I suspect this isn't AI, but simply a library of loops which the platform puts together to create "tracks"! It would not surprise me if lots of people end up with the same or very similar results. I am sure soon enough, someone with enough time on their hands will catch these platforms out!
@bennycesa You're describing how AI works. It doesn't 'create' anything out of nothing, there has to be always something and somewhere that AI can pull things from. The trick is how complex and vast the database is and how "intelligent' the program is so it never pulls the same things over and over
@@PauloARod No, that's not AI! If it's creating the actual music compositional parts from scratch everytime because it's been taught how to do so (by analysing lots of music) and on the requirements set out by the user, then ok, but if it's simply presenting some pre-produced musical loops, that is not AI! If the loops are tagged (say dark, happy, dreamy etc), then it simply pulls loops based on tags, I'd be embarrassed to call that AI! For example, ChatGPT analyses the answers to questions from multiple sources and decides on the best answer, some of which have been very poor as it's still early days, but it's deciding based on the analysis, this in turn is simply (most likely) triggering pre-produced loops!
@@PauloARod That's not how AI works, the problem is what we are calling AI is not actually AI. It is actually machine learning. A true AI would be able to come up with something all by itself. What we have now is software that uses the internet to pull data and use that to create or say what it was programmed to say or create.
I totally agree. This sounds more like a glorified bunch of scripting of VST plugins and pre-recorded loops, rather than machine learning (which is often called AI these days). For more modern examples of AI/ML music, check out Suno or Udio for example. Sonauto is even harder to control, but sounds more natural if you keep trying. You cannot customize anything. It's like rolling dice. But they really generate unique stuff from scratch every time.
I’m totally against any form of creative ai and I hope that someone starts a music platform where ai isn’t allowed just like Cara for art. Give ai generated data metadata so it can be tagged . AI will get exponentially better in the future where it will become harder to distinguish from being artificial. A lot of commercials are already generated by ai and I don’t like it at all.
I certainly think there's a point here, but it's notoriously difficult to uninvent (unvent?) technology. I'm still hoping that autotune will suddenly vanish into thin air.
Maybe I'm wrong, but SoundRaw does not generate music, it compiles from ready-made packages. There's a similar thing in the Magix Music Maker. I tried to do this a year ago, but it takes money to constantly buy new pacts, otherwise all the "generated" music will be similar to each other. There's a real generation of music, it's SUNO/UDIO. They use a model, this is different from making music from loops.But it's also a problem because it's a cat in a bag
Future ai mixing tech will be easy! Extract stems , analyze , match , process , boom finish! Im not extending to ai mastering cuz that’s been around 10 years now!
It could be inspiring to extract some loops from the generated songs and rework them ! But I read the license agreement and we are far from what they advertise : they specified it can’t be registered through labels, or even music distribution services like Tunecore, Distrokid (3.ii of the license agreement). They also say writer’s share and publisher’s share will be shared with them and they will have 50%… and give them the right to “publicly perform, license and otherwise exploit the User Compositions” (which include even a song using a small sample of what they generated) (5.a) and they need to be credited (7.a)
Really, it's basically acid loops but the program chooses them for you and gives you a AI singer too... well, other ones do. If you brought a loop library you would get the same thing but you'd have to put them together entirely. These AI song generators will be great for light hearted use. If someone isn't that worried about it to be made by a genius and is for a advert or a RUclips backing track... or even a silly song to use as a bit. Ideally this is for people not making music, people who just want to get royalty free music that is not completely copy and paste. I can imagine the beat at 6:30 would be a great backing track for a Anime song lol
I am checking reviews about apps which might make my job easier, and are treated to butthurt "musicians" ranting totally unrelated to the product concerned, about how technology is all bad😂
For now… (not really because suno ai is already really good). But let’s see how much these will improve in just one year. You wont be able tell the difference.
