Bytes[i]; }
+	};
+}

include/bitstream/stream/serialize_traits.h

@@ -0,0 +1,12 @@
+#pragma once
+
+namespace bitstream
+{
+	/**
+	 * @brief A class for specializing trait serialization functions
+	 * @tparam Trait Make a specialization on this type
+	 * @tparam Void Use std::enable_if here if you need to, otherwise leave empty
+	*/
+	template<typename Trait, typename Void = void>
+	struct serialize_traits;
+}

include/bitstream/stream/stream_traits.h

@@ -0,0 +1,96 @@
+#pragma once
+
+#include "byte_buffer.h"
+
+#include <cstddef>
+#include <cstdint>
+#include <limits>
+#include <type_traits>
+
+namespace bitstream
+{
+	struct fixed_policy
+	{
+		/**
+		 * @brief Construct a stream pointing to the given byte array with @p num_bytes size
+		 * @param bytes The byte array to serialize to/from. Must be 4-byte aligned and the size must be a multiple of 4
+		 * @param num_bytes The number of bytes in the array
+		*/
+		fixed_policy(void* buffer, uint32_t num_bits) noexcept :
+			m_Buffer(static_cast<uint32_t*>(buffer)),
+			m_NumBitsSerialized(0),
+			m_TotalBits(num_bits) {}
+
+		/**
+		 * @brief Construct a stream pointing to the given @p buffer
+		 * @param buffer The buffer to serialize to/from
+		 * @param num_bits The maximum number of bits that we can read
+		*/
+		template<size_t Size>
+		fixed_policy(byte_buffer<Size>& buffer, uint32_t num_bits) noexcept :
+			m_Buffer(reinterpret_cast<uint32_t*>(buffer.Bytes)),
+			m_NumBitsSerialized(0),
+			m_TotalBits(num_bits) {}
+
+		/**
+		 * @brief Construct a stream pointing to the given @p buffer
+		 * @param buffer The buffer to serialize to/from
+		*/
+		template<size_t Size>
+		fixed_policy(byte_buffer<Size>& buffer) noexcept :
+			m_Buffer(reinterpret_cast<uint32_t*>(buffer.Bytes)),
+			m_NumBitsSerialized(0),
+			m_TotalBits(Size * 8) {}
+
+		uint32_t* get_buffer() const noexcept { return m_Buffer; }
+
+		// TODO: Transition sizes to size_t
+		uint32_t get_num_bits_serialized() const noexcept { return m_NumBitsSerialized; }
+
+		bool can_serialize_bits(uint32_t num_bits) const noexcept { return m_NumBitsSerialized + num_bits <= m_TotalBits; }
+
+		uint32_t get_total_bits() const noexcept { return m_TotalBits; }
+
+		bool extend(uint32_t num_bits) noexcept
+		{
+			if (!can_serialize_bits(num_bits))
+				return false;
+
+			m_NumBitsSerialized += num_bits;
+			return true;
+		}
+
+		uint32_t* m_Buffer;
+		// TODO: Transition sizes to size_t
+		uint32_t m_NumBitsSerialized;
+		uint32_t m_TotalBits;
+	};
+
+	template<typename T>
+	struct growing_policy
+	{
+		growing_policy(T& container) noexcept :
+			m_Buffer(container),
+			m_NumBitsSerialized(0) {}
+
+		uint32_t* get_buffer() const noexcept { return m_Buffer.data(); }
+
+		uint32_t get_num_bits_serialized() const noexcept { return m_NumBitsSerialized; }
+
+		bool can_serialize_bits(uint32_t num_bits) const noexcept { return true; }
+
+		uint32_t get_total_bits() const noexcept { return (std::numeric_limits<uint32_t>::max)(); }
+
+		bool extend(uint32_t num_bits)
+		{
+			m_NumBitsSerialized += num_bits;
+			uint32_t num_bytes = (m_NumBitsSerialized - 1) / 8U + 1;
+			m_Buffer.resize(num_bytes);
+			return true;
+		}
+
+		T& m_Buffer;
+
+		uint32_t m_NumBitsSerialized;
+	};
+}

include/bitstream/traits/array_traits.h

@@ -0,0 +1,165 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/meta.h"
+#include "../utility/parameter.h"
+
+#include "../stream/serialize_traits.h"
+
+#include "../traits/bool_trait.h"
+#include "../traits/integral_traits.h"
+
+#include <cstdint>
+
+namespace bitstream
+{
+    /**
+     * @brief Wrapper type for subsets of arrays
+     * @tparam T The type of the array
+    */
+	template<typename T, typename = T>
+	struct array_subset;
+
+	/**
+	 * @brief A trait used for serializing a subset of an array of objects
+	 * @tparam T The type of the object in the array
+	 * @tparam Trait 
+	*/
+	template<typename T, typename Trait>
+	struct serialize_traits<array_subset<T, Trait>>
+	{
+    private:
+        template<uint32_t Min, uint32_t Max, typename Stream>
+        static bool serialize_difference(Stream& stream, int& previous, int& current, uint32_t& difference)
+        {
+            bool use_bits;
+            if constexpr (Stream::writing)
+                use_bits = difference <= Max;
+            BS_ASSERT(stream.template serialize<bool>(use_bits));
+            if (use_bits)
+            {
+                using bounded_trait = bounded_int<uint32_t, Min, Max>;
+
+                BS_ASSERT(stream.template serialize<bounded_trait>(difference));
+                if constexpr (Stream::reading)
+                    current = previous + difference;
+                previous = current;
+                return true;
+            }
+
+            return false;
+        }
+
+        template<typename Stream>
+        static bool serialize_index(Stream& stream, int& previous, int& current, int max_size)
+        {
+            uint32_t difference;
+            if constexpr (Stream::writing)
+            {
+                BS_ASSERT(previous < current);
+                difference = current - previous;
+                BS_ASSERT(difference > 0);
+            }
+
+            // +1 (1 bit)
+            bool plus_one;
+            if constexpr (Stream::writing)
+                plus_one = difference == 1;
+            BS_ASSERT(stream.template serialize<bool>(plus_one));
+            if (plus_one)
+            {
+                if constexpr (Stream::reading)
+                    current = previous + 1;
+                previous = current;
+                return true;
+            }
+
+            // [+2,5] -> [0,3] (2 bits)
+            if (serialize_difference<2, 5>(stream, previous, current, difference))
+                return true;
+
+            // [6,13] -> [0,7] (3 bits)
+            if (serialize_difference<6, 13>(stream, previous, current, difference))
+                return true;
+
+            // [14,29] -> [0,15] (4 bits)
+            if (serialize_difference<14, 29>(stream, previous, current, difference))
+                return true;
+
+            // [30,61] -> [0,31] (5 bits)
+            if (serialize_difference<30, 61>(stream, previous, current, difference))
+                return true;
+
+            // [62,125] -> [0,63] (6 bits)
+            if (serialize_difference<62, 125>(stream, previous, current, difference))
+                return true;
+
+            // [126,MaxObjects+1] 
+            BS_ASSERT(stream.template serialize<uint32_t>(difference, 126, max_size));
+            if constexpr (Stream::reading)
+                current = previous + difference;
+            previous = current;
+            return true;
+        }
+
+    public:
+		/**
+		 * @brief Writes a subset of the array @p values into the writer
+		 * @tparam Compare A function type which returns a bool
+		 * @tparam ...Args The types of any additional arguments
+		 * @param writer The stream to write to
+		 * @param values The array of objects to serialize
+		 * @param max_size The size of the array
+		 * @param compare A function which returns true if the object should be written, false otherwise
+		 * @param ...args Any additional arguments to use when serializing each individual object
+		 * @return Success
+		*/
+        template<typename Stream, typename Compare, typename... Args>
+        typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, T* values, int max_size, Compare compare, Args&&... args) noexcept
+		{
+			int prev_index = -1;
+			for (int index = 0; index < max_size; index++)
+			{
+				if (!compare(values[index]))
+					continue;
+
+                BS_ASSERT(serialize_index(writer, prev_index, index, max_size));
+
+				BS_ASSERT(writer.template serialize<Trait>(values[index], std::forward<Args>(args)...));
+			}
+
+            BS_ASSERT(serialize_index(writer, prev_index, max_size, max_size));
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes a subset of a serialized array into @p values
+		 * @tparam ...Args The types of any additional arguments
+		 * @param reader The stream to read from
+		 * @param values The array of objects to read into
+		 * @param max_size The size of the array
+		 * @param compare Not used, but kept for compatibility with the serialize write function
+		 * @param ...args Any additional arguments to use when serializing each individual object
+		 * @return Success
+		*/
+        template<typename Stream, typename... Args>
+        typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, T* values, int max_size, Args&&... args) noexcept
+		{
+			int prev_index = -1;
+            int index = 0;
+			while (true)
+			{
+                BS_ASSERT(serialize_index(reader, prev_index, index, max_size));
+
+				if (index == max_size)
+					break;
+
+				BS_ASSERT(reader.template serialize<Trait>(values[index], std::forward<Args>(args)...));
+			}
+
+			return true;
+		}
+	};
+}

include/bitstream/traits/bool_trait.h

@@ -0,0 +1,74 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/meta.h"
+#include "../utility/parameter.h"
+
+#include "../stream/serialize_traits.h"
+
+namespace bitstream
+{
+	/**
+	 * @brief A trait used to serialize a boolean as a single bit
+	*/
+	template<>
+	struct serialize_traits<bool>
+	{
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, in<bool> value) noexcept
+		{
+			uint32_t unsigned_value = value;
+
+			return writer.serialize_bits(unsigned_value, 1U);
+		}
+
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, out<bool> value) noexcept
+		{
+			uint32_t unsigned_value;
+
+			BS_ASSERT(reader.serialize_bits(unsigned_value, 1U));
+
+			value = unsigned_value;
+
+			return true;
+		}
+	};
+
+	/**
+	 * @brief A trait used to serialize multiple boolean values
+	*/
+	template<size_t Size>
+	struct serialize_traits<bool[Size]>
+	{
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, const bool* values) noexcept
+		{
+			uint32_t unsigned_value;
+			for (size_t i = 0; i < Size; i++)
+			{
+				unsigned_value = values[i];
+				BS_ASSERT(writer.serialize_bits(unsigned_value, 1U));
+			}
+
+			return writer.serialize_bits(unsigned_value, 1U);
+		}
+
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, bool* values) noexcept
+		{
+			uint32_t unsigned_value;
+			for (size_t i = 0; i < Size; i++)
+			{
+				BS_ASSERT(reader.serialize_bits(unsigned_value, 1U));
+
+				values[i] = unsigned_value;
+			}
+
+			return true;
+		}
+	};
+}

