3 ms·
the problem is it sounds awful, like a 64kbps MP3 or worse Perhaps AI can be trained to create music in different ways than generating spectrograms and convert
by PcChip 4y ago
the problem is it sounds awful, like a 64kbps MP3 or worse
Perhaps AI can be trained to create music in different ways than generating spectrograms and converting them to audio?
- w_for_wumbo 4y agoIt doesn't need to sound good at all for it to be useful. Like with the AI Art creation, it can be a starting point for artists to play around and rapidly try different concepts, and then interpret the concept using high quality tools to create something really quite remarkable. It's all about empowering artists to explore more possibilities.
- stevehiehn 4y agoExactly! UX/Pipelines/Integrations are the next logical step. It's my belief that samples will essentially be 'free' very soon. We will see DAW plugins/integrations that contextually offer samples to the composer. I'm confident in this because that's what I'm working on.
- magicalhippo 4y agoI haven't been paying close attention to the AI field but seeing someone painting with NVIDIA Canvas blew me away[1]. I can totally see how it can help prototyping or exploring ideas and visions. [1]: https://www.youtube.com/watch?v=uLlhyKxygrI https://www.youtube.com/watch?v=uLlhyKxygrI