For longer samples (~ >= 10s) things seem to be OK but for shorter samples e.g. 5s or smaller the first iteration plays fine but subsequent iterations speed up the speech (as if speaking faster). Also I think lesser sampled pieces have more issues i.e. 16000 Hz sample plays worse than 32000 and upwards.
Need help here on why is this happening !
I have audio samples that can help you check this on your side. Please find them at GitHub - pankajidwallet/audio_samples: audio_samples to help debug the wav issue
Although I’m not sure why the audio loop is not working for the linux raw recording sample, if you’re open to using an API, you could use this meeting bot API that shows you how to output audio/video in the bot instead.
@chunsiong.zoom I have already done that, used the correct sampling rates in code when sending.
If you use the clips I have provided and run the loop a few times with respective sampling rates, you will see the issue.