include/bitstream/traits/enum_trait.h

@@ -0,0 +1,81 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/meta.h"
+#include "../utility/parameter.h"
+
+#include "../stream/serialize_traits.h"
+
+#include "../traits/integral_traits.h"
+
+namespace bitstream
+{
+	/**
+	 * @brief Wrapper type for compiletime known integer bounds
+	 * @tparam T
+	*/
+	template<typename T, std::underlying_type_t<T> = (std::numeric_limits<T>::min)(), std::underlying_type_t<T> = (std::numeric_limits<T>::max)()>
+	struct bounded_enum;
+
+	/**
+	 * @brief A trait used to serialize an enum type with runtime bounds
+	*/
+	template<typename T>
+	struct serialize_traits<T, typename std::enable_if_t<std::is_enum_v<T>>>
+	{
+        using value_type = std::underlying_type_t<T>;
+
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, T value, value_type min = 0, value_type max = (std::numeric_limits<value_type>::max)()) noexcept
+		{
+			value_type unsigned_value = static_cast<value_type>(value);
+
+			return writer.template serialize<value_type>(unsigned_value, min, max);
+		}
+
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, T& value, value_type min = 0, value_type max = (std::numeric_limits<value_type>::max)()) noexcept
+		{
+			value_type unsigned_value;
+
+			BS_ASSERT(reader.template serialize<value_type>(unsigned_value, min, max));
+
+			value = static_cast<T>(unsigned_value);
+
+			return true;
+		}
+	};
+
+	/**
+	 * @brief A trait used to serialize an enum type with compiletime bounds
+	*/
+	template<typename T, std::underlying_type_t<T> Min, std::underlying_type_t<T> Max>
+	struct serialize_traits<bounded_enum<T, Min, Max>, typename std::enable_if_t<std::is_enum_v<T>>>
+	{
+		using value_type = std::underlying_type_t<T>;
+		using bound_type = bounded_int<value_type, Min, Max>;
+
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, T value) noexcept
+		{
+			value_type unsigned_value = static_cast<value_type>(value);
+
+			return writer.template serialize<bound_type>(unsigned_value);
+		}
+
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, T& value) noexcept
+		{
+			value_type unsigned_value;
+
+			BS_ASSERT(reader.template serialize<bound_type>(unsigned_value));
+
+			value = static_cast<T>(unsigned_value);
+
+			return true;
+		}
+	};
+}

include/bitstream/traits/float_trait.h

@@ -0,0 +1,102 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/meta.h"
+#include "../utility/parameter.h"
+
+#include "../stream/serialize_traits.h"
+
+#include <cstdint>
+#include <cstring>
+
+namespace bitstream
+{
+	/**
+	 * @brief A trait used to serialize a float as-is, without any bound checking or quantization
+	*/
+	template<>
+	struct serialize_traits<float>
+	{
+		/**
+		 * @brief Serializes a whole float into the writer
+		 * @param writer The stream to write to
+		 * @param value The float to serialize
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, in<float> value) noexcept
+		{
+			uint32_t tmp;
+			std::memcpy(&tmp, &value, sizeof(float));
+
+			BS_ASSERT(writer.serialize_bits(tmp, 32));
+
+			return true;
+		}
+
+		/**
+		 * @brief Serializes a whole float from the reader
+		 * @param reader The stream to read from
+		 * @param value The float to serialize to
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, float& value) noexcept
+		{
+			uint32_t tmp;
+
+			BS_ASSERT(reader.serialize_bits(tmp, 32));
+
+			std::memcpy(&value, &tmp, sizeof(float));
+
+			return true;
+		}
+	};
+
+	/**
+	 * @brief A trait used to serialize a double as-is, without any bound checking or quantization
+	*/
+	template<>
+	struct serialize_traits<double>
+	{
+		/**
+		 * @brief Serializes a whole double into the writer
+		 * @param writer The stream to write to
+		 * @param value The double to serialize
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, in<double> value) noexcept
+		{
+			uint32_t tmp[2];
+			std::memcpy(tmp, &value, sizeof(double));
+
+			BS_ASSERT(writer.serialize_bits(tmp[0], 32));
+			BS_ASSERT(writer.serialize_bits(tmp[1], 32));
+
+			return true;
+		}
+
+		/**
+		 * @brief Serializes a whole double from the reader
+		 * @param reader The stream to read from
+		 * @param value The double to serialize to
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, double& value) noexcept
+		{
+			uint32_t tmp[2];
+
+			BS_ASSERT(reader.serialize_bits(tmp[0], 32));
+			BS_ASSERT(reader.serialize_bits(tmp[1], 32));
+
+			std::memcpy(&value, tmp, sizeof(double));
+
+			return true;
+		}
+	};
+}

include/bitstream/traits/integral_traits.h

@@ -0,0 +1,233 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/bits.h"
+#include "../utility/meta.h"
+#include "../utility/parameter.h"
+
+#include "../stream/serialize_traits.h"
+
+#include <cstdint>
+#include <limits>
+#include <type_traits>
+
+namespace bitstream
+{
+	/**
+	 * @brief Wrapper type for compiletime known integer bounds
+	 * @tparam T 
+	*/
+	template<typename T, T = (std::numeric_limits<T>::min)(), T = (std::numeric_limits<T>::max)()>
+	struct bounded_int;
+
+#pragma region const integral types
+	/**
+	 * @brief A trait used to serialize integer values with compiletime bounds
+	 * @tparam T A type matching an integer value
+	 * @tparam Min The lower bound. Inclusive
+	 * @tparam Max The upper bound. Inclusive
+	*/
+	template<typename T, T Min, T Max>
+	struct serialize_traits<bounded_int<T, Min, Max>, typename std::enable_if_t<std::is_integral_v<T> && !std::is_const_v<T>>>
+	{
+		static_assert(sizeof(T) <= 8, "Integers larger than 8 bytes are currently not supported. You will have to write this functionality yourself");
+
+		/**
+		 * @brief Writes an integer into the @p writer
+		 * @param writer The stream to write to
+		 * @param value The value to serialize
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, in<T> value) noexcept
+		{
+			static_assert(Min < Max);
+
+			BS_ASSERT(value >= Min && value <= Max);
+
+			constexpr uint32_t num_bits = utility::bits_in_range(Min, Max);
+
+			static_assert(num_bits <= sizeof(T) * 8);
+
+			if constexpr (sizeof(T) > 4 && num_bits > 32)
+			{
+				// If the given range is bigger than a word (32 bits)
+				uint32_t unsigned_value = static_cast<uint32_t>(value - Min);
+				BS_ASSERT(writer.serialize_bits(unsigned_value, 32));
+
+				unsigned_value = static_cast<uint32_t>((value - Min) >> 32);
+				BS_ASSERT(writer.serialize_bits(unsigned_value, num_bits - 32));
+			}
+			else
+			{
+				// If the given range is smaller than or equal to a word (32 bits)
+				uint32_t unsigned_value = static_cast<uint32_t>(value - Min);
+				BS_ASSERT(writer.serialize_bits(unsigned_value, num_bits));
+			}
+
+			return true;
+		}
+
+		/**
+		 * @brief Reads an integer from the @p writer into @p value
+		 * @param reader The stream to read from
+		 * @param value The value to serialize
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, T& value) noexcept
+		{
+			static_assert(Min < Max);
+
+			constexpr uint32_t num_bits = utility::bits_in_range(Min, Max);
+
+			static_assert(num_bits <= sizeof(T) * 8);
+
+			if constexpr (sizeof(T) > 4 && num_bits > 32)
+			{
+				// If the given range is bigger than a word (32 bits)
+				value = 0;
+				uint32_t unsigned_value;
+
+				BS_ASSERT(reader.serialize_bits(unsigned_value, 32));
+				value |= static_cast<T>(unsigned_value);
+
+				BS_ASSERT(reader.serialize_bits(unsigned_value, num_bits - 32));
+				value |= static_cast<T>(unsigned_value) << 32;
+
+				value += Min;
+			}
+			else
+			{
+				// If the given range is smaller than or equal to a word (32 bits)
+				uint32_t unsigned_value;
+				BS_ASSERT(reader.serialize_bits(unsigned_value, num_bits));
+
+				value = static_cast<T>(unsigned_value) + Min;
+			}
+
+			BS_ASSERT(value >= Min && value <= Max);
+
+			return true;
+		}
+	};
+#pragma endregion
+
+#pragma region integral types
+	/**
+	 * @brief A trait used to serialize integer values with runtime bounds
+	 * @tparam T A type matching an integer value
+	*/
+	template<typename T>
+	struct serialize_traits<T, typename std::enable_if_t<std::is_integral_v<T> && !std::is_const_v<T>>>
+	{
+		static_assert(sizeof(T) <= 8, "Integers larger than 8 bytes are currently not supported. You will have to write this functionality yourself");
+
+		/**
+		 * @brief Writes an integer into the @p writer
+		 * @param writer The stream to write to
+		 * @param value The value to serialize
+		 * @param min The minimum bound that @p value can be. Inclusive
+		 * @param max The maximum bound that @p value can be. Inclusive
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, in<T> value, T min, T max) noexcept
+		{
+			BS_ASSERT(min < max);
+            
+            BS_ASSERT(value >= min && value <= max);
+
+			uint32_t num_bits = utility::bits_in_range(min, max);
+
+			BS_ASSERT(num_bits <= sizeof(T) * 8);
+
+			if constexpr (sizeof(T) > 4)
+			{
+				if (num_bits > 32)
+				{
+					// If the given range is bigger than a word (32 bits)
+					uint32_t unsigned_value = static_cast<uint32_t>(value - min);
+					BS_ASSERT(writer.serialize_bits(unsigned_value, 32));
+
+					unsigned_value = static_cast<uint32_t>((value - min) >> 32);
+					BS_ASSERT(writer.serialize_bits(unsigned_value, num_bits - 32));
+
+					return true;
+				}
+			}
+
+			// If the given range is smaller than or equal to a word (32 bits)
+			uint32_t unsigned_value = static_cast<uint32_t>(value - min);
+			BS_ASSERT(writer.serialize_bits(unsigned_value, num_bits));
+
+			return true;
+		}
+
+		/**
+		 * @brief Reads an integer from the @p reader into @p value
+		 * @param reader The stream to read from
+		 * @param value The value to read into
+		 * @param min The minimum bound that @p value can be. Inclusive
+		 * @param max The maximum bound that @p value can be. Inclusive
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, T& value, T min, T max) noexcept
+		{
+			BS_ASSERT(min < max);
+
+			uint32_t num_bits = utility::bits_in_range(min, max);
+
+			BS_ASSERT(num_bits <= sizeof(T) * 8);
+
+			if constexpr (sizeof(T) > 4)
+			{
+				if (num_bits > 32)
+				{
+					// If the given range is bigger than a word (32 bits)
+					value = 0;
+					uint32_t unsigned_value;
+
+					BS_ASSERT(reader.serialize_bits(unsigned_value, 32));
+					value |= static_cast<T>(unsigned_value);
+
+					BS_ASSERT(reader.serialize_bits(unsigned_value, num_bits - 32));
+					value |= static_cast<T>(unsigned_value) << 32;
+
+					value += min;
+
+					BS_ASSERT(value >= min && value <= max);
+
+					return true;
+				}
+			}
+
+			// If the given range is smaller than or equal to a word (32 bits)
+			uint32_t unsigned_value;
+			BS_ASSERT(reader.serialize_bits(unsigned_value, num_bits));
+
+			value = static_cast<T>(unsigned_value) + min;
+            
+            BS_ASSERT(value >= min && value <= max);
+
+			return true;
+		}
+
+		/**
+		 * @brief Writes or reads an integer into the @p stream
+		 * @param stream The stream to serialize to/from
+		 * @param value The value to serialize
+		 * @return Success
+		*/
+		template<typename Stream, typename U>
+		static bool serialize(Stream& stream, U&& value) noexcept
+		{
+			return serialize_traits<bounded_int<T>>::serialize(stream, std::forward<U>(value));
+		}
+	};
+#pragma endregion
+}

