diff --git a/lib/Makefile.MinGW b/lib/Makefile.MinGW
index 1d657db2e..4d7cd6fcd 100644
--- a/lib/Makefile.MinGW
+++ b/lib/Makefile.MinGW
@@ -20,7 +20,7 @@ CFLAGS = -I. -fbounds-check -mno-stack-arg-probe
 %.o: %.F90
 	${FC} ${FFLAGS} -c $<
 
-all:    libjt9.a jt9sim.exe jt9.exe jt9code.exe jt9test.exe
+all:    libjt9.a jt9sim.exe jt9.exe jt9code.exe wsjt24d.exe
 
 OBJS1 = pctile.o graycode.o sort.o ssort.o \
 	unpackmsg.o igray.o unpackcall.o unpackgrid.o \
@@ -30,8 +30,9 @@ OBJS1 = pctile.o graycode.o sort.o ssort.o \
 	symspec.o timf2.o analytic.o db.o genjt9.o \
 	packbits.o unpackbits.o encode232.o interleave9.o \
 	entail.o fano232.o gran.o spec9.o sync9.o decode9.o \
-	peakdt9.o peakdf9.o fil3.o redsync.o decoder.o \
-	grid2n.o n2grid.o timer.o 
+	fil3.o redsync.o decoder.o grid2n.o n2grid.o timer.o \
+	decode9a.o getlags.o afc9.o fchisq.o twkfreq.o downsam9.o \
+	peakdt9.o symspec2.o
 
 libjt9.a: $(OBJS1)
 	ar cr libjt9.a $(OBJS1) 
@@ -53,8 +54,13 @@ jt9code.exe: $(OBJS4) libjt9.a
 	$(FC) -o jt9code.exe $(OBJS4) libjt9.a
 
 OBJS5 = jt9test.o 
-jt9test.exe: $(OBJS5) libjt9.a
-	$(FC) -o jt9test.exe $(OBJS5) libjt9.a ../libfftw3f_win.a 
+jt9test.exe: $(OBJS5) libjt9.a 
+	$(FC) -o jt9test.exe $(OBJS5) libjt9.a ../libfftw3f_win.a
+
+OBJS6 = wsjt24d.o wsjt24.o sync24.o decode24.o ps24.o flat1.o \
+	xcor24.o slope.o peakup.o interleave24.o genmet24.o
+wsjt24d.exe: $(OBJS6) libjt9.a 
+	$(FC) -o wsjt24d.exe $(OBJS6) libjt9.a ../libfftw3f_win.a
 
 INCPATH = -I'c:/QtSDK/Desktop/Qt/4.7.4/mingw/include/QtCore' \
 	 -I'c:/QtSDK/Desktop/Qt/4.7.4/mingw/include' \
diff --git a/lib/decode24.f90 b/lib/decode24.f90
new file mode 100644
index 000000000..0e708dcad
--- /dev/null
+++ b/lib/decode24.f90
@@ -0,0 +1,167 @@
+subroutine decode24(dat,npts,dtx,dfx,flip,mode,mode4,decoded,ncount,   &
+     nlim,deepmsg,qual,submode)
+
+! Decodes JT65 data, assuming that DT and DF have already been determined.
+
+  parameter (MAXAVE=120)
+  real dat(npts)                        !Raw data
+  character decoded*22,deepmsg*22
+  character*72 c72
+  character submode*1
+  real*8 dt,df,phi,f0,dphi,twopi,phi1,dphi1
+  complex*16 cz,cz1,c0,c1
+  integer*1 symbol(207)
+  real*4 rsymbol(207)
+  integer*1 data1(13)                   !Decoded data (8-bit bytes)
+  integer   data4a(9)                   !Decoded data (8-bit bytes)
+  integer   data4(12)                   !Decoded data (6-bit bytes)
+  integer amp,delta
+  integer mettab(0:255,0:1)             !Metric table
+  integer fano
+  integer nch(7)
+  integer npr2(207)
+!  common/ave/ppsave(64,63,MAXAVE),nflag(MAXAVE),nsave,iseg(MAXAVE)
+  data mode0/-999/
+  data nsum/0/,rsymbol/207*0.0/
+  data npr2/                                                         &
+       0,0,0,0,1,1,0,0,0,1,1,0,1,1,0,0,1,0,1,0,0,0,0,0,0,0,1,1,0,0,  &
+       0,0,0,0,0,0,0,0,0,0,1,0,1,1,0,1,1,0,1,0,1,1,1,1,1,0,1,0,0,0,  &
+       1,0,0,1,0,0,1,1,1,1,1,0,0,0,1,0,1,0,0,0,1,1,1,1,0,1,1,0,0,1,  &
+       0,0,0,1,1,0,1,0,1,0,1,0,1,0,1,1,1,1,1,0,1,0,1,0,1,1,0,1,0,1,  &
+       0,1,1,1,0,0,1,0,1,1,0,1,1,1,1,0,0,0,0,1,1,0,1,1,0,0,0,1,1,1,  &
+       0,1,1,1,0,1,1,1,0,0,1,0,0,0,1,1,0,1,1,0,0,1,0,0,0,1,1,1,1,1,  &
+       1,0,0,1,1,0,0,0,0,1,1,0,0,0,1,0,1,1,0,1,1,1,1,0,1,0,1/
+
+  data nch/1,2,4,9,18,36,72/
+  save mettab,mode0,nsum,rsymbol
+
+  if(mode.ne.mode0) call genmet24(mode,mettab)
+  mode0=mode
+  twopi=8*atan(1.d0)
+  dt=2.d0/11025             !Sample interval (2x downsampled data)
+  df=11025.d0/2520.d0
+  nsym=206
+  amp=15
+  istart=nint(dtx/dt)              !Start index for synced FFTs
+  if(istart.lt.0) istart=0
+  idfbest=0
+  nchips=0
+  ich=0
+
+! Should amp be adjusted according to signal strength?