@@mixbustv i bet the plugins you are talking about were not AI generated and were not having millions of dollars being ploughed into improving them at a phenomenal rate! The more people laugh at AI music now, the more unfunny they will think it is in a years time when people (including yourself) cannot tell the difference!......this is also not true AI...only udio and suno are currently utilising true AI and some of the stuff being produced is incredible, and improving every day!!!.....add to that the fact that AI will make the production and mastering of tracks fully automated and its all over!......you cannot compare any trend from the past in music to AI!!!....nothing remotely comes close!
I think it would be really good if you could export the songs as "Midi" files, so that someone who isn't that familiar with music can get some inspiration. But having entire songs created is for lazy people. And you miss out on all the fun you have when you produce something, when you suddenly have a great idea for a melody, your foot starts tapping and your mood improves... But it is a really good ringtone generator.
Also, it seems like this music is pretty formulaic. With every measure it changes to a new chord and then repeats that every 4 measures. It’s like they’ll take a popular song that;s already been written, change the melody by changing the order of the original, and then calling that original? Ok but guess what? Is has the same FEEL as the original that it was ripped from! Sorry… not for me! Stay ORIGINAL folks! As usual, thank you David! :)
A.I that's saves producers time like gain staging, organising projects etc is fine IMO. But so called 'A.I' that creates entire songs with a click is just lazy and will result in all music sounding the same.
this thing sucks, but on a side note, i have noticed that suno understood prosody on at least two occasions where i uploaded some of my own song lyrics into the app.
I think it a song generator for starting off point if your not sure what your doing but, I never use samples or any AI for own music there's a big law suit for using AI's that you don't own and samples that people don't understand for the use of there material. I produce and play music so I think things like this intrigue us producers, if any of them.
I'm not afraid of AI at all. If it can't replace analog hardware, it'll never replace composition. :D On a side note, how good are the HRK ST552 tape emulators?
Hi David. Bro please help me. I got pro tools this week and I was Fl studio user. Fl studio don't allow to zoom in the audio waveform as deep as Pro tools and other DAW so I never noticed this problem with pitch correction plugins. Everytime I render my vocals with Autotune or Waves tune or any other pitch correction VST the audio waveform is rendered with few samples of delay, a lil bit out of sync that you can only see if you zoom really deep. I use Windows 10, Asio4all 48khz, the plugin delay compensation is working fine, only when I use Auto tune I have this problem, no matter if I just consolidate the track inside the DAW or if I export it nothing changes. It's a little delay but It makes vocal groups sound really bad and out of sync.
This is a pretty big limitation of FL Studio, one thing you can try is changing the “Time Base” settings to over 900 pulses in the settings. Quick Google search will set you up. This will give more granularity with timing waveforms: “Timebase (PPQ) - Sets the event resolution for the current project in Pulses Per Quarter-note (PPQ). In other words, how finely the time-line of the project is represented and so notes/clips/events placed on that time-line. The default setting is 96 with higher PPQ allowing finer control but it also uses more CPU.”
@@nashse7en I don't know if this will work, but try to load a fruity delay 3 plugin, on the default setting turn down anything that produces a "delay" effect and simply drag the time knob until you have the needed ms delay. You could also look into manual Plugin Delay Compensation settings on the mixer insert, but that gets messy real quick as other plugins start factoring in.
@@nashse7en You'll have to turn off "tempo sync" to get the millisecond settings, probably drag feedback to zero, modulation to near zero etc. It'll take some tweaking but it might work
I don't know how youtube is monetising these Ai music channels, uploading 10 songs a day to the Platform and getting 1000 subs in record time. Ai music with no soul!
Writing a great song is something that even the most skilled song writers labor over and only hope to achieve. This will never change in my mind. No-one cares about everything else. It's filler.
Suno is a better example than this. Try that one and tell me it's not at least a little creepy. I've been making music for 15 years and have been laughing the whole way along this AI generated music wagon ride but suno makes me feel like they are heading in the right direction and it is discomforting. I never thought in my lifetime i'd see music generated and actually sound pleasant even for a second. Who knows what the future will hold but this feels like a threat to average artists and producers because of how devalued music has already become. Commercials most likely will use this tech, rappers who just want a quick beat and don't want to have to pay anyone, ect. Personally, if you aren't the best at what you do, AI will most likely replace you in anything in the future. If you are of some best in the business, I wouldn't be too worried, but if you are starting out or do audio as a hobby, expect to never get noticed without major connections or major talent and advanced abilities to still stand above the tech. Just my opinion.
