[Openal] AL_SEC_OFFSET is rounded to nearest 0.5s after play/stop/play

[nic at weartronics.com]

Dear OpenAL list,

To answer my own question - This problem, and many other very strange 
problems I encountered are solved by using the "Generic Software" device 
instead of "Generic Hardware" device. So, it seems there are some 
problems with the hardware implementation for NVidia SoundStorm on the 
NForce chipset.

But one problem persists. I am attaching only one buffer to my source, 
and I turn looping on. I note that AL_SEC_OFFSET only returns the 
correct time if the buffer size exactly matches the sampling rate 
(a.k.a. a one-second buffer). If the buffer size is half of the sampling 
rate, AL_SEC_OFFSET returns double the expected time, and if the buffer 
size is double the sampling rate, AL_SEC_OFFSET returns half the 
expected time. I am not varying the pitch or frequency etc... only the 
buffer size. Is there some explanation for this?

Thank you,

Nic

nic at weartronics.com wrote:
> Hi OpenAL list,
>
> If I play a source once and get AL_SEC_OFFSET whilst it is playing I 
> get good results (granularity of AL_SEC_OFFSET is very small). If I 
> stop the source, then play it again, suddenly AL_SEC_OFFSET is always 
> rounded to the nearest 0.5 seconds. I get the same result with 
> AL_SAMPLE_OFFSET (rounded to nearest 22050 samples) and AL_BYTE_OFFSET 
> (rounded to nearest XXX bytes).
>
> Any clues? Thank you!
>
> Nic
> _______________________________________________
> Openal mailing list
> Openal at opensource.creative.com
> http://opensource.creative.com/mailman/listinfo/openal
>
>
>
>