include/bitstream/traits/quantization_traits.h

@@ -0,0 +1,113 @@
+#pragma once
+#include "../quantization/bounded_range.h"
+#include "../quantization/half_precision.h"
+#include "../quantization/smallest_three.h"
+#include "../utility/assert.h"
+#include "../utility/meta.h"
+#include "../utility/parameter.h"
+
+#include "../stream/serialize_traits.h"
+
+#include <cstdint>
+
+namespace bitstream
+{
+	/**
+	 * @brief A trait used to serialize a single-precision float as half-precision
+	*/
+	template<>
+	struct serialize_traits<half_precision>
+	{
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+        static serialize(Stream& stream, in<float> value) noexcept
+		{
+			uint32_t int_value = half_precision::quantize(value);
+            
+			BS_ASSERT(stream.serialize_bits(int_value, 16));
+
+			return true;
+		}
+
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+        static serialize(Stream& stream, out<float> value) noexcept
+		{
+			uint32_t int_value;
+            
+			BS_ASSERT(stream.serialize_bits(int_value, 16));
+            
+			value = half_precision::dequantize(int_value);
+
+			return true;
+		}
+	};
+
+	/**
+	 * @brief A trait used to quantize and serialize a float to be within a given range and precision
+	*/
+	template<>
+	struct serialize_traits<bounded_range>
+	{
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& stream, in<bounded_range> range, in<float> value) noexcept
+		{
+			uint32_t int_value = range.quantize(value);
+            
+			BS_ASSERT(stream.serialize_bits(int_value, range.get_bits_required()));
+
+			return true;
+		}
+
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& stream, in<bounded_range> range, out<float> value) noexcept
+		{
+			uint32_t int_value;
+
+			BS_ASSERT(stream.serialize_bits(int_value, range.get_bits_required()));
+
+			value = range.dequantize(int_value);
+
+			return true;
+		}
+	};
+
+	/**
+	 * @brief A trait used to quantize and serialize quaternions using the smallest-three algorithm
+	*/
+	template<typename Q, size_t BitsPerElement>
+	struct serialize_traits<smallest_three<Q, BitsPerElement>>
+	{
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& stream, in<Q> value) noexcept
+		{
+			quantized_quaternion quantized_quat = smallest_three<Q, BitsPerElement>::quantize(value);
+
+			BS_ASSERT(stream.serialize_bits(quantized_quat.m, 2));
+			BS_ASSERT(stream.serialize_bits(quantized_quat.a, BitsPerElement));
+			BS_ASSERT(stream.serialize_bits(quantized_quat.b, BitsPerElement));
+			BS_ASSERT(stream.serialize_bits(quantized_quat.c, BitsPerElement));
+
+			return true;
+		}
+
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& stream, out<Q> value) noexcept
+		{
+			quantized_quaternion quantized_quat;
+
+			BS_ASSERT(stream.serialize_bits(quantized_quat.m, 2));
+			BS_ASSERT(stream.serialize_bits(quantized_quat.a, BitsPerElement));
+			BS_ASSERT(stream.serialize_bits(quantized_quat.b, BitsPerElement));
+			BS_ASSERT(stream.serialize_bits(quantized_quat.c, BitsPerElement));
+
+			value = smallest_three<Q, BitsPerElement>::dequantize(quantized_quat);
+
+			return true;
+		}
+	};
+}