Well technically an artist takes everything they've heard, seen, and experienced to create something of their own. I have written things that can have similarities to other music I've listened to throughout the years without even knowing it. We all know the industry would love to own this tech, hide it and use it for themselves if they can. You can't tell me I'm wrong because business is business.
@@mixbustv You’re right. AI model progression is very exponential though. The last 2 years of AI have shown DRAMATIC increases in photo generation, sound, and video. Once a large corp with lots of money and processing power hops onto music AI, it will get crazy. I hope it takes much longer than I’m guessing though, I don’t really like the idea of it :/
![Twenty One Pilots - “The Line” (from Arcane Season 2) [Official Music Video]](http://i.ytimg.com/vi/E2Rj2gQAyPA/mqdefault.jpg)
Get 50% off with code MIXBUS when you subscribe to any plan on.soundraw.io/MIXBUS
I suspect this isn't AI, but simply a library of loops which the platform puts together to create "tracks"! It would not surprise me if lots of people end up with the same or very similar results. I am sure soon enough, someone with enough time on their hands will catch these platforms out!
@bennycesa You're describing how AI works. It doesn't 'create' anything out of nothing, there has to be always something and somewhere that AI can pull things from. The trick is how complex and vast the database is and how "intelligent' the program is so it never pulls the same things over and over
@@PauloARod No, that's not AI! If it's creating the actual music compositional parts from scratch everytime because it's been taught how to do so (by analysing lots of music) and on the requirements set out by the user, then ok, but if it's simply presenting some pre-produced musical loops, that is not AI! If the loops are tagged (say dark, happy, dreamy etc), then it simply pulls loops based on tags, I'd be embarrassed to call that AI! For example, ChatGPT analyses the answers to questions from multiple sources and decides on the best answer, some of which have been very poor as it's still early days, but it's deciding based on the analysis, this in turn is simply (most likely) triggering pre-produced loops!
I got the same feeling, its more of a customizable selection of presets
@@PauloARod That's not how AI works, the problem is what we are calling AI is not actually AI. It is actually machine learning. A true AI would be able to come up with something all by itself. What we have now is software that uses the internet to pull data and use that to create or say what it was programmed to say or create.
I totally agree. This sounds more like a glorified bunch of scripting of VST plugins and pre-recorded loops, rather than machine learning (which is often called AI these days). For more modern examples of AI/ML music, check out Suno or Udio for example. Sonauto is even harder to control, but sounds more natural if you keep trying. You cannot customize anything. It's like rolling dice. But they really generate unique stuff from scratch every time.
I’m totally against any form of creative ai and I hope that someone starts a music platform where ai isn’t allowed just like Cara for art. Give ai generated data metadata so it can be tagged . AI will get exponentially better in the future where it will become harder to distinguish from being artificial. A lot of commercials are already generated by ai and I don’t like it at all.
I Agree with that for sure, and it's the starting of the AI era which can only end badly...
I certainly think there's a point here, but it's notoriously difficult to uninvent (unvent?) technology. I'm still hoping that autotune will suddenly vanish into thin air.
But when starts AI ?
When you played the cinematic/epic one I cracked up. Guess my job is safe for a while.
for a while....lets see how funny it is in a years time!
Maybe I'm wrong, but SoundRaw does not generate music, it compiles from ready-made packages. There's a similar thing in the Magix Music Maker. I tried to do this a year ago, but it takes money to constantly buy new pacts, otherwise all the "generated" music will be similar to each other. There's a real generation of music, it's SUNO/UDIO. They use a model, this is different from making music from loops.But it's also a problem because it's a cat in a bag
Future ai mixing tech will be easy! Extract stems , analyze , match , process , boom finish! Im not extending to ai mastering cuz that’s been around 10 years now!
