Anicet Mbida 06:53, January 31, 2023

Anicet Mbida delivers to us every morning what is best in terms of innovation.

This Tuesday, he is interested in artificial intelligence in the service of music.

We are now able to create a melody by typing a few words on a keyboard.

The innovation of the day is a new feat of artificial intelligence.

After the homework of the students, it is now music that we are able to create by typing a few words on a keyboard.

It's called MusicML and it's the music version of ChatGPT.

That is to say, it is enough to describe the type of music that one wishes to obtain, and it will compose it in a few seconds.

An example: you can simply ask him “create an 80s club atmosphere for me”. 

And if we don't like it, we tell him "start again" and he will make another suggestion.

You can also be more specific in your request.

And say to him for example "make me a reggae style song with a melodious guitar and very expressive voices"

Obviously the voices say anything.

It just generates something musically similar.

Could we guide it with a melody that we already have in mind?

You can hum "Bella Ciao" and ask her "do the same to me new age style with bongos and flute..."

It's quite amazing because we can also give indications on the atmosphere and the emotions that the music should convey...

What is meant is that was generated with the description of Picasso's painting Guernica.

Suffering, chaos, violence, desolation...

So rather than rummaging for hours in catalogs of ready-made music, with this kind of tool we could create much more personalized sounds.

And is it available?

Can we try it?

No not yet.

Unlike ChatGPT, they try to avoid controversy.

Especially around copyright and plagiarism.

Because we can talk about artificial intelligence, these tools do not invent anything.

They only make a subtle copy/copy of everything they have been made to swallow.