Closed luarvique closed 4 months ago
Thanks for the report, I'd be happy to accept a PR for this issue.
I have not got to the point of finding the exact culprit inside multimon-ng. Will try looking at it this upcoming week, but if you know where it may be, don't wait for me =)
This issue is related to: #197 #91 #70
Fixed with pull request #214. Came down to incorrect subsample tracking between separate portions of incoming data. The same issue might be present in some other decoders, I have not checked.
I am seeing the same subsampling issue in demod_fmsfsk.c. You may want to check the rest.
Thanks for the report, I'd be happy to accept a PR for this issue.
The PR is waiting for your review and approval. I will wait fixing remaining decoders (such as FMSFSK) until the current PR is merged in.
PR got merged, so closing this one. Will make similar fixes to other decoders that use the same subsampling method.
When running multimon-ng 1.1.9 executable and feeding it IQ input via a pipe, POCSAG decoders spew garbage unless the data is written in chunks of at least 512 complete IQ samples at a time. The FLEX decoder appears to be unaffected though.
The problem was found when running a background POCSAG/FLEX decoder in OpenWebRX+ as opposed to running it interactively. In the interactive mode, OpenWebRX+ has a Squelch() module in its DSP chain that buffers data in 1024-sample chunks. In the background mode, where Squelch() is not present and data arrives in smaller increments, POCSAG breaks.
I can obviously implement buffering at the input to multimon-ng, but it probably better be addressed in the multimon-ng code as well.