⚠ SDRangel is intended for the power user. We expect you to already have some experience with SDR applications and digital signal processing in general. SDRangel might be a bit overwhelming for you however you are encouraged to use the discussion group above to look for help. You can also find more information in the [Wiki](https://github.com/f4exb/sdrangel/wiki).
SDRangel is a near real time application that is demanding on CPU power and clock speeds for low latency. Recent (2015 or later) Core i7 class CPU is recommended preferably with 4 HT CPU cores (8 logical CPUs or more) with nominal clock over 2 GHz and at least 8 GB RAM. Modern Intel processors will include a GPU suitable for proper OpenGL support. On the other hand SDRangel is not as demanding as recent computer games for graphics and CPU integrated graphics are perfectly fine. USB-3 ports are also preferable for high speed, low latency USB communication. It may run on beefy SBCs and was tested successfully on a Udoo Ultra.
The server variant does not need a graphical display and therefore OpenGL support. It can run on smaller devices check the Wiki for hardware requirements and a list of successful cases.
- legacy: the modified code from the parent application [hexameron rtl-sdrangelove](https://github.com/hexameron/rtl-sdrangelove) before a major redesign of the code was carried out and sync was lost.
Since version 3 transmission or signal generation is supported for BladeRF, HackRF (since version 3.1), LimeSDR (since version 3.4) and PlutoSDR (since version 3.7.8) using a sample sink plugin. These plugins are:
Since version 4 a REST API is available to interact with the SDRangel application. More details are provided in the server instance documentation in the `sdrsrv` folder.
Since version 4 the `sdrangelsrv` binary launches a server mode SDRangel instance that runs wihout the GUI. More information is provided in the Readme file of the `sdrsrv` folder.
Since version 4.1 the previously separated project SDRdaemon has been modified and included in SDRangel. The `sdrangelsrv` headless variant can be used for this purpose using the Daemon source or sink channels.
The audio devices with Qt are supported through pulseaudio and unless you are using a single sound chip (or card) with a single output port or you are an expert with pulseaudio config files you may get into trouble when trying to route the audio to a different output port. These notes are a follow-up of issue #31 with my own experiments with HDMI audio output on the Udoo x86 board. So using this example of HDMI output you can do the following:
- Install pavucontrol. It is included in most distributions for example:
- Check the audio config with alsamixer. On the Udoo x86 the HDMI output depends on the S/PDIF control and it occasionally gets muted when the HDMI monitor is turned off or goes to sleep. So in any case make sure nothing is muted there.
- Open pavucontrol and open the last tab (rightmost) called 'Configuration'
- Select HDMI from the profiles list in the 'Configuration' tab
- Then in the 'Output devices' tab the HDMI / display port is selected as it is normally the only one. Just double check this
- In SDRangel's Preferences/Audio menu make sure something like `alsa_output.pci-0000_00_1b.0.hdmi-stereo` is selected. The default might also work after the pulseaudio configuration you just made.
In case you cannot see anything related to HDMI or your desired audio device in pavucontrol just restart pulseaudio with `pulseaudio -k` (`-k` kills the previous instance before restarting) and do the above steps again.
[Airspy R2](https://airspy.com/airspy-r2/) and [Airspy Mini](https://airspy.com/airspy-mini/) are supported through the libairspy library that should be installed in your system for proper build of the software and operation support. Add `libairspy-dev` to the list of dependencies to install.
If you use your own location for libairspy install directory you need to specify library and include locations. Example with `/opt/install/libairspy` with the following defines on `cmake` command line:
Please note that if you are using a recent version of libairspy (>= 1.0.6) the dynamic retrieval of sample rates is supported. To benefit from it you should modify the `plugins/samplesource/airspy/CMakeLists.txt` and change line `add_definitions(${QT_DEFINITIONS})` by `add_definitions("${QT_DEFINITIONS} -DLIBAIRSPY_DYN_RATES")`. In fact both lines are present with the last one commented out.
Be also aware that the lower rates (2.5 MS/s or 5 MS/s with modified firmware) are affected by a noise artifact so 10 MS/s is preferable for weak signal work or instrumentation. A decimation by 64 was implemented to facilitate narrow band work at 10 MS/s input rate.
If you use your own location for libairspyhf install directory you need to specify library and include locations. Example with `/opt/install/libairspyhf` with the following defines on `cmake` command line:
It is recommended to add `-DRX_SAMPLE_24BIT=ON` on the cmake command line to activate the Rx 24 bit DSP build and take advantage of improved dynamic range when using decimation.
