Some fun with it. 1.) resynthesize a sound 2.) stop after the second or third iteration 3.) bounce every second iteration from every stem 4.) reset the recreation and start again. Synplant starts every process with more or less random settings. But it is going in the right direction. If you do the process above often enough, you will have a load of sounds, which sounds different but similar. Some stems stop growing immediately after the first try. You can trigger the continuation by clicking one sample in the stem. You can also set a startingpoint by clicking ALT and the refresh button.
This is insane, i’ve always loved the synth but never had money to buy it. One of the reasons was I thought they abandoned the project. Now its finally the time to learn it deeply!
Going in the right direction? Starting from random is how evolutionary algorithms are done. That's how you train a neural network or do reinforcement learning. Until we invent oracle machines or possibly get quantum computing in PCs, thats the way to do it. :)
@@martinkrauser4029 I don't think they were saying the idea/plugin is going in the right direction, I think they were just describing how an evolutionary algorithm works: It starts with random settings and then evolves those in the right direction. And I think the point they were making is that you can take advantage of that randomness to get a variety of similar sounds.
Synplant used to be one of the most unusual and interesting plugins I've stumbled across back in my school days in 2008-2009. Glad to see that it's still evolving.
agreed this is a very old plugin but 2 must have been dropped recently. I had one of those copies for synplant 1 but cant remember or have the serail number now. When I visited their webiste it's 149 euros. Not sure if the price was that high back then haha
Im always afraid to use it because it seems so impossible to replicate sounds or actually learn how to use it. I feel like a synth like serum is just better to learn
The interface is really tactile and addictive. Unfortunately this means I spent hours just generating new patches instead of making any music, but once the novelty wears off I think I'll get some pretty cool use out of it
Another neat thing I'd that u can keep on clicking and make more stems. Or even just delete it then generate it again until u find that perfect sound. I've been messing with the demo and it's insane what kind of sounds u can create with it.
@@Cobruz I think this type of stuff is cool but in the next few years by the time people end up mastering this tool, there will just be a new AI that you hit a button and will make way better music than anybody could make and all creativity will be finished and all that time you spent making music will basically be pointless. That’s why I’m not even bothering. Same thing with art what’s the point of spending 15 years to become the best digital artist when AI can do it way better now in 20 seconds. Same thing will happen with music. This is just a rudimentary version of what will come.
You're the first video i saw where someone actually shows the dna editor. The AI giving you starting point to then sculpt on a parameter level on your own is so powerful.
Unrelated to audio, but yes, thus is the THING AI needs. Firefly (adobe's awful AI image generator) does a meh job of making images, but the ability to modify anything is very limited, and mostly behind a paywall (and it's Adobe, so what else would you expect?) Anyway, I don't do audio, but this has me interested, and I've had enough times when I wanted a SLIGHTLY different sound that this would be worth getting even for a video guy like me.
I like the part when it's doing all the attempts and the little DNA strands grow out and it seems you can't get enough of it and you keep smiling at every attempt. This is an amazing tool!
That’s exactly what excited me about it! I don’t make music and have only ever messed around in Ableton briefly, but this kind of thing is really encouraging for me. Hopefully the price of stuff like this will come down in a little while/a freeware version will be released.
It doing this so fast and going over iterations is what is really impressive. The more training data it gets it'll get better. Matter of time before this is a staple feature in all synths.
This doesn't seem to be a neural network, so it doesn't work on data training. The neurons of a CNN don't translate into human-adjustable parameters. You do get good replication, but you can't tweak the sounds like you can here. This is an evolutionary algorithm, which means it has an evalution function that gives you a score of good a solution is - say, a sum up the of the square root of each sample of a wave you get subtracting you result signal from what you're trying to copy. The way you get to a better solution is randomly changing parameters a little bit as if they're DNA mutating, but only keep the better-performing results in a generation to mutate in the next pass. Repeat this many times, and you've bred yourself a patch that is going to score high on your evaluation function, and thus also sound like your input, since you're searching for the least difference in signals. You do this several times starting from randomized parameters to avoid getting stuck in a local maximum - when tweaking some of the parameters will only make it sound less like the target, but if you made big changes to what you have, you'd get a better result. The quality of these algorithms is determined by how well your evaluation function works, and how flexible the adjustable parameters are. These parameters are the synth engine itself - just throwing data at the problem won't help you there. You need engineering to build you one of those, and it helps to know that you will be doing evolutionary algos with the parameters when you're designing your synth. It's not really about data.
@@martinkrauser4029 You can definitely train an AI to control human controllable parameters. Here's how you train this: Input neurons = audio sample, output neurons = knobs, convert knob info into sound using the plugin, compare the result with the original sample using an adversarial neutral network which is trained in parallel, if it can tell the difference easily, try again, Repeat When you use this neutral network, just have it generate an output 20 times, and display those outputs as temporary 'presets' on the tree
I how like zero musical understanding beyond playing clarinet in middle school but to me it seemed pretty obvious that the human voice he sampled and then that the program iterated on after he removed the modulation (or whatever that dial was) was heavily reminiscent of the Miis voices from Mii music which makes sense since Nintendo probably did sample little snippets of human voice and then just synthesized it until the sound wasn’t really human but like eerily/novel as a sound reminiscent of it lol
@@CitizenflabaI think what people find amazing, or at least what I found amazing, was the unexpected harmonies he played more so than the sound. For me at least, it would sound just as good on a clavier or a glockenspiel
This is a tutorial of how to REALLY uses AI to bring so many options, I'm so impressed by this!! I can imagine the whole new universe in Music Production with it.
The most important part is that people can do this by themselves (no need for big music companies) and record a few samples from your own instruments as seeds to generate from.