include/bitstream/traits/string_traits.h

@@ -0,0 +1,344 @@
+#pragma once
+#include "../utility/assert.h"
+#include "../utility/bits.h"
+#include "../utility/meta.h"
+#include "../utility/parameter.h"
+
+#include "../stream/serialize_traits.h"
+
+#include <cstdint>
+#include <string>
+
+namespace bitstream
+{
+	/**
+	 * @brief Wrapper type for compiletime known string max_size
+	*/
+	template<typename T, size_t I>
+	struct bounded_string;
+
+#pragma region const char*
+	/**
+	 * @brief A trait used to serialize bounded c-style strings
+	*/
+	template<>
+	struct serialize_traits<const char*>
+	{
+		/**
+		 * @brief Writes a c-style string into the @p writer
+		 * @param writer The stream to write to
+		 * @param value The string to serialize
+		 * @param max_size The maximum expected length of the string, including the null terminator
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, const char* value, uint32_t max_size) noexcept
+		{
+			uint32_t length = static_cast<uint32_t>(std::char_traits<char>::length(value));
+
+			BS_ASSERT(length < max_size);
+            
+            if (length == 0)
+                return true;
+            
+			uint32_t num_bits = utility::bits_to_represent(max_size);
+            
+			BS_ASSERT(writer.serialize_bits(length, num_bits));
+
+			return writer.serialize_bytes(reinterpret_cast<const uint8_t*>(value), length * 8);
+		}
+
+		/**
+		 * @brief Read a c-style string from the @p reader into @p value
+		 * @param reader The stream to read from
+		 * @param value A pointer to the buffer that should be read into. The size of this buffer should be at least @p max_size
+		 * @param max_size The maximum expected length of the string, including the null terminator
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, char* value, uint32_t max_size) noexcept
+		{
+			uint32_t num_bits = utility::bits_to_represent(max_size);
+
+			uint32_t length;
+			BS_ASSERT(reader.serialize_bits(length, num_bits));
+
+			BS_ASSERT(length < max_size);
+
+			if (length == 0)
+			{
+				value[0] = '\0';
+				return true;
+			}
+
+			BS_ASSERT(reader.serialize_bytes(reinterpret_cast<uint8_t*>(value), length * 8));
+
+			value[length] = '\0';
+
+			return true;
+		}
+	};
+
+	/**
+	 * @brief A trait used to serialize bounded c-style strings with compiletime bounds
+	 * @tparam MaxSize The maximum expected length of the string, including the null terminator
+	*/
+	template<size_t MaxSize>
+	struct serialize_traits<bounded_string<const char*, MaxSize>>
+	{
+		/**
+		 * @brief Writes a c-style string into the @p writer
+		 * @param writer The stream to write to
+		 * @param value The string to serialize
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, const char* value) noexcept
+		{
+			uint32_t length = static_cast<uint32_t>(std::char_traits<char>::length(value));
+
+			BS_ASSERT(length < MaxSize);
+
+			if (length == 0)
+				return true;
+
+			constexpr uint32_t num_bits = utility::bits_to_represent(MaxSize);
+
+			BS_ASSERT(writer.serialize_bits(length, num_bits));
+
+			return writer.serialize_bytes(reinterpret_cast<const uint8_t*>(value), length * 8);
+		}
+
+		/**
+		 * @brief Read a c-style string from the @p reader into @p value
+		 * @param reader The stream to read from
+		 * @param value A pointer to the buffer that should be read into. The size of this buffer should be at least @p max_size
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, char* value) noexcept
+		{
+			constexpr uint32_t num_bits = utility::bits_to_represent(MaxSize);
+
+			uint32_t length;
+			BS_ASSERT(reader.serialize_bits(length, num_bits));
+
+			BS_ASSERT(length < MaxSize);
+
+			if (length == 0)
+			{
+				value[0] = '\0';
+				return true;
+			}
+
+			BS_ASSERT(reader.serialize_bytes(reinterpret_cast<uint8_t*>(value), length * 8));
+
+			value[length] = '\0';
+
+			return true;
+		}
+	};
+#pragma endregion
+
+#ifdef __cpp_char8_t
+	/**
+	 * @brief A trait used to serialize bounded c-style UTF-8 strings
+	*/
+	template<>
+	struct serialize_traits<const char8_t*>
+	{
+		/**
+		 * @brief Writes a c-style UTF-8 string into the @p writer
+		 * @param writer The stream to write to
+		 * @param value The string to serialize
+		 * @param max_size The maximum expected length of the string, including the null terminator
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, const char8_t* value, uint32_t max_size) noexcept
+		{
+			uint32_t length = static_cast<uint32_t>(std::char_traits<char8_t>::length(value));
+
+			BS_ASSERT(length < max_size);
+
+			if (length == 0)
+				return true;
+
+			uint32_t num_bits = utility::bits_to_represent(max_size);
+
+			BS_ASSERT(writer.serialize_bits(length, num_bits));
+
+			return writer.serialize_bytes(reinterpret_cast<const uint8_t*>(value), length * 8);
+		}
+
+		/**
+		 * @brief Read a c-style UTF-8 string from the @p reader into @p value
+		 * @param reader The stream to read from
+		 * @param value A pointer to the buffer that should be read into. The size of this buffer should be at least @p max_size
+		 * @param max_size The maximum expected length of the string, including the null terminator
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, char8_t* value, uint32_t max_size) noexcept
+		{
+			uint32_t num_bits = utility::bits_to_represent(max_size);
+
+			uint32_t length;
+			BS_ASSERT(reader.serialize_bits(length, num_bits));
+
+			BS_ASSERT(length < max_size);
+
+			if (length == 0)
+			{
+				value[0] = '\0';
+				return true;
+			}
+
+			BS_ASSERT(reader.serialize_bytes(reinterpret_cast<uint8_t*>(value), length * 8));
+
+			value[length] = '\0';
+
+			return true;
+		}
+	};
+#endif
+
+#pragma region std::basic_string
+	/**
+	 * @brief A trait used to serialize any combination of std::basic_string
+	 * @tparam T The character type to use
+	 * @tparam Traits The trait type for the T type
+	 * @tparam Alloc The allocator to use
+	*/
+	template<typename T, typename Traits, typename Alloc>
+	struct serialize_traits<std::basic_string<T, Traits, Alloc>>
+	{
+		/**
+		 * @brief Writes a string into the @p writer
+		 * @param writer The stream to write to
+		 * @param value The string to serialize
+		 * @param max_size The maximum expected length of the string, excluding the null terminator
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, in<std::basic_string<T, Traits, Alloc>> value, uint32_t max_size) noexcept
+		{
+			uint32_t length = static_cast<uint32_t>(value.size());
+
+			BS_ASSERT(length <= max_size);
+
+			uint32_t num_bits = utility::bits_to_represent(max_size);
+
+			BS_ASSERT(writer.serialize_bits(length, num_bits));
+
+			if (length == 0)
+				return true;
+
+			return writer.serialize_bytes(reinterpret_cast<const uint8_t*>(value.c_str()), length * sizeof(T) * 8);
+		}
+
+		/**
+		 * @brief Reads a string from the @p reader into @p value
+		 * @param reader The stream to read from
+		 * @param value The string to read into. It will be resized if the read string won't fit
+		 * @param max_size The maximum expected length of the string, excluding the null terminator
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, out<std::basic_string<T, Traits, Alloc>> value, uint32_t max_size)
+		{
+			uint32_t num_bits = utility::bits_to_represent(max_size);
+
+			uint32_t length;
+			BS_ASSERT(reader.serialize_bits(length, num_bits));
+
+			BS_ASSERT(length <= max_size);
+
+			if (length == 0)
+			{
+				value->clear();
+				return true;
+			}
+
+			value->resize(length);
+
+			BS_ASSERT(reader.serialize_bytes(reinterpret_cast<uint8_t*>(value->data()), length * sizeof(T) * 8));
+
+			return true;
+		}
+	};
+
+	/**
+	 * @brief A trait used to serialize any combination of std::basic_string with compiletime bounds
+	 * @tparam T The character type to use
+	 * @tparam Traits The trait type for the T type
+	 * @tparam Alloc The allocator to use
+	 * @tparam MaxSize The maximum expected length of the string, excluding the null terminator
+	*/
+	template<typename T, typename Traits, typename Alloc, size_t MaxSize>
+	struct serialize_traits<bounded_string<std::basic_string<T, Traits, Alloc>, MaxSize>>
+	{
+		/**
+		 * @brief Writes a string into the @p writer
+		 * @param writer The stream to write to
+		 * @param value The string to serialize
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_writing_t<Stream>
+		static serialize(Stream& writer, in<std::basic_string<T, Traits, Alloc>> value) noexcept
+		{
+			uint32_t length = static_cast<uint32_t>(value.size());
+
+			BS_ASSERT(length <= MaxSize);
+
+			constexpr uint32_t num_bits = utility::bits_to_represent(MaxSize);
+
+			BS_ASSERT(writer.serialize_bits(length, num_bits));
+
+			if (length == 0)
+				return true;
+
+			return writer.serialize_bytes(reinterpret_cast<const uint8_t*>(value.c_str()), length * sizeof(T) * 8);
+		}
+
+		/**
+		 * @brief Reads a string from the @p reader into @p value
+		 * @param reader The stream to read from
+		 * @param value The string to read into. It will be resized if the read string won't fit
+		 * @return Success
+		*/
+		template<typename Stream>
+		typename utility::is_reading_t<Stream>
+		static serialize(Stream& reader, out<std::basic_string<T, Traits, Alloc>> value)
+		{
+			constexpr uint32_t num_bits = utility::bits_to_represent(MaxSize);
+
+			uint32_t length;
+			BS_ASSERT(reader.serialize_bits(length, num_bits));
+
+			BS_ASSERT(length <= MaxSize);
+
+			if (length == 0)
+			{
+				value->clear();
+				return true;
+			}
+
+            value->resize(length);
+
+			BS_ASSERT(reader.serialize_bytes(reinterpret_cast<uint8_t*>(value->data()), length * sizeof(T) * 8));
+
+			return true;
+		}
+	};
+#pragma endregion
+}

include/bitstream/utility/assert.h

@@ -0,0 +1,18 @@
+#pragma once
+
+#ifdef BS_DEBUG_BREAK
+#if defined(_WIN32) // Windows
+#define BS_BREAKPOINT() __debugbreak()
+#elif defined(__linux__) // Linux
+#include <csignal>
+#define BS_BREAKPOINT() std::raise(SIGTRAP)
+#else // Non-supported
+#define BS_BREAKPOINT() throw
+#endif
+
+#define BS_ASSERT(...) if (!(__VA_ARGS__)) { BS_BREAKPOINT(); return false; }
+#else // BS_DEBUG_BREAK
+#define BS_ASSERT(...) if (!(__VA_ARGS__)) { return false; }
+
+#define BS_BREAKPOINT() throw
+#endif // BS_DEBUG_BREAK

include/bitstream/utility/bits.h

@@ -0,0 +1,26 @@
+#pragma once
+
+#include <cstddef>
+#include <cstdint>
+
+namespace bitstream::utility
+{
+	constexpr inline uint32_t bits_to_represent(uintmax_t n)
+	{
+		uint32_t r = 0;
+
+		if (n >> 32) { r += 32U; n >>= 32U; }
+		if (n >> 16) { r += 16U; n >>= 16U; }
+		if (n >> 8) { r += 8U; n >>= 8U; }
+		if (n >> 4) { r += 4U; n >>= 4U; }
+		if (n >> 2) { r += 2U; n >>= 2U; }
+		if (n >> 1) { r += 1U; n >>= 1U; }
+
+		return r + static_cast<uint32_t>(n);
+	}
+
+	constexpr inline uint32_t bits_in_range(intmax_t min, intmax_t max)
+	{
+		return bits_to_represent(static_cast<uintmax_t>(max) - static_cast<uintmax_t>(min));
+	}
+}

include/bitstream/utility/crc.h

@@ -0,0 +1,47 @@
+#pragma once
+
+#include <array>
+#include <cstdint>
+
+namespace bitstream::utility
+{
+	inline constexpr auto CHECKSUM_TABLE = []()
+	{
+		constexpr uint32_t POLYNOMIAL = 0xEDB88320;
+
+		std::array<uint32_t, 0x100> table{};
+
+		for (uint32_t i = 0; i < 0x100; ++i)
+		{
+			uint32_t item = i;
+			for (uint32_t bit = 0; bit < 8; ++bit)
+				item = ((item & 1) != 0) ? (POLYNOMIAL ^ (item >> 1)) : (item >> 1);
+			table[i] = item;
+		}
+
+		return table;
+	}();
+
+	inline constexpr uint32_t crc_uint32(const uint8_t* bytes, uint32_t size)
+	{
+		uint32_t result = 0xFFFFFFFF;
+
+		for (uint32_t i = 0; i < size; i++)
+			result = CHECKSUM_TABLE[(result & 0xFF) ^ *(bytes + i)] ^ (result >> 8);
+
+		return ~result;
+	}
+
+	inline constexpr uint32_t crc_uint32(const uint8_t* checksum, const uint8_t* bytes, uint32_t size)
+	{
+		uint32_t result = 0xFFFFFFFF;
+
+		for (uint32_t i = 0; i < 4; i++)
+			result = CHECKSUM_TABLE[(result & 0xFF) ^ *(checksum + i)] ^ (result >> 8);
+
+		for (uint32_t i = 0; i < size; i++)
+			result = CHECKSUM_TABLE[(result & 0xFF) ^ *(bytes + i)] ^ (result >> 8);
+
+		return ~result;
+	}
+}