+
+! Compute soft symbols using differential BPSK demodulation
+  c0=0.                                !### C0=amp ???
+  k=istart
+  fac=1.e-4
+  phi=0.d0
+  phi1=0.d0
+  ang0=0.
+
+40 ich=ich+1
+  nchips=nch(ich)
+  nspchip=1260/nchips
+  k=istart
+  phi=0.d0
+  phi1=0.d0
+  fac2=1.e-8 * sqrt(float(mode4))
+  do j=1,nsym+1
+     if(flip.gt.0.0) then
+        f0=1270.46 + dfx + (npr2(j)-1.5)*mode4*df
+        f1=1270.46 + dfx + (2+npr2(j)-1.5)*mode4*df
+     else
+        f0=1270.46 + dfx + (1-npr2(j)-1.5)*mode4*df
+        f1=1270.46 + dfx + (3-npr2(j)-1.5)*mode4*df
+     endif
+     dphi=twopi*dt*f0
+     dphi1=twopi*dt*f1
+     sq0=0.
+     sq1=0.
+     do nc=1,nchips
+        phi=0.d0
+        phi1=0.d0
+        c0=0.
+        c1=0.
+        do i=1,nspchip
+           k=k+1
+           phi=phi+dphi
+           phi1=phi1+dphi1
+           cz=dcmplx(cos(phi),-sin(phi))
+           cz1=dcmplx(cos(phi1),-sin(phi1))
+           if(k.le.npts) then
+              c0=c0 + dat(k)*cz
+              c1=c1 + dat(k)*cz1
+           endif
+        enddo
+        sq0=sq0 + real(c0)**2 + aimag(c0)**2
+        sq1=sq1 + real(c1)**2 + aimag(c1)**2
+     enddo
+     sq0=fac2*sq0
+     sq1=fac2*sq1
+     rsym=amp*(sq1-sq0)
+     r=rsym+128.
+     if(r.gt.255.0) r=255.0
+     if(r.lt.0.0) r=0.0
+     i4=nint(r)
+     if(i4.gt.127) i4=i4-256
+     if(j.ge.1) then
+        symbol(j)=i4
+        rsymbol(j)=rsymbol(j) + rsym
+     endif
+  enddo
+  
+!###  The following does simple message averaging:
+  nsum=nsum+1
+  do j=1,207
+     r=rsymbol(j)/nsum + 128.
+     if(r.gt.255.0) r=255.0
+     if(r.lt.0.0) r=0.0
+     i4=nint(r)
+     if(i4.gt.127) i4=i4-256
+     symbol(j)=i4
+  enddo
+!###
+  
+  nbits=72+31
+  delta=100
+  limit=100000
+  ncycles=0
+  ncount=-1
+  call interleave24(symbol(2),-1)         !Remove the interleaving
+
+  call fano232(symbol(2),nbits,mettab,delta,limit,data1,ncycles,metric,ncount)
+  nlim=ncycles/nbits
+
+  if(ncount.ge.0) go to 100
+  if(mode.eq.7 .and. nchips.lt.mode4) go to 40
+
+100 do i=1,9
+     i4=data1(i)
+     if(i4.lt.0) i4=i4+256
+     data4a(i)=i4
+  enddo
+!  call cs_lock('decode24')
+  write(c72,1100) (data4a(i),i=1,9)
+1100 format(9b8.8)
+  read(c72,1102) data4
+1102 format(12b6)
+!  call cs_unlock
+
+  decoded='                      '
+  submode=' '
+  if(ncount.ge.0) then
+     call unpackmsg(data4,decoded)
+     submode=char(ichar('A')+ich-1)
+  endif
+  if(decoded(1:6).eq.'000AAA') then
+     decoded='***WRONG MODE?***'
+     ncount=-1
+  endif
+
+  qual=0.
+  deepmsg='                      '
+
+! Save symbol spectra for possible decoding of average.