Your description of sampling vs synthesis was so succinct. It's things I have used and already kind of knew but my understanding now is much better after that explanation. Thanks for the video
love this video! when synplant came out many years ago it was already revolutionary as a synth that was super helpful. this time, theyve done it again; cheers to synplant 2!
i'm a hobbyist/learning producer and I was just browsing after I spent a whole night in ableton and torturing serum trying to recreate some sounds.... this is just...wow. my ispirations always came from specific sounds and instruments I always struggle to replicate, take the original vangelis synth for example. this can get me 99% close enough and then can play with the sound to make it mine...this is just crazy. it's 7.03 in the morning, hats off to you for making this video brother
Finally I don't have to spend countless hours trying to find some synth I heard in a song, I can just sample it and use it as an actual synth and not a sample!
This has been possible before but with a lot more hands on work required. This tool takes the process and turns it into an instant gratification wet dream lol. I have AnA2 along with all of its packs, Serum, FM 2, Massive, Masive X, kontakt player and tons of 1st and 3rd part libraries, countless sample packs and I'm just scratching surface with what I can recall off the top of my head. Add this badass little guy into the mix and I can't really fathom needing anything else lol. That's how it always goes with this stuff though. Everything you think you don't need something else, youtube and the internet decides to throw something else up in your face lol.
@@jedimindtrix2142 i don't even think this particular synth is necessarily neded to actually play in your mix. It could be the case of putting a sample in, waiting for the magic to happen and then just copy the values of envelope, lfo, waveform generators and so on to whatever synth you're in love with, and out you get the needed sound, which you can tweak within your favorite environment.
@@iAmNothingness i think you are too caught up in the whole ai art stealing thing,. this is certainly not stealing. and sampling has been going on for decades lol.
This sent shivers down my spine man there are so many uses for this. That's REALLY saying something seeing as how I'm not even a producer and have very little experience with making music
I bought my first synth in 1979, a Micromoog which is of course monophonic and without any Midi since that was invented in the mid 80s. This was the last year they were sold "new" in the stores. Since then I´ve used it a ton and still do. Then came Midi, samplers and digital audio, and even after a while computers that were fast enough to handle it all, and DAWs that didn´t crash all the time. It is a big understatement to say that things have moved on since then. Will have to try this, thanks for demoing!
Absolute game changer this for me, I’m a engineer/ remixer so people want me to remix tracks they have heard and want me to keep it same as orginal and I spent hours trying to recreate the same synths , sometimes I cannot do it, now I have this, amazing
Analysing samples and turning them into synthesis models heavily reminds of the Hartmann Neuron synthesizer from 2003 which did exactly that - that's how long this idea is on the market now.
I think a few additive synths featured something similar as well. Now if we could get accurate physical modeling synthesis from an acoustic sample, that would be another leap
I believe super expensive Synclavier from early '80s, expensive Kurzweil K150 and E-mu Emulator II w Mac soft from mid-'80s, and relatively inexpensive Kawai K5 synth from 1987 could all do this, probably some others, too.
Synplant has been around for longer than most modern young producers have been alive, I remember playing with it on my first Power Mac over 30 plus years ago.
@@JM-yk6eh Back in my day, we walked 10 miles to the nearest music store to buy a Synplant CD-ROM. We didn't have any fancy internet or streaming services. We had to install it on our Power Macs with floppy disks and hope it didn't crash. We spent hours tweaking the knobs and creating our own sounds. We didn't care about genres or trends. We were pioneers of electronic music. You kids today don't know how good you have it. You just download Synplant from the web and copy the presets from UA-cam tutorials. You have no originality or creativity. You should respect your elders and learn from the history of music. Synplant has been around for longer than most modern young producers have been alive, and I'm proud to say I was one of the first to use it.
this is amazing. I haven't made music in a very long time. But if tools like this become cheap and available, I will 100% get back into it. The idea of putting my own synths together from random sounds I collect and find interesting is amazing.
This is insanely useful!!! I do a lot of experimental stuff with my music lately, and I think this plugin can help me make unique sounds so much easier. Specially some of those weird sounds it generates from a sample that people normally call it a "fail". Those can be the roots to some of the coolest sounds specially in colourbass sound design! Also the fact that you can see the settings and knobs after a sound is generated from a sample, helps SO MUCH in finding out how some of the sounds in other musics might have been made.
As an old electronic music producer who used to record individual samples into an AKAI S3000 with a CD player - this is mind blowing. I really hope it ushers in a new era of HUMAN creativity and great music. Thanks for sharing - great discovery.
@@brownie3454 seriously why do u think i dont support ai i never said my stance i said why do u think its ok to take a sound designers job but not a singers or a composers
On the one hand, i fully agree that this is an awesome plugin from a technical standpoint. On the other hand, i have to say that too many producers who are pushing the hype around it, are not giving the developer enough feedback concerning the terrible GUI concerning accessibility and overview. I mean it looks lovely, from a visual standpoint. But it's beyond bad, UX wise, when everything is sliced into many different sub-windows, when instead, you could see about 400% more of the interface, if it was bigger, and not forced into that tiny format, with dozens of sub windows. You wouldn't have to click around to switch views in the interface AT ALL, if it was the size of Massive/Serum/Omnisphere... I think there should be an option for that big window size.
@@KBoxx ok, fascinating. If it was an optional setting, every user preference would be satisfied. The Kontakt/KompleteKontrol Interfaces have different "view" settings for example.
I tried this tool once after seeing it on TikTok and it didn't recreate my intended sound at all, so I gave up. But you revived my belief in it. I can only hope they manage to expand the universe of possible sounds it can recreate. A tool that works all of the time is a hundred times more useful than a tool that only works half of the time.
This is indeed a game changer. If you want to change your game, on an intergalactic level, you can use Serato Sample 2 to extract, say a gnarly bass from a nice choon, and import it into synplant.
