semantics Re: variable block size ( was RE: [music-dsp] [OT] Stupid C Question )

robert bristow-johnson rbj at surfglobal.net
Tue Dec 28 16:50:23 EST 2004


on 12/28/2004 14:24, David Lowenfels at dfl at alum.mit.edu wrote:

> 
> On Dec 27, 2004, at 5:21 PM, Christopher Weare wrote:
>> 
>> I agree that processing blocks of samples helps performance but defeats
>> what is trying to be accomplished.  The moment block processing comes
>> in
>> to play, barriers must be erected.
>> 
>> Try building an IIR from primitive blocks processing 10 samples at a
>> whack.
>> 
>> -chris
> 
> Can't this easily be fixed by allowing variable blocksizes? And if you
> keep them limited to powers of two, you won't have the mess of
> partially overlapping frames.
> 
> In PD, you can make a sub-page and give a blocksize command that makes
> it run at a different rate than the main program page.

may i suggest that we use the word "block" to describe the block of samples
(or "chunk" or "vector" or "buffer") that are processed and that we use the
word "module" to describe the "microscopic" processing algorithms that are
processing these samples.  otherwise we might have to call these blocks of
samples "whacks" and the number of samples in the block the "whacksize".


-- 

r b-j                  rbj at audioimagination.com

"Imagination is more important than knowledge."





More information about the music-dsp mailing list