☞ From version 3.12.0 the Linux binaries are built with the 24 bit Rx option.
[BladeRF1](https://www.nuand.com/bladerf-1) is supported through the libbladeRF library that should be installed in your system for proper build of the software and operation support. Add `libbladerf-dev` to the list of dependencies to install. Note that libbladeRF v2 is used since version 4.2.0 (git tag 2018.08).
If you compile and use your own location for libbladeRF install directory you need to specify library and include locations. Example with `/opt/install/libbladerf` with the following defines on `cmake` command line:
☞ Please note that if you use your own library the FPGA image `hostedx40.rbf` or `hostedx115.rbf` shoud be placed in e.g. `/opt/install/libbladeRF/share/Nuand/bladeRF` unless you have flashed the FPGA image inside the BladeRF.
The plugins used to support BladeRF classic are specific to this version of the BladeRF:
[BladeRF 2 micro](https://www.nuand.com/bladerf-2-0-micro/) is also supported using libbladeRF library that should be installed and configured in the same way as for BladeRF1.
The plugins used to support BladeRF 2 micro are specific to this version of the BladeRF:
Both Pro and Pro+ are supported with the plugins in fcdpro and fcdproplus respectively. For the Pro+ the band filter selection is not effective as it is handled by the firmware using the center frequency.
The control interface is based on qthid and has been built in the software in the fcdhid library. You don't need anything else than libusb support. Library fcdlib is used to store the constants for each dongle type.
The Pro+ has trouble starting. The sound card interface is not recognized when you just plug it in and start SDRAngel. The workaround is to start qthid then a recording program like Audacity and start recording in Audacity. Then just quit Audacity without saving and quit qthid. After this operation the Pro+ should be recognized by SDRAngel until you unplug it.
[HackRF](https://greatscottgadgets.com/hackrf/) is supported through the libhackrf library that should be installed in your system for proper build of the software and operation support. Add `libhackrf-dev` to the list of dependencies to install. Please note that you will need a recent version (2015.07.2 at least) of libhackrf that supports the sequential listing of devices and the antenna power control (bias tee). To be safe anyway you may choose to build and install the Github version: `https://github.com/mossmann/hackrf.git`. Note also that the firmware must be updated to match the library version as per instructions found in the HackRF wiki.
If you use your own location for libhackrf install directory you need to specify library and include locations. Example with `/opt/install/libhackrf` with the following defines on `cmake` command line:
HackRF is better used with a sampling rate of 4.8 MS/s and above. The 2.4 and 3.2 MS/s rates are considered experimental and are way out of specs of the ADC. You may or may not achieve acceptable results depending on the unit. A too low sampling rate will typically create ghost signals (images) and/or raise the noise floor.
<p>⚠ Version 18.04.1 of LimeSuite is used in the builds and corresponding gateware loaded in the LimeSDR should be is used (2.16 for LimeSDR-USB and 1.24 for LimeSDR-Mini). If you compile from source version 18.04.1 of LimeSuite must be used.</p>
[The Perseus](http://microtelecom.it/perseus/) is supported with [my fork of libperseus-sdr library](https://github.com/f4exb/libperseus-sdr.git) on the `fixes` branch which is the default. There are a few fixes from the original mainly to make it work in a multi-device context.
Please note that the Perseus input plugin will be built only if the 24 bit Rx DSP chain is activated in the compilation with the `-DRX_SAMPLE_24BIT=ON` option on the cmake command line.
If you use your own location for libperseus-sdr install directory you need to specify library and include locations. Example with `/opt/install/libperseus` with the following defines on `cmake` command line: `-DLIBPERSEUS_INCLUDE_DIR=/opt/install/libperseus/include -DLIBPERSEUS_LIBRARIES=/opt/install/libperseus/lib/libperseus-sdr.so`.
☞ From version 3.12.0 the Linux binaries are built with the 24 bit Rx option and Perseus input plugin.
[PlutoSDR](https://wiki.analog.com/university/tools/pluto) is supported with the libiio interface. This library should be installed in your system for proper build of the software and operation support. Add `libiio-dev` to the list of dependencies to install. Be aware that version 0.10 is needed and is not available yet in all distributions. You may have to compile it from source instead.