As someone who’s very new to this, I often have an idea for how I want something to sound but I just don’t understand all the stuff to adjust to achieve that. This tool would probably help me discover what the heck is going on in a sample for me to like it so much lol
The plugin is an a very interesting take on biomorphism. Seems like its breaking down the soundwave into its constituents and then generating new sounds. Getting close to soundwave physic's. I'm sold on it. These days its getting extremely interesting in all the sound design gear there is and how deep its getting. The immersive sound these days is amazing. I've been doing some projects with the audio in VR games. I absolutely love the effects you can get from it. Some of the VR tech for creating music i think will be a game changer in the future also.
Alright, if you really want some cool and/or avante-garde shit- Drop your sample in genopatch, then record the sounds it makes as it generates the new patches. Even in that one complex sample in the video at 6:47 which didn’t “work” (so to speak … I thought it sounded badass), the rapid sequence of sounds it cycled through while trying to re-create the original sample was awesome imo. So even when it “fails,” Synplant2 has the potential to potentially succ-seed (get it?). I love this damn synth, and your quick review of it LSE2 was exactly what so many other reviewers have missed. Cheers!
Don't be so shortsighted. AI will obviously destroy music production as a career, just like every other job, it's just a matter of time. We should be protesting this technology but instead we're buying into the message the tech companies are feeding us, that this tech "empowers" humans. Yeah it'll empower some of us for like 10 years but then it's game over.
@@ts4gvAgreed, it's all a race between big companies. Developing smarter and better AI every day without limitations. This all is but a money game for them, that will for sure not end well.
@@ts4gv it's just the natural circle of life, we won't be here forever but we just created something that will, I'm not afraid at all, AI will never have something that only humans have, that is creativity, as long as we have this we're safe.... but if we loose creativity then yeah we fucked.
@@thesis_gaia7960 why couldn't an AI be creative? What is it exactly about the human brain that a sufficiently complex machine couldn't possibly replicate? I'd argue that there is no distinction. We'll have machines faster, more capable, and more creative than the human brain within our lifetimes. Mass unemployment ensues. Our economic models will not be updated in time to prepare us. Then there's the alignment problem - there's no guarantee that a superintelligent machine's interests line up with our own interests, and we have no idea how to even assess what a machine *really* wants to do, or how to fundamentally instill human values into them... A superintelligent machine that we've incorporated into every aspect of our lives that doesn't care about human life will find a clever & unstoppable way to kill us all the moment we stand in its way Any pro-AI content is unethical and absurd. This is the most important existential issue of our time by a huge margin.
@@ts4gv dude? The fact that AI art works analyzing the art style of the artist to replicate it just answer your question, but if you're not satisfied, AI can't think for itself, everything about it is coded to emulate consciousness and human behavior. And it doesn't matter how afraid we are about it, if it keeps generating money it will never be stopped, all you have to do is *be human* and adapt to it.
I've tried this synth and there's something to mention about it. Although it can reproduce many single-layer sounds, it's limited to that. Most samples come as either multi-layer or with embedded effects like distortion, reverbs, modulations or delays. In such case, the Synplant has quite a difficult time to reproduce the sound to at least acceptable accuracy.
Lol in a year, copyright stans will be crying that "this is stealing the essence of sounds people worked so hard to create!1!!1!” I, on the other hand, am here for the creativity this will bring about!
Haven't done music since a while but I remember spending hours messing with synplant you can make some very organic stuff there, now I want to try synplant 2 that DNA extraction is so cool
That must be the closest we're getting to the "thought-to-file" headset we've been dreaming of. Absolutely mindblowing 🤯 Lazy producers who have barely created anything in the last 10 years ! ASSEMBLE ! THIS IS OUR TIME !!
@@Nahhh868 Hahaha, unfortunately I can relate. But honestly, "productively procrastinating" - that's good. 😄 I'm gonna use this and create a vocal stem. And I'm not gonna procrastinate. Are you on Soundcloud? I could message you there and send you the result. 😉
Amazing plug in. Pro tip. In ableton, map as many parametres as you can, then map to lfo, or use random to experiment with sounds you like to create completely new fresh ones.
Doesn't work for dubstep haha! It's so cool how you see it iterating as it tries to get closer to the original sound. We need more AI tools like this, that helps us to reach our goals. Cool stuf.
That is one of the craziest VST’s I’ve seen yet! I’m wondering what if we gave that same AI a really really powerful synth like pigments, the sounds it can come up with would be insane.
Hell yes. Also more of this is coming in probably most synths, if you generate a bunch of sounds and record the settings of any synth you could train a neural net to do this (not like, super easily, but definitely do-able)
This was before AI got so mainstreamed, Synplant has been around at least since 2010, I used it on a lot of tracks in 2011. So praise to the guys who made synplant, since there was no AI back then, not to speak of at least ❤🎉
You are enjoyable to watch. This is your chance to become one of THE guys when it comes to showcasing AI music tools. This segment is definitely not going anywhere anytime soon. Good winds.
It wont be like that, it will just be itunes producing on the fly specifically for each user. a stream of endless music made by machine. People wont even listen to the same songs anymore.
My A-level computer science project was trying to achieve something similar to this for FM synthesis but using a more procedural approach rather than AI. It didn't go very well mostly because I didn't put very much effort into it, and didn't have a very good grasp of the maths, but it was an idea I had been interested in for a few years. It sort of worked for replicating very basic 2-op FM tones. Hopefully I will get back to it eventually. This thing obviously has a lot more potential though for complex sounds.
Honest question, how is this "AI" as opposed to just software? be specific in your answer please. The term is being abused so much lately it means nothing
"Genopatch crafts synth patches from audio recordings, using AI to find optimal synth settings based on your source sample." Is literally the first sentence about the product on their website, link in description
I remember an old VH1 show where they invited bands to get back together. As I remember, Berlin had trouble re-creating the patches because the synth player's synth was gone and so were all the presets. You could recreate them easily with this.
