AI disqualifies anyone as a musician! It's like playback.

[enroe]

... People are looking at this as some binary choice, and it’s not. Before the day of the computer, most of us were forced to pick a focus or two ...

Thanks for your detailed explanations, zerocrossing!

Actually, I do think it is a clear-cut decision — and it is a pact
with the devil; after all, that is precisely what this thread is all
about.

However — with the two points that must be distinguished —
and which are simply ignored by many here:

1. As long as you use AI as a tool, you are doing exactly that:
AI serves as a tool — be it for sound shaping, EQing,
compression, or mastering. There is absolutely nothing wrong
with that.

2. If you are not a composer, but rather a performing artist —
that is, a conductor, an instrumentalist who reproduces all
manner of music, and so on — then this entire discussion does
not concern you.

-----------------------------------------------------------

And now we come to the crucial point: If you use AI for
composition – that is, to completely recreate an entire song or
entire song sections – then you are indeed crossing a red line.
Because then it is no longer "YOU" who is creating the song,
but the AI!

If you don't play drums yourself, why not a. hire a drummer? Or
b. learn or edit drum tracks yourself?

If you aren't coming up with the text yourself, why not a. ask a
friend or acquaintance? Or b. learn to write texts yourself?

As soon as you delegate something genuinely creative to AI, you
have brought the AI ​​on board — perhaps even put it at the wheel.
You can no longer claim, then, that *you* created the song — or
that you are expressing *your* feelings through it. You have
crossed the red line; you are no longer truly the master of your
own house. Instead, you morph into something more like a
master of prompting within a virtual world.

Perhaps, however, this is precisely the path to the future: AI
creates songs because it is cheaper — and not necessarily
worse. Executives at major record labels, as well as newcomers
in startups, will take a close look at this — and just like that —
after only a short time, if not indeed already, we will see a tsunami
of AI-generated songs bearing down upon us.

A small (but creative and talented) composer like zerocrossing
will then be completely overlooked. As an honest and conscious
person, I therefore appeal to us all to draw this one clear line –
at least for ourselves personally.

[zerocrossing]

If you don't play drums yourself, why not a. hire a drummer? Or
b. learn or edit drum tracks yourself?

Learning to play drums is a great example. A drum kit takes a lot of years of practice to become proficient. It takes a fair outlay of money to buy a good drum kit, including all the mics needed to capture it. It takes a lot of money to afford a place that has room for a drum kit and properly treat the room. You could bypass some of this by buying a small digital kit, but you still have to commit to learning another instrument. Hiring a drummer? More money, if you want a good one.

Does that make sense if you're a singer/songwriter who's somewhat proficient in guitar or piano and who wants to hear fleshed out arrangements of their music? Not really.

As soon as you delegate something genuinely creative to AI, you
have brought the AI ​​on board — perhaps even put it at the wheel.
You can no longer claim, then, that *you* created the song — or
that you are expressing *your* feelings through it. You have
crossed the red line; you are no longer truly the master of your
own house. Instead, you morph into something more like a
master of prompting within a virtual world.

What's "truly creative," though? It's a slippery slope, for sure. Often song writers have a copyright for the basic words, melody and chord structure of a song, but the band might have arrangement copyrights, or performance copyrights. Maybe you're a pianist who doesn't want to write a song, so you have Suno generate one, and you add your piano part. What does that mean? You've avoided paying a label for rights to the music, and instead payed Suno, but now at least your song can get posted that shows off your piano skills. You're still a musician in this scenario, just not a composer.

Perhaps, however, this is precisely the path to the future: AI
creates songs because it is cheaper — and not necessarily
worse. Executives at major record labels, as well as newcomers
in startups, will take a close look at this — and just like that —
after only a short time, if not indeed already, we will see a tsunami
of AI-generated songs bearing down upon us.

I think that's already happening, but I also think that it's a fad. I honestly don't know what a lot of AI music sounds like, but I do see a lot of AI visual art, and I can almost always pick it out. Like 95% of the time, I'd estimate. I don't know why I can, but I can. Even the most "realistic" stuff has a vibe that calls itself out. I imagine the music is similar.

A small (but creative and talented) composer like zerocrossing
will then be completely overlooked. As an honest and conscious
person, I therefore appeal to us all to draw this one clear line –
at least for ourselves personally.

HA! The jokes on you! I have been completely overlooked as a musician my entire life! No AI necessary!

Thanks for the complement, though.

[VOODOO U]

Christ. It doesn't take "a lot of years of practice to become proficient" with drumming. Even If you're aiming for a Neil Pert style you can do it with disciplined continual play and it will happen in less than 90 days unless you're too lazy or lacking passion due to "life getting in the way" (which doesn't mean it takes 'years').
I got my first kit at 9 years of age, took my first lesson and within that hour I was drumming to Joan Jett and Twisted Sister right there at the instructor's home. One hour. Mind you it wasn't anything fancy-filled but I could've backed up a funk band right then and there no issues.

