Just to confirm that there is no neural network involved, the name is alluding to the network of interconnections between blocks that you explore when using "synaptic" or "synaptic slide" modes. I'll make this clearer in the documentation as it's a good point (I think this was less confusing back in 2015 before all the current AI hype). Thanks for the awesome video!
Cool video Alessandro! Btw you can load more than one sample at once, using the directory import feature, it imports an entire folder of samples in in one go. Could save some time!
A long time ago, maybe around the 2002 ish era some people demonstrated how you could take existing video (i.e. a Michael Jackson interview was their pool of data) and it would analyze the audio similarities. Then a person could speak a given sentence, and it would find sonically similar aspects, chop up the video (i.e. the Michael Jackson interview) and "have michael jackson speak for you". Probably the same idea as this ^ but less parameters. It was cool! Unfortunately I can't find this video. but they showed the diagram of the spectral breakdowns and matching similarities from sections of one large bank/database crossed over to another.
I remember this video. Was heavily into pure data at the time. I think there was some open source code around at the time that you could use to try this on your own content.
I don’t think it’s a prank. I’ve gotten quite musical results from it with not a ton of messing around. Using the different window shapes and overlap sizes.
I already recreated a bunch of my favorite drum loops with this, and created new insane sounding versions of samples I already have and ones that are completely unique
Nice demo. "Glitchy as hell" sums up this tool (in its current incarnation) perfectly. I hope they evolve more ways to easily tame and soften the resulting sounds.
dunno if this is a goal for samplebrain but for a long time i’ve thought it would be really cool to have something that had a source and destination buffer, and used only audio from source, chopped, stretched, and layered to recreate source as closely as possible within given restrictions, like if the minimum grain size is a single sample then it’s going to not change much, while if there’s a maximum length and dest uses a sample in source, it could make it less likely to just paste the whole sample in. and then layering would try to match the difference between dest and result, which would then cause result to get closer and closer to dest
Thanks so much for this overview, it’s a 4king ballistic tool…. What more can we expect from the super lateral vision brain of Aphex Twin. MangletastiQ
@@AlessandroMastroianni thanks for reply - I do this and still get the pretty short left and right .wav files. Also if having to hold down record, I cant amend any of the parameters!? Surely theres a better way
@@Jacob-yb6bv Your own minds are setting the boundaries here. Pretty much anything, pointless or not, can become a really cool paperweight in the correct context. But unlike a paperweight these types of glitches can send your creativity afloat, an adventure that's not for everyone. I wish you all the best on your own journey, glitched or not. :)
I didn’t, I’m guessing it picked up that my system has dark theme on (I’m on Big Sur) and it loaded that way, and I can’t find a way to change that. In fact o don’t like it because some parts are unreadable.
You can have a look at my other video on this but there isn’t much to it other than unpacking the installer and tell your mac that it’s from a trusted developer
@@samhein321 Yeah I downloaded the mac one but the issue seems to be that my mac thinks it's a virus and there's no option in my settings to allow for foreign downloads
Everything sounds to be running at the same speed with no variation. I don’t see the point of this - there must be more tweak ability otherwise this just seems like a moat that Aphex has put around his castle for fans to drown in as they try and figure out how he really makes his sounds. I’ve also been working on a ‘superdrummachine’ for about 15 years in PureData and hopefully it sounds a lot better than this!
I have no idea why these videos exist beside promoting the plugin.. No one seem to have ANY idea what they are doing (can't blame them...) with it. It's like Richard is playing a game of 4D chess against all these "internet-guys" at the same time, making them look like fools.
Just to confirm that there is no neural network involved, the name is alluding to the network of interconnections between blocks that you explore when using "synaptic" or "synaptic slide" modes. I'll make this clearer in the documentation as it's a good point (I think this was less confusing back in 2015 before all the current AI hype). Thanks for the awesome video!
Thank you so much for stepping in and making up for my ignorance :)
It’s an amazing piece of software
Really great project dave!🧠🧠🧠
Essentially what you are trying to say is that it's pretty dope.
Used for 10 minutes, blown away. Had great results using drum loops as Target.
Ottimo video, è la prima volta che ne vedo uno tuo ma credo proprio che non sia l'ultima ;)
Devo assolutamente provare SampleBrain
Cool video Alessandro! Btw you can load more than one sample at once, using the directory import feature, it imports an entire folder of samples in in one go. Could save some time!
Yes haha I found out about it after shooting… wish I knew it about it sooner
A long time ago, maybe around the 2002 ish era
some people demonstrated how you could take existing video (i.e. a Michael Jackson interview was their pool of data) and it would analyze the audio similarities.
Then a person could speak a given sentence, and it would find sonically similar aspects, chop up the video (i.e. the Michael Jackson interview) and "have michael jackson speak for you".
Probably the same idea as this ^ but less parameters.
It was cool! Unfortunately I can't find this video. but they showed the diagram of the spectral breakdowns and matching similarities from sections of one large bank/database crossed over to another.
Now I want to see this video!
Lucky for you both there’s a link at the bottom of the samplebrain GitHub, it was called sCrAmBlEd?HaCkZ! ruclips.net/video/eRlhKaxcKpA/видео.html
I remember this video. Was heavily into pure data at the time. I think there was some open source code around at the time that you could use to try this on your own content.
