You can take an MP3 and encode it into MIDI. The results are horrifyingly mesmerising. Can you hear the vocal?
(Stolen from https://www.tumblr.com/red3blog/135098280942/formeldeharv-i-put-all-i-want-for-christmas-is )
…and here's the MIDI version "enhanced" with Adobe's AI Podcast tool.
There is a deeply strange male voice in there at times.
(Thanks to @lazerwalker for the inspirations - https://xoxo.zone/@lazerwalker/111269073594821019)
cursed fact: Adobe Podcast's "Enhance AI", a tool for noise removal and voice boosting, firmly believes that any audio you give it *must* have human speech. If you upload, say, vocal-free chiptunes playing on a Game Boy, it will *find* the speech. https://lzrwlkr.me/3S424J7
@Edent So this is doing pitch-to-MIDI analysis? I kinda thought this would be a raw data to MIDI thing which would sound very, VERY weird.
On the other side, you can use tools to embed images in MP3s, as famously used in the soundtrack to the game FEZ. https://venturebeat.com/games/fez-hidden-images/
I was messing around with midis in win3.11 and this outdated meme is the result. The bass drops are my favorite part. The synth used in this is apparently "SB16 Extended Music Synthesis" according to Windows. Basically default soundblaster (emulated by dosbox) settings I suppose. I don't actually know shit about midis. edit: oh my god this was featured in a jpara video LMAO https://youtu.be/Pa-pZlh5sKw?t=371 "ew" -JasonParadise, 2021
@Edent Nice. Reminds me of Chopsticks, the talking piano.
@Edent OMG, I’m adding that to my Christmas mix *now*
Interesting that I can’t hear the vocal at *all*, even though I’m very familiar with the song and am trying to hear Mariah in that mess. Normally the brain hears what it expects to hear (viz the “Yanny” & “Laurel “effect, or “Green needle” & “Brainstorm”).
@Edent This is insane, horrible and mesmerizing all at the same time.
Really curious though, does anyone know how I can actually hear the vocals? Is it really possible to synthesize speech with only piano sounds?