And what makes you think that mixing which is 100000 times more complex and complicated will be any better? Lol makes zero sense.
It could be inspiring to extract some loops from the generated songs and rework them ! But I read the license agreement and we are far from what they advertise : they specified it can’t be registered through labels, or even music distribution services like Tunecore, Distrokid (3.ii of the license agreement). They also say writer’s share and publisher’s share will be shared with them and they will have 50%… and give them the right to “publicly perform, license and otherwise exploit the User Compositions” (which include even a song using a small sample of what they generated) (5.a) and they need to be credited (7.a)
Really, it's basically acid loops but the program chooses them for you and gives you a AI singer too... well, other ones do.
If you brought a loop library you would get the same thing but you'd have to put them together entirely.
These AI song generators will be great for light hearted use.
If someone isn't that worried about it to be made by a genius and is for a advert or a RUclips backing track... or even a silly song to use as a bit.
Ideally this is for people not making music, people who just want to get royalty free music that is not completely copy and paste.
I can imagine the beat at 6:30 would be a great backing track for a Anime song lol
I want ai to do my laundry or cook my tea so I can make music, not the other way around.
I am checking reviews about apps which might make my job easier, and are treated to butthurt "musicians" ranting totally unrelated to the product concerned, about how technology is all bad😂
Imagine being a producer complaining that this is gonna take their job…
Complaints about these types of tools is more of a mask off moment then anything else. It is really just saying how incompetent you are
For now… (not really because suno ai is already really good). But let’s see how much these will improve in just one year. You wont be able tell the difference.
That's what they said about plugins lol
@CraddyMusic
@@mixbustv i bet the plugins you are talking about were not AI generated and were not having millions of dollars being ploughed into improving them at a phenomenal rate! The more people laugh at AI music now, the more unfunny they will think it is in a years time when people (including yourself) cannot tell the difference!......this is also not true AI...only udio and suno are currently utilising true AI and some of the stuff being produced is incredible, and improving every day!!!.....add to that the fact that AI will make the production and mastering of tracks fully automated and its all over!......you cannot compare any trend from the past in music to AI!!!....nothing remotely comes close!
Beautiful music!!!
I think it would be really good if you could export the songs as "Midi" files, so that someone who isn't that familiar with music can get some inspiration. But having entire songs created is for lazy people. And you miss out on all the fun you have when you produce something, when you suddenly have a great idea for a melody, your foot starts tapping and your mood improves... But it is a really good ringtone generator.
Hell!!! midi is what these companies siphon off from other people's hard work and feed it to AI to regurgitate it back to us
Totally agree @matthias
@@4thumbsdown I don't think you know what midi is.
That doesnt sound like AI i think they just intelligently mash up loops
Is there an ai program that mixes your own raw tracks? I’d be interested in seeing/hearing that.
Also, it seems like this music is pretty formulaic. With every measure it changes to a new chord and then repeats that every 4 measures. It’s like they’ll take a popular song that;s already been written, change the melody by changing the order of the original, and then calling that original? Ok but guess what? Is has the same FEEL as the original that it was ripped from! Sorry… not for me! Stay ORIGINAL folks! As usual, thank you David! :)
Izotope neutron!??
Just out of curiosity, what audio converters do you use?
Dangrous, Motu and Mergin
A.I that's saves producers time like gain staging, organising projects etc is fine IMO. But so called 'A.I' that creates entire songs with a click is just lazy and will result in all music sounding the same.
What you described is presets, there no need for AI for those tasks, we had them for 40 years
this thing sucks, but on a side note, i have noticed that suno understood prosody on at least two occasions where i uploaded some of my own song lyrics into the app.
I think it a song generator for starting off point if your not sure what your doing but, I never use samples or any AI for own music there's a big law suit for using AI's that you don't own and samples that people don't understand for the use of there material. I produce and play music so I think things like this intrigue us producers, if any of them.
I'm not afraid of AI at all. If it can't replace analog hardware, it'll never replace composition. :D On a side note, how good are the HRK ST552 tape emulators?
