Forums > MaxMSP

elastic


Feb 25 2014 | 1:38 pm

hi,
does somebody elastic with max6?
does it work correctly?

thank you
laurent


dtr
Feb 25 2014 | 2:20 pm

elastic?

Feb 25 2014 | 3:57 pm

Hi
I’m assuming you mean elasticx~.
If so, this external was built for a 32-bit architecture, an older version of Max; if you have 64-bit Max6 it will not work. Download the 32-bit version of Max6 instead perhaps?

Brendan

Feb 25 2014 | 8:24 pm

I gave up on it some time ago due to complete lack of support after running into some problems. Switched to the IRCAM SuperVP, which works great.

http://anasynth.ircam.fr/home/english/software/supervp

Feb 26 2014 | 2:33 am

I got SuperVP after having issues with elasticx~, but it sounds completely different. Good for some things, but unfortunately completely wrong for most of what I want to do.(It sounds much more "electronic" than elasticx to me, which might be fine for some things, but is way off for my sound).

So I’ve persevered with elasticx~, and under most circumstances it’s ok for what I want to do. In other circumstances, like live performance (!!!!), it has a tendency to blow up (and stop audio working), and require a complete restart. So if you _do get it working proceed with caution!
(Just playing through a prerecorded buffer seems to be safe. It’s when there are other processor intensive things going on that it seems to become unstable)
And as Brendan said, you’ll need to work in 32bit mode.

It’s a real shame that it has these problems, and I guess it never made enough money for the guys who wrote it to carry on with it. None of the other time stretching patches sound quite like it (though they each have their own distinctive, interesting sounds), so it’s a pity that it appears to be end-of-line.

Feb 26 2014 | 4:02 am

I completely agree with David re the subjective sound of elasticx~ and also its denoument; how does it compare with grainstretch~ I wonder David?

Brendan

Feb 26 2014 | 7:16 am

hi Brendan,

Well of course I had to try that out. grainstretch~ is much closer to the sound I like than SVP is, but it still lacks a certain je ne sais quoi. I think it sounds a little more pulsed, less overlap between the grains (less smooth), and other things I’m not really clear on (maybe a bit too "clean"). There’s also the issue of it looping constantly, and so having to create a restart/muting scheme for it. But I’m going to add it to my multi-method "granular workspace" patch as a source of a slightly different sound.

Feb 26 2014 | 7:19 am

Just found the grainsize & fade parameters. Hmm. Maybe i can get closer to the elasticx~ sound than I thought. Maybe i also need to have more of a play with the parameters in SVP.
(edit) and the graincloud settings. Hmm – I’m getting more impressssed with grainstretch~.

Feb 26 2014 | 12:26 pm

thank you very much

but i work in 64-bit,
where i can find grainstrech?

Feb 26 2014 | 12:53 pm

Search the forum, or look on maxobjects.com

Feb 27 2014 | 4:17 am

[OT -ish]

David
I’d be interested to see/hear your own Max efforts in this sphere, if you’ve attempted DIY stretchy/freezy goodness recently. I tend to adhere to phase-locked overlapping fat sine windows, and despite numerous failed algorithms I cannot decorrelate grainsize from pitch. The simple algorithm I use is clearly illustrated in the attached patch; perhaps you could [learn from or] improve it?

Cheers
Brendan

Feb 27 2014 | 4:20 am

upload error:

Attachments:
  1. 2014-Grains.zip
Feb 27 2014 | 7:07 am

Hello Meister,
I would love to learn more about your patch, techniques and theory applied there, etc.
Do you have documentation/comments about it?

thanks

Feb 27 2014 | 7:31 am

Hi loadmess
without my derailing the thread, I can summarise it here. The idea of driving a [play~] object and it’s associated grain window, with a core phasor, and then overlapping each grain via phase offsetting goes back to Nathan Wolek, Chris Dobrian et al. The idea of a single central phasor – to avoid phase de-synchronisation during parameter changes – was offered here by Peter McCulloch. The fat sine window is my own implementation, but is also suggested by Dobrian and others. The allpass filter(s) on the output are also my own idea, in that they subjectively blur AM arefacts, and also allow for user-defined "reverb" too.

If you want a more detailed description, then by all means pm me

Best
Brendan

Feb 27 2014 | 7:38 am

ps

there are several developments/enhancements yet to be implemented . . .

