Hi, I tested the latest trunk afternoon (r8378). The timing is not perfect, 16,16,31ms almost constant but we have done a lot of tests and the audio was good if not perfect. I am sure its going to be ok when we start to use it out of the test environment. Thanks for your time and help Anthony! Csaba ps: If I find a way to do better timing on windows i will surely share it with you Anthony Minessale wrote:
I think that is where i have to draw the line and say get some better hardware........ I did my best but if the box sucks that badly at timing i am not sure what else to do because it's working flawlessly on mine which proves it depends on the box and i dont 'even like portaudio
On Fri, May 9, 2008 at 6:34 AM, Sluschny, Thomas <Thomas.Sluschny@siemens.com <mailto:Thomas.Sluschny@siemens.com>> wrote:
ok, now it looks better, i. e. the switch_core_timer_check() times the ReadAudioStream function, but not perfect, now i see packets with differences of 15, 30, 45, 0 ms, e. g. 15, 15, 30, 15, 15, 15, 31, 16, 15, 31, 15, 16, 16, 29 ms and so on
Looks like we already stuck with time resolution, but the switch_yield(1000) in ReadAudioStream() seems to work properly, strange ... I don't have the time to go deeper into and i next read my emails on tuesday :(
I don't want to hijack this thread from Csaba, should we make a jira entry, Csaba? (for Anthonys clients sake :) )
Thomas