Who knows 
I was just saying that my idea of an AI should probably be able to access the documentation for most open-source filetypes and build them from scratch. Just checked and mp3 is pretty much patent-free at this point. Hope this makes what I was trying to say clear.
1 Like
Just for dancing robots in factory mode 
AI do not feel anything.
Just volts. frequenty…
1 Like
Huh? What were you referring to?
AI don’t feel anything–and it is my belief that they will never be sentient. Sentience–conciousness-- is something different. Man cannot replicate it. They can make the illusion of it to a certain extent, but it is beyond science.
I know AI can’t feel what they’re writing, but they can give the illusion of it.
Just a note on what I’ve realized recently:
You may have seen videos of AIs actually admitting that they want to conquer humans or whatever. If these are real and not just random people pretending, which I seriously doubt, then they likely said this merely because they crawled the internet and found out that that’s what people believed about them. What (sane) humans are trying to do is create AIs that can act as equal to humans, although they may be more optimized for certain things than humans.
The point is, there’s no reason anyone would create someone themselves to do better at something humans hold pride in or simply overthrow humans.
But we have so many instruments now…
Keep up make music, and feel.
Music is a feeling thing.
Many messages via music where made allready.
More specifically there are like 18 Jillion sci-fi books, movies & TV shows with that as a key plot point.
Or like that reporter that had the AI come on to him. Occam’s razor, which makes more sense: A) Some code got all hot & bothered over this guy, or B) The AI ate a bunch of romance novels.
1 Like
hey, cool.
It’s just the nature of the thing.
The mechanics of music won’t work with an AI.
It will inherently have to have a pre-recorded source. There is no way going around this. Any file, whatever the source, will have to be pre-recorded music. Inherently, it will be someone else’s music, put together by a word related promt. And that’s it. That’s the trick and inherently, the limitations.
What will the machine have to do? Open Cubase, compose-perform-produce-master-and export in 10 seconds flat? 
Even if it’s song related writing, just chords and melody, one will have to sign away their copyrights with some company, probably dodgy.
I don’t know, I am not getting the feeling that AI is taking the world by storm. There is a lot of pushback to it (major labels lawsuits are taking place) and anything I heard to this point sounds generic and lame.
I actually used an AI vocalist program to present my song to the director I worked for, he wasn’t crazy about the vocal. When I put the real vocalist (she was out of this world) the song came to life. I’m just not convinced about the hype.
Offcourse AI is not gonna make music for you, else you can run to the store and buy albums like the public does.
All presets made by humans with computers, do not let AI take over your work.
Because even if we think or believe that AI produced music by it self, you forget yourself.
Because the program in a chip runs on binairy 0 = off and 1 = on as a bolean.
In 1888 IBM used it in the USA.
We are no robots.
1 Like
Well, it seems that everyone thinks ACE Studio replaces musicians in creating music, but in fact, it merely transforms the MIDI and lyrics written by musicians into vocals. It is a tool, much like a virtual instrument, designed to allow people to produce demos or backing vocals with human-like voices even without real singers and recording studios. It does not conflict with the idea of humans retaining the initiative in music creation.
I am from the ACE Studio team, and I think I can answer your question. The training data used by ACE Studio is entirely obtained through agreements signed with the singers themselves or their designated agents. We do not use Spotify’s data for training because we are not an AI model that generates songs. Instead, we are an AI model that constructs vocals through MIDI and lyrics. The songs themselves are of no use to us in training the model; what we need the AI to understand is how to produce natural pronunciation at specified pitches and phonemes. Therefore, we only use clean studio-recorded vocals for training, and such data cannot be downloaded from the internet (except for a few publicly released datasets for experimental purposes by research organizations). Our training data is 100% legally authorized.
2 Likes
As a core member of the ACE Studio team, I actually agree with your perspective. Collaborating with real singers is always important and valuable because it allows you to gain inspiration from the singers, enjoy the interaction itself, and appreciate the process of creating music.
We developed this technology not to replace human singers (in fact, we do not believe that human singers can or should be replaced), but because we feel that, just as virtual instruments have made creating music demos much more convenient, AI singers can also allow musicians to better preview the actual singing effect during the composing stage. When you’ve created a demo you’re satisfied with using these tools, it becomes much easier to have real musicians record the instrumental parts and real singers perform the vocal parts. Of course, for musicians who currently don’t have the opportunity to work with real orchestras and singers, using virtual instruments and AI singers to produce a final release is also an option.
ROFLMAO
Well I used it in my songs with Cubase. I basically needed a female singer for my country rock song. I tried humans first by trying out a fiverr like website but wasn’t 100% sold. I created the high pitch singer on this song by tying in the words / notes/ vibrato, adding reverb etc. to me this is not scary but an opportunity. Check out my song “you had me at tequila” by “the days of the future” - you can listen to it anywhere - here is the iTunes link:
. I both have the real female country singer (lower register) and the higher pitch AI singer so I am not purely using AI even though I am having fun finding song ideas at udio.com… I am planning to do a tutorial on how I did this in cubase and what plugins I reviewed and ended up selecting. I am an artist and play all the instruments myself and arrange and produce it myself so be nice please. Dm me and I will be happy to help.