Feb 27 2014 | 3:19 pm

hey Brendan,

I have to admit I’ve never delved _that deep into the programming – the closest I’ve gotten to that is making some mods to the core of sakonda’s patch, which is where I got started with this. The "granularworkspace" patch is a sort of combination of all of the granular methods & associated controls that I’ve been using in assorted patches, all brought together in one place for creating sounds that I then assemble in Mixbus or Cubase – as opposed to the aforementioned "assorted patches", which are for live work.

Having said that – I’ll definitely dl your patch & have a look! thanks.

Feb 28 2014 | 3:14 am

hey Brendan – that’s really rather good! I like the sound – very smooth. I see what you mean about the pitch moving (down and back up) when you change grain size, and also changing pitch seems to move the playhead. Whether or not that’s an issue (apart from perfectionism :-) ) I think depends on how you’re going to use it. For the kinds of thing I do none of that would (at least as far as I can imagine it at the moment) be a problem. Hope you don’t mind me adding it to my "workspace"!

Feb 28 2014 | 8:06 am

Yes, I like the smoothness too, which of course is a characteristic of the phase-locked grains. I have previously got decorrelation between pitch and size, but without the phase-locking. I’ve really been round the houses with this one, and the version you have is the "best" so far; varying pitch shouldn’t affect the playhead position (as it’s phase-locked) – it will just slightly alter the grain size, but I can fix that . . .

(this workspace of yours must be looking pretty busy right now)

Feb 28 2014 | 9:48 am

:-)

6 elasticx~/grainstretch~
3 svp
A differently controlled elasticx
A sakonda
1 granny (that’s your object :-) )

And not shown is a patterner module that creates either random pulses within a buffer,
or straight pulses/euclidean rhythms playback of short samples. (That’s how i created the rhythmic sections in the last couple of electroacoustic poesetics pieces I’ve put up on my soundcloud page – soundcloud.com/davidestevens if you’re interested. Check out Morgenkuss, December Sun (both long) and the extremely short Unbecomingun.)

Attachments:
  1. granularworkspace

    granularworkspace.jpg

Mar 03 2014 | 2:42 am

@loadmess, and perhaps David too

here is a descriptive and annotated example of my hybrid granular synthesis/playback engine. "Hybrid" because it features a few control dimensions from each approach, without offering comprehensive control of the parameters common to both – that is to say it doesn’t offer control of grain envelope or panorama, but it does do stretching and repitching quite well.

I have also decorrelated pitch from grain size, and added some pitch mod control too btw.

Brendan

Attachments:
  1. genGrain-2014.zip
Mar 03 2014 | 9:20 am

thanks for this Brendan.

presumably I can just plop this in in place of the previous version and it’ll behave the same?

Mar 03 2014 | 12:17 pm

Yes David
it should drop right in, there are no major changes besides the addition of 2 controls for pitch randomisation. Furthermore, inside this [p pitchCalc] subpatch I changed the static grain size from 80ms to 40ms. This will make it sound a little grainy-er. Set this value to >80ms for smoothyness (yes, it’s a word!).

Please do feel free to change/distribute/improve etc

Best regards
Brendan

Mar 03 2014 | 12:34 pm

This has been bugging me for days, and I just got it – DiracLE~ !
I knew there was another quality time-stretching/pitch-shifting object out there, but couldn’t remember the name!
http://timorozendal.com/ – any use to you?
Cheers
Roger

Mar 03 2014 | 12:57 pm

This is the sound I am chasing. But I wouldn’t know how to incorporate externals for pitching etc, such as DiracLE.

Might just change my handle to "wheel-reinventor"; grainstretch~ is indeed pretty damn perfect.

Mar 03 2014 | 1:53 pm

hmm. Just tried out DiracLE~, and it _is very smooth. In fact, way too smooth for my taste. I can still understand the source material! I’m liking genGrain~, but I have to say that elasticx~ is _still my preferred sound.

Mar 04 2014 | 4:50 am

It is a bit smooth, isn’t it?!
When it comes to roughness, the one I miss is Topher Lafata’s stretch~ object – proper Old Skool drum-n’-bass bad time-stretching, but I don’t think it even made it to UB, let alone 64-bit Max.
Puckette’s overlap sample looper patch ( http://msp.ucsd.edu/techniques/latest/book-html/node37.html ) comes somewhere close in roughness, especially with a low precession value. Attached is a rough translation,
Cheers
Roger

Viewing 26 posts - 1 through 26 (of 26 total)

Forums > MaxMSP