If you use your own location for libiio install directory you need to specify library and include locations. Example with `/opt/install/libiio` with the following defines on `cmake` command line: `-DLIBIIO_INCLUDE_DIR=/opt/install/libiio/include -DLIBIIO_LIBRARY=/opt/install/libiio/lib/libiio.so`. In openSUSE the lib directory path would be: `-DLIBIIO_LIBRARY=/opt/install/libiio/lib64/libiio.so`.
RTL-SDR based dongles are supported through the librtlsdr library that should be installed in your system for proper build of the software and operation support. Add `librtlsdr-dev` to the list of dependencies to install.
If you use your own location for librtlsdr install directory you need to specify library and include locations. Example with `/opt/install/librtlsdr` with the following defines on `cmake` command line:
[SDRplay RSP1](https://www.sdrplay.com/rsp1/) is supported through the [libmirisdr-4](https://github.com/f4exb/libmirisdr-4) library found in this very same Github space. There is no package distribution for this library and you will have to clone it, build and install it in your system. However Debian packages of SDRangel contain a pre-compiled version of this library.
If you use your own location for libmirisdr-4 install directory you need to specify library and include locations with cmake. For example with `/opt/install/libmirisdr` the following defines must be added on `cmake` command line:
The [File source plugin](https://github.com/f4exb/sdrangel/tree/dev/plugins/samplesource/filesource) allows the playback of a recorded IQ file. Such a file is obtained using the recording feature. Click on the record button on the left of the main frequency dial to toggle recording. The file has a fixed name `test_<n>.sdriq` created in the current directory where `<n>` is the device tab index.
Note that this plugin does not require any of the hardware support libraries nor the libusb library. It is always available in the list of devices as `FileSource[0]` even if no physical device is connected.
The `.sdriq` format produced are the 2x2 bytes I/Q samples with a header containing the center frequency of the baseband, the sample rate and the timestamp of the recording start. Note that this header length is a multiple of the sample size so the file can be read with a simple 2x2 bytes I/Q reader such as a GNU Radio file source block. It will just produce a short glitch at the beginning corresponding to the header data.
The [File sink plugin](https://github.com/f4exb/sdrangel/tree/dev/plugins/samplesink/filesink) allows the recording of the I/Q baseband signal produced by a transmission chain to a file in the `.sdriq` format thus readable by the file source plugin described just above.
Note that this plugin does not require any of the hardware support libraries nor the libusb library. It is always available in the list of devices as `FileSink[0]` even if no physical device is connected.
The [Test source plugin](https://github.com/f4exb/sdrangel/tree/master/plugins/samplesource/testsource) is an internal continuous wave generator that can be used to carry out test of software internals.
The [SDRdaemon source input plugin](https://github.com/f4exb/sdrangel/tree/dev/plugins/samplesource/sdrdaemonsource) is the client side of an instance of SDRangel running the Daemon Sink channel plugin. On the "Data" line you must specify the local address and UDP port to which the remote server connects and samples will flow into the SDRangel application (default is `127.0.0.1`port `9090`). It uses the meta data to retrieve the sample flow characteristics such as sample rate and receiving center frequency. The remote is entirely controlled by the REST API. On the "API" line you can specify the address and port at which the remote REST API listens. However it is used just to display basic information about the remote.
The data blocks transmitted via UDP are protected against loss with a Cauchy MDS block erasure codec. This makes the transmission more robust in particular with WiFi links.
There is an automated skew rate compensation in place. During rate readjustment streaming can be suspended or signal glitches can occur for about one second.
This plugin will be built only if the [CM256cc library](https://github.com/f4exb/cm256cc) is installed in your system. For CM256cc if you install it in a non standard directory you will then have to specify the include and library paths on the cmake command line. Say if you install cm256cc in `/opt/install/cm256cc` you will have to add `-DCM256CC_INCLUDE_DIR=/opt/install/cm256cc/include/cm256cc -DCM256CC_LIBRARIES=/opt/install/cm256cc/lib/libcm256cc.so` to the cmake commands.
Note that this plugin does not require any of the hardware support libraries nor the libusb library. It is always available in the list of devices as `SDRdaemonSource[0]` even if no physical device is connected.
The [SDRdaemon sink output plugin](https://github.com/f4exb/sdrangel/tree/dev/plugins/samplesink/sdrdaemonsink) is the client side of and instance of SDRangel running the Daemon Source channel plugin. On the "Data" line you must specify the distant address and UDP port to which the plugin connects and samples from the SDRangel application will flow into the transmitter server (default is `127.0.0.1`port `9090`). The remote is entirely controlled by the REST API. On the "API" line you can specify the address and port at which the remote REST API listens. The API is pinged regularly to retrieve the status of the data blocks queue and allow rate control to stabilize the queue length. Therefore it is important to connect to the API properly (The status line must return "API OK" and the API label should be lit in green).
