David Henningsson wrote:

On 08/15/2011 10:31 PM, Pierre-Louis Bossart wrote:

...

• Increasing the number of packets/urbs solves my power issue but not the

synchronization issue. If I reduce the number of urbs to reduce the interrupt rate, then the accuracy of the hw_pointer is decreased big time and it becomes difficult to synchronize with video.

I think the same thing is a problem for quite a few other devices as well

Yes; sometimes, the period interrupt is the only information from which the driver can derive the position.

I wonder if we need some kind of "pointer granularity" variable to be exported through the ALSA API?

This is certainly possible, and by putting it into the hw_params, it would be possible to model any dependencies between granularity and period size. However, this isn't quite at the top of my steadily-growing ToDo list.

a call to hw_pointer could return hw_pointer and time,

After calling snd_pcm_status(), you have snd_pcm_status_get_avail() and snd_pcm_status_get_(h)tstamp().

and then PulseAudio etc could use that for extrapolation

IIRC PA does exactly this if it detects that the pointer values don't increase smoothly.

Regards, Clemens