mirror of
https://github.com/saitohirga/WSJT-X.git
synced 2024-09-16 18:26:34 -04:00
4fea8ebd0c
Both decoders now have slightly better performance and faster execution. The rare "duplicate decodes" in JT9 were eliminated. On Windows, at least, calls to f90 routine system_clock() do not provide correct wall time increments. Changed to using secnds() instead. git-svn-id: svn+ssh://svn.code.sf.net/p/wsjt/wsjt/branches/wsjtx@4571 ab8295b8-cf94-4d9e-aec4-7959e3be5d79
92 lines
2.4 KiB
Fortran
92 lines
2.4 KiB
Fortran
subroutine symspec2(c5,nz3,nsps8,nspsd,fsample,freq,drift,snrdb,schk, &
|
|
i1SoftSymbolsScrambled)
|
|
|
|
! Compute soft symbols from the final downsampled data
|
|
|
|
complex c5(0:4096-1)
|
|
complex z
|
|
integer*1 i1SoftSymbolsScrambled(207)
|
|
real aa(3)
|
|
real ss2(0:8,85)
|
|
real ss3(0:7,69)
|
|
include 'jt9sync.f90'
|
|
data scale/10.0/
|
|
|
|
aa(1)=-1500.0/nsps8
|
|
aa(2)=0.
|
|
aa(3)=0.
|
|
do i=0,8 !Loop over the 9 tones
|
|
if(i.ge.1) call twkfreq(c5,c5,nz3,fsample,aa)
|
|
m=0
|
|
k=-1
|
|
do j=1,85 !Loop over all symbols
|
|
z=0.
|
|
do n=1,nspsd !Sum over 16 samples
|
|
k=k+1
|
|
z=z+c5(k)
|
|
enddo
|
|
ss2(i,j)=real(z)**2 + aimag(z)**2 !Symbol speactra, data and sync
|
|
if(i.ge.1 .and. isync(j).eq.0) then
|
|
m=m+1
|
|
ss3(i-1,m)=ss2(i,j) !Symbol speactra, data only
|
|
endif
|
|
enddo
|
|
enddo
|
|
|
|
call chkss2(ss2,freq,drift,schk)
|
|
if(schk.lt.2.0) then
|
|
i1SoftSymbolsScrambled=0
|
|
go to 900
|
|
endif
|
|
|
|
ss=0.
|
|
sig=0.
|
|
do j=1,69
|
|
smax=0.
|
|
do i=0,7
|
|
smax=max(smax,ss3(i,j))
|
|
ss=ss+ss3(i,j)
|
|
enddo
|
|
sig=sig+smax
|
|
ss=ss-smax
|
|
enddo
|
|
ave=ss/(69*7) !Baseline
|
|
call pctile(ss2,9*85,35,xmed)
|
|
ss3=ss3/ave
|
|
sig=sig/69. !Signal
|
|
t=max(1.0,sig - 1.0)
|
|
snrdb=db(t) - 61.3
|
|
|
|
m0=3
|
|
k=0
|
|
do j=1,69
|
|
smax=0.
|
|
do i=0,7
|
|
if(ss3(i,j).gt.smax) smax=ss3(i,j)
|
|
enddo
|
|
|
|
do m=m0-1,0,-1 !Get bit-wise soft symbols
|
|
if(m.eq.2) then
|
|
r1=max(ss3(4,j),ss3(5,j),ss3(6,j),ss3(7,j))
|
|
r0=max(ss3(0,j),ss3(1,j),ss3(2,j),ss3(3,j))
|
|
else if(m.eq.1) then
|
|
r1=max(ss3(2,j),ss3(3,j),ss3(4,j),ss3(5,j))
|
|
r0=max(ss3(0,j),ss3(1,j),ss3(6,j),ss3(7,j))
|
|
else
|
|
r1=max(ss3(1,j),ss3(2,j),ss3(4,j),ss3(7,j))
|
|
r0=max(ss3(0,j),ss3(3,j),ss3(5,j),ss3(6,j))
|
|
endif
|
|
|
|
k=k+1
|
|
i4=nint(scale*(r1-r0))
|
|
if(i4.lt.-127) i4=-127
|
|
if(i4.gt.127) i4=127
|
|
i4=i4+128
|
|
if(i4.le.127) i1SoftSymbolsScrambled(k)=i4
|
|
if(i4.ge.128) i1SoftSymbolsScrambled(k)=i4-256
|
|
enddo
|
|
enddo
|
|
|
|
900 return
|
|
end subroutine symspec2
|