0
0

Hi Brett, I noticed that FMOD 3.5 has a buffer of 1024 samples (taken from the parameter of a dsp callback), while FMOD 3.4 had a buffer of 1104 samples. Everything at 44100 Hz. What happened there? With some calculations, I saw that the mixing buffer is only 23.2 ms long!
Why ??

  • You must to post comments
0
0

I don’t know ofcourse, but I would say that it gives less latency, but more important, functions such as calculating the spectrum (fft) can have a lot of optimalization if done on a power of 2 samples.
Maybe it’s even faster for the soundcard buffers…

  • You must to post comments
0
0

That’s quite strange, so…

I got this behaviour on my Atlhon, that’s really not a Pocket PC…

I know something about optimizations with FFT. Really I know that FFT MUST be used with buffers that are power of 2. However this isn’t the problem.

  • You must to post comments
0
0

Uh! Got it… :smile:
I’m not really a master in English! Sorry!

However I should say that documents that come with fmod say that buffers are still 25 ms. For a syncrhonization routine this isn’t really a help. :sad:

About FFT, I read some docs, a lot of time ago, from Don Cross. He said that buffers must be always a power of 2, else the filter can’t be applied. What kind of FFT you used?

Thanks! And excuse me for my bad english…

  • You must to post comments
Showing 3 results
Your Answer

Please first to submit.