Dear pd-list,
I have a query regarding recording/saving sound files in a 16kHz samplerate while pd is set to 48kHz.
I have created a patch that utilizes Whisper ((https://github.com/ggerganov/whisper.cpp)) for automatic speech recognition. The shell object is used to call Whisper, and the result is piped back to pd. This process works well when pd is set to 16kHz. (Whisper only supports 16kHz files.) Although it worked in real-time with a bit of tweaking.
Unfortunately, I am unable to find an optimal solution to this issue as block~ cannot be used due to its inability to handle a sampling factor of 3. I attempted using pd~ by setting the samplerate to 16kHz and piping audio from my host pd, which has been set at 48kHz, but this also does not lead to a good result.
Do you have any suggestions or solutions to this issue?
Best regards, d