I haven't patched them in yet
It sounds like a music generator. (Which is a good thing)
Sorry i just can't get on board . Ideas maybe but it just sounds pretty generic . But I'm a rock guy . To each his own
I recalled Emotionaliser.
I can't imagine an actual professional, creative person using this. It didn't sound good to me, either.
Hi David. Bro please help me. I got pro tools this week and I was Fl studio user. Fl studio don't allow to zoom in the audio waveform as deep as Pro tools and other DAW so I never noticed this problem with pitch correction plugins. Everytime I render my vocals with Autotune or Waves tune or any other pitch correction VST the audio waveform is rendered with few samples of delay, a lil bit out of sync that you can only see if you zoom really deep. I use Windows 10, Asio4all 48khz, the plugin delay compensation is working fine, only when I use Auto tune I have this problem, no matter if I just consolidate the track inside the DAW or if I export it nothing changes. It's a little delay but It makes vocal groups sound really bad and out of sync.
This is a pretty big limitation of FL Studio, one thing you can try is changing the “Time Base” settings to over 900 pulses in the settings. Quick Google search will set you up. This will give more granularity with timing waveforms:
“Timebase (PPQ) - Sets the event resolution for the current project in Pulses Per Quarter-note (PPQ). In other words, how finely the time-line of the project is represented and so notes/clips/events placed on that time-line. The default setting is 96 with higher PPQ allowing finer control but it also uses more CPU.”
@@TRXST.ISSUES Thanks bro, now the zoom is deeper. But the track still have a few samples of delay.
@@nashse7en I don't know if this will work, but try to load a fruity delay 3 plugin, on the default setting turn down anything that produces a "delay" effect and simply drag the time knob until you have the needed ms delay.
You could also look into manual Plugin Delay Compensation settings on the mixer insert, but that gets messy real quick as other plugins start factoring in.
@@nashse7en You'll have to turn off "tempo sync" to get the millisecond settings, probably drag feedback to zero, modulation to near zero etc. It'll take some tweaking but it might work
@@nashse7en Maybe even turn Dry completely off.
I don't know how youtube is monetising these Ai music channels, uploading 10 songs a day to the Platform and getting 1000 subs in record time. Ai music with no soul!
Bot farms
AI is going to leave artist jobless............. Not!
Writing a great song is something that even the most skilled song writers labor over and only hope to achieve. This will never change in my mind. No-one cares about everything else. It's filler.
Suno is a better example than this. Try that one and tell me it's not at least a little creepy. I've been making music for 15 years and have been laughing the whole way along this AI generated music wagon ride but suno makes me feel like they are heading in the right direction and it is discomforting. I never thought in my lifetime i'd see music generated and actually sound pleasant even for a second. Who knows what the future will hold but this feels like a threat to average artists and producers because of how devalued music has already become. Commercials most likely will use this tech, rappers who just want a quick beat and don't want to have to pay anyone, ect. Personally, if you aren't the best at what you do, AI will most likely replace you in anything in the future. If you are of some best in the business, I wouldn't be too worried, but if you are starting out or do audio as a hobby, expect to never get noticed without major connections or major talent and advanced abilities to still stand above the tech. Just my opinion.
AI doesn't create music, it circumvents copyright laws
Well technically an artist takes everything they've heard, seen, and experienced to create something of their own. I have written things that can have similarities to other music I've listened to throughout the years without even knowing it. We all know the industry would love to own this tech, hide it and use it for themselves if they can. You can't tell me I'm wrong because business is business.
There are currently only two real LLMs available; the rest, like Soudraw, are not true AI.
time to sell analogue gear
Does another person get the same progressions if they choose the same genres and instruments?
Not sure, I assume not
ai 🤮
Give it 5 years.
That's what they kept saying about plugins.. Every 5 years
@@mixbustv You’re right. AI model progression is very exponential though. The last 2 years of AI have shown DRAMATIC increases in photo generation, sound, and video. Once a large corp with lots of money and processing power hops onto music AI, it will get crazy. I hope it takes much longer than I’m guessing though, I don’t really like the idea of it :/
I recalled Emotionaliser.