Coalescence: neural concatenative multisampler

VST, AU, AAX, CLAP, etc. Plugin Virtual Instruments Discussion
RELATED
PRODUCTS

Post

Announcing the release of Coalescence today! This is an instrument that uses a neural network (SOM) to organize sample slices into clusters based on their spectral features. This device offers you three different playback modes that take advantage of this clustering to create concatenative sampling. Additionally, you can route external audio to control what sound is playing based on the similarity. This is a really powerful yet compact sampler that focuses more on being a versatile and expressive instrument rather than accurately sorting a sample library (and that is the key difference between this device and similar ones on the market). It can be used in all kinds of contexts, from conventional to experimental, too much to list briefly!

Teaser video:


Walkthrough video:


You don't need to be connected to the internet to use the device, no servers, the machine learning is done internally on a light neural network.


There is sooo much involved in this device so you'd need to check the product page for more info:
https://dillonbastan.com/store/maxforli ... oalescence

Image

Hope it is fun and helpful!

Post

such a creative device. instant buy. unique sampling instrument.

Post

I missed info about it being Ableton only in the first post.

Post

Is it based on MuBu the work of Ircam’s Diemo Schwartz?

Post

No its not. I coded all the analysis and SOM (neural network) in gen, I dont use externals as my work is commercial.

Post

Bummer it's M4L only.
Duality without regard to physicality

Post

I bought it just to be supportive, been waiting for a concatenative synth using evolutionary algorithms for 15 years. Kinda surprised no one ever made one based on my rants. I have many Ableton licenses I've never activated from buying gear, maybe I'll try it out now. If anyone especially wants my license you're welcome to it, $30 well spent to encourage advancement of our artform regardless.

Post

I usually use Logic Pro X for my sessions, but I upgraded an old copy of Ableton Live to version 11 so that I could have access to Max and the enormous selection of M4L devices. I do wish this was available as an AU or VST, but I'll gladly switch DAWs to use unique software like this. Much respect to the developer. Thank you for thinking outside of the box and creating a device that is loads of fun for people like me, who enjoy experimenting for hours!! :tu:

Post

runagate wrote: Tue Dec 12, 2023 12:10 am I bought it just to be supportive, been waiting for a concatenative synth using evolutionary algorithms for 15 years. Kinda surprised no one ever made one based on my rants. I have many Ableton licenses I've never activated from buying gear, maybe I'll try it out now. If anyone especially wants my license you're welcome to it, $30 well spent to encourage advancement of our artform regardless.
Well I do have this device which is an evolutionary granular synth which you can use multiple sources so you couuuld describe it as an evolutionary concatenative sampler.. but maybe not
https://dillonbastan.com/store/maxforli ... ain-forest

And thanks for the support :)

Post

morphex wrote: Tue Dec 12, 2023 12:59 am I usually use Logic Pro X for my sessions, but I upgraded an old copy of Ableton Live to version 11 so that I could have access to Max and the enormous selection of M4L devices. I do wish this was available as an AU or VST, but I'll gladly switch DAWs to use unique software like this. Much respect to the developer. Thank you for thinking outside of the box and creating a device that is loads of fun for people like me, who enjoy experimenting for hours!! :tu:
thanks!

Post

Wow! (bookmarked page) Inspiring stuff, man. When I upgrade from Live 9 (suite) to a newer version, I will absolutely be getting into a number of your products.
I'm not a musician, but I've designed sounds that others use to make music. http://soundcloud.com/obsidiananvil

Post

Intriguing.
You can be creative in any right place on Earth, and not only in the wealthiest cities. Bring the world feelings from everywhere, and not only feelings of capitalistic or jail environment.
― Aleksey Vaneev


https://linuxdaw.org

Post

bought. will be fun to see to where i can go with it.
I have there some funny ideas creating specific folders.
For example cut up some speech/single words by hand upfront.
add into one folder. load.
"Plugin has turned Drug now"....and the business knows it.

Post

Funky40 wrote: Wed Dec 13, 2023 6:04 am bought. will be fun to see to where i can go with it.
I have there some funny ideas creating specific folders.
For example cut up some speech/single words by hand upfront.
add into one folder. load.
Indeed! 8) I thought about doing likewise with short lines from old (public domain) sci-fi movies, and TV shows. Could prove quite entertaining, if nothing else. :tu:
I'm not a musician, but I've designed sounds that others use to make music. http://soundcloud.com/obsidiananvil

Post

Cool!! Keep in mind the neural network organizes slices by their spectral features PER SPECTRAL FRAME (1024 samples), so if you are looking for it to cluster sound by spectral features over multiple seconds it won't work like that (it'll still cluster them but maybe not meaningfully for your use case). For individual vowels/consonances and other speech sounds should work better.

Post Reply

Return to “Instruments”