+
+  return
+end subroutine decode24
diff --git a/lib/flat1.f90 b/lib/flat1.f90
new file mode 100644
index 000000000..526ab9bd6
--- /dev/null
+++ b/lib/flat1.f90
@@ -0,0 +1,30 @@
+subroutine flat1(psavg,s2,nh,nsteps,nhmax,nsmax)
+
+  real psavg(nh)
+  real s2(nhmax,nsmax)
+  real x(8192),tmp(33)
+
+  nsmo=33
+  ia=nsmo/2 + 1
+  ib=nh - nsmo/2 - 1
+  do i=ia,ib
+     call pctile(psavg(i-nsmo/2),nsmo,50,x(i))
+  enddo
+  do i=1,ia-1
+     x(i)=x(ia)
+  enddo
+  do i=ib+1,nh
+     x(i)=x(ib)
+  enddo
+
+  do i=1,nh
+     psavg(i)=psavg(i)/x(i)
+     do j=1,nsteps
+        s2(i,j)=s2(i,j)/x(i)
+     enddo
+  enddo
+
+  return
+end subroutine flat1
+
+      
diff --git a/lib/genmet24.f90 b/lib/genmet24.f90
new file mode 100644
index 000000000..c15342d3f
--- /dev/null
+++ b/lib/genmet24.f90
@@ -0,0 +1,27 @@
+subroutine genmet24(mode,mettab)
+
+! Return appropriate metric table for soft-decision convolutional decoder.
+
+  real bias                         !bias for integer table
+  integer scale                     !scale factor for integer table
+! Metric table (RxSymbol,TxSymbol)
+  integer mettab(0:255,0:1)
+
+  bias=0.5
+  scale=10
+  if(mode.eq.7) then  !Non-coherent 2FSK
+     open(19,file='dmet_10_-1_3.dat',status='old')
+  else
+     print*,'Unsupported mode:',mode,' in genmet.'
+     stop 'genmet'
+  endif
+
+  do i=0,255
+     read(19,*) junk,d0,d1
+     mettab(i,0)=nint(scale*(d0-bias))
+     mettab(i,1)=nint(scale*(d1-bias))
+  enddo
+
+  return
+end subroutine genmet24
+
diff --git a/lib/interleave24.f90 b/lib/interleave24.f90
new file mode 100644
index 000000000..3cdd9c525
--- /dev/null
+++ b/lib/interleave24.f90
@@ -0,0 +1,43 @@
+subroutine interleave24(id,ndir)
+  integer*1 id(0:205),itmp(0:205)
+  integer j0(0:205)
+  logical first
+  data first/.true./
+  save first,j0
+
+  if(first) then
+     k=-1
+     do i=0,255
+        m=i
+        n=iand(m,1)
+        n=2*n + iand(m/2,1)
+        n=2*n + iand(m/4,1)
+        n=2*n + iand(m/8,1)
+        n=2*n + iand(m/16,1)
+        n=2*n + iand(m/32,1)
+        n=2*n + iand(m/64,1)
+        n=2*n + iand(m/128,1)
+        if(n.le.205) then
+           k=k+1
+           j0(k)=n
+        endif
+     enddo
+     first=.false.
+  endif
+
+  if(ndir.eq.1) then
+     do i=0,205
+        itmp(j0(i))=id(i)
+     enddo
+  else
+     do i=0,205
+        itmp(i)=id(j0(i))
+     enddo
+  endif
+
+  do i=0,205
+     id(i)=itmp(i)
+  enddo
+
+  return
+end subroutine interleave24
diff --git a/lib/peakup.f90 b/lib/peakup.f90
new file mode 100644
index 000000000..8e5aedaf7
--- /dev/null
+++ b/lib/peakup.f90
@@ -0,0 +1,8 @@
+subroutine peakup(ym,y0,yp,dx)
+
+  b=(yp-ym)/2.0
+  c=(yp+ym-2.0*y0)/2.0
+  dx=-b/(2.0*c)
+
+  return
+end subroutine peakup
diff --git a/lib/ps24.f90 b/lib/ps24.f90
new file mode 100644
index 000000000..241e49fa2
--- /dev/null
+++ b/lib/ps24.f90
@@ -0,0 +1,27 @@
+subroutine ps24(dat,nfft,s)
+
+  parameter (NMAX=2520+2)
+  parameter (NHMAX=NMAX/2-1)
+  real dat(nfft)
+  real dat2(NMAX)
+  real s(NHMAX)
+  complex c(0:NMAX)
+  equivalence(dat2,c)
+
+  nh=nfft/2
+  do i=1,nh
+     dat2(i)=dat(i)/128.0       !### Why 128 ??
+  enddo
+  do i=nh+1,nfft
+     dat2(i)=0.
+  enddo
+
+  call four2a(c,nfft,1,-1,0)
+  
+  fac=1.0/nfft
+  do i=1,nh
+     s(i)=fac*(real(c(i))**2 + aimag(c(i))**2)
+  enddo
+
+  return
+end subroutine ps24
diff --git a/lib/slope.f90 b/lib/slope.f90
new file mode 100644
index 000000000..5581ade14
--- /dev/null
+++ b/lib/slope.f90
@@ -0,0 +1,40 @@
+subroutine slope(y,npts,xpk)
+
+! Remove best-fit slope from data in y(i).  When fitting the straight line,
+! ignore the peak around xpk +/- 2.
+
+  real y(npts)
+  real x(100)
+
+  do i=1,npts
+     x(i)=i
+  enddo
+
+  sumw=0.
+  sumx=0.
+  sumy=0.
+  sumx2=0.