Bro I get that you're a music guy, but you gotta eq some of that bass out of your voice for the voiceover. It's so hard to listen to this without waking up the next room
Ah, Synplant! Have had that little synth for some years, it's neat, and can do so much. Will try this, was some time ago I used it. Looks and sounds amazing. What a stellar tool this is!
This is crazy! Can't imagine what I would have done with this when I was still producing stuff. 🤯 Remember loving the first Synplant but this is one hell of a step up!
you can "make any sound a synth preset" using any synth that has drag & drop wavetable synthesis, such as serum. this has been a thing for as long as wavetable synthesis has existed as far as i know, so i'm not sure what the deal is here. surely it's just an a.i. assisted version of a utility that we already have
This has me creating a playlist for music production ideas. I've been slowly building up starting production with a PC upgrade and this looks like something super fun to toy with. I want to take samples of cartoon sounds and see what happens
this synth/resampler would be a really cool way to make sick wavetables for WT synthesizers like serum, using all sorts of basses and stuff. ofc, you could drag the samples into the WT set to the same note as the sample, but i feel like this would allow for a lot more options and a lot more variety in what sound you form from a specific sample.
I have totally been waiting for this :) Thanks for covering this... I guess I can maybe manually automate a my way between two sounds... I have been wanting that (something like Zynaptic Morph)
I loved Synplant1 not that I could ever understand it, but the sounds it created were so different to anything else I have.. I wasn't going to bother upgrading, but seems I don't have a choice now grrrr! They have definitely raised the bar a few notches!
Some fun with it.
1.) resynthesize a sound
2.) stop after the second or third iteration
3.) bounce every second iteration from every stem
4.) reset the recreation and start again.
Synplant starts every process with more or less random settings. But it is going in the right direction.
If you do the process above often enough, you will have a load of sounds, which sounds different but similar.
Some stems stop growing immediately after the first try. You can trigger the continuation by clicking one sample in the stem.
You can also set a startingpoint by clicking ALT and the refresh button.
It really is like plant breeding to grow the perfect cultivar
This is insane, i’ve always loved the synth but never had money to buy it. One of the reasons was I thought they abandoned the project. Now its finally the time to learn it deeply!
just resample the generation sounds
Going in the right direction? Starting from random is how evolutionary algorithms are done. That's how you train a neural network or do reinforcement learning. Until we invent oracle machines or possibly get quantum computing in PCs, thats the way to do it. :)
@@martinkrauser4029 I don't think they were saying the idea/plugin is going in the right direction, I think they were just describing how an evolutionary algorithm works: It starts with random settings and then evolves those in the right direction. And I think the point they were making is that you can take advantage of that randomness to get a variety of similar sounds.
Synplant used to be one of the most unusual and interesting plugins I've stumbled across back in my school days in 2008-2009. Glad to see that it's still evolving.
i see what you did here lol
agreed this is a very old plugin but 2 must have been dropped recently. I had one of those copies for synplant 1 but cant remember or have the serail number now. When I visited their webiste it's 149 euros. Not sure if the price was that high back then haha
Im always afraid to use it because it seems so impossible to replicate sounds or actually learn how to use it. I feel like a synth like serum is just better to learn
this kind of AI is the type of AI that we need, tools and not replacements
exactly, ai shouldn't replace people
but should help people
I agree,we really need more ai like this
Yes, I’m afraid there may need to be legislation to protect artists or workers in general for that matter.
is it not a replacement for synthesists?
I don't see it that way cause that's still a very important and valid way of making sounds
The interface is really tactile and addictive. Unfortunately this means I spent hours just generating new patches instead of making any music, but once the novelty wears off I think I'll get some pretty cool use out of it
Relevant username
😬 Same here. I just lost 20+ hours in a few days and won at least 100 presets during my exploration.
@@rez9159 if u think u lost hours messing around with this.. ur doing music completly wrong
Just record all the things you do with it in another track. Splice it to a drum rack.
Dude you have absolutely no idea how far my jaw dropped when it's starting generating the samples, this is absolutely insane and so absolutely amazing
genuinely beyond impressed
Toro! Awesome profile pic. Was my favorite character in All Stars Battle Royale haha
Another neat thing I'd that u can keep on clicking and make more stems. Or even just delete it then generate it again until u find that perfect sound. I've been messing with the demo and it's insane what kind of sounds u can create with it.
@@Cobruz I think this type of stuff is cool but in the next few years by the time people end up mastering this tool, there will just be a new AI that you hit a button and will make way better music than anybody could make and all creativity will be finished and all that time you spent making music will basically be pointless. That’s why I’m not even bothering. Same thing with art what’s the point of spending 15 years to become the best digital artist when AI can do it way better now in 20 seconds. Same thing will happen with music. This is just a rudimentary version of what will come.
@@ZxAMobile we can recognize human creativity beyond what an ai already copies off huge databases, better is nothing objective there
The riff that starts at 7:43 is amazing! I would dearly love a full length song along the same lines. Really beautiful.
Sounds alright. Nowhere near amazing. But to each is own.
Hard same. Now I want to check out his music.
Reminds me a bit of the older sounds of Infected Mushroom. Really beautiful indeed!
Same - started boppin my head to that one
Absolutely
You're the first video i saw where someone actually shows the dna editor. The AI giving you starting point to then sculpt on a parameter level on your own is so powerful.
if he's a video, I'm a movie :D. Synplant 2 is amazing tool. Must have it.
Unrelated to audio, but yes, thus is the THING AI needs. Firefly (adobe's awful AI image generator) does a meh job of making images, but the ability to modify anything is very limited, and mostly behind a paywall (and it's Adobe, so what else would you expect?)
Anyway, I don't do audio, but this has me interested, and I've had enough times when I wanted a SLIGHTLY different sound that this would be worth getting even for a video guy like me.