[BONES]

Actually, I do think it is a clear-cut decision — and it is a pact with the devil; after all, that is precisely what this thread is all
about.

It's a pact you should seriously consider because I think it could only make your music better.

1. As long as you use AI as a tool, you are doing exactly that:
AI serves as a tool — be it for sound shaping, EQing, compression, or mastering. There is absolutely nothing wrong with that.

OTOH, when you carry on like this, you become the tool.

2. If you are not a composer, but rather a performing artist —
that is, a conductor, an instrumentalist who reproduces all
manner of music, and so on — then this entire discussion does
not concern you.

Oh, but it does. The AI does some of our vocals, both lead and backing, on stage. There will also be plenty of AI instrumental backing on stage, too. Of course, I also record backing vox to use on stage for some older songs and most of the music is sequenced, so it's not all AI, but there will be plenty of it to be heard at future gigs. And the best part of it is that as we finish songs for the new album, we are running them through an AI song checker and so far it's suggesting that there is less than a 5% chance that any of the finished songs were not made entirely by humans.

You want to make a distinction because it suits you to but AI can be equally pervasive on stage as it is ITB and even other AI's can't tell the difference.

And now we come to the crucial point: If you use AI for
composition – that is, to completely recreate an entire song or
entire song sections – then you are indeed crossing a red line.

No, we're not, because that's not where we draw our red line. Our red line would definitely prevent us from churning out soulless krap as bad as the stuff you put out but it doesn't stop us using Ai to make better music.

Because then it is no longer "YOU" who is creating the song,
but the AI!

No it's not, it is my bandmate directing the AI. It's a co-writing effort, or a directorial effort. I think I'd rather be a director than a composer, wouldn't you?

[Scotty]

We are running them through an AI song checker and so far it's suggesting that there is less than a 5% chance that any of the finished songs were not made entirely by humans.

Hell Bones to eliminate the 5% chance of being discovered as an AI wanker below are 5 usable prompts that will help you further fake your punk sensibilities. These are guaranteed to work so you won't have to check to see if the GPU output is in stealth mode. Stay tuned on how to fake the look and attitude associated with your chosen genre. It is important that we keep it "real".

Raw Basement Hardcore

“Early 80s basement hardcore punk recorded live in one take on cheap cassette equipment, blown-out drums, distorted bass, harsh guitar buzz, exhausted shouted vocals, uneven tempo, gang shouts in the chorus, room mic bleed, clipping microphones, DIY rehearsal space energy, raw and chaotic, no polished mastering.”

Lo-Fi Street Punk

“Dirty street punk with sneering male vocals, loose drumming, cheap amp distortion, sloppy guitar timing, shouted gang vocals, cigarette-smoke atmosphere, recorded in a cramped garage with old analog gear, aggressive but imperfect performance, tape hiss, rough mix, anti-commercial energy.”

DIY Garage Punk

“Lo-fi garage punk recorded on a broken 4-track recorder, cracked vocals, overdriven guitar amps, primitive drumming, messy transitions, distorted room sound, cheap microphones, reckless live performance feel, imperfect tuning, chaotic but catchy hooks, raw underground punk aesthetic.”

Crust Punk / D-Beat Style

“Bleak crust punk with blown-out production, aggressive shouted vocals, filthy guitar tone, relentless d-beat drums, rehearsal-room acoustics, overloaded tape saturation, raw political energy, chaotic mix, clipping cymbals, rough live-band feel, intentionally unpolished and abrasive.”

1977 DIY Punk Single

“1977-style DIY punk rock single recorded fast and cheaply, nervous vocalist, buzzing guitar amps, thin bass tone, primitive drum recording, short angry song structure, mono-ish mix, tape imperfections, spontaneous energy, amateur but passionate performance, underground club atmosphere.”

“DIY garage punk recorded in a damp basement, cracked shouted vocals, distorted rehearsal amps, sloppy drumming, cassette tape hiss, raw anti-commercial energy.”

You can also experiment with deliberately contradictory prompts because AI models sometimes produce more interesting results when forced between opposing aesthetics, like:

“catchy but collapsing”
“melodic yet hostile”
“tight hooks with messy performance”
“anthemic chorus recorded badly”

Those contradictions often create outputs that feel more human and less algorithmically “perfect.”

[Zeisner]

As long as you use AI as a tool, you are doing exactly that: AI serves as a tool — be it for sound shaping, EQing, compression, or mastering. There is absolutely nothing wrong with that.