+  sumxy=0.
+  sumy2=0.
+
+  do i=1,npts
+     if(abs(i-xpk).gt.2.0) then
+        sumw=sumw + 1.0
+        sumx=sumx + x(i)
+        sumy=sumy + y(i)
+        sumx2=sumx2 + x(i)**2
+        sumxy=sumxy + x(i)*y(i)
+        sumy2=sumy2 + y(i)**2
+     endif
+  enddo
+
+  delta=sumw*sumx2 - sumx**2
+  a=(sumx2*sumy - sumx*sumxy) / delta
+  b=(sumw*sumxy - sumx*sumy) / delta
+
+  do i=1,npts
+     y(i)=y(i)-(a + b*x(i))
+  enddo
+
+  return
+end subroutine slope
diff --git a/lib/sync24.f90 b/lib/sync24.f90
new file mode 100644
index 000000000..4045b1caf
--- /dev/null
+++ b/lib/sync24.f90
@@ -0,0 +1,195 @@
+subroutine sync24(dat,jz,DFTolerance,NFreeze,MouseDF,mode,mode4,    &
+     dtx,dfx,snrx,snrsync,ccfblue,ccfred1,flip,width)
+
+! Synchronizes JT4 data, finding the best-fit DT and DF.  
+
+  parameter (NFFTMAX=2520)         !Max length of FFTs
+  parameter (NHMAX=NFFTMAX/2)      !Max length of power spectra
+  parameter (NSMAX=525)            !Max number of half-symbol steps
+  integer DFTolerance              !Range of DF search
+  real dat(jz)
+  real psavg(NHMAX)                !Average spectrum of whole record
+  real s2(NHMAX,NSMAX)             !2d spectrum, stepped by half-symbols
+  real ccfblue(-5:540)             !CCF with pseudorandom sequence
+  real ccfred(-450:450)            !Peak of ccfblue, as function of freq
+  real ccfred1(-224:224)           !Peak of ccfblue, as function of freq
+  real tmp(1260)
+  save
+
+! Do FFTs of twice symbol length, stepped by half symbols.  Note that 
+! we have already downsampled the data by factor of 2.
+
+  nsym=207
+  nfft=2520
+  nh=nfft/2
+  nq=nfft/4
+  nsteps=jz/nq - 1
+  df=0.5*11025.0/nfft
+  psavg(1:nh)=0.
+
+! Compute power spectrum for each step and get average
+  do j=1,nsteps
+     k=(j-1)*nq + 1
+!         call limit(dat(k),nfft)
+     call ps24(dat(k),nfft,s2(1,j))
+!         if(mode65.eq.4) call smooth(s2(1,j),nh)
+     psavg(1:nh)=psavg(1:nh) + s2(1:nh,j)
+  enddo
+
+  call flat1(psavg,s2,nh,nsteps,NHMAX,NSMAX)        !Flatten the spectra
+! Find the best frequency channel for CCF
+!      famin= 670.46
+!      fbmax=1870.46
+  famin=200.
+  fbmax=2700.
+
+  fa=famin
+  fb=fbmax
+  if(NFreeze.eq.1) then
+     fa=max(famin,1270.46+MouseDF-DFTolerance)
+     fb=min(fbmax,1270.46+MouseDF+DFTolerance)
+  else
+     fa=max(famin,1270.46+MouseDF-600)
+     fb=min(fbmax,1270.46+MouseDF+600)
+  endif
+  ia=fa/df
+  ib=fb/df
+  if(mode.eq.7) then
+     ia=ia - 3*mode4
+     ib=ib - 3*mode4
+  endif
+  i0=nint(1270.46/df)
+  lag1=-5
+  lag2=59
+  syncbest=-1.e30
+  syncbest2=-1.e30
+
+  ccfred=0.
+  do i=ia,ib
+     call xcor24(s2,i,nsteps,nsym,lag1,lag2,mode4,ccfblue,ccf0,lagpk0,flip)
+     j=i-i0
+     if(mode.eq.7) j=j + 3*mode4
+     if(j.ge.-372 .and. j.le.372) ccfred(j)=ccf0
+
+! Find rms of the CCF, without the main peak
+     call slope(ccfblue(lag1),lag2-lag1+1,lagpk0-lag1+1.0)
+     sync=abs(ccfblue(lagpk0))
+     ppmax=psavg(i)-1.0
+
+! Find the best sync value
+     if(sync.gt.syncbest2) then
+        ipk2=i
+        lagpk2=lagpk0
+        syncbest2=sync
+     endif
+
+! We are most interested if snrx will be more than -30 dB.
+     if(ppmax.gt.0.2938) then            !Corresponds to snrx.gt.-30.0
+        if(sync.gt.syncbest) then
+           ipk=i
+           lagpk=lagpk0
+           syncbest=sync
+        endif
+     endif
+  enddo
+
+! If we found nothing with snrx > -30 dB, take the best sync that *was* found.