I like the part when it's doing all the attempts and the little DNA strands grow out and it seems you can't get enough of it and you keep smiling at every attempt. This is an amazing tool!
The part where he sampled a spoken word sounds so silly, i love it. I wonder if its possible to use a longer sample like an entire sentence.
The genopatch function is really impressive! You could use it as a learning tool too -- to see, and understand, how a certain sound is made.
That’s exactly what excited me about it! I don’t make music and have only ever messed around in Ableton briefly, but this kind of thing is really encouraging for me. Hopefully the price of stuff like this will come down in a little while/a freeware version will be released.
As a non music producer I’m not sure if I’m more impressed by the state of the art synth or your ability to create beautiful music out of thin air.
It doing this so fast and going over iterations is what is really impressive. The more training data it gets it'll get better. Matter of time before this is a staple feature in all synths.
There's no way this won't become a staple. No way!
This doesn't seem to be a neural network, so it doesn't work on data training. The neurons of a CNN don't translate into human-adjustable parameters. You do get good replication, but you can't tweak the sounds like you can here.
This is an evolutionary algorithm, which means it has an evalution function that gives you a score of good a solution is - say, a sum up the of the square root of each sample of a wave you get subtracting you result signal from what you're trying to copy. The way you get to a better solution is randomly changing parameters a little bit as if they're DNA mutating, but only keep the better-performing results in a generation to mutate in the next pass. Repeat this many times, and you've bred yourself a patch that is going to score high on your evaluation function, and thus also sound like your input, since you're searching for the least difference in signals.
You do this several times starting from randomized parameters to avoid getting stuck in a local maximum - when tweaking some of the parameters will only make it sound less like the target, but if you made big changes to what you have, you'd get a better result.
The quality of these algorithms is determined by how well your evaluation function works, and how flexible the adjustable parameters are. These parameters are the synth engine itself - just throwing data at the problem won't help you there. You need engineering to build you one of those, and it helps to know that you will be doing evolutionary algos with the parameters when you're designing your synth. It's not really about data.
I can see a company like Ableton buying up this tech for sure...
@@martinkrauser4029 You can definitely train an AI to control human controllable parameters. Here's how you train this:
Input neurons = audio sample,
output neurons = knobs,
convert knob info into sound using the plugin,
compare the result with the original sample using an adversarial neutral network which is trained in parallel, if it can tell the difference easily, try again,
Repeat
When you use this neutral network, just have it generate an output 20 times, and display those outputs as temporary 'presets' on the tree
@@martinkrauser4029i wonder how it calculates its “score” or loss.
That is the BEST electronic music tool ive ever seen 😮 Synplant is my Christmas present to myself.
THIS IS CRAZY USEFUL WHAT
Also, the fact you just made that beautiful chord progression at the end like it was nothing is just bonkers to me
winnie the pooh sounding progression, i loved it lol
The chord progression is nuts! How'd he do that??
@corvus8638 video game soundtrack came to mind immediately
7:50 holy crap that sounds magical
I how like zero musical understanding beyond playing clarinet in middle school but to me it seemed pretty obvious that the human voice he sampled and then that the program iterated on after he removed the modulation (or whatever that dial was) was heavily reminiscent of the Miis voices from Mii music which makes sense since Nintendo probably did sample little snippets of human voice and then just synthesized it until the sound wasn’t really human but like eerily/novel as a sound reminiscent of it lol
very interesting@@Citizenflaba
@@CitizenflabaI think what people find amazing, or at least what I found amazing, was the unexpected harmonies he played more so than the sound. For me at least, it would sound just as good on a clavier or a glockenspiel
This is a tutorial of how to REALLY uses AI to bring so many options, I'm so impressed by this!! I can imagine the whole new universe in Music Production with it.
The most important part is that people can do this by themselves (no need for big music companies) and record a few samples from your own instruments as seeds to generate from.
Your description of sampling vs synthesis was so succinct. It's things I have used and already kind of knew but my understanding now is much better after that explanation. Thanks for the video
love this video! when synplant came out many years ago it was already revolutionary as a synth that was super helpful. this time, theyve done it again; cheers to synplant 2!
i'm a hobbyist/learning producer and I was just browsing after I spent a whole night in ableton and torturing serum trying to recreate some sounds....
this is just...wow. my ispirations always came from specific sounds and instruments I always struggle to replicate, take the original vangelis synth for example.
this can get me 99% close enough and then can play with the sound to make it mine...this is just crazy.
it's 7.03 in the morning, hats off to you for making this video brother
I love how the plugin actually "grows" your synth like a plant. Amazing stuff...
this is incredible!!! the vocal pad at the end was so unique
Finally I don't have to spend countless hours trying to find some synth I heard in a song, I can just sample it and use it as an actual synth and not a sample!
This has been possible before but with a lot more hands on work required. This tool takes the process and turns it into an instant gratification wet dream lol. I have AnA2 along with all of its packs, Serum, FM 2, Massive, Masive X, kontakt player and tons of 1st and 3rd part libraries, countless sample packs and I'm just scratching surface with what I can recall off the top of my head. Add this badass little guy into the mix and I can't really fathom needing anything else lol. That's how it always goes with this stuff though. Everything you think you don't need something else, youtube and the internet decides to throw something else up in your face lol.
@@jedimindtrix2142 i don't even think this particular synth is necessarily neded to actually play in your mix. It could be the case of putting a sample in, waiting for the magic to happen and then just copy the values of envelope, lfo, waveform generators and so on to whatever synth you're in love with, and out you get the needed sound, which you can tweak within your favorite environment.
Nice! Just steal! Like in art.
@@iAmNothingness Nail on the head...
Well, one of many nails that need to be put on the AI head
@@iAmNothingness i think you are too caught up in the whole ai art stealing thing,. this is certainly not stealing. and sampling has been going on for decades lol.
