21 comments

[ 42.0 ms ] story [ 1286 ms ] thread
Apparently there is no app for this? long live websites!
One more step on the way to the day when you can get help identifying that old tune that's been running thru your head for days.
I sing in tune but the tabulature that it wrote for the melody i sang was way off. Also im singing in B flat and the tabulature shows it to be in F sharp / G flat.
Ok I'm picturing one of those scenes from the original Star Trek where the computer melts down in a blaze of sparks when I sing to it.
Getting

> A Permissions-Policy header is blocking the microphone Your server is sending a header that switches this feature off for the whole page. Look for Permissions-Policy: microphone=() in your host config and allow microphone=(self) instead.

Despite allowing microphone.

(comment deleted)
My guess is this is a warning generated by a library trying desperately to reach the webmaster.

The server response header needs to be fixed:

> permissions-policy: camera=(), microphone=(), geolocation=(), payment=()

It's just really funny/confusing because now visitors to the site are seeing it, and the message is not meant for visitors.

It works in Safari because Safari does not support the header:

https://caniuse.com/mdn-http_headers_permissions-policy_micr...

What technology is this using? pYIN? Deep learning? I don't see any indication.
Looks good - is there a way to play back?
I'm afraid the computer would refuse to process my "singing" voice...
Is a printable note?

Edit: emoji fail.

Disclosure: it is a favor for my little sister. I made it for her in 5min without much testing on recognition accuracy. Which was enough to blow her mind ;)

Given your feedback, i might do another loop on this.

Please do. My kid has been messing with it but seems to be getting some notes wrong. Im sure some tweaking will make it behave correctly and I'd appreciate it.
Try ~crappy~ regular headset microphone and/or whistling for some basic edge cases :D

It worked quite well to start with, but then after a few bars deteriorated to mostly outputting pauses, and a stream of short notes (without pauses) when singing legato. Could be due to the aforementioned microphone, or is there some internal state that does not get cleared up when one clears up the score? I tried twiddling some of the variables that seemed related, but no noticeable improvement.

I really like the general idea, but accuracy was underwhelming. The biggest critique would be the note values / durations.
Reminds me of https://scorecloud.com/ "ScoreCloud Songwriter takes recordings of voice and instruments and turns it into Lead Sheets, with melody, lyrics and chords." except there is no live demo (but a free download for W$ and Mac, and an iPhone app) (didn't try)

(It's built in Common Lisp (LispWorks))

Very nice. Unfortunately with tunes with notes that are identical but with different sylabes, it's not separated clearly, but I believe this is unavoidable unless there is a speech recognition too. Very nice tool.
Interesting, but hard to tell how accurate it is. Converting to midi and having a playback from the transcription would help you tell how close it was.
I let my wife (degree in vocal performance) sing into it and what was transcribed was not accurate at all.