Thank you!!!! My problem was that I was just connecting the output unit to a mixer, then grabbing the audio data and timeStamp through a render notify (which must be the output timestamp). I followed your instruction and set it up with kAudioOutputUnitProperty_SetInputCallback and am now getting predictable (and sensible) timestamps. Now all I have to do is go through everything and un-hack my hacky latency compensation.
Thank you, Dave On Mon, Apr 13, 2015 at 10:36 AM, Dan Klingler <[email protected]> wrote: > > How can a future time stamp represent a buffer of samples from the > microphone that has already been captured. > > > You’re correct, host time for an input buffer should be less than > mach_absolute_time(). For input, you should look at the timestamp that’s > passed to you as part of your input callback (the one set on the AU with > kAudioOutputUnitProperty_SetInputCallback). I would expect this host time > to be less than mach_absolute_time(). > > For input, you’re the one calling AudioUnitRender (from the input > callback), so you should call AudioUnitRender with the timestamp you get in > the input callback. > > Dan
_______________________________________________ Do not post admin requests to the list. They will be ignored. Coreaudio-api mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/coreaudio-api/archive%40mail-archive.com This email sent to [email protected]
