Re: [Audacity-devel] Fwd: [Audacity-quality] Fwd: [Mixxx-devel] Using librubberband for tempo and pitch stretching.

[Vaughan J. <va...@au...>]

Thanks for the research on librubberband. Seems we should try it, but 
only after we get a new release out, so we have time to play with it. I 
think Mixxx using it is a pretty strong endorsement, but it's still 
unknown to us.

- V


On 6/23/2013 4:36 PM, Gale Andrews wrote:
>
> | From Richard Ash <ri...@au...>
> | Sat, 22 Jun 2013 22:05:54 +0100
> | Subject: [Audacity-devel] Fwd: [Audacity-quality] Fwd: [Mixxx-devel] Using librubberband for tempo and pitch stretching.
>> On Thu, 20 Jun 2013 14:58:47 -0700
>> Vaughan Johnson <va...@au...> wrote:
>>
>>> Hi. I didn't get any response on this on -quality, but do any devs
>>> know about librubberband?
>> No, sorry - never (knowingly) used anything that uses it.
>>
>> Worth pointing out we are I think at least one version behind on
>> Soundtouch these days.
>
> I feel - given we now have Sliding Time Scale and PaulStretch
> for more extreme stretch - that the main issue we have with
> SoundTouch is the length changes.
>
>>From some off lists I had with Olli (the SoundTouch author)
> then Steve a while ago, there is no prospect of SoundTouch
> "fixing" the length problem because the emphasis of the library
> is running it on complete files, not modifying short selections
> where the proportional length change can be dramatic.
>
> We could perhaps "fix" it by inputting more samples then
> trimming - or repeating samples at the end to fit. But as
> Steve pointed out these changes would affect Preview
> globally, so probably not what we want.
>
> There are some worthwhile quality improvements in the
> latest SoundTouch library though.
>
>>From the librubberband docs:
> http://breakfastquay.com/rubberband/technical.html
>
>   "sample-accurate stretching when run offline"
>
> "Rubber Band implements pitch-shifting using the standard
>   technique of resampling before or after a time stretch, with
>   the combination of resampling in one direction and stretching
>   in the other resulting in  returning the audio to its original
>   duration but with different pitch.  This method has the
>   advantage of making the amount of pitch adjustment extremely
>   accurate and stable, even for small changes."
>
>
>
>
> Gale
>
>