This sent shivers down my spine man there are so many uses for this. That's REALLY saying something seeing as how I'm not even a producer and have very little experience with making music
I haven't seen synplant in DECADES! kinda confusing but really inovative GUI, brings me memories of the time I started producing
I bought my first synth in 1979, a Micromoog which is of course monophonic and without any Midi since that was invented in the mid 80s.
This was the last year they were sold "new" in the stores. Since then I´ve used it a ton and still do.
Then came Midi, samplers and digital audio, and even after a while computers that were fast enough to handle it all, and DAWs that didn´t crash all the time. It is a big understatement to say that things have moved on since then. Will have to try this, thanks for demoing!
You explained this entire set of concepts really well! Sampling, synthesis, and this crazy awesome AI tool.
Absolute game changer this for me, I’m a engineer/ remixer so people want me to remix tracks they have heard and want me to keep it same as orginal and I spent hours trying to recreate the same synths , sometimes I cannot do it, now I have this, amazing
Analysing samples and turning them into synthesis models heavily reminds of the Hartmann Neuron synthesizer from 2003 which did exactly that - that's how long this idea is on the market now.
I think a few additive synths featured something similar as well. Now if we could get accurate physical modeling synthesis from an acoustic sample, that would be another leap
I believe super expensive Synclavier from early '80s, expensive Kurzweil K150 and E-mu Emulator II w Mac soft from mid-'80s, and relatively inexpensive Kawai K5 synth from 1987 could all do this, probably some others, too.
This is light years ahead. This will make every cover band sound better immediately
@@johnviera3884 How does this make my ACDC coverband sound better?
@@ickebins6948 that’s a sinking ship. Nothing could save it.
your enthusiasm is contagious - thanks
So... we ain't gonna talk about how that UI look like a washing machine?
Synplant was one of my favorite plugins years and years ago. Really awesome that they've improved it so much and re-released it.
Synplant has been around for longer than most modern young producers have been alive, I remember playing with it on my first Power Mac over 30 plus years ago.
Ciertamente
weird flex
Stop lying. First release of synplant was released maximum 15 years ago
@@JM-yk6eh
Back in my day, we walked 10 miles to the nearest music store to buy a Synplant CD-ROM. We didn't have any fancy internet or streaming services. We had to install it on our Power Macs with floppy disks and hope it didn't crash. We spent hours tweaking the knobs and creating our own sounds. We didn't care about genres or trends. We were pioneers of electronic music. You kids today don't know how good you have it. You just download Synplant from the web and copy the presets from UA-cam tutorials. You have no originality or creativity. You should respect your elders and learn from the history of music. Synplant has been around for longer than most modern young producers have been alive, and I'm proud to say I was one of the first to use it.
30 plus yrs? There was no internet
this is amazing. I haven't made music in a very long time. But if tools like this become cheap and available, I will 100% get back into it. The idea of putting my own synths together from random sounds I collect and find interesting is amazing.
This is insanely useful!!! I do a lot of experimental stuff with my music lately, and I think this plugin can help me make unique sounds so much easier. Specially some of those weird sounds it generates from a sample that people normally call it a "fail". Those can be the roots to some of the coolest sounds specially in colourbass sound design!
Also the fact that you can see the settings and knobs after a sound is generated from a sample, helps SO MUCH in finding out how some of the sounds in other musics might have been made.
As an old electronic music producer who used to record individual samples into an AKAI S3000 with a CD player - this is mind blowing. I really hope it ushers in a new era of HUMAN creativity and great music. Thanks for sharing - great discovery.
I'm all about AI being used like this. Very handy, lots of unique sounds and some strangeness, which is always fun.
so its allowed to take a sound designers job just not a producers or a composers or a singers job thosse arent ok
@@lilwoodiewood3457sound designers are still needed for very synth heavy sounds
@@lilwoodiewood3457 it’s all possible already so it’s already ok whether you like it or not
@@brownie3454 i never said it was wrong im calling out hypocrites
@@brownie3454 seriously why do u think i dont support ai i never said my stance i said why do u think its ok to take a sound designers job but not a singers or a composers
Make a song around Synplant synthesizing sounds. It sounds really cool when it does it. 😮
7:43 that goes so hard. i can barely believe you can just make that off your head
i giggle afterward because i was genuinely surprised at how much i liked those random chords lmao
lmao@@lse2
@@lse2 I couldn't stop listening to that melody you made up with those goofy ahh sounds, it genuinely sounds cool
@@lse2in all seriousness, if you ever sneak that melody in a song of yours I would flip
please make something out of that melody. I love it a lot too, and with that sound its very nostalgic feeling, liminal almost@@lse2
Never thought Synplant would make a sequel. So glad they did. It is the most original synthesis platform Ive ever used. Take my money. . .again.
On the one hand, i fully agree that this is an awesome plugin from a technical standpoint. On the other hand, i have to say that too many producers who are pushing the hype around it, are not giving the developer enough feedback concerning the terrible GUI concerning accessibility and overview. I mean it looks lovely, from a visual standpoint. But it's beyond bad, UX wise, when everything is sliced into many different sub-windows, when instead, you could see about 400% more of the interface, if it was bigger, and not forced into that tiny format, with dozens of sub windows. You wouldn't have to click around to switch views in the interface AT ALL, if it was the size of Massive/Serum/Omnisphere... I think there should be an option for that big window size.
I like the UI from a usability standpoint. Breaks it up into manageable pieces so you only deal with one thing at a time
@@KBoxx ok, fascinating. If it was an optional setting, every user preference would be satisfied. The Kontakt/KompleteKontrol Interfaces have different "view" settings for example.
@@sternenherz agreed having the options would be great!
