[LAU] Timestretch without sample damage.

J M Needham J.M.Needham at bath.ac.uk
Thu Jan 3 09:15:23 EST 2008


Well, using Audacity, I found silence at the end of the sample, and using
Rubberband I found a small chunk of silence around 10ms in from the start
of my sample. That was a wav which was reduced in time from around 3
seconds to 2.666 using cl. Would you like me to make the samples
available?

On Thu, 3 Jan 2008, Chris Cannam wrote:

> On 03/01/2008, J M Needham <J.M.Needham at bath.ac.uk> wrote:
> > The audacity time stretcher is great, rubberband is even better but they
> > both put chunks of silence into the samples when I use them.
>
> Do you mean silence within the sample somewhere, or do you mean that
> the duration of the resulting sample isn't exactly correct and so
> there is silence at the end?
>
> The former would be very strange and quite wrong, particularly if it
> happened with both Audacity (which uses SoundTouch) and Rubber Band,
> as the two use completely different algorithms.  Since both run the
> stretcher offline, they also shouldn't depend on any realtime
> processing constraints.
>
> The latter effect would be expected using Audacity, as SoundTouch
> isn't sample-accurate, but it shouldn't happen when using the Rubber
> Band command-line utility.
>
>
> Chris
>




More information about the Linux-audio-user mailing list