[LAD] [LAU] Help on rubberband

[Chris Cannam]

On Wed, Jan 13, 2010 at 6:40 PM, gerald mwangi <gerald.mwangi at gmx.de> wrote:
> Hi, has someone looked at the code? I really need an answer to the question.

Sorry, I didn't notice this earlier on LAD because the subject line
("Is TerminatorX development stalled?") bore no relationship to the
substance of the query.  No slight to TerminatorX intended, but I
wouldn't normally look at messages on LAD about it!

> I have a question to for the devs out there though: It seems to me that i
> have to run process() a few times with a fixed blocksize before,
> getRequiredSamples() returns something >0 in Realtime-mode. All other
> options are default options. Is this true?

You have to supply process() with a certain amount of data before
anything becomes available at output, yes.  If you are using a small
fixed block size, then yes it is true that you will likely have to
call process() several times.  You might like to consider querying
getLatency() first, after setting the initial ratios, and "priming"
the stretcher with that number of zeroed samples before you start.


Chris