Not that my opinion on this should matter to you, but that makes sense and I don't see anything wrong at all with doing that for what I personally would be getting out of either creating or listening to that kind of music. Thanks for the explanation. It actually would probably be a pretty good way to understand a genre too, like you were trying to do. I personally kinda dislike AI for art, like I said in my post, but I don't think the way you use it is taking away from the creative side of things. Again, not that you should care about that anyways.. just saying... it of course is for you to decide how and why you do things.

 

Clearly you don't know how Suno works. You can still produce Ai content. The Ai is just a tool in the process. I'm a long standing Suno user. The OP actually did give an explanation of how he produces the music while still using the Ai and sometimes retain some of the the generated content. So yet, it's produced. The promp is at the beginning stage so the Ai can generate demos and ideas.

 

The first things needed to make a track are the idea, the musical composition, the arrangement, etc. It spits that out by way of an example track immediately if you want. If you stem separate that output, convert audio->midi, someone could be doing nothing but creating presets and assigning different samples. That is paint by numbers no matter what you are telling yourself. Someone can tell you they used 10% or 90% and it does not matter, because the minute you know it is done with AI you should just assume all of it was done with it.

Not that I care. But if you look on Soundcloud for example; it's full of that stuff, and also "cover versions" made electronically. It's like production Karaoke. If that is how you want to practice, it doesn't bother me. I'm never going to listen to it either way.

 

In the past, people visited forums to have discussions with other people, and those who answered "usually" did so from personal experience, now even those who don't know shit about certain topics, throw the question in chatgpt, and paste paragraphs of bullshit.. What can you do? This is the present/future!

 

This is where you see the stupidity and cultural decline.

 

That's what i'm telling since the start of the thread. The end result can be completly different and non-conventional if you work the output the right way. The output of AI is already on a song format, that's why a lot of people thinks you're not properly workin, but you can mangle it on zillion ways to create a track that no one have yet, not even the AI.

Some times can work real quick and easy. And yet if i thought i couldn't take any advantage on it, it would be easier to just leave the boat and let it sink by itself.

 

Last edited: Jun 27, 2025 at 5:32 PM

Can you elaborate on the countless ways you can manipulate it?
It's best to imagine you're trying to sell a beginner's Suno.

 

There have always been dull people, with or without GPT, so nothing new. This topic is about audio production techniques with the help of Suno: anyone who continues to troll with the usual BS about ethics and blablablah is out of place: that’s not useful to the discussion. I guess they should take their rants to Facebook and free up some space for us

 

Besides the Audio to MIDI that it's already mentioned and can open a new range of possibilities you can:

Extract the stems on the Software itself (not on RX or any other third party to avoid artifacts), download as .wav and start layreing that basic stem with quality synths as UVI Falcon.
Create texture with percussive and background AI stems using tools like RC-20 from sound sources you wouldn't find easily on sample packs (especially regional styles like middle east or latin music)
Change drastically it's tonal characteristics using MultiFX softwares like Shaperbox and Murda Melodies.
Feed some VSTi like Synplant or Factoid to craft 100% new thing that will ressamble nothing to the original if that's your wish.
Use the Output as LoopFX for transitions or intros to transpose the song to another atmosphere while you keep your composition intact.
Send the output stems to an Analog Rack while you record the other real intruments on the same analog chain if you're a musician and have instruments and gear at your disposal.
Do hard reamp and IR work on guitar samples to give em character.
Creating drumbreaks and fills with some tool like Stutter or Shaperbox.

Just to mention a few. I'm avoiding to provide those kind of information about creating music with AI as a tool, because there's plenty of information out there at the same time i don't wanna endorse it. If you don't miss a thing like @Djord Emer i'm not start advocating the AI workflow on you. Just use if you think you can take any advantage from it. My workflow is basically treating Suno as a .wav Scaler and texture maker.

 

I'm in this thread because i'd like to hear something produced with these techniques, and find this kind of stuff interesting. Sorry the previous OT...

 

@Dom_Perignon, you are beyond ignorant of your fellow human beings. You downgrade criticism as blah, chatter, and a troll. Devaluing others in order to increase your own value is your true nature.

 

Thanks @Riddim Machine for the expert and detailed description. Great job!

I only do preset sound design and sound collages, but of course I want
to be informed about what's going on and how everything works.

 

It's an open thread, the OP does not invite specific discussion of any such techniques other than to claim 2 weeks worth of work now take two days; which by the way, I'd really love to hear an album done in 2 weeks. Maybe in an elevator. So that is trolling from post #1. Compound that with the fact anyone who cannot already do all these great AI "techniques" probably should find a website full of rank amateurs to discuss those topics with.

And finally, some rando who has been here for 3 whole months isn't telling anyone what to discuss. We'll post whatever we like.

 

yes, I was talking about you

 

"Another little OT " I don't know much about these things, but do any of you use them offline with your GPU? Or do you only use online sites? Do you also train these ai locally?

 

If all you're doing is using AI as a kickstarter for ideas and heavily editing some of the stems while layering a bunch of other VSTi and live instruments on top, it's just not my type of workflow but it's a way of working most def. That being said, I thought you were going to mention the Kontakt instruments (which sounded more interesting, but still a hassle).