Recording

The managed capture API collects audio in memory while your application does other work. Native capture buffers are drained on a background thread, so the common API returns one PCM value instead of exposing chunks:

from pyalsoft import start_recording, stop_recording

recording = start_recording()
input("Speak now, then press Enter to stop... ")
captured = stop_recording(recording)

start_recording() uses the default input device and records 48 kHz, mono, 16-bit PCM unless told otherwise. Use list_capture_devices() to select a specific input. For a known duration, record() is the blocking equivalent.

Captured and generated PCM values can be passed directly to play():

from pyalsoft import play

sound = play(captured)

See the runnable record_and_play.py example for the complete flow.