The data blocks sent via UDP are protected against loss with a Cauchy MDS block erasure codec. This makes the transmission more robust in particular with WiFi links.
This plugin will be built only if the [CM256cc library](https://github.com/f4exb/cm256cc) IS installed in your system. For CM256cc if you install it in a non standard directory you will then have to specify the include and library paths on the cmake command line. Say if you install cm256cc in `/opt/install/cm256cc` you will have to add `-DCM256CC_INCLUDE_DIR=/opt/install/cm256cc/include/cm256cc -DCM256CC_LIBRARIES=/opt/install/cm256cc/lib/libcm256cc.so` to the cmake commands.
Note that this plugin does not require any of the hardware support libraries nor the libusb library. It is always available in the list of devices as `SDRdaemonSink[0]` even if no physical device is connected.
It is based on the [DSDcc](https://github.com/f4exb/dsdcc) C++ library which is a rewrite of the original [DSD](https://github.com/szechyjs/dsd) program. So you will need to have DSDcc installed in your system. Please follow instructions in [DSDcc readme](https://github.com/f4exb/dsdcc/blob/master/Readme.md) to build and install DSDcc. If you install it in a custom location say `/opt/install/dsdcc` you will need to add these defines to the cmake command: `-DLIBDSDCC_INCLUDE_DIR=/opt/install/dsdcc/include/dsdcc -DLIBDSDCC_LIBRARIES=/opt/install/dsdcc/lib/libdsdcc.so`
If you have one or more serial devices interfacing the AMBE3000 chip in packet mode you can use them to decode AMBE voice frames. For that purpose you will need to compile with [SerialDV](https://github.com/f4exb/serialDV) support. Please refer to this project Readme.md to compile and install SerialDV. If you install it in a custom location say `/opt/install/serialdv` you will need to add these defines to the cmake command: `-DLIBSERIALDV_INCLUDE_DIR=/opt/install/serialdv/include/serialdv -DLIBSERIALDV_LIBRARY=/opt/install/serialdv/lib/libserialdv.so` Also your user must be a member of group `dialout` (Ubuntu/Debian) or `uucp` (Arch) to be able to use the dongle.
Although such serial devices work with a serial interface at 400 kb in practice maybe for other reasons they are capable of handling only one conversation at a time. The software will allocate the device dynamically to a conversation with an inactivity timeout of 1 second so that conversations do not get interrupted constantly making the audio output too choppy. In practice you will have to have as many devices connected to your system as the number of conversations you would like to be handled in parallel.
Alternatively you can use [mbelib](https://github.com/szechyjs/mbelib) but mbelib comes with some copyright issues (see next). If you have mbelib installed in a custom location, say `/opt/install/mbelib` you will need to add these defines to the cmake command: `-DLIBMBE_INCLUDE_DIR=/opt/install/mbelib/include -DLIBMBE_LIBRARY=/opt/install/mbelib/lib/libmbe.so`
Possible copyright issues apart (see next) the audio quality with the DVSI AMBE chip is much better.
While DSDcc is intended to be patent-free, `mbelib` that it uses describes functions that may be covered by one or more U.S. patents owned by DVSI Inc. The source code itself should not be infringing as it merely describes possible methods of implementation. Compiling or using `mbelib` may infringe on patents rights in your jurisdiction and/or require licensing. It is unknown if DVSI will sell licenses for software that uses `mbelib`.
Possible copyright issues apart the audio quality with the DVSI AMBE chip is much better.
If you are not comfortable with this just do not install DSDcc and/or mbelib and the plugin will not be compiled and added to SDRangel. For packaged distributions just remove from the installation directory:
- For Linux distributions: `plugins/channel/libdemoddsd.so`
**⚠** The udev rules are not set by the package installation so you will have to set it manually in order to be able to access the various SDR hardware. The `udev-rules` folder contains the rules file and the `install.sh` script that you can run as sudo to install all rules files. You may also adapt the script to copy only the required files.
The default CPU governor is now `powersave` which exhibits excessive CPU usage when running SDRangel. In the case of benchmarking and maybe high throughput usage it is recommended to switch to `performance` before running SDRangel by running the command: `sudo cpupower frequency-set --governor performance`. You can turn it back to `powersave` any time by running: `sudo cpupower frequency-set --governor powersave`. It is normal that with a lower CPU frequency the relative CPU usage rises for the same actual load. If not impairing operation this is normal and overall beneficial for heat and power consumption.
