Personal opinion, but I wish autotune would have been ever invented.. I can see it as effect sometimes, but man that crying in puke in your mouth..

 

Because people taking their first tentative steps into singing are already fragile and lacking confidence. It isn't always immediately obvious that a singer is autotuned, especially not to a beginner, they will probably think that's the standard they have to be at, which is essentially impossible. Beginners might not even know what autotune is and certainly not realise that they've been conned by their favourite artistes.

 

What is Art made of? what are its ingredients? In my opinion it is basically made up of rational use of skill and irrational creative imagination. Where technology is helping us a lot is in our skills (or lack of), speaking about music:

• Do you suck at playing? it doesn't matter, here you have a sequencer/daw
• Do you suck at composing/arranging? It doesn't matter, here you have an auto-composer (Cthulhu, Scaler, any arpeggiator..)
• Do you suck at singing? Here's your Autotune - Melodyne

The tools' list to help the artist's skills can be much longer, but without the second part, the most important one, Imagination, the art will not come out. I'm working with AI quite a bit, and from what I see, so far I haven't found any spark of imagination in it.
The dream of technological man has always been to build the machine with a soul, any science fiction enthusiast knows this topic well, it has been covered a lot by novelists (Any novel from Asimov, Philip Dick' s "Blade Runner" or even the Frankenstein saga to name a few). For now, the project of a "machine with a soul" still remains a dream, in my opinion, it won't be ChatGPT, Microsoft Copilot, or even Suno.ai that will make this dream come true: there's still a long way to go

 

Auto-composers, Auto-tune, Auto-EQ, Auto-harmonize, auto-scale.... The list is getting bigger every year....

The great thing last century was that you knew instantly who was great, who was just good and who needed to do more work.
Kanye has already made millions and openly admitted he cannot play any instrument well. So with the all the 'auto' anything out there, are people composing, creating and using their imagination? That is a big fat NO. The computer already has built-in algorithms to choose from, built-in decision-making based on what is selected.

TRUE imagination and creativity requires individual inspiration, not spoon-feeding from a any source. If something else is giving you ideas, logic says beyond doubt, that it is not your own imagination and unfettered inspiration feeding you, the computer is. If a person cannot play an instrument, they should stop fucking pretending they are a musician because they are NOT. They are a computer user who uses computer tools to generate something they could not play on an instrument. Congratulations Kanye lovers, you gave millions to someone who is a shit musician.
If Kanye's AI was truly intelligent, his computer would demand 90% of the money.

Milli-Vanilli got globally shamed, shunned and removed pre-2000 by doing what everyone on Tik-Tok does - mime to something they did not sing - now videos have already been mastered, pretending to be live. So A.I lovers are basically saying, it is ok to bullshit that you're a decent musician and singer when neither is true. SAD TIMES.

 

Last edited by a moderator: Dec 28, 2023

^ that's why I love small jazz groups live
The tech is still interesting though

 

Most people don't know how music works and Interestingly enough, don't want to know because ...... .

If they know, they won't use these platforms.

 

Note, not always the discouragement to sing is due to pitch, but rather, how they sound to themselves on a sound player. I remember when I was a teenager and wanted to encourage my friends to sing. My dream was always to have a band made up of close friends, that is, best friends. But I would say to them, "Let's record something, just sing, cheer up!" So, I would turn on the recorder; I didn't have a PC at that time, few people had PCs. So, I would turn on the cassette recorder, and we would record ourselves. Then, when playing back their voices, they would say, "What? Is that my voice? I sound horrible; I have a 'shrill' voice or an 'ugly' voice." And this happens due to a physical phenomenon. Someone explained it to me once, and I don't remember it entirely, but it's something like this: we hear our voice from the "inside," and it has a deeper and more resonant sound because our voice resonates in our cavities. But when we emit our voice, it's not the same as what we perceive from ourselves.

So, the disappointment for many singers is that they believe in their minds that they sound like "X" thing, but then they hear themselves, and this goes beyond pitch. Now, voice cloning will surely help in this because they can adapt their voice. For example, I could mix my voice with someone else's, let's say McCartney, and put it through training. I would get a Rick McCartney, meaning it would be my voice with shades of McCartney (in fact, I tried it out of curiosity). It's similar to Auto-Tune; they will deceive themselves, perceiving something they are not.

The first step in everything (I believe) is self-acceptance. This is how I sound; this is my voice, and I want to give it my style. Moreover, people will recognize you for your voice, for being yourself.

 

I know it's likely to improve but I gave it a whirl and It seems to construct some rather boring, imo, popular cookie cutter stuff. Anything more imaginative I requested it couldn't do well. Even some basic stuff like Coldwave, for example, it really sucked at it.

I don't feel threatened at all. I ain't skeered.

Ultimately I don't care either because I'm still going to make music the way I've always made music.

