On 13 May 2012 16:16, Bruce Perens <[email protected]> wrote:
> On 05/13/2012 06:32 AM, Andreas Weller wrote:
>
>> It seems codec2 runs fine on armv6l - but BCM2835 SoC is too slow for
>> realtime operation.
>>
> We have to do the fixed-point implementation before smaller processors
> will work.
>
> BTW - The windows version of fdmdv_demod crashes whenever I run it on my
>> Win7 VM.
>>
> A stack trace (with symbols) might be useful.
>
>
Actually I found the culprit line. It seems for whatever reason compiling
to windows (I'm using cygwin) doesn't like the float array being so big:
float rx_fdm_log[FDMDV_MAX_SAMPLES_PER_FRAME*MAX_FRAMES];
I adjusted the value to be lower.
#define MAX_FRAMES 50*35 /* 1 minute at 50 symbols/s */
Which is 35 seconds instead of one minute. This will generate messages to
stdout after 35 seconds, which will destroy any hope of realtime piping.
So, I fixed that too:
fprintf(stderr,"MAX_FRAMES exceed in Octave log, log
truncated\n");
instead of normal printf. That way stderr can be re-diverted. In this
fashion I managed to decode my test modulation all the way to playback on
windows with the following command:
./fdmdv_demod.exe Codec2TT-MOD.raw -|./c2dec.exe 1400 - -|play -b 16 -r
8000 -t raw -s -
In theory I could set record up to record live. However I am not sure what
fdm_demod will do with frames with no codec2.. Hopefully output nothing.
But, I've not tried it.
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Freetel-codec2 mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