This is the archive of the complete binary distribution that expands to the `sdrangel` directory. You can install it anywhere you like and click on `sdrangel.exe` to start.
<b>⚠ Windows distribution was provided as a by product thanks to the Qt toolchain. The platform of choice to run SDRangel is definitely Linux and very little support can be given for this Windows distribution.</b>
By default all Rx DSP processes use 16 bit samples coded on int16 fields. In order to use 24 bit samples coded on int32 fields you can specify `-DRX_SAMPLE_24BIT=ON` on the cmake command line. This will give more dynamic range when the number of bits with decimation exceeds 16 bits:
Then do `sudo apt-get update` and go to the next step. Alternatively if you have an older version of ffmpeg suite already installed just do `sudo apt-get dist-upgrde`.
- Note1: if you are on Leap you will need a more recent g++ compiler so in place of `gcc-c++` use `gcc6-c++` or `gcc7-c++` then add the following in the cmake command: `-DCMAKE_C_COMPILER=/usr/bin/gcc-7 -DCMAKE_CXX_COMPILER=/usr/bin/g++-7` (for gcc 7) and then `-DCMAKE_INSTALL_PREFIX:PATH=...` for the custom install path (not `-DCMAKE_INSTALL_PREFIX=...`)
- Note2 for udev rules: installed udev rules for BladeRF and HackRF are targeted at Debian or Ubuntu systems that have a plugdev group for USB hotplug devices. This is not the case in openSUSE. To fix it you can either:
- Note3: A package has been created in openSUSE thanks to Martin, see: [sdrangel](https://build.opensuse.org/package/show/hardware:sdr/sdrangel). It is based on the latest release on master branch.
Tested with the 15.09 version with LXDE desktop (community supported). The exact desktop environment should not matter anyway. Prerequisites should be similar for Arch and all derivatives.
- Note2: Two package are avaliable in the AUR (thanks Mikos!), [sdrangel](https://aur.archlinux.org/packages/sdrangel), which provides the lastest tagged release (stable), and [sdrangel-git](https://aur.archlinux.org/packages/sdrangel-git), which builds the latest commit from this repository (unstable).
This is a rather long story and one may prefer using the software distribution instead. However the brave may follow [this link](ReadmeWindowsBuild.md)
<h1>Mac O/S</h1>
A Mac O/S build was contributed from version 2.0.1. Please be aware that this is still experimental.
<h1>Android</h1>
Despite several attempts and the presence of Android related stuff still present in the .pro files there is no Android build available. An APK can be built but Qt fails miserably at porting applications other than its ridiculously simple examples. When multi-threading is involved a lot like in SDRangel this simply crashes at the very beginning of the application when starting the event loop.
Contributors welcome!
<h1>Software installation on Linux flavours</h1>
Simply do `make install` or `sudo make install` depending on you user rights on the target installation directory. On most systems the default installation directory is `/usr/local` a custom installation directory can be specified with the `-DCMAKE_INSTALL_PREFIX=...` option on the `cmake` command line as usual with cmake.
You can uninstall the software with `make uninstall` or `sudo make uninstall` from the build directory (it needs the `install_manifest.txt` file in the same directory and is automatically created by the `make install`command). Note that this will not remove the possible empty directories.
<h1>Known Issues</h1>
- The message queuing model supports a n:1 connection to an object (on its input queue) and a 1:1 connection from an object (on its output queue). Assuming a different model can cause insidious disruptions.
- As the objects input and output queues can be publicly accessed there is no strict control of which objects post messages on these queues. The correct assumption is that messages can be popped from the input queue only by its holder and that messages can be pushed on the output queue only by its holder.
- Objects managing more than one message queue (input + output for example) do not work well under stress conditions. Output queue removed from sample sources but this model has to be revised throughout the application.
- SDRdaemon FEC plugin: it has trouble doing the first connection or reconnecting to another device. The best option is to try then acknowledge the error message and restart SDRangel.
<h1>Limitations</h1>
- Your hardware. Still SDRangel is relatively conservative on computer resources.
Since version 3.3.2 the "todos" are in the form of tickets opened in the Issues section with the label "feature". When a specific release is targeted it will appear as a milestone. Thus anyone can open a "feature" issue to request a new feature.