include/bitstream/utility/endian.h

@@ -0,0 +1,89 @@
+#pragma once
+
+#include <cstdint>
+
+#if defined(__cpp_lib_endian) && __cpp_lib_endian >= 201907L
+#include <bit>
+#else // __cpp_lib_endian
+#ifndef BS_LITTLE_ENDIAN
+// Detect with GCC 4.6's macro.
+#if defined(__BYTE_ORDER__)
+#if (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__)
+#define BS_LITTLE_ENDIAN true
+#elif (__BYTE_ORDER__ == __ORDER_BIG_ENDIAN__)
+#define BS_LITTLE_ENDIAN false
+#else
+#error "Unknown machine byteorder endianness detected. Need to manually define BS_LITTLE_ENDIAN."
+#endif
+// Detect with GLIBC's endian.h.
+#elif defined(__GLIBC__)
+#include <endian.h>
+#if (__BYTE_ORDER == __LITTLE_ENDIAN)
+#define BS_LITTLE_ENDIAN true
+#elif (__BYTE_ORDER == __BIG_ENDIAN)
+#define BS_LITTLE_ENDIAN false
+#else
+#error "Unknown machine byteorder endianness detected. Need to manually define BS_LITTLE_ENDIAN."
+#endif
+// Detect with _LITTLE_ENDIAN and _BIG_ENDIAN macro.
+#elif defined(_LITTLE_ENDIAN) && !defined(_BIG_ENDIAN)
+#define BS_LITTLE_ENDIAN true
+#elif defined(_BIG_ENDIAN) && !defined(_LITTLE_ENDIAN)
+#define BS_LITTLE_ENDIAN false
+// Detect with architecture macros.
+#elif defined(__sparc) || defined(__sparc__) || defined(_POWER) || defined(__powerpc__) || defined(__ppc__) || defined(__hpux) || defined(__hppa) || defined(_MIPSEB) || defined(_POWER) || defined(__s390__)
+#define BS_LITTLE_ENDIAN false
+#elif defined(__i386__) || defined(__alpha__) || defined(__ia64) || defined(__ia64__) || defined(_M_IX86) || defined(_M_IA64) || defined(_M_ALPHA) || defined(__amd64) || defined(__amd64__) || defined(_M_AMD64) || defined(__x86_64) || defined(__x86_64__) || defined(_M_X64) || defined(__bfin__)
+#define BS_LITTLE_ENDIAN true
+#elif defined(_MSC_VER) && (defined(_M_ARM) || defined(_M_ARM64))
+#define BS_LITTLE_ENDIAN true
+#else
+#error "Unknown machine byteorder endianness detected. Need to manually define BS_LITTLE_ENDIAN."
+#endif
+#endif // BS_LITTLE_ENDIAN
+#endif // __cpp_lib_endian
+
+
+#if defined(_WIN32)
+#include <intrin.h>
+#endif
+
+namespace bitstream::utility
+{
+    inline constexpr bool little_endian()
+    {
+#ifdef BS_LITTLE_ENDIAN
+#if BS_LITTLE_ENDIAN
+        return true;
+#else // BS_LITTLE_ENDIAN
+        return false;
+#endif // BS_LITTLE_ENDIAN
+#else // defined(BS_LITTLE_ENDIAN)
+        return std::endian::native == std::endian::little;
+#endif // defined(BS_LITTLE_ENDIAN)
+    }
+
+    inline uint32_t endian_swap32(uint32_t value)
+    {
+#if defined(_WIN32)
+        return _byteswap_ulong(value);
+#elif defined(__linux__)
+        return __builtin_bswap32(value);
+#else
+        const uint32_t first = (value << 24) & 0xFF000000;
+        const uint32_t second = (value << 8) & 0x00FF0000;
+        const uint32_t third = (value >> 8) & 0x0000FF00;
+        const uint32_t fourth = (value >> 24) & 0x000000FF;
+
+        return first | second | third | fourth;
+#endif // _WIN32 || __linux__
+    }
+
+    inline uint32_t to_big_endian32(uint32_t value)
+    {
+        if constexpr (little_endian())
+            return endian_swap32(value);
+        else
+            return value;
+    }
+}

include/bitstream/utility/meta.h

@@ -0,0 +1,105 @@
+#pragma once
+
+#include "../stream/serialize_traits.h"
+
+#include <type_traits>
+
+namespace bitstream::utility
+{
+	// Check if type has a serializable trait
+	template<typename Void, typename T, typename Stream, typename... Args>
+	struct has_serialize : std::false_type {};
+
+	template<typename T, typename Stream, typename... Args>
+	struct has_serialize<std::void_t<decltype(serialize_traits<T>::serialize(std::declval<Stream&>(), std::declval<Args>()...))>, T, Stream, Args...> : std::true_type {};
+
+	template<typename T, typename Stream, typename... Args>
+	using has_serialize_t = std::void_t<decltype(serialize_traits<T>::serialize(std::declval<Stream&>(), std::declval<Args>()...))>;
+
+	template<typename T, typename Stream, typename... Args>
+	constexpr bool has_serialize_v = has_serialize<void, T, Stream, Args...>::value;
+
+
+	// Check if stream is writing or reading
+	template<typename T, typename R = bool>
+	using is_writing_t = std::enable_if_t<T::writing, R>;
+
+	template<typename T, typename R = bool>
+	using is_reading_t = std::enable_if_t<T::reading, R>;
+
+
+	// Check if type is noexcept, if it exists
+	template<typename Void, typename T, typename Stream, typename... Args>
+	struct is_serialize_noexcept : std::false_type {};
+
+	template<typename T, typename Stream, typename... Args>
+	struct is_serialize_noexcept<std::enable_if_t<has_serialize_v<T, Stream, Args...>>, T, Stream, Args...> :
+		std::bool_constant<noexcept(serialize_traits<T>::serialize(std::declval<Stream&>(), std::declval<Args>()...))> {};
+
+	template<typename T, typename Stream, typename... Args>
+	constexpr bool is_serialize_noexcept_v = is_serialize_noexcept<void, T, Stream, Args...>::value;
+
+
+	// Get the underlying type without &, &&, * or const
+	template<typename T>
+	using base_t = typename std::remove_const_t<std::remove_pointer_t<std::decay_t<T>>>;
+
+
+	// Meta functions for guessing the trait type from the first argument
+	template<typename Void, typename Trait, typename Stream, typename... Args>
+	struct deduce_trait
+	{
+		using type = Trait;
+	};
+
+	// Non-const value
+	template<typename Trait, typename Stream, typename... Args>
+	struct deduce_trait<std::enable_if_t<
+		!std::is_pointer_v<std::decay_t<Trait>> &&
+		has_serialize_v<base_t<Trait>, Stream, Trait, Args...>>,
+		Trait, Stream, Args...>
+	{
+		using type = base_t<Trait>;
+	};
+
+	// Const value
+	template<typename Trait, typename Stream, typename... Args>
+	struct deduce_trait<std::enable_if_t<
+		!std::is_pointer_v<std::decay_t<Trait>> &&
+		has_serialize_v<std::add_const_t<base_t<Trait>>, Stream, Trait, Args...>>,
+		Trait, Stream, Args...>
+	{
+		using type = std::add_const_t<base_t<Trait>>;
+	};
+
+	// Non-const pointer
+	template<typename Trait, typename Stream, typename... Args>
+	struct deduce_trait<std::enable_if_t<
+		std::is_pointer_v<std::decay_t<Trait>> &&
+		has_serialize_v<std::add_pointer_t<base_t<Trait>>, Stream, Trait, Args...>>,
+		Trait, Stream, Args...>
+	{
+		using type = std::add_pointer_t<base_t<Trait>>;
+	};
+
+	// Const pointer
+	template<typename Trait, typename Stream, typename... Args>
+	struct deduce_trait<std::enable_if_t<
+		std::is_pointer_v<std::decay_t<Trait>> &&
+		has_serialize_v<std::add_pointer_t<std::add_const_t<base_t<Trait>>>, Stream, Trait, Args...>>,
+		Trait, Stream, Args...>
+	{
+		using type = std::add_pointer_t<std::add_const_t<base_t<Trait>>>;
+	};
+
+	template<typename Trait, typename Stream, typename... Args>
+	using deduce_trait_t = typename deduce_trait<void, Trait, Stream, Args...>::type;
+
+
+	// Shorthands for deduced type_traits
+	template<typename Trait, typename Stream, typename... Args>
+	using has_deduce_serialize_t = has_serialize_t<deduce_trait_t<Trait, Stream, Args...>, Stream, Trait, Args...>;
+
+	template<typename Trait, typename Stream, typename... Args>
+	constexpr bool is_deduce_serialize_noexcept_v = is_serialize_noexcept_v<deduce_trait_t<Trait, Stream, Args...>, Stream, Trait, Args...>;
+}

include/bitstream/utility/parameter.h

@@ -0,0 +1,124 @@
+#pragma once
+
+#include "assert.h"
+
+#include <utility>
+#include <type_traits>
+
+#ifdef __cpp_constexpr_dynamic_alloc
+#define BS_CONSTEXPR constexpr
+#else // __cpp_constexpr_dynamic_alloc
+#define BS_CONSTEXPR
+#endif // __cpp_constexpr_dynamic_alloc
+
+namespace bitstream
+{
+#ifdef BS_DEBUG_BREAK
+	template<typename T>
+	class out
+	{
+	public:
+		BS_CONSTEXPR out(T& value) noexcept :
+			m_Value(value),
+			m_Constructed(false) {}
+
+		out(const out&) = delete;
+
+		out(out&&) = delete;
+
+		BS_CONSTEXPR ~out()
+		{
+			if (!m_Constructed)
+				BS_BREAKPOINT();
+		}
+
+		template<typename U, typename = std::enable_if_t<std::is_assignable_v<T&, U>>>
+		BS_CONSTEXPR out& operator=(U&& arg) noexcept(std::is_nothrow_assignable_v<T&, U>)
+		{
+			m_Value = std::forward<U>(arg);
+
+			m_Constructed = true;
+
+			return *this;
+		}
+
+		BS_CONSTEXPR T* operator->() noexcept
+		{
+			m_Constructed = true;
+			return &m_Value;
+		}
+
+		BS_CONSTEXPR T& operator*() noexcept
+		{
+			m_Constructed = true;
+			return m_Value;
+		}
+
+	private:
+		T& m_Value;
+		bool m_Constructed;
+	};
+#else
+	template<typename T>
+	class out
+	{
+	public:
+		BS_CONSTEXPR out(T& value) noexcept :
+			m_Value(value) {}
+
+		out(const out&) = delete;
+
+		out(out&&) = delete;
+
+		template<typename U, typename = std::enable_if_t<std::is_assignable_v<T&, U>>>
+		BS_CONSTEXPR out& operator=(U&& arg) noexcept(std::is_nothrow_assignable_v<T&, U>)
+		{
+			m_Value = std::forward<U>(arg);
+
+			return *this;
+		}
+
+		BS_CONSTEXPR T* operator->() noexcept { return &m_Value; }
+
+		BS_CONSTEXPR T& operator*() noexcept { return m_Value; }
+
+	private:
+		T& m_Value;
+	};
+#endif
+
+	/**
+	 * @brief Passes by const or const reference depending on size
+	*/
+	template<typename T>
+	using in = std::conditional_t<(sizeof(T) <= 16 && std::is_trivially_copy_constructible_v<T>), std::add_const_t<T>, std::add_lvalue_reference_t<std::add_const_t<T>>>;
+
+	/**
+	 * @brief Passes by reference
+	*/
+	template<typename Stream, typename T>
+	using inout = std::conditional_t<Stream::writing, in<T>, std::add_lvalue_reference_t<T>>;
+
+
+	/**
+	 * @brief Test type
+	*/
+	template<typename Lambda>
+	class finally
+	{
+	public:
+		constexpr finally(Lambda func) noexcept :
+			m_Lambda(func) {}
+
+		~finally()
+		{
+			m_Lambda();
+		}
+
+	private:
+		Lambda m_Lambda;
+	};
+
+	template<typename Lambda>
+	finally(Lambda func) -> finally<Lambda>;
+}

include/encoder/FECEncoder.h

@@ -5,7 +5,7 @@
 #include <stdexcept>
 #include <vector>
 
-#include "bitstream.h"
+#include "bitstream/bitstream.h"
 