+  if(syncbest.lt.-10.) then
+     ipk=ipk2
+     lagpk=lagpk2
+     syncbest=syncbest2
+  endif
+
+! Peak up in frequency to fraction of channel
+!      call peakup(psavg(ipk-1),psavg(ipk),psavg(ipk+1),dx)
+!      if(dx.lt.-1.0) dx=-1.0
+!      if(dx.gt.1.0) dx=1.0
+  dx=0.
+  dfx=(ipk+dx-i0)*df
+  if(mode.eq.7) dfx=dfx + 3*mode4*df
+
+! Peak up in time, at best whole-channel frequency
+  call xcor24(s2,ipk,nsteps,nsym,lag1,lag2,mode4,ccfblue,ccfmax,lagpk,flip)
+  xlag=lagpk
+  if(lagpk.gt.lag1 .and. lagpk.lt.lag2) then
+     call peakup(ccfblue(lagpk-1),ccfmax,ccfblue(lagpk+1),dx2)
+     xlag=lagpk+dx2
+  endif
+
+! Find rms of the CCF, without the main peak
+  call slope(ccfblue(lag1),lag2-lag1+1,xlag-lag1+1.0)
+  sq=0.
+  nsq=0
+  do lag=lag1,lag2
+     if(abs(lag-xlag).gt.2.0) then
+        sq=sq+ccfblue(lag)**2
+        nsq=nsq+1
+     endif
+  enddo
+  rms=sqrt(sq/nsq)
+  snrsync=abs(ccfblue(lagpk))/rms - 1.1    !Empirical
+
+  dt=2.0/11025.0
+  istart=xlag*nq
+  dtx=istart*dt
+  snrx=-99.0
+!  ppmax=psavg(ipk)-1.0
+  if(ipk.ge.1 .and. ipk.le.1260) ppmax=psavg(ipk)-1.0           !###
+
+  if(ppmax.gt.0.0001) then
+     snrx=db(ppmax*df/2500.0) + 7.5        !Empirical
+     if(mode.eq.7) snrx=snrx + 3.0         !Empirical
+  endif
+  if(snrx.lt.-33.0) snrx=-33.0
+
+! Compute width of sync tone to outermost -3 dB points
+!      call pctile(ccfred(ia-i0),tmp,ib-ia+1,45,base)
+!      i1=max(-224,ia-i0)
+!      i2=min(224,ib-i0)
+  i1=max(-450,ia-i0)
+  i2=min(450,ib-i0)
+!###  call pctile(ccfred(i1),tmp,i2-i1+1,45,base)
+  call pctile(ccfred(i1),i2-i1+1,45,base)
+
+  jpk=ipk-i0
+  if(abs(jpk).gt.450) then
+     print*,'sync24 a:',jpk,ipk,i0
+     snrsync=0.
+     go to 999
+  else
+     stest=base + 0.5*(ccfred(jpk)-base) ! -3 dB
+  endif
+  do i=-10,0
+     if(jpk+i.ge.-371) then 
+        if(ccfred(jpk+i).gt.stest) go to 30
+     endif
+  enddo
+  i=0
+30 continue
+  if(abs(jpk+i-1).gt.450 .or. abs(jpk+i).gt.450) then
+     print*,'sync24 b:',jpk,i
+  else
+!         x1=i-1+(stest-ccfred(jpk+i-1))/(ccfred(jpk+i)-ccfred(jpk+i-1))
+     x1=i-0.5
+  endif
+
+  do i=10,0,-1
+     if(jpk+i.le.371) then
+        if(ccfred(jpk+i).gt.stest) go to 32
+     endif
+  enddo
+  i=0
+! 32   x2=i+1-(stest-ccfred(jpk+i+1))/(ccfred(jpk+i)-ccfred(jpk+i+1))
+32 x2=i+0.5
+  width=x2-x1
+  if(width.gt.1.2) width=sqrt(width**2 - 1.44)
+  width=df*width
+  width=max(0.0,min(99.0,width))
+
+  do i=-224,224
+     ccfred1(i)=ccfred(i)
+  enddo
+
+999 return
+end subroutine sync24
+
diff --git a/lib/wsjt24.f90 b/lib/wsjt24.f90
new file mode 100644
index 000000000..dc630e7bd
--- /dev/null
+++ b/lib/wsjt24.f90
@@ -0,0 +1,200 @@
+subroutine wsjt24(dat,npts,cfile6,NClearAve,MinSigdB,                  &
+     DFTolerance,NFreeze,mode,mode4,Nseg,MouseDF,NAgain,               &
+     idf,lumsg,lcum,nspecial,ndf,NSyncOK,ccfblue,ccfred,ndiag)
+
+! Orchestrates the process of decoding JT4 messages, using data that 
+! have been 2x downsampled.  
+! No message averaging and no deep search, at present.
+
+  parameter (MAXAVE=120)
+  real dat(npts)                                     !Raw data
+  real*4 ccfblue(-5:540)                             !CCF in time
+  real*4 ccfred(-224:224)                            !CCF in frequency
+  integer DFTolerance
+  logical first
+  logical lcum
+  character decoded*22,cfile6*6,special*5,cooo*3
+  character*22 avemsg1,avemsg2,deepmsg
+  character*77 line,ave1,ave2
+  character*1 csync,c1
+  character*12 mycall
+  character*12 hiscall
+  character*6 hisgrid
+  character submode*1
+  real*4 ccfbluesum(-5:540),ccfredsum(-224:224)
+!  common/ave/ppsave(64,63,MAXAVE),nflag(MAXAVE),nsave,iseg(MAXAVE) !For msg avg
+  integer nflag(MAXAVE),iseg(MAXAVE)
+  data first/.true./,ns10/0/,ns20/0/
+  save
+
+  if(first) then
+     nsave=0
+     first=.false.
