google-research.github.io/seanet/musiclm/examples

We introduce MusicLM, a model generating high-fidelity music from text descriptions such as "a calming violin melody backed by a distorted guitar riff". MusicLM casts the process of conditional music generation as a hierarchical sequence-to-sequence modeling task, and it generates music at 24 kHz that remains consistent over several minutes. Our experiments show that MusicLM outperforms previous systems both in audio quality and adherence to the text description. Moreover, we demonstrate that MusicLM can be conditioned on both text and a melody in that it can transform whistled and hummed melodies according to the style described in a text caption. To support future research, we publicly release MusicCaps, a dataset composed of 5.5k music-text pairs, with rich text descriptions provided by human experts.

not perfect by any means but its impressive how smooth the generation is esp if youre only a casual listener not knowing what to look for

who will be the first rapper to hop on an AI-generated beat

when u dont have to pay the producer....

we need AI mixing and mastering now

wasn't Jai Paul working on something like this a long time ago where like you could upload stems and it would auto-master/mix it and he used jasmine as proof of concept?

Russ is in shambles right now.

@S

the first one under "Story mode" sounds demonic

what if he jus tryna steal ideas

That melodic techno sounds like 2013

jaipaul.bronze.ai

ok this hard

yeah this is what i was thinking of

Embrace it… Empower yourself… Obliterate your enemy.

its so hard to believe that some of these are AI man its over lol

These b**** ass Tech niggas really tryna get rid of artist in every field

Isn’t Landr that too (mastering)

F*** AI

wasnt familiar with this but googled it and yeah seems to be. not surprised

They’re taking er jerbs

So many big words in op