 /**
  * @class FECEncoder
@@ -30,26 +30,30 @@ public:
      * @param data The input BitStream to be encoded.
      * @return The encoded BitStream.
      */
-    BitStream encode(const BitStream& data) {
-        BitStream input_data(data);
-        BitStream output_data;
+    void encode(bitstream::growing_bit_writer<std::vector<uint8_t>>& output_data, bitstream::fixed_bit_reader& input_data) {
+        std::vector<uint8_t> intermediate_buffer;
+        bitstream::growing_bit_writer<std::vector<uint8_t>> intermediate_data(intermediate_buffer);
+
+        while (input_data.get_remaining_bits() > 0) {
+            uint32_t bit;
+            input_data.serialize_bits(bit, 1);
 
-        while (input_data.hasNext()) {
-            uint8_t bit = input_data.getNextBit();
             // Shift the input bit into the shift register
             shift_register = ((shift_register << 1) | bit) & 0x7F;
 
             // Calculate T1 and T2 using the generator polynomials
-            uint8_t t1 = calculateT1();
-            uint8_t t2 = calculateT2();
+            uint32_t t1 = calculateT1();
+            uint32_t t2 = calculateT2();
 
             // Append T1 and T2 to the encoded data
-            output_data.putBit(t1);
-            output_data.putBit(t2);
+            intermediate_data.serialize_bits(t1, 1);
+            intermediate_data.serialize_bits(t2, 1);
         }
 
+        bitstream::fixed_bit_reader intermediate_reader(intermediate_buffer.data(), intermediate_data.get_num_bits_serialized());
+
         // Apply repetition or puncturing based on baud rate and operation mode
-        return adjustRate(output_data);
+        return adjustRate(output_data, intermediate_reader);
     }
 
 private:
@@ -62,7 +66,9 @@ private:
      * @return The calculated T1 bit.
      */
     uint8_t calculateT1() {
-        return (shift_register >> 6) ^ ((shift_register >> 4) & 0x01) ^ ((shift_register >> 3) & 0x01) ^ ((shift_register >> 1) & 0x01) ^ (shift_register & 0x01);
+        return (shift_register >> 6) ^ ((shift_register >> 4) & 0x01) ^
+               ((shift_register >> 3) & 0x01) ^ ((shift_register >> 1) & 0x01) ^
+               (shift_register & 0x01);
     }
 
     /**
@@ -70,7 +76,9 @@ private:
      * @return The calculated T2 bit.
      */
     uint8_t calculateT2() {
-        return (shift_register >> 6) ^ ((shift_register >> 5) & 0x01) ^ ((shift_register >> 4) & 0x01) ^ ((shift_register >> 1) & 0x01) ^ (shift_register & 0x01);
+        return (shift_register >> 6) ^ ((shift_register >> 5) & 0x01) ^
+               ((shift_register >> 4) & 0x01) ^ ((shift_register >> 1) & 0x01) ^
+               (shift_register & 0x01);
     }
 
     /**
@@ -78,32 +86,46 @@ private:
      * @param encoded_data The encoded BitStream to be adjusted.
      * @return The adjusted BitStream.
      */
-    BitStream adjustRate(const BitStream& encoded_data) {
-        BitStream adjusted_data;
+    void adjustRate(bitstream::growing_bit_writer<std::vector<uint8_t>>& adjusted_data, bitstream::fixed_bit_reader& encoded_data) {
+        size_t repetition_factor = getRepetitionFactor();
 
-        if ((baud_rate == 300 || baud_rate == 150 || baud_rate == 75) && is_frequency_hopping) {
-            // Repetition for frequency-hopping operation at lower baud rates
-            size_t repetition_factor = (baud_rate == 300) ? 2 : (baud_rate == 150) ? 4 : 8;
-            for (size_t i = 0; i < encoded_data.getMaxBitIndex(); i += 2) {
-                for (size_t j = 0; j < repetition_factor; j++) {
-                    adjusted_data.putBit(encoded_data.getBitVal(i));
-                    adjusted_data.putBit(encoded_data.getBitVal(i + 1));
-                }
+        if (repetition_factor == 1) {
+            while (encoded_data.get_remaining_bits() > 0) {
+                uint32_t bit;
+                encoded_data.serialize_bits(bit, 1);
+                adjusted_data.serialize_bits(bit, 1);
             }
-        } else if ((baud_rate == 300 || baud_rate == 150) && !is_frequency_hopping) {
-            // Repetition for fixed-frequency operation at lower baud rates
-            size_t repetition_factor = (baud_rate == 300) ? 2 : 4;
-            for (size_t i = 0; i < encoded_data.getMaxBitIndex(); i += 2) {
-                for (size_t j = 0; j < repetition_factor; j++) {
-                    adjusted_data.putBit(encoded_data.getBitVal(i));
-                    adjusted_data.putBit(encoded_data.getBitVal(i + 1));
-                }
-            }
-        } else {
-            adjusted_data = encoded_data;
+            return;
         }
 
-        return adjusted_data;
+        while (encoded_data.get_remaining_bits() >= 2) {
+            uint32_t t1, t2;
+            encoded_data.serialize_bits(t1, 1);
+            encoded_data.serialize_bits(t2, 1);
+            
+            for (size_t j = 0; j < repetition_factor; j++) {
+                adjusted_data.serialize_bits(t1, 1);
+                adjusted_data.serialize_bits(t2, 1);
+                
+            }
+        }
+    }
+
+    size_t getRepetitionFactor() const {
+        if (is_frequency_hopping) {
+            switch (baud_rate) {
+                case 300: return 2;
+                case 150: return 4;
+                case 75: return 8;
+                default: return 1;
+            }
+        } else {
+            switch (baud_rate) {
+                case 300: return 2;
+                case 150: return 4;
+                default: return 1;
+            }
+        }
     }
 };
 

include/encoder/Interleaver.h

@@ -6,7 +6,7 @@
 #include <stdexcept>
 #include <vector>
 
-#include "bitstream.h"
+#include "bitstream/bitstream.h"
 
 /**
  * @class Interleaver
@@ -34,30 +34,43 @@ public:
      * @param input_data The input BitStream to be interleaved.
      * @return A new BitStream containing the interleaved data.
      */
-    std::vector<uint8_t> interleaveStream(const BitStream& input_data) {
-        BitStream data = input_data;
-        BitStream interleaved_data;
+    std::vector<uint8_t> interleaveStream(bitstream::fixed_bit_reader& input_data) {
+        std::vector<uint8_t> interleaved_buffer;
+
+        bitstream::growing_bit_writer<std::vector<uint8_t>> interleaved_data(interleaved_buffer);
+
         size_t chunk_size = rows * columns;
         size_t input_index = 0;
 
-        while (input_index < data.getMaxBitIndex()) {
-            size_t end_index = std::min(input_index + chunk_size, data.getMaxBitIndex());
-            BitStream chunk = data.getSubStream(input_index, end_index);
-            if (chunk.getMaxBitIndex() > rows * columns) {
-                throw std::invalid_argument("Input data exceeds interleaver matrix size in loadChunk()");
+        std::vector<uint8_t> chunk_data((chunk_size + 7) / 8, 0);
+
+        while (input_data.get_remaining_bits() >= chunk_size) {
+            std::fill(chunk_data.begin(), chunk_data.end(), 0);
+
+            if (!input_data.serialize_bytes(chunk_data.data(), chunk_size)) {
+                throw std::runtime_error("Failed to serialize chunk from input data");
             }
-            BitStream interleaved_chunk = interleaveChunk(chunk);
-            interleaved_data += interleaved_chunk;
-            input_index = end_index;
+
+            bitstream::fixed_bit_reader chunk_reader(chunk_data.data(), chunk_size);
+
+            interleaveChunk(interleaved_data, chunk_reader);
         }
 
         // Apply puncturing for 2400 bps in frequency-hopping mode (Rate 2/3)
         if (baud_rate == 2400 && is_frequency_hopping) {
-            return applyPuncturing(interleaved_data);
+            std::vector<uint8_t> punctured_buffer;
+
+            bitstream::growing_bit_writer<std::vector<uint8_t>> punctured_writer(punctured_buffer);
+            bitstream::fixed_bit_reader interleaved_reader(interleaved_buffer.data(), interleaved_buffer.size() * 8);
+
+            applyPuncturing(punctured_writer, interleaved_reader);
+
+            interleaved_buffer = punctured_buffer;
         }
 
-        std::vector<uint8_t> final_interleaved_data = groupSymbols(interleaved_data);
-        return final_interleaved_data;
+        bitstream::fixed_bit_reader final_reader(interleaved_buffer.data(), interleaved_buffer.size() * 8);
+
+        return groupSymbols(final_reader);
     }
 
     /**
@@ -65,7 +78,7 @@ public:
      * @return The number of bits needed for a complete flush.
      */
     size_t getFlushBits() const {
-        return rows * columns;
+        return (interleave_setting == 0) ? 0 : (rows * columns);
     }
 
 private:
@@ -84,21 +97,16 @@ private:
      * @param input_data The input BitStream to be grouped into symbols.
      * @return A vector of grouped symbols.
      */
-    std::vector<uint8_t> groupSymbols(BitStream& input_data) {
+    std::vector<uint8_t> groupSymbols(bitstream::fixed_bit_reader& input_data) {
         std::vector<uint8_t> grouped_data;
-        size_t max_index = input_data.getMaxBitIndex();
         size_t bits_per_symbol = (baud_rate == 2400) ? 3 : (baud_rate == 1200 || (baud_rate == 75 && !is_frequency_hopping)) ? 2 : 1;
-        size_t current_index = 0;
+        
+        while (input_data.get_remaining_bits() >= bits_per_symbol) {
+            uint32_t symbol = 0;
 