While not exactly what you're talking about, googles magenta can do something similar
I don’t think it’s a prank. I’ve gotten quite musical results from it with not a ton of messing around. Using the different window shapes and overlap sizes.
Indeed
lmao
I already recreated a bunch of my favorite drum loops with this, and created new insane sounding versions of samples I already have and ones that are completely unique
It’s not a prank. I hear where he used something like this in his more recent SoundCloud songs.
great video! thank you!
CataRT made by Ircam was doing this for ages - concatenative synthesis, sound mosaicing, corpus based synthesis.
I guess now we know how he made the garbled voices in Syro.
😂
maxmsp, supercollider and player pro on mac
5:45 there's a directory button, dump your samples into one folder and use that to import
I came for Samplebrain, great video. Then I realized you are behind the Dronar instruments. So awesome. Subbed!
Nice demo. "Glitchy as hell" sums up this tool (in its current incarnation) perfectly. I hope they evolve more ways to easily tame and soften the resulting sounds.
really nice!
Don’t know if “nice” is the right word for it 😂
dunno if this is a goal for samplebrain but for a long time i’ve thought it would be really cool to have something that had a source and destination buffer, and used only audio from source, chopped, stretched, and layered to recreate source as closely as possible within given restrictions, like if the minimum grain size is a single sample then it’s going to not change much, while if there’s a maximum length and dest uses a sample in source, it could make it less likely to just paste the whole sample in. and then layering would try to match the difference between dest and result, which would then cause result to get closer and closer to dest
Sei un genio, grazie!!!!
Thanks so much for this overview, it’s a 4king ballistic tool…. What more can we expect from the super lateral vision brain of Aphex Twin. MangletastiQ
Avocado does similar things...this looks fun tho!
I don’t know this one!
great walkthrough, whats the best way you think to record/export much longer samples - so you can record in app automation etc?
You can just hit record and play for as long as you like I think, the buffer will record everything. I have a 5 minutes file I did recently..
@@AlessandroMastroianni thanks for reply - I do this and still get the pretty short left and right .wav files. Also if having to hold down record, I cant amend any of the parameters!? Surely theres a better way
@@elshuffles melda production have a free recorder VST that literally just records whatever is happening in the DAW to a wav file. You could try that?
Awesome modular snippet!
Kontact was Majora's Mask vibes
Subscribing! Thanx!
Thank you! 🙏
@@AlessandroMastroianni 🤝
I get a message after hitting open anyway that basically says samplebrain quit unexpectedly. Anyway to fix this?
this neural network is truly based.
mystery
I feel quite ignorant about this. Is the purpose of this to mix samples to create sounds to use as notes for making electronic music?
you should have the program bigger and you camera smaller.
Was really expecting some "samples" that could really be used instead of blips beeps and noise.
Someone thinks the way I do. This is basically pointless.
@@Jacob-yb6bv Your own minds are setting the boundaries here. Pretty much anything, pointless or not, can become a really cool paperweight in the correct context. But unlike a paperweight these types of glitches can send your creativity afloat, an adventure that's not for everyone. I wish you all the best on your own journey, glitched or not. :)
Piano chords can be had elsewhere
how did you give it a dark theme tho
I didn’t, I’m guessing it picked up that my system has dark theme on (I’m on Big Sur) and it loaded that way, and I can’t find a way to change that. In fact o don’t like it because some parts are unreadable.
It's a real shame the dry signal still comes through degraded.
Can anyone help me install this? Not sure how to get it up and running 🙏🏼
I'm on mac osx btw
You can have a look at my other video on this but there isn’t much to it other than unpacking the installer and tell your mac that it’s from a trusted developer
@@AlessandroMastroianni Maybe I have some kind of firewall up? I'm not seeing the installer in the zip file 🤷🏼♂
@@AlessandroMastroianni I ran all the exec files and went through each folder but didn't see an installer?
@@AkashGamana yeah files from github are often aimed towards people with programming/computer savvy knowledge.
Maybe I'm stupid but how do you actually open the app?
Did you scroll down the page to see the windows and mac downloads, instead of downloading the one at top?
The top one doesn't have an exe file
@@samhein321 Yeah I downloaded the mac one but the issue seems to be that my mac thinks it's a virus and there's no option in my settings to allow for foreign downloads
Granular synthesis did this 20 years ago. It’s complete rubbish.
🦾🦾🦾
You’ve been had mate.
@@AlessandroMastroianni yeah, no one would pay for it.
Everything sounds to be running at the same speed with no variation. I don’t see the point of this - there must be more tweak ability otherwise this just seems like a moat that Aphex has put around his castle for fans to drown in as they try and figure out how he really makes his sounds. I’ve also been working on a ‘superdrummachine’ for about 15 years in PureData and hopefully it sounds a lot better than this!
could you link your superdrummachine please? really curious about that
Can almost understand you 😮
🤔 ρгό𝔪σŞｍ
I have no idea why these videos exist beside promoting the plugin..
No one seem to have ANY idea what they are doing (can't blame them...) with it.
It's like Richard is playing a game of 4D chess against all these "internet-guys" at the same time, making them look like fools.
Richard? You know him?
What a shame....looks like he’s part of that club too.
The what club?
This is rubbish.
NOTHING special...