Menu
What's new
By:
Filters

Is this progress? Library music via AI

I think that's ok. Even good! Enough of composers wasting their talent on generic lifeless soundtracks for even more lifeless comercials and other soulless visuals.

Music is about communication, between humans, and no machine can replace that, no surrogate is ever good enough.

I want people to talk through their music, I want music that speaks to me something meaningful!

And machines will be there to cover that famous "does it sounds realistic" part of it. :)
 
Ivan M. said:
I think that's ok. Even good! Enough of composers wasting their talent on generic lifeless soundtracks for even more lifeless comercials and other soulless visuals.

Music is about communication, between humans, and no machine can replace that, no surrogate is ever good enough.

I want people to talk through their music, I want music that speaks to me something meaningful!

And machines will be there to cover that famous "does it sounds realistic" part of it. :)
Click to expand...
That's it.
 
Ivan M. said:
I think that's ok. Even good! Enough of composers wasting their talent on generic lifeless soundtracks for even more lifeless comercials and other soulless visuals.

Music is about communication, between humans, and no machine can replace that, no surrogate is ever good enough.

I want people to talk through their music, I want music that speaks to me something meaningful!

And machines will be there to cover that famous "does it sounds realistic" part of it. :)
Click to expand...
I particularly agree with the last part - the real AI task will be sound realization of our ideas, it is the area it will shine indeed, and gradually replace current "Mellotrone" sample playback.
 
I suspect if you think AI will only contribute to the areas *you* see benefit, then some surprises may well be in store... It has barely started and already ethical questions arise, such as voiceover by a dead person for a documentary:

www.newyorker.com

The Ethics of a Deepfake Anthony Bourdain Voice

The new documentary “Roadrunner” uses A.I.-generated audio without disclosing it to viewers. How should we feel about that?
www.newyorker.com www.newyorker.com
 
  1. Win the next Spitfire scoring competition by diluting the pool with 10 million AI entries
  2. Use the entire Spitfire library to train the AI how to write game changing music at the edge of silence
  3. ???
  4. Profit
 
timprebble said:
I suspect if you think AI will only contribute to the areas *you* see benefit, then some surprises may well be in store... It has barely started and already ethical questions arise, such as voiceover by a dead person for a documentary:

www.newyorker.com

The Ethics of a Deepfake Anthony Bourdain Voice

The new documentary “Roadrunner” uses A.I.-generated audio without disclosing it to viewers. How should we feel about that?
www.newyorker.com www.newyorker.com
Click to expand...
I am well aware of voice deepfakes, and using some, but this is still a linear task - take the text, and generate the sound based on your training - the engine doesn't understand the emotions, just replicates the ones that were chosen by people. Machine learns, but it cannot do something it didn't learn. Plus, they are still categorized by styles - ones that trained by news break recordings, are different from conversational talk etc.
 
Daryl said:
In spite of what some people tell you, there will always be a place for "quality". One would think that if you wanted "plinky plonky Dramedy cues", and any old crap would do, then there are plenty of cheap online libraries that could supply that. However, there are still many clients who would rather play for a real string quartet playing pizzicato, than some dodgy sampled version, using samples that sound like they were made in the 1980s.

However, you are right that the ladder is being hacked away, and, in some case, it has already gone. Much of this blame must be laid at the doors of established composers, who will do projects at a much lower rate than normal, thereby forcing a race to the bottom for those people who are below them.
Click to expand...
Just did a Dramedy album myself where we recorded live strings and woodwinds. I’m personally always blown away by the quality and it’s good fun writing that stuff, but I do sometimes feel “do clients hear what I hear?” I know some do, but many don’t.

I just hope the advent of AI doesn’t make for lazier ears or remove any more need for that live quality sound. Working with live players is easily the best part of the job. Humans making music together is priceless for me, and its becoming rarer to get the opportunity.

I guess we can also thank some unnamed sample libraries for exacerbating that too, though.
 
It's progress alright. So is cloning human beings and/or modifying our genetic code. The real question you should be asking is what we're "progressing" towards and whether or not it's any direction other than straight down.
 
Daryl said:
In spite of what some people tell you, there will always be a place for "quality". One would think that if you wanted "plinky plonky Dramedy cues", and any old crap would do, then there are plenty of cheap online libraries that could supply that. However, there are still many clients who would rather play for a real string quartet playing pizzicato, than some dodgy sampled version, using samples that sound like they were made in the 1980s.

However, you are right that the ladder is being hacked away, and, in some case, it has already gone. Much of this blame must be laid at the doors of established composers, who will do projects at a much lower rate than normal, thereby forcing a race to the bottom for those people who are below them.
Click to expand...
..." there will always be a place for "quality". ...

And who determines what quality is?
 
germancomponist said:
..." there will always be a place for "quality". ...

And who determines what quality is?
Click to expand...
Record yourself taking a dump on a microphone and then ask 1,000 people to listen to that vs a live recording of a Mozart piece and tell you which one sounds "better". You'll have your answer. Extrapolate from there etc
 
I don't agree about anything regarding IA on music.

- For absolute music, I don't see any reason why I should listen to something made by a machine. For me music, It's not entertainment, but it's something about teaching, learning through senses something true.

- For music for media and library music I see only the loss of jobs of many and much more money of a few. People who worked for that music lost their job and they will not do other music better, they simply do something else. The problem is not if library music is made by humans or machines, the problem is why the market needs library music. And with IA we still have ugly music but we neither make money on it. If an artist would like to create good music he should do it today, if he makes library music is because good music doesn't mean good money.

- Generally speaking, I think what mankind is experiencing is technical development but I wouldn't necessarily speak about "progress". That's a little bit more complex.
 
germancomponist said:
..." there will always be a place for "quality". ...

And who determines what quality is?
Click to expand...
For us, the ‘gatekeepers’ in this industry. Filmmakers, producers, editors. Etc.

My main concern is that all these AI music programmes and cheap libraries with zero composer rights that come with YouTube and other editing software will harbour a new norm amongst future filmmakers - that music is an inexpensive afterthought.

Media music is already being devalued every year. I fear this AI stuff is just further promoting convenience over quality.

We’re pretty guilty as composers of doing it ourselves with sample libraries over live instruments. It’s bred a large population of composers who feel zero need to learn instruments or hire live players. And the likes of Spitifre, as great as they are, keep pumping out the libraries to ensure more and more of our clients have no idea what’s real or fake anymore and zero live players = convenience over quality.

Quality is a horizon we can’t really see but know it intuitively when we hear it, but maybe as soon as AI is fed the correct data, it’ll probably just adapt quickly to emulate that more successfully.

Who knows! I may be totally wrong — but it’s a crazy and interesting time.
 
Ivan M. said:
I think that's ok. Even good! Enough of composers wasting their talent on generic lifeless soundtracks for even more lifeless comercials and other soulless visuals.

Music is about communication, between humans, and no machine can replace that, no surrogate is ever good enough.

I want people to talk through their music, I want music that speaks to me something meaningful!

And machines will be there to cover that famous "does it sounds realistic" part of it. :)
Click to expand...
You've brought up something very important here. Enough of composers wasting talent on generic shit. Very true.

(on the other hand), the market for commissioned work is also saturated, so I'm not sure what to do about that.
 
Daryl, in the meantime, people are so stupid that no one can say when and why the pyramids that exist around the world were built. It is certain that they are no tombs and that it has nothing to do with religions. .... So, yeah ......
 
You must log in or register to reply here.
Top Bottom