-        while ((current_index + bits_per_symbol) < max_index) {
-            uint8_t symbol = 0;
+            input_data.serialize_bits(symbol, bits_per_symbol);
 
-            for (int i = 0; i < bits_per_symbol; i++) {
-                symbol = (symbol << 1) | input_data.getBitVal(current_index + i);
-            }
-
-            grouped_data.push_back(symbol);
-            current_index += bits_per_symbol;
+            grouped_data.push_back(static_cast<uint8_t>(symbol));
         }
 
         return grouped_data;
@@ -109,16 +117,16 @@ private:
      * @param input_data The input BitStream chunk.
      * @return A BitStream representing the interleaved chunk.
      */
-    BitStream interleaveChunk(const BitStream& input_data) {
+    void interleaveChunk(bitstream::growing_bit_writer<std::vector<uint8_t>>& interleaved_writer, bitstream::fixed_bit_reader& input_data) {
         loadChunk(input_data);
-        return fetchChunk();
+        return fetchChunk(interleaved_writer);
     }
 
     /**
      * @brief Loads bits from the input BitStream into the interleaver matrix.
      * @param data The input BitStream to load.
      */
-    void loadChunk(const BitStream& data) {
+    void loadChunk(bitstream::fixed_bit_reader& data) {
         size_t row = 0;
         size_t col = 0;
         size_t index = 0;
@@ -127,13 +135,19 @@ private:
 
         // Load bits into the matrix
         std::fill(matrix.begin(), matrix.end(), 0);  // Clear previous data
-        while (index < data.getMaxBitIndex() && col < columns) {
+
+        while (data.get_remaining_bits() > 0 && col < columns) {
             size_t matrix_idx = row * columns + col;
             if (matrix_idx >= matrix.size()) {
                 throw std::out_of_range("Matrix index out of bounds in loadChunk()");
             }
 
-            matrix[matrix_idx] = data.getBitVal(index++);
+            uint32_t bit = 0;
+            if (!data.serialize_bits(bit, 1)) {
+                throw std::runtime_error("Failed to read bit from chunk_reader in loadChunk()");
+            }
+            matrix[matrix_idx] = static_cast<uint8_t>(bit);
+
             row = (row + row_increment) % rows;
 
             if (row == 0) {
@@ -146,21 +160,23 @@ private:
      * @brief Fetches bits from the interleaver matrix in the interleaved order.
      * @return A BitStream containing the fetched interleaved data.
      */
-    BitStream fetchChunk() {
-        BitStream fetched_data;
+    void fetchChunk(bitstream::growing_bit_writer<std::vector<uint8_t>>& interleaved_writer) {
         size_t row = 0;
         size_t col = 0;
 
         size_t column_decrement = (baud_rate == 75 && interleave_setting == 2) ? 7 : 17;
 
-        // Fetch bits from the matrix
-        while (fetched_data.getMaxBitIndex() < rows * columns) {
+        for (size_t i = 0; i < rows * columns; i++) {
             size_t matrix_idx = row * columns + col;
             if (matrix_idx >= matrix.size()) {
                 throw std::out_of_range("Matrix index out of bounds in fetchChunk()");
             }
 
-            fetched_data.putBit(matrix[matrix_idx]);
+            uint32_t bit = static_cast<uint32_t>(matrix[matrix_idx]);
+            if (!interleaved_writer.serialize_bits(bit, 1)) {
+                throw std::runtime_error("Failed to write bit to interleaved_writer in fetchChunk()");
+            }
+
             row++;
 
             if (row == rows) {
@@ -170,8 +186,6 @@ private:
                 col = (col + columns - column_decrement) % columns;
             }
         }
-
-        return fetched_data;
     }
 
 
@@ -210,14 +224,18 @@ private:
      * @param interleaved_data The interleaved data to be punctured.
      * @return A BitStream containing punctured data.
      */
-    BitStream applyPuncturing(const BitStream& interleaved_data) {
-        BitStream punctured_data;
-        for (size_t i = 0; i < interleaved_data.getMaxBitIndex(); i++) {
-            if ((i % 4) != 1) { // Skip every fourth bit (the second value of T2)
-                punctured_data.putBit(interleaved_data.getBitVal(i));
+    void applyPuncturing(bitstream::growing_bit_writer<std::vector<uint8_t>>& punctured_data, bitstream::fixed_bit_reader& interleaved_data) {
+        size_t bit_index = 0;
+
+        while (interleaved_data.get_remaining_bits() > 0) {
+            uint32_t bit = 0;
+            interleaved_data.serialize_bits(bit, 1);
+
+            if ((bit_index % 4) != 3) {
+                punctured_data.serialize_bits(bit, 1);
             }
+            bit_index++;
         }
-        return punctured_data;
     }
 };
 

include/encoder/ModemController.h

@@ -6,7 +6,7 @@
 #include <memory>
 #include <vector>
 
-#include "bitstream.h"
+#include "bitstream/bitstream.h"
 #include "FECEncoder.h"
 #include "Interleaver.h"
 #include "MGDDecoder.h"
@@ -40,7 +40,7 @@ public:
      * @param data The input data stream to be transmitted. The `is_voice` parameter controls whether the modem treats it as binary file data,
      *             or a binary stream from the MELPe (or other) voice codec.
      */
-    ModemController(const size_t _baud_rate, const bool _is_voice, const bool _is_frequency_hopping, const size_t _interleave_setting, BitStream _data)
+    ModemController(const size_t _baud_rate, const bool _is_voice, const bool _is_frequency_hopping, const size_t _interleave_setting)
         : baud_rate(_baud_rate),
           is_voice(_is_voice),
           is_frequency_hopping(_is_frequency_hopping),
@@ -49,7 +49,6 @@ public:
           scrambler(),
           fec_encoder(baud_rate, is_frequency_hopping),
           interleaver(baud_rate, interleave_setting, is_frequency_hopping),
-          input_data(std::move(_data)),
           mgd_decoder(baud_rate, is_frequency_hopping),
           modulator(baud_rate, 48000, 0.5, is_frequency_hopping) {}
 
@@ -58,36 +57,47 @@ public:
      * @return The scrambled data ready for modulation.
      * @note The modulated signal is generated internally but is intended to be handled externally.
      */
-    std::vector<int16_t> transmit() {
-        // Step 1: Append EOM Symbols
-        BitStream eom_appended_data = appendEOMSymbols(input_data);
+    std::vector<int16_t> transmit(bitstream::fixed_bit_reader& input_data) {
+        // Step 1: Append EOM Symbols using a uint32_t aligned output buffer
+        std::vector<uint8_t> output_buffer;
+        bitstream::growing_bit_writer<std::vector<uint8_t>> output_writer(output_buffer);
+        appendEOMSymbols(output_writer, input_data);
 
+        // Step 2: Handle Baud Rate Specific Encoding
         std::vector<uint8_t> processed_data;
         if (baud_rate == 4800) {
-            processed_data = splitTribitSymbols(eom_appended_data);
+            // For 4800 baud, perform tribit symbol splitting
+            bitstream::fixed_bit_reader eom_appended_reader(output_buffer.data(), output_writer.get_num_bits_serialized());
+            processed_data = splitTribitSymbols(eom_appended_reader);
         } else {
-            // Step 2: FEC Encoding
-            BitStream fec_encoded_data = fec_encoder.encode(eom_appended_data);
+            // Step 3: FEC Encoding
+            bitstream::fixed_bit_reader eom_appended_reader(output_buffer.data(), output_writer.get_num_bits_serialized());
+            std::vector<uint8_t> fec_encoded_buffer;
+            bitstream::growing_bit_writer<std::vector<uint8_t>> fec_encoded_writer(fec_encoded_buffer);
+            fec_encoder.encode(fec_encoded_writer, eom_appended_reader);
 
-            // Step 3: Interleaving
-            processed_data = interleaver.interleaveStream(fec_encoded_data);
+            // Step 4: Interleaving
+            bitstream::fixed_bit_reader fec_encoded_reader(fec_encoded_buffer.data(), fec_encoded_writer.get_num_bits_serialized());
+            processed_data = interleaver.interleaveStream(fec_encoded_reader);
         }
 
+        // Step 5: MGD Decoding
         std::vector<uint8_t> mgd_decoded_data = mgd_decoder.mgdDecode(processed_data);
 
-        // Step 4: Symbol Formation. This function injects the sync preamble symbols. Scrambling is built-in.
+        // Step 6: Symbol Formation (including sync preamble and scrambling)
         std::vector<uint8_t> symbol_stream = symbol_formation.formSymbols(mgd_decoded_data);
 
+        // Step 7: Modulation
         std::vector<int16_t> modulated_signal = modulator.modulate(symbol_stream);
 
         return modulated_signal;
     }
 
+
 private:
     size_t baud_rate;                ///< The baud rate for the modem.
     bool is_voice;                   ///< Indicates if the data being transmitted is voice.
     bool is_frequency_hopping;       ///< Indicates if frequency hopping is used.
-    BitStream input_data;            ///< The input data stream.
     size_t interleave_setting;       ///< The interleave setting to be used.
     size_t sample_rate;
 
@@ -106,38 +116,38 @@ private:
      *          the FEC encoder and interleaver matrices. The function calculates the number of flush bits required
      *          based on the FEC and interleaver settings.
      */
-    BitStream appendEOMSymbols(const BitStream& input_data) const {
-        BitStream eom_data = input_data;
+    void appendEOMSymbols(bitstream::growing_bit_writer<std::vector<uint8_t>>& output_data, bitstream::fixed_bit_reader& input_data) const {
+        while (input_data.get_num_bits_serialized() < input_data.get_total_bits()) {
+            uint32_t value;
+            uint32_t bits_to_read = std::min(32U, input_data.get_remaining_bits());
+            input_data.serialize_bits(value, bits_to_read);
+            output_data.serialize_bits(value, bits_to_read);
+        }
+
         // Append the EOM sequence (4B65A5B2 in hexadecimal)
-        BitStream eom_sequence({0x4B, 0x65, 0xA5, 0xB2}, 32);
-        eom_data += eom_sequence;
+        uint32_t eom_sequence = 0x4B65A5B2;
+        output_data.serialize_bits(eom_sequence, 32);
 
