PYNQ: PYTHON PRODUCTIVITY FOR ZYNQ

PYNQ Audio Streaming

I want to create a real time speech to text on the PYNQ. I have found the PYNQ Audio module accessible using the Base Overlay, but there doesn’t seem to be a built in way to access the audio recording as a stream. I’ve thought of buffering the output of the .record(time) function using multiple threads but it seems I would be dropping some audio time in doing this.

Is there a built in way to achieve this real time audio stream using existing libraries or should I look towards messing with the Audio IP in Vivado?

Are you using z1 or z2? If I remember correctly, z2 already offers a bypass method which allows you to stream audio out. Then you can play with the audio C driver in pynq/lib/_pynq/_audio. If you are not satisfied with the performance for some processing, you can try adding new vivado IP blocks.