+     ave1=' '
+     ave2=' '
+     ccfblue=0.
+     ccfred=0.
+     if(nspecial.eq.999) go to 900        !Silence compiler warning
+  endif
+
+  naggressive=0
+  if(ndepth.ge.2) naggressive=1
+  nq1=3
+  nq2=6
+  if(naggressive.eq.1) nq1=1
+
+  if(NClearAve.ne.0) then
+     nsave=0                        !Clear the averaging accumulators
+     ns10=0
+     ns20=0
+     ave1=' '
+     ave2=' '
+  endif
+  if(MinSigdB.eq.99 .or. MinSigdB.eq.-99) then
+     ns10=0                         !For Include/Exclude ?
+     ns20=0
+  endif
+
+! Attempt to synchronize: look for sync pattern, get DF and DT.
+  call sync24(dat,npts,DFTolerance,NFreeze,MouseDF,mode,             &
+       mode4,dtx,dfx,snrx,snrsync,ccfblue,ccfred,flip,width)
+
+  csync=' '
+  decoded='                      '
+  deepmsg='                      '
+  special='     '
+  cooo='   '
+  ncount=-1             !Flag for RS decode of current record
+  ncount1=-1            !Flag for RS Decode of ave1
+  ncount2=-1            !Flag for RS Decode of ave2
+  NSyncOK=0
+  nqual1=0
+  nqual2=0
+
+  if(nsave.lt.MAXAVE .and. (NAgain.eq.0 .or. NClearAve.eq.1)) nsave=nsave+1
+  if(nsave.le.0) go to 900          !Prevent bounds error
+  
+  nflag(nsave)=0                    !Clear the "good sync" flag
+  iseg(nsave)=Nseg                  !Set the RX segment to 1 or 2
+  nsync=nint(snrsync-3.0)
+  nsnr=nint(snrx)
+  if(nsnr.lt.-30 .or. nsync.lt.0) nsync=0
+  nsnrlim=-33
+  if(nsync.lt.MinSigdB .or. nsnr.lt.nsnrlim) go to 200
+
+! If we get here, we have achieved sync!
+  NSyncOK=1
+  nflag(nsave)=1            !Mark this RX file as good
+  csync='*'
+  if(flip.lt.0.0) then
+     csync='#'
+     cooo='O ?'
+  endif
+
+  call decode24(dat,npts,dtx,dfx,flip,mode,mode4,decoded,           &
+       ncount,nlim,deepmsg,qual,submode)
+
+200 kvqual=0
+  if(ncount.ge.0) kvqual=1
+  nqual=qual
+  if(ndiag.eq.0 .and. nqual.gt.10) nqual=10
+  if(nqual.ge.nq1 .and.kvqual.eq.0) decoded=deepmsg
+
+  ndf=nint(dfx)
+  if(flip.lt.0.0 .and. (kvqual.eq.1 .or. nqual.ge.nq2)) cooo='OOO'
+  if(kvqual.eq.0.and.nqual.ge.nq1.and.nqual.lt.nq2) cooo(2:3)=' ?'
+  if(decoded.eq.'                      ') cooo='   '
+  do i=1,22
+     c1=decoded(i:i)
+     if(c1.ge.'a' .and. c1.le.'z') decoded(i:i)=char(ichar(c1)-32)
+  enddo
+  jdf=ndf+idf
+
+!  call cs_lock('wsjt24')
+  write(line,1010) cfile6,nsync,nsnr,dtx-1.0,jdf,nint(width),         &
+       csync,special,decoded(1:19),cooo,kvqual,nqual,submode,nlim
+1010 format(a6,i3,i5,f5.1,i5,i3,1x,a1,1x,a5,a19,1x,a3,i4,i4,1x,a1,i8)
+
+! Blank all end-of-line stuff if no decode
+  if(line(31:40).eq.'          ') line=line(:30)
+
+!  if(lcum) write(21,1011) line
+
+! Write decoded msg unless this is an "Exclude" request:
+  if(MinSigdB.lt.99) write(*,1011) line
+1011 format(a77)
+
+!  if(nsave.ge.1) call avemsg65(1,mode65,ndepth,                      &
+!       avemsg1,nused1,nq1,nq2,neme,mycall,hiscall,hisgrid,qual1,     &
+!       ns1,ncount1)
+!  if(nsave.ge.1) call avemsg65(2,mode65,ndepth,                      &
+!       avemsg2,nused2,nq1,nq2,neme,mycall,hiscall,hisgrid,qual2,     &
+!       ns2,ncount2)
+  nqual1=qual1
+  nqual2=qual2
+  if(ndiag.eq.0 .and. nqual1.gt.10) nqual1=10
+  if(ndiag.eq.0 .and. nqual2.gt.10) nqual2=10
+  nc1=0
+  nc2=0
+  if(ncount1.ge.0) nc1=1
+  if(ncount2.ge.0) nc2=1
+
+! Write the average line
+!      if(ns1.ge.1 .and. ns1.ne.ns10) then
+
+!  if(ns1.ge.1) then