         // Append additional zeros to flush the FEC encoder and interleaver
         size_t fec_flush_bits = 144; // FEC encoder flush bits
         size_t interleave_flush_bits = interleaver.getFlushBits();
         size_t total_flush_bits = fec_flush_bits + ((interleave_setting == 0) ? 0 : interleave_flush_bits);
-        while ((eom_data.getMaxBitIndex() + total_flush_bits) % interleave_flush_bits)
-            total_flush_bits++;
         
-        size_t total_bytes = (total_flush_bits + 7) / 8; // Round up to ensure we have enough bytes to handle all bits.
-        BitStream flush_bits(std::vector<uint8_t>(total_bytes, 0), total_flush_bits);
-        eom_data += flush_bits;
-
-        return eom_data;
+        size_t current_bit_index = output_data.get_num_bits_serialized();
+        size_t alignment_bits_needed = (interleave_flush_bits - (current_bit_index + fec_flush_bits) % interleave_flush_bits) % interleave_flush_bits;
+        total_flush_bits += alignment_bits_needed;
     }
 
-    std::vector<uint8_t> splitTribitSymbols(const BitStream& input_data) {
+    std::vector<uint8_t> splitTribitSymbols(bitstream::fixed_bit_reader& input_data) {
         std::vector<uint8_t> return_vector;
-        size_t max_index = input_data.getMaxBitIndex();
-        size_t current_index = 0;
+        size_t total_bits = input_data.get_total_bits();
+        size_t num_bits_serialized = input_data.get_num_bits_serialized();
 
-        while (current_index + 2 < max_index) {
-            uint8_t symbol = 0;
-            for (int i = 0; i < 3; i++) {
-                symbol = (symbol << 1) | input_data.getBitVal(current_index + i);
-            }
-            return_vector.push_back(symbol);
-            current_index += 3;
+        while (num_bits_serialized + 3 <= total_bits) {
+            uint32_t symbol = 0;
+            input_data.serialize_bits(symbol, 3);
+            return_vector.push_back(static_cast<uint8_t>(symbol));
+            num_bits_serialized = input_data.get_num_bits_serialized();
         }
 
         return return_vector;

include/utils/bitstream.h

@@ -1,220 +0,0 @@
-#ifndef BITSTREAM_H
-#define BITSTREAM_H
-
-#include <algorithm>
-#include <cstdint>
-#include <stdexcept>
-#include <vector>
-
-/**
- * @class BitStream
- * @brief A class to represent a stream of bits with bit-level read and write access.
- *
- * The BitStream class provides functionality to manipulate a byte stream at the bit level.
- * It derives from std::vector<uint8_t> to utilize the benefits of a byte vector while providing
- * additional methods for bit manipulation.
- */
-class BitStream : public std::vector<uint8_t> {
-public:
-    /**
-     * @brief Default constructor.
-     */
-    BitStream() : bit_index(0), max_bit_idx(0) {}
-
-    /**
-     * @brief Constructs a BitStream from an existing vector of bytes.
-     * @param data The byte stream to be used for initializing the BitStream.
-     */
-    BitStream(const std::vector<uint8_t>& data) : std::vector<uint8_t>(data), bit_index(0), max_bit_idx(data.size() * 8) {}
-
-    /**
-     * @brief Constructs a BitStream from an existing vector of bytes with a specified bit size.
-     * @param data The byte stream to be used for initializing the BitStream.
-     * @param size_in_bits The number of bits to consider in the stream.
-     */
-    BitStream(const std::vector<uint8_t>& data, size_t size_in_bits) : std::vector<uint8_t>(data), bit_index(0), max_bit_idx(size_in_bits) {}
-
-    /**
-     * @brief Copy constructor from another BitStream.
-     * @param data The BitStream to copy from.
-     */
-    BitStream(const BitStream& data) : std::vector<uint8_t>(data), bit_index(0), max_bit_idx(data.max_bit_idx) {}
-
-    /**
-     * @brief Constructs a BitStream from a substream of another BitStream.
-     * @param other The original BitStream.
-     * @param start_bit The starting bit index of the substream.
-     * @param end_bit The ending bit index of the substream (exclusive).
-     * @throws std::out_of_range if start or end indices are out of bounds.
-     */
-    BitStream(const BitStream& other, size_t start_bit, size_t end_bit) : bit_index(0) {
-        if (start_bit >= other.max_bit_idx || end_bit > other.max_bit_idx || start_bit > end_bit) {
-            throw std::out_of_range("BitStream substream indices are out of range.");
-        }
-        max_bit_idx = end_bit - start_bit;
-        for (size_t i = start_bit; i < end_bit; i++) {
-            putBit(other.getBitVal(i));
-        }
-    }
-
-    /**
-     * @brief Reads the next bit from the stream.
-     * @return The next bit (0 or 1).
-     * @throws std::out_of_range if no more bits are available in the stream.
-     */
-    int getNextBit() {
-        if (bit_index >= max_bit_idx) {
-            throw std::out_of_range("No more bits available in the stream.");
-        }
-
-        int bit = getBitVal(bit_index++);
-        return bit;
-    }
-
-    /**
-     * @brief Gets the value of a bit at a specific index.
-     * @param idx The index of the bit to be retrieved.
-     * @return The value of the bit (0 or 1).
-     * @throws std::out_of_range if the bit index is out of range.
-     */
-    int getBitVal(const size_t idx) const {
-        if (idx >= max_bit_idx) {
-            throw std::out_of_range("Bit index out of range in getBitVal.");
-        }
-
-        size_t byte_idx = idx / 8;
-        size_t bit_idx = idx % 8;
-
-        uint8_t tmp = this->at(byte_idx);
-        uint8_t mask = 0x80 >> bit_idx;
-        uint8_t result = tmp & mask;
-
-        return result ? 1 : 0;
-    }
-
-    /**
-     * @brief Checks if there are more bits available in the stream.
-     * @return True if there are more bits available, otherwise false.
-     */
-    bool hasNext() const {
-        return bit_index < max_bit_idx;
-    }
-
-    /**
-     * @brief Sets a specific bit value in the stream.
-     * @param idx The index of the bit to set.
-     * @param val The value to set the bit to (0 or 1).
-     *
-     * This function ensures that the stream has enough bytes to accommodate
-     * the given bit index. If the bit index is out of bounds, the stream is
-     * resized accordingly.
-     */
-    void setBitVal(const size_t idx, uint8_t val) {
-        size_t byte_idx = idx / 8;
-        size_t bit_idx = idx % 8;
-        uint8_t mask = 0x80 >> bit_idx;
-
-        if (byte_idx >= this->size()) {
-            this->resize(byte_idx + 1, 0);
-        }
-
-        if (val == 0) {
-            this->at(byte_idx) &= ~mask;
-        } else {
-            this->at(byte_idx) |= mask;
-        }
-    }
-
-    /**
-     * @brief Appends a bit to the end of the stream.
-     * @param bit The value of the bit to append (0 or 1).
-     *
-     * This function keeps track of the current bit index and appends bits
-     * sequentially. If necessary, the stream is resized to accommodate the new bit.
-     */
-    void putBit(uint8_t bit) {
-        size_t byte_idx = max_bit_idx / 8;
-        if (byte_idx >= this->size()) {
-            this->push_back(0);
-        }
-        size_t bit_idx = max_bit_idx % 8;
-        setBitVal(max_bit_idx, bit);
-        max_bit_idx += 1;
-    }
-
-    /**
-     * @brief Resets the bit index to the beginning of the stream.
-     */
-    void resetBitIndex() {
-        bit_index = 0;
-    }
-
-    /**
-     * @brief Returns the maximum bit index value (total number of bits in the stream).
-     * @return The total number of bits in the stream.
-     */
-    size_t getMaxBitIndex() const {
-        return max_bit_idx;
-    }
-
-    BitStream& operator=(const BitStream& other) {
-        this->clear();
-        this->resize(other.size());
-        std::copy(other.begin(), other.end(), this->begin());
-        this->bit_index = other.bit_index;
-        this->max_bit_idx = other.max_bit_idx;
-        return *this;
-    }
-
-    /**
-     * @brief Adds the contents of another BitStream to the current BitStream.
-     * @param other The BitStream to be added.
-     * @return Reference to the current BitStream after adding.
-     */
-    BitStream& operator+=(const BitStream& other) {
-        size_t other_max_bit_idx = other.getMaxBitIndex();
-        for (size_t i = 0; i < other_max_bit_idx; i++) {
-            this->putBit(other.getBitVal(i));
-        }
-
-        return *this;
-    }
-
-    /**
-     * @brief Gets a substream from the current BitStream.
-     * @param start_bit The starting bit index of the substream.
-     * @param end_bit The ending bit index of the substream (exclusive).
-     * @return A new BitStream containing the specified substream.
-     * @throws std::out_of_range if start or end indices are out of bounds.
-     */
-    BitStream getSubStream(size_t start_bit, size_t end_bit) const {
-        if (start_bit >= max_bit_idx || end_bit > max_bit_idx || start_bit > end_bit) {
-            throw std::out_of_range("BitStream substream indices are out of range.");
-        }
-        BitStream substream;
-        for (size_t i = start_bit; i < end_bit; i++) {
-            substream.putBit(getBitVal(i));
-        }
-        return substream;
-    }
-
-    /**
-     * @brief Returns the current bit index in the stream.
-     * @return The current bit index.
-     */
-    size_t getCurrentBitIndex() const {
-        return bit_index;
-    }
-
-private:
-    size_t bit_index; ///< The current bit index in the stream.
-    size_t max_bit_idx; ///< The total number of bits in the stream.
-};
-
-BitStream operator+(const BitStream& lhs, const BitStream& rhs) {
-    BitStream result = lhs;
-    result += rhs;
-    return result;
-}
-
-#endif