Skip to content

Comment on Hybrid-Net: Real-time audio source separation, generate lyrics, chords, beat

Comments

I tried it on one of my own tunes:

https://lamucal.ai/songs/adrian-holovaty/adrian-holovaty-the...

The beats/chords were consistently a full beat off, and the chords were probably only 50% right. I chose this tune because (to my ears) the harmony is pretty clear.

Compare this to my own manually created transcription of the same tune, and it's night-and-day difference:

https://www.soundslice.com/slices/tpbwc/

Beat detection and chord detection are hard problems, likely due to a lack of diverse training data. Chordify (another site that does this, which has been around for ages) has roughly similar performance.

Full disclosure: I run Soundslice, a website built around synced sheet music, in which there's no automatic transcription involved (maybe someday, but the tech isn't good enough yet!). I've been following these developments for 15+ years.

Thanks for your feedback. We're currently in the process of adjusting our dataset and model to address issues with chords and rhythm. We're looking forward to providing you with a better experience in the future.

Adrian, I made a Google Colab notebook to try a different beat detection algorithm with your tune. The results sound pretty good to me!

You can listen here:

https://colab.research.google.com/drive/1Pqgc9s-nBKxU_3Ap6K0...

sounds like your site could be a great training set, someone should reach out to you sounds like a good business opportunity.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.