[Openal-devel] (no subject)

[Ramy Sadek]

Hello all, 

 

Implementing AL_BYTE_OFFSET leads me to some confusion. Consider 3 files
FileA, FileB and FileC, which are in differing formats, MONO8, MONO16
and STERO16 respectively. If these 3 files are queued into a buffer,
what is the meaning of calls like 

 

alSourcei(source, AL_BYTE_OFFSET, num) ? 

 

Depending on the order of the files, <num> bytes represents a different
number of samples (and therefore different amounts of time).
Specifically, <num> refers to twice as many samples in the MONO8 FileA
as it does in the 16-bit files FileB and FileC.

 

Furthermore, if implementations are using longer words to represent
samples internally (e.g. 32 bit float samples), the meaning of
AL_BYTE_OFFSET is even less clear. 

 

What is the intended behavior in these cases? 

 

Do implementations need to track all calls to bufferData, to reconcile
the original calls with the values passed in during AL_BYTE_OFFSET
calls? This seems like a great deal of implementation overhead so I'm
hoping I missed step, and you guys can set me right. 

 

Thanks!

 

-Ramy

 

 

 

____________________________________
Ramy Sadek
Lead Software Engineer
Audio Research
Institute for Creative Technologies
University of Southern California
310.448.0353
sadek at ict.usc.edu

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://localhost.localdomain/pipermail/openal-devel/attachments/20060210/ea6a3977/attachment.html