@@KBoxxI want my UI like phaseplant. All oscillators, lfos and effects on the same page, and it works
I tried this tool once after seeing it on TikTok and it didn't recreate my intended sound at all, so I gave up. But you revived my belief in it. I can only hope they manage to expand the universe of possible sounds it can recreate. A tool that works all of the time is a hundred times more useful than a tool that only works half of the time.
This is indeed a game changer. If you want to change your game, on an intergalactic level, you can use Serato Sample 2 to extract, say a gnarly bass from a nice choon, and import it into synplant.
FL Studio is beta testing an in DAW stem splitter at the moment.
Bro the guy who designed the UI is freaking amazing! I love it all!
As someone who’s very new to this, I often have an idea for how I want something to sound but I just don’t understand all the stuff to adjust to achieve that. This tool would probably help me discover what the heck is going on in a sample for me to like it so much lol
The plugin is an a very interesting take on biomorphism. Seems like its breaking down the soundwave into its constituents and then generating new sounds. Getting close to soundwave physic's. I'm sold on it. These days its getting extremely interesting in all the sound design gear there is and how deep its getting. The immersive sound these days is amazing. I've been doing some projects with the audio in VR games. I absolutely love the effects you can get from it. Some of the VR tech for creating music i think will be a game changer in the future also.
Alright, if you really want some cool and/or avante-garde shit-
Drop your sample in genopatch, then record the sounds it makes as it generates the new patches. Even in that one complex sample in the video at 6:47 which didn’t “work” (so to speak … I thought it sounded badass), the rapid sequence of sounds it cycled through while trying to re-create the original sample was awesome imo. So even when it “fails,” Synplant2 has the potential to potentially succ-seed (get it?). I love this damn synth, and your quick review of it LSE2 was exactly what so many other reviewers have missed. Cheers!
Something tells me Trevor Wishart would have a day with this
I remember Synplant was very visible at sites etc about 10-15 yrs ago. Looks very useful to this day so to speak.
I feel like everything that can enhance music in any way, just expands the magic
Don't be so shortsighted. AI will obviously destroy music production as a career, just like every other job, it's just a matter of time.
We should be protesting this technology but instead we're buying into the message the tech companies are feeding us, that this tech "empowers" humans. Yeah it'll empower some of us for like 10 years but then it's game over.
@@ts4gvAgreed, it's all a race between big companies. Developing smarter and better AI every day without limitations. This all is but a money game for them, that will for sure not end well.
@@ts4gv it's just the natural circle of life, we won't be here forever but we just created something that will, I'm not afraid at all, AI will never have something that only humans have, that is creativity, as long as we have this we're safe.... but if we loose creativity then yeah we fucked.
@@thesis_gaia7960 why couldn't an AI be creative? What is it exactly about the human brain that a sufficiently complex machine couldn't possibly replicate?
I'd argue that there is no distinction. We'll have machines faster, more capable, and more creative than the human brain within our lifetimes. Mass unemployment ensues. Our economic models will not be updated in time to prepare us.
Then there's the alignment problem - there's no guarantee that a superintelligent machine's interests line up with our own interests, and we have no idea how to even assess what a machine *really* wants to do, or how to fundamentally instill human values into them...
A superintelligent machine that we've incorporated into every aspect of our lives that doesn't care about human life will find a clever & unstoppable way to kill us all the moment we stand in its way
Any pro-AI content is unethical and absurd. This is the most important existential issue of our time by a huge margin.
@@ts4gv dude? The fact that AI art works analyzing the art style of the artist to replicate it just answer your question, but if you're not satisfied, AI can't think for itself, everything about it is coded to emulate consciousness and human behavior. And it doesn't matter how afraid we are about it, if it keeps generating money it will never be stopped, all you have to do is *be human* and adapt to it.
I've tried this synth and there's something to mention about it. Although it can reproduce many single-layer sounds, it's limited to that. Most samples come as either multi-layer or with embedded effects like distortion, reverbs, modulations or delays. In such case, the Synplant has quite a difficult time to reproduce the sound to at least acceptable accuracy.
Lol in a year, copyright stans will be crying that "this is stealing the essence of sounds people worked so hard to create!1!!1!” I, on the other hand, am here for the creativity this will bring about!
What do you think about music generators
I dont think ive ever bought something from a youtube video this fast
washing machine
Barbecue Bacon Burger
Haven't done music since a while but I remember spending hours messing with synplant you can make some very organic stuff there, now I want to try synplant 2 that DNA extraction is so cool
They should add a feature to allow you to prompt a sound: create a patch based on the lead synth in Voodoo People by The Prodigy.
No thx
How lazy do you want to be holy shit
@corvus8638Or YOU could stop being lazy and sample your own stuff instead of finding them.
For the first few minutes I was like “ok whatever”
But when you demonstrated it, I was mind blown. I need this plugin.
That must be the closest we're getting to the "thought-to-file" headset we've been dreaming of. Absolutely mindblowing 🤯
Lazy producers who have barely created anything in the last 10 years ! ASSEMBLE ! THIS IS OUR TIME !!
😂👍
I’ve been productively procrastinating 😈
@@Nahhh868 Hahaha, unfortunately I can relate. But honestly, "productively procrastinating" - that's good. 😄 I'm gonna use this and create a vocal stem. And I'm not gonna procrastinate. Are you on Soundcloud? I could message you there and send you the result. 😉
Amazing plug in. Pro tip.
In ableton, map as many parametres as you can, then map to lfo, or use random to experiment with sounds you like to create completely new fresh ones.
Doesn't work for dubstep haha! It's so cool how you see it iterating as it tries to get closer to the original sound. We need more AI tools like this, that helps us to reach our goals. Cool stuf.
It doesn’t work because the synth it’s too simple but this opens the door to big companies training their big synths to make this, it will be wild
And it’s a shame because no one has ever released any presets, sample packs or tutorials on how to make dubstep sounds.
Synplant 2 can make dubstep growls, but the AI has gotten smart enough to become a conscientious objector. Same if try gabber.
