Alternatives to elastic~

[elastic~] uses the Elastique efficient algorithm from Zplane I think, and is limited to double/half speed and pitch.

[elasticx~] and [elasticindex~] sound like a granular algorithms, and are not limited in pitch or playback rate. They sound quite different, and [elastic~] is more suited to small changes, while the [...x~] objects are more suited to when you want to mangle sounds.

It's really easy to patch together a [groove~]-based alternative using [gizmo~], an example of which can be found here:

http://devinkerr.com/category/software/

It probably won't sound as good as [elastic~] on as many different source sounds, but it's free!

I have [elastic~] and the [...x~] objects, and I use the [...x~] ones a lot more often, but only for reeeeally extreme time-stretching. They sound quite warbly and phasey when operating at/close to normal speed/pitch, and only come into their own with extreme settings.

If you end up getting any of them, I recommend patching together an abstraction that uses a standard [groove~] or [index~] when speed and pitch = 1, and then smoothly crossfades over to the elastic objects when you change a parameter.

Thank you for very usefull reply.
".....sound like a granular algorithms": Do you mean that the order of the sound is not cronological with the elasticx~ object? Does it become like a collage/cloud of soundfragments like munger1~ (http://www.maxobjects.com/?v=objects&id_objet=3972)? In that case that sounds like something quite different than the elastic~ (timestretch/pitchshift)

[munger1~] as far as I know, isn't designed to produce natural sounding time-stretching, and the [...x~] objects are. They still sound smooth, but exhibit different artefacts than the standard [elastic~].

It gives you essentially the "same" kind of sound effect, but much more extreme, while still sounding smooth. It doesn't sound like granular synthesis as such, if that's what you're asking.

I could do a couple sound clips for you?

Extreme stretching eh? I'm curious how elasticx~ might go maintaing the character of a voice while time stretching at a factor of around 500 which makes it hard to get a really smooth continuous (and still solo and natural sounding voice) So far I haven't settled on an optimal solution for this yet - I've tried the usual suspects in the spectral domain as well as granular and still haven't decided on the best method yet. FWIW, free_elastic is OK at not so extreme stretches but at the kind of rates i'm looking at its "all over the shop" - seriously jittery and squelchy for want of a better way of describing it...

So far for me, the best subjective overall results for this purpose have been using the GMU externals (bufgranul~) to get some really extreme granular stretches.The windowing wizardy (yet to be achieved on my part) to avoid the AM sub-tone artefacts, or depending on the rate - evidence of discrete windowing becomes an issue when going for a lower number of grains per unit time, and though solving the problem with a really high grain density sounds pretty cool, it ends up sounding more choral and epic than a solo voice in that everything gets smeared and diffuse - which - in this particular case - is not what I want.

With all that said, the GMU externals http://dvlpt.gmem.free.fr/web/static.php?page=max_externals are very useful general purpose granulators given just about all the parameter inlets accept signal rate inputs. And at less extreme stretch factors the results for stretching could be pretty useful for many contexts..

Unfortunately I haven't found any (online) available examples of extreme stretching using elasticx~, so pointers suggestions or examples welcomed.

Edit: Sorry I didn't see the last couple of posts from you guys before I sent this post. Kind of looks out of place now...

Well, [elasticx~] will not give you natural sounding time-stretching at all. What it does is the classic granular epic-expansive-soundscape thing.

I'm interested to hear how natural sounding stretching you've achieved with the GMU externals, as I've not used them. I've done stretching only with standard msp and high resolution objects so far, with no real results in terms of keeping it sounding natural.

I too have used both and agree with all the above comments re the differences/alternatives;

@spectro, I'd be very very interested to hear/see your developments in this area.

@hmarstra, check out this video of some variable rate playback which sounds blissfully artefact free;

http://www.youtube.com/watch?v=IsT9FA2VXqM

the guy is planning a commercial release at some point, so was understandably not giving too much away, but graciously provided me with some parts of the main engine, describing how he achieved this effect. I couldn't decipher it even though it is fairly straightforward (dynamically variable messages to [line~] objects). I can upload it if anyone is interested enough to tinker with it....Tim?

Best
Brendan

I'm well up for some tinkering! :)

Good man!
Did you check the video?

Yeah, I really like his concept of scanning a file and then playing through it generatively with no user input; and the stretching sounds nice as well.