AI tools suck. They don't make things better but worse. Just look at all those vibe-coded AI "assistants" that keep popping up here, giving you "advice" like "You need a midrange EQ boost to reach -10 dB LUFS" without understanding what LUFS really is nor the story of the song (Is it even supposed to have that in-your-face loudness of a pop track? Does it even belong to the genre the AI "thinks" it does?). Just stupid stochastic parrots designed to give you the illusion of mixing/mastering. Machine learning is no substitute for trained ears listening to good speakers in a good environment (and of course good hardware/software). There is no shortcut to excellence.

The equivalent of mixing/mastering with AI is to take lots of random tracks from the web (without listening of course), doing a spectral analysis with Audacity, exporting the data via CSV and then using Excel to calculate the median for each frequency bin using that data. Then you twist your stock EQ knobs until your track's spectrum matches the calculated medians. Still without listening of course. Because the median is always the optimum... right? Would you call that mixing or mastering? If so, then prompting "Happy uplifting pop song" in Suno counts as composing and performing as well. Anything else would be a double standard.

[BONES]

I do see a lot of AI visual art, and I can almost always pick it out. Like 95% of the time, I'd estimate. I don't know why I can, but I can. Even the most "realistic" stuff has a vibe that calls itself out. I imagine the music is similar.

I see no problem with that at all. In fact, I preferred MidJourney when it was completely obvious that it was doing it. It had a style that really worked for me. Things like these -

f69c248a-2e16-4854-9f86-a722b5399478 - a nighttime cyberpunk city with clean high rises in the background and grunge industrial in the foreground, photorealism, 8lk, tokyo,.png

df434193-4def-4039-bc1c-b1f89f3c50d2 - a dark grunge cluttered urban decay foreground, with futuristic cyberpunk towers behind it, by Artur Sadios,.png

6656c4de-1035-4393-971f-1e972520c021 - a cyberpunk city with clean high rises in the background and grunge industrial in the foreground.png

Now, in their pursuit of perfection, it's lost of lot of its quirkiness which has ruined it for me.

HA! The jokes on you! I have been completely overlooked as a musician my entire life! No AI necessary!

That's the thing that's true for all of us and it's mostly for the simple reason that, whether you can admit it to yourself or not, we're not good enough. We do what we do, it keeps us happy and amused but if it was actually good, we'd be doing a lot better from it. I think having the drive to succeed is also a very big part of it. You can overcome most obstacles to success with the 10,000 hour thing but unless you have the drive to push your barrow to the top, or find someone (manager, parent) who will do it for you, you'll never get anywhere. That's simply how the world works. You have to want it more than the other person.

Christ. It doesn't take "a lot of years of practice to become proficient" with drumming.

Speak for yourself. I could do the full 10,000 hours and I'd still be shit at it because I have no hand-eye co-ordination to speak of. I am hopeless. It's the same for me at work. I've been earning a living as a graphic artist for 30+ years but I can't draw to save my life. If it wasn't for computers, I couldn't hold down a job at all. Meanwhile, my bandmate is the opposite, he can pick up anything in no time at all. He's a natural sportsman, he picked up the drums pretty much instantly and I need to practice about ten times more than he does, just to be half as good as he is at playing keys.

I got my first kit at 9 years of age, took my first lesson and within that hour I was drumming to Joan Jett and Twisted Sister right there at the instructor's home. One hour. Mind you it wasn't anything fancy-filled but I could've backed up a funk band right then and there no issues.
/quote]
But how would you have gone if you'd been the one setting the beat?

Machine learning is no substitute for trained ears listening to good speakers in a good environment (and of course good hardware/software). There is no shortcut to excellence.

Of course there is, you outsource it. You pay someone else to do or you give it over to an AI to do for you. The Master Assistant in Ozone, for example, is a great way to get a good starting point from which to do a bit of tweaking. I let it do all the technical stuff, which leaves me to make the more creative decisions around genre/style.

The equivalent of mixing/mastering with AI is to take lots of random tracks from the web (without listening of course), doing a spectral analysis with Audacity, exporting the data via CSV and then using Excel to calculate the median for each frequency bin using that data. Then you twist your stock EQ knobs until your track's spectrum matches the calculated medians.

Sure, but who can be bothered doing all that themselves when the AI will do it for them in seconds?

Would you call that mixing or mastering? If so, then prompting "Happy uplifting pop song" in Suno counts as composing and performing as well. Anything else would be a double standard.

So what? What is it you think is so f**king special about being a performer or a composer? I've been getting sequencers to do most of the music performance for 40 years, it's a bit late to start worrying about it now. And for the last 20-odd years I've left much of the creative side of composition to my bandmate. I don't have an ego to feed, I just want to hear/have good music to play on stage. I love playing covers, I love playing our original music but I don't think I have ever been as excited about any of that as I am about performing our new AI material to an audience. Why? For one simple reason - it's the best stuff we've ever had to work with.

[VOODOO U]

I am hopeless.

Good. Stay hopeless. Makes room for those of us who get out of pathetic mode and make it happen.