View Single Post
Posts: 3 | Thanked: 3 times | Joined on Dec 2010
#82
Originally Posted by mr id View Post
Always interested in metrics. What speed are you running your n900 at, and what was the latency you were getting? I know it's sizeable (it's never going to be real-time and pulseaudio has an impact on this as well
I'm runnin my device at 900Mhz at a low voltage (xlu?)
This is why you should do the rewrite! Simple code and direct Alsa access.
Originally Posted by mr id View Post
Definitely not a bug, sounds like a feature request to me Thanks a lot for the numbers.
Its maybe a feature request, but that's not important now.
No problem, if you need more help on specific tasks as measuring latency and so on, PM me.
I gues tuning pulseaudio would posibly help too. (run htop and sort on +time, and you'll see its on second place after X (opviously)
Originally Posted by mr id View Post
As I said I'm pretty sure that the pole ordering is different between the devices but I would love to be proved wrong - did you get a chance yet?
No sorry, we talked about it at work, but I forgot. But i read some others had dissapointing results, it's like someone did that on purpose
Originally Posted by mr id View Post
I've also been playing around with HEN mode and USB ADC devices but we're missing some kernel/module magic to get the audio devices registered and available. They enumerate and show up in the logs correctly but they aren't registered for audio (they also burn power at an alarming rate).

Thanks again for your feedback.
NP, as this is a interesting piece of sw