 

Wow, just tried it (suno.ai) and it surprised me quite a bit, hahaha. I just entered a psychedelic title and it created a song from that. Not bad at all for being an algorithm or whatever is behind it. What I mean by this is that it's interesting. From my perspective, I wouldn't call it 'creating' as I said before, but it's a good starting point. I would say it's like going on Spotify and hitting shuffle with songs, but here you can input lyrics. Interesting

Edit:
I was noticing that the AI struggles to provide cohesion or structure. For example, out of ten generated songs, two may have interesting parts that a musician could refine, while the others might not be as compelling. Nevertheless, if two out of ten show promise, it's a good starting point. However, the challenge lies in improving the overall cohesion. There's an issue where a melody starts with the voice and, as it rises in tone, instead of reaching a conclusion, it leaves the melody open-ended. This lack of cohesion is similar to saying, "Wait, is something missing here?" An example of what I mean is, let's say I have Am F E and resolve it with Am; the AI might generate Am F E and then go back to F, disrupting the cohesion as it should have resolved on Am. This inconsistency can also be observed in the melody. So, for a less inspired musician, it might serve as a starting point to generate ideas.

 

Last edited by a moderator: Dec 29, 2023

Yep. You tried it. It out out a garbage song. That's prove of: garbage in - garbage out. So the AI renders you and your abilities obsolet. See where that goes?

 

AI guys are master in wasting people's time.

 

Suno is quite inconsistent, there is a lot of trial & error to do, adjusting the lyrics and the musical genre: be prepared to press the create button at least 50 times before listening to something almost-interesting. (obviously I bought the premium). Yes, it often comes up with nonsense chord progressions, in fact, what I do when I finally have an acceptable track is to break down the instrumental and vocal stems, and then replay it with my synths, giving it a more logical sense of arrangement. I noticed that the algorithm always tends to create lyrics with a positive sense: for some reason I can't make it sing something introspective, suffered, sorrowful.
What I would like to have is the complete list of template genres that the AI has in the database.

 

Well, initially, what you're mentioning is something creative, as you don't just stick with what's generated; you break it down and create something new. In fact, this has been happening for a while. Do you know how many "composers" take things from other groups to later reformulate and create something new? That's why I mentioned at the beginning, and maybe it was misunderstood there, that one thing is pressing a button and generating something and sticking with that; that's not creation, that's generating something. In fact, I use AI for many things, including voice cloning, instrument cloning, etc. I've even cloned my voice for days when I'm tired from singing too much (I sing in the metro). I sing in a low tone, and then with cloning, I guide myself on how it would sound to sing it on the day I recover my voice, and so on.

Having said that, what I saw on the website is that it seems to generate chords more according to the lyrics than the style or genre. For example, if I select rock punk and a romantic lyric, it generates a romantic pop song instead of a rock punk song with romantic lyrics, do you understand? I think there's an error there. Another thing that could be added or modified is, in addition to the style or genre, for example, you input a lyric and then input the name of the artist, let's say Queen. Then the AI could generate a song according to the database it has of Queen's music. That would be interesting, but going a bit further, you could input Queen+Nirvana, and you could hear how a song mixing these artists or bands would sound, because it's important to note that many bands or artists don't have a defined musical genre but do have a unique "genius." As I said before, I find it interesting and will always support all technology or AI, but it should be for creating, meaning a parallel work between humans and AI. In fact, ChatGPT itself, if you ask whether AI will replace humans, it says the intention is to work together. In conclusion, it seemed interesting to me. If they gave me 50 credits, it generates two songs of 1 minute each, but it's enough for a musician who needs ideas or filler.

 

If you're that damn delicate, maybe it's not yours! Art is hard, the music business is even harder. Do something else where you feel safe! Because you will fail anyway and you may lose money (their money). Because to do anything meaningful, you have to fail a billion times.

 

Are you aware that many great singers have imposter syndrome? Also, producers know that singers have fragile egos and when you think about it they're opening themselves up through their songs so they're naturally hypervigilant to criticism. There's nothing worse than an arrogant singer, it's even called lead singer syndrome (LSS) and it's a real thing. That sounds like what you're describing.

You also fail to consider that I was talking about beginners, not seasoned pros who have been down the well trodden track of hardship for decades (which I have). The business toughens you up, you usually don't start out that way. We all start out as dreamers.

 

What are you talking about?
If somebody has a syndrome (any), he/she should see a qualified specialists. You are not the one, I suppose. And the luck of autotune doesn't fix syndromes anyway.
On the contrary, such things will help you: if you cannot sing live, just record it. If you are not confident with your n word (whatever), mix it with a DAW. If you are not sure about pitch control, use autotune. And finish the song.

 

More importantly, with the AI help, you can see your idea finished within minutes. It may be better or worse, but finished. What artificial intelligence excels is prototyping. You can start with an idea, and within minutes the artificial intelligence will create hundreds of takes for you to choose from. And this is more important than having many projects started but never finished.

 

You seem to have everything figured out. Good for you!

 

That's it https://en.wikipedia.org/wiki/Automaton

 

On December 25, a science fiction novel 'The land of machine memories' created entirely by artificial intelligence (using 66 prompts and taking up to three hours) wins an award in China.

In April 2023 photographer Boris Eldagsen won the Sony World Photography Awards for his AI-created image 'The Electrician', but declined the award in protest against AI.

Tesla, the robot revolt: two years ago an engineer was attacked and injured by Optimus (the humanoid robot) but the story was kept secret. A tragedy averted, witnesses said, speaking of 'a trail of blood' left on the factory floor.

Guys nightmares advance quickly and never come alone... draw your own conclusions.

 

Last edited: Dec 29, 2023