+!     if(ns1.lt.10) write(ave1,1021) cfile6,1,nused1,ns1,avemsg1,nc1,nqual1
+!1021 format(a6,i3,i4,'/',i1,20x,a19,i8,i4)
+!     if(ns1.ge.10 .and. nsave.le.99) write(ave1,1022) cfile6,        &
+!          1,nused1,ns1,avemsg1,nc1,nqual1
+!1022 format(a6,i3,i4,'/',i2,19x,a19,i8,i4)
+!     if(ns1.ge.100) write(ave1,1023) cfile6,1,nused1,ns1,            &
+!          avemsg1,nc1,nqual1
+!1023 format(a6,i3,i4,'/',i3,18x,a19,i8,i4)
+!     if(lcum .and. (avemsg1.ne.'                  '))                &
+!          write(21,1011) ave1
+!     ns10=ns1
+!  endif
+
+! If Monitor segment #2 is available, write that line also
+!      if(ns2.ge.1 .and. ns2.ne.ns20) then     !***Why the 2nd part?? ***
+!  if(ns2.ge.1) then
+!     if(ns2.lt.10) write(ave2,1021) cfile6,2,nused2,ns2,avemsg2,nc2,nqual2
+!     if(ns2.ge.10 .and. nsave.le.99) write(ave2,1022) cfile6,       &
+!          2,nused2,ns2,avemsg2,nc2,nqual2
+!     if(ns2.ge.100) write(ave2,1023) cfile6,2,nused2,ns2,avemsg2,nc2,nqual2
+!     if(lcum .and. (avemsg2.ne.'                  '))               &
+!          write(21,1011) ave2
+!     ns20=ns2
+!  endif
+
+  if(ave1(31:40).eq.'          ') ave1=ave1(:30)
+  if(ave2(31:40).eq.'          ') ave2=ave2(:30)
+!  write(12,1011) ave1
+!  write(12,1011) ave2
+!  call flush(12)
+ 
+  if(lumsg.ne.6) end file 11
+!  call cs_unlock
+  
+900 continue
+
+  ccfbluesum=ccfbluesum + ccfblue
+  ccfredsum=ccfredsum + ccfred
+
+! This was for testing message averaging:
+
+!  rewind 71
+!  rewind 72
+!  do i=-5,540
+!     write(71,3001) 0.2 + i*0.057143,ccfbluesum(i)
+!3001 format(2f12.3)
+!  enddo
+!  do i=-224,224
+!     write(72,3001) i*2.1875,ccfredsum(i)
+!  enddo
+!  call flush(71)
+!  call flush(72)
+
+  return
+end subroutine wsjt24
diff --git a/lib/wsjt24d.f90 b/lib/wsjt24d.f90
new file mode 100644
index 000000000..fe429b6f3
--- /dev/null
+++ b/lib/wsjt24d.f90
@@ -0,0 +1,36 @@
+program wsjt24d
+
+  real*4 dat(60*11025/2)
+  character*6 cfile6
+  character*12 arg
+  real ccfblue(-5:540)        !X-cor function in JT65 mode (blue line)
+  real ccfred(450)            !Average spectrum of the whole file
+
+  nargs=iargc()
+  if(nargs.ne.2) then
+     print*,'Usage: wspr24d ifile1 ifile2'
+     go to 999
+  endif
+  call getarg(1,arg)
+  read(arg,*) ifile1
+  call getarg(2,arg)
+  read(arg,*) ifile2
+
+  open(50,file='vk7mo.dat',form='unformatted',status='old')
+
+  do ifile=1,ifile2
+     read(50,end=999) jz,cfile6,NClearAve,MinSigdB,DFTolerance,NFreeze,  &
+          mode,mode4,Nseg,MouseDF2,NAgain,idf,lumsg,lcum,nspecial,ndf,   &
+          NSyncOK,dat(1:jz)
+     if(ifile.lt.ifile1) cycle
+
+!     write(*,3000) ifile,cfile6,jz,mode,mode4,idf
+!3000 format(i3,2x,a6,i10,3i5)
+
+    call wsjt24(dat,jz,cfile6,NClearAve,MinSigdB,DFTolerance,             &
+         NFreeze,mode,mode4,Nseg,MouseDF2,NAgain,idf,lumsg,lcum,nspecial, &
+         ndf,NSyncOK,ccfblue,ccfred,ndiag)
+    if(ifile.ge.ifile2) exit
+  enddo
+
+999 end program wsjt24d
diff --git a/lib/xcor24.f90 b/lib/xcor24.f90
new file mode 100644
index 000000000..b76093b2a
--- /dev/null
+++ b/lib/xcor24.f90
@@ -0,0 +1,94 @@
+subroutine xcor24(s2,ipk,nsteps,nsym,lag1,lag2,mode4,ccf,ccf0,lagpk,flip)
+
+! Computes ccf of a row of s2 and the pseudo-random array pr2.  Returns
+! peak of the CCF and the lag at which peak occurs.  For JT65, the 
+! CCF peak may be either positive or negative, with negative implying
+! the "OOO" message.
