Sakonda Granular 2.5 for Live Input?

Any help would be greatly appreciated.
I basically want the live input method to sound as smooth as the sample read version does.
Thanks!

k

I used to have a patch that did this, but it's long been separated into separate subpatchers.

However, it's not difficult. Just use a record~ object in loop record mode to record into the named buffer~ that the granulation part of the patch is reading. You can add a feedback path to the recorder by using play~ sync'd slightly behind the record~.
Obviously the record position and grain play position will cross from time to time, and you'll get clicks. That's either part of the sound, or you could use something like tap.click~ to reduce it.

It's honestly easy to do (I figured it out in my early maxing days!), but if you're really stuck, I'll see if I can dig out the older versions of the patch.

David

this shouldn't be a problem at all, i used sakodna's patch as a basis for my first few live-sampling granulators. you've left out your version the abstraction grain2.5live~ so i can't use the patch you attached to see what's wrong with it! attach it so we can see...
n

your patch worked nicely for me although I had to create a copy of grain2.5live~ by changing grain2.5~'s references to the buffers you set. I also shortened the time of the buffer from (I think it was) 50 seconds, down to about 5, or 2.5 or something.
Worked a treat..

All above advice stands...

Enjoy.
DF

Hi guys,

Thanks very much for your input!
I figured it out...it was just a dumb oversight on my part...
I had the wave~ objects referencing the record buffer instead of the window buffer.
I've cleaned up the patch and have reposted this functional version. If you come up with any suggestions for improving the input/record architecture please let me know. I'd love to hear any ideas.

Thanks again!

k

koriifuraa wrote on Wed, 25 February 2009 14:43
I've cleaned up the patch and have reposted this functional version. If you come up with any suggestions for improving the input/record architecture please let me know. I'd love to hear any ideas.
Thanks again!
k

Nice. Thanks for sharing!

working on creating my own live version of this patch. so far i've managed to condense it down to use less sends/returns and subpatches, reduced extraneous features and swapped the rslider with waveform~.

i have a question though. would converting the grain2.5~ abstraction into a poly~ make the patch any more efficient? i saw this method being used in the granularized example patch. not sure how this will work with each grain having a different phase though and i've never used poly~ before.

hey cudnylon,
I'm on the same tip, re-jigging the sakonda patch for my own usage, I'm just trying to figure out how to swap the rslider with waveform~, could you possibly point me in the right direction?
T'would be much appreciated.

Cheers, DF

plug the third outlet of waveform~ (selection start) into the left inlet of the "- 0" object.

for position randomness, subtract the selection start from selection end. expr object is best here. (ex: selection start = 30, selection end = 200, 200 - 30 = position randomness of 170). plug this value directly into the position randomness flonum.

now you can remove the rslider, "p slider_adjuster" object, and the "+ 5" object. if you need any more assistance let me know.

DF wrote on Thu, 16 July 2009 12:53hey cudnylon,
I'm on the same tip, re-jigging the sakonda patch for my own usage, I'm just trying to figure out how to swap the rslider with waveform~, could you possibly point me in the right direction?
T'would be much appreciated.

Cheers, DF