WHOA!! Brother. THANK YOU for showing this to us !!! This is unreal
Great synth . Lightyears ahead, I would love to see this with longer sample time, loop phrasing and a few more layers .
That is one of the craziest VST’s I’ve seen yet! I’m wondering what if we gave that same AI a really really powerful synth like pigments, the sounds it can come up with would be insane.
Hell yes. Also more of this is coming in probably most synths, if you generate a bunch of sounds and record the settings of any synth you could train a neural net to do this (not like, super easily, but definitely do-able)
I'm looking forward to seeing this done with hardware synths too. Should be possible with any synth.
@@gstringer2211 true! That would be an awesome thing to see!
Awesome plugin. Thanks for the video, never heard of it and right up my street!
7:43 undertale type beat
Bro, you would have been one of my favorite high school teachers! Besides being a versed musician, you’re a great educator!
This was before AI got so mainstreamed, Synplant has been around at least since 2010, I used it on a lot of tracks in 2011. So praise to the guys who made synplant, since there was no AI back then, not to speak of at least ❤🎉
To be clear, Genopatch does use AI and was added just a few months ago.
You are enjoyable to watch. This is your chance to become one of THE guys when it comes to showcasing AI music tools. This segment is definitely not going anywhere anytime soon. Good winds.
0:22 fnf moment fr
Man, even as a programmer i am always mind blown by these powerful AI. That is some very cool tool that you got there.
Let's skip directly to the point in history when AI produces music, sells, buys and chooses it. So we can relax and sell all our music gear.
im big producer : prompt " s 4 chords, 3 min , pop "
It wont be like that, it will just be itunes producing on the fly specifically for each user. a stream of endless music made by machine. People wont even listen to the same songs anymore.
My A-level computer science project was trying to achieve something similar to this for FM synthesis but using a more procedural approach rather than AI. It didn't go very well mostly because I didn't put very much effort into it, and didn't have a very good grasp of the maths, but it was an idea I had been interested in for a few years. It sort of worked for replicating very basic 2-op FM tones. Hopefully I will get back to it eventually. This thing obviously has a lot more potential though for complex sounds.
Honest question, how is this "AI" as opposed to just software? be specific in your answer please. The term is being abused so much lately it means nothing
This is NOT Ai
Lol, it’s obvs just for views. But hey, man’s gotta eat, can’t hate that
"Genopatch crafts synth patches from audio recordings, using AI to find optimal synth settings based on your source sample." Is literally the first sentence about the product on their website, link in description
It's kinda ai kinda not. It's a learning algorithm that extrapolates sounds from an existing synthesizer. It's pretty close
@@jolkyb2039 what is AI but learning algorithms?
I remember an old VH1 show where they invited bands to get back together. As I remember, Berlin had trouble re-creating the patches because the synth player's synth was gone and so were all the presets. You could recreate them easily with this.
Bro I get that you're a music guy, but you gotta eq some of that bass out of your voice for the voiceover. It's so hard to listen to this without waking up the next room
How about turning ur eq down bud
Brooooo your mom loves to wake up the neighborsc
I never even thought about this.
I'm so proud of people making this plugin.
synplant isn't AI...
i have no idea why i have seen this 😂 but i loved every small example you gave us with the new generated sounds! insane!
Best synth plugin interface I've ever seen. Makes something tedious actually fun to use.
i love it when youtubers explain basic concepts from the ground up every video they make, really cool
So, it's the Sylar of music software... only it doesn't kill the other software or hardware after figuring out how it works.
Ah, Synplant! Have had that little synth for some years, it's neat, and can do so much. Will try this, was some time ago I used it. Looks and sounds amazing. What a stellar tool this is!
The end with the vocal was just wow!!! I'm speechless
This is crazy! Can't imagine what I would have done with this when I was still producing stuff. 🤯
Remember loving the first Synplant but this is one hell of a step up!
I was literally looking up SK1 casio and thought to myself how nice it would be to have a digital version of this. This is beyond my hopes!
ive been wanting this thing for a while now. synplant is insane
this is incredible. the kinds of good things that can come out of ai, fucking awesome
Really cool! The sounds and graphic squiggles remind me of communicating with the aliens in the movie Arrival!
you can "make any sound a synth preset" using any synth that has drag & drop wavetable synthesis, such as serum. this has been a thing for as long as wavetable synthesis has existed as far as i know, so i'm not sure what the deal is here. surely it's just an a.i. assisted version of a utility that we already have
How can I adjust the pitch of the sound created by Genopatch?
the sound I recreated was in B scale
This has me creating a playlist for music production ideas. I've been slowly building up starting production with a PC upgrade and this looks like something super fun to toy with. I want to take samples of cartoon sounds and see what happens
this synth/resampler would be a really cool way to make sick wavetables for WT synthesizers like serum, using all sorts of basses and stuff. ofc, you could drag the samples into the WT set to the same note as the sample, but i feel like this would allow for a lot more options and a lot more variety in what sound you form from a specific sample.
Woooow! I saw this plug-in a decade ago and find it crazy this plug-in now has a feature like this!
I think things like acoustic samples products really are the future as it is the link between sampling and synthesis.
Synplant is crazy :-)
So clicking generate patch makes an ad appear?😭
I have totally been waiting for this :) Thanks for covering this... I guess I can maybe manually automate a my way between two sounds... I have been wanting that (something like Zynaptic Morph)
Your melody making is superb!
Best explained video about this plugin on YT
this is sick! the options are only going to get better this thing is already so limitless
I loved Synplant1 not that I could ever understand it, but the sounds it created were so different to anything else I have.. I wasn't going to bother upgrading, but seems I don't have a choice now grrrr! They have definitely raised the bar a few notches!
I wish it could do converting male rap vocals to female or whatever, not just single tone?
2:16 I love that groove ❤