PyAudio 中从麦克风到扬声器的实时流式传输

Question

I'd like to stream audio in real-time from mic to speaker using PyAudio, with an opportunity to read / modify / write the sample buffers as they go by.

What is the idiomatically correct way to do this in PyAudio?

I understand that in callback mode, the output stream driving the speaker wants to "pull" samples in its callback function. Similarly, the input stream consuming samples from the microphone wants to "push" samples in its callback function. I also understand that callbacks run in their own threads, and that the docs say:

| Do not call Stream.read() or Stream.write() if using non-blocking operation.

Given those constraints, it's not clear how to connect a microphone's stream to a speaker's stream. (And I understand the complexities if the microphone and speaker clocks are not synchronized.)

Assuming that the microphone and speaker clocks ARE synchronized, how would you stream from mic to speaker?

Answer 1

[As soon as I hit [send], the following occurred to me. Let me know if its the right idiom for PyAudio...]

You could allocate three frames ahead of time and cycle through them: one gets passed to the microphone callback, one is available for read/modify/write processing, the third gets passed to the speaker callback. (You might actually need four frames to allow for latency and delays.)

I didn't see any documentation on the format of the frames themselves; are they just arrays of ints (for 16 bit sample data)? [UPDATE: "out_data is a byte array whose length should be the (frame_count * channels * bytes-per-channel)"] So that's easy...