+
+  parameter (NHMAX=1260)           !Max length of power spectra
+  parameter (NSMAX=525)            !Max number of half-symbol steps
+  real s2(NHMAX,NSMAX)             !2d spectrum, stepped by half-symbols
+  real a(NSMAX)
+  real ccf(-5:540)
+  integer npr2(207)
+  real pr2(207)
+  logical first
+  data lagmin/0/                    !Silence g77 warning
+  data first/.true./
+  data npr2/                                                        &
+       0,0,0,0,1,1,0,0,0,1,1,0,1,1,0,0,1,0,1,0,0,0,0,0,0,0,1,1,0,0, &
+       0,0,0,0,0,0,0,0,0,0,1,0,1,1,0,1,1,0,1,0,1,1,1,1,1,0,1,0,0,0, &
+       1,0,0,1,0,0,1,1,1,1,1,0,0,0,1,0,1,0,0,0,1,1,1,1,0,1,1,0,0,1, &
+       0,0,0,1,1,0,1,0,1,0,1,0,1,0,1,1,1,1,1,0,1,0,1,0,1,1,0,1,0,1, &
+       0,1,1,1,0,0,1,0,1,1,0,1,1,1,1,0,0,0,0,1,1,0,1,1,0,0,0,1,1,1, &
+       0,1,1,1,0,1,1,1,0,0,1,0,0,0,1,1,0,1,1,0,0,1,0,0,0,1,1,1,1,1, &
+       1,0,0,1,1,0,0,0,0,1,1,0,0,0,1,0,1,1,0,1,1,1,1,0,1,0,1/
+  save
+
+  if(first) then
+     do i=1,207
+        pr2(i)=2*npr2(i)-1
+     enddo
+     first=.false.
+  endif
+
+  do j=1,nsteps
+     n=2*mode4
+     if(mode4.eq.1) then
+        a(j)=max(s2(ipk+n,j),s2(ipk+3*n,j)) - max(s2(ipk  ,j),s2(ipk+2*n,j))
+     else
+        kz=mode4/2
+        ss0=0.
+        ss1=0.
+        ss2=0.
+        ss3=0.
+        wsum=0.
+        do k=-kz+1,kz-1
+           w=float(kz-iabs(k))/mode4
+           wsum=wsum+w
+           if(ipk+k.lt.1 .or. ipk+3*n+k.gt.1260) then
+              print*,'xcor24:',ipk,n,k
+           else
+              ss0=ss0 + w*s2(ipk    +k,j)
+              ss1=ss1 + w*s2(ipk+  n+k,j)
+              ss2=ss2 + w*s2(ipk+2*n+k,j)
+              ss3=ss3 + w*s2(ipk+3*n+k,j)
+           endif
+        enddo
+        a(j)=(max(ss1,ss3) - max(ss0,ss2))/sqrt(wsum)
+     endif
+  enddo
+
+  ccfmax=0.
+  ccfmin=0.
+  do lag=lag1,lag2
+     x=0.
+     do i=1,nsym
+        j=2*i-1+lag
+        if(j.ge.1 .and. j.le.nsteps) x=x+a(j)*pr2(i)
+     enddo
+     ccf(lag)=2*x                        !The 2 is for plotting scale
+     if(ccf(lag).gt.ccfmax) then
+        ccfmax=ccf(lag)
+        lagpk=lag
+     endif
+
+     if(ccf(lag).lt.ccfmin) then
+        ccfmin=ccf(lag)
+        lagmin=lag
+     endif
+  enddo
+
+  ccf0=ccfmax
+  flip=1.0
+  if(-ccfmin.gt.ccfmax) then
+     do lag=lag1,lag2
+        ccf(lag)=-ccf(lag)
+     enddo
+     lagpk=lagmin
+     ccf0=-ccfmin
+     flip=-1.0
+  endif
+
+  return
+end subroutine xcor24
diff --git a/mainwindow.cpp b/mainwindow.cpp
index f84628e83..a258381c2 100644
--- a/mainwindow.cpp
+++ b/mainwindow.cpp
@@ -1,4 +1,4 @@
-//--------------------------------------------------------------- MainWindow
+//---------------------------------------------------------------- MainWindow
 #include "mainwindow.h"
 #include "ui_mainwindow.h"
 #include "devsetup.h"
diff --git a/wsjtx.iss b/wsjtx.iss
index 11c6126df..05bf0060a 100644
--- a/wsjtx.iss
+++ b/wsjtx.iss
@@ -1,6 +1,6 @@
 [Setup]
 AppName=wsjtx
-AppVerName=wsjtx Version 0.5 r2788
+AppVerName=wsjtx Version 0.5 r2791
 AppCopyright=Copyright (C) 2001-2012 by Joe Taylor, K1JT
 DefaultDirName=c:\wsjtx
 DefaultGroupName=wsjtx