Audio doesn't play for one session on one computer

I am collaborating with a partner on a SayMore project. One of the 319 sessions cannot play the audio file (or annotation segments) on my computer. But on my partner’s computer, it does play them. When I copy this particular session from his computer to mine, my computer still does not play these recordings. This is the only session that has this problem. What could be causing this?

Is it possible that this recording is different in some way? Perhaps a different “container” format or a different codec? Does it play using your default player (or any player) outside of SayMore? It’s just possible that the file was corrupted/damaged during copying. It could have happened to land in a place on your hard disk that is bad. You could try doing a full “surface scan” of your drive to check for bad spots. Could also try temporarily renaming the bad session folder and re-copying it again, and see if that fixes it.

I discovered that I have worked with this audio file in ELAN. So when I navigate to the Folder containing this audio file, I can open it in ELAN and it plays there. But it won’t play in SayMore, except on my partner’s machine.

Can you email me a copy of it? Or better yet, upload the entire session folder to Google drive or some other file-sharing service so I can try it? Maybe also send me a copy of a file that does work for you so I can see if I can see any notable difference.

Okay, I just shared a Drive folder with you that includes the session that doesn’t play, and another one that does. Thanks!

Thanks. I am able to reproduce the problem here. I’m looking into what the issue might be.

The immediately obvious difference between the two is that the one that plays correctly was recorded at 48000 Hz and the one that doesn’t was recorded at 44100 Hz. Otherwise, they seem to be the same (codec, number of channels, etc.).

In researching this, I discovered that FFmpeg can convert the file to a file (48000Hz or 44100Hz), which Mplayer (the program SayMore uses to do playback) can play. I also re-remembered that SayMore has a nifty secret feature that enables you to see the output from Mplayer. If you hold down the SHIFT key while clicking the Project menu, there is a command Show Mplayer Debug Window. Then if you press the play button, a bunch of potentially quasi-helpful output will appear in that window. I fed the MPlayer output from the bad file into Chat GPT so it could make sense of it. Here’s my summary:
File Identification and Probing:
MPlayer identified the file as an MPEG-PS format. When it attempted to load the file as an MPEG-PS stream, it failed to find a sequence header, which typically contains important information about the stream.
Audio Decoding Attempts:
MPlayer tried multiple decoders to process the audio stream. Each decoder failed with various errors:
mpg123: Failed with an “Error reading the stream” (code 18).
ffmpeg: Also failed with similar stream reading errors.
libmad: Failed because it couldn’t sync the MAD frame.
hwmpa: Failed due to being unable to sync the MPA frame.
The decoder [spdifmpa] was finally selected as a pass-through decoder for MPEG audio. It identified the audio format as 48000 Hz, 2 ch, mpeg2, 768.0 kbit (which I think is actually incorrect since FFmpeg said it was 44.1 KHz).
However, the audio was in big-endian MPEG-2 format, which was not supported by the filter system, resulting in errors when attempting to reinitialize the audio filter.
I suspect the bottom-line problem is that the version of MPlayer (from Feb 2022) in SayMore is possibly too old to know how to deal with this file, which was recorded later in 2022. MPLayer is no longer actively maintained, but there is a fork of it called MPV which is. I probably need to look into changing SayMore to use MPV.
Meanwhile, I can either send you the converted version of the file that will work or you can use FFmpeg yourself to convert it. (MPlayer has an older version of FFmpeg built into it, but SayMore comes with a more recent version.)