The @CadenzaChalleng I work on has just launched a novel series of ML/DSP music challenges. Tasks: listening over headphones and in the car. This is about improving the music for the millions of people who have hearing loss @SalfordScience cadenzachallenge.org/

Diff. between our simulated train and measured eval dataset:
* Talkers speaking differently.
* Real room acoustics.
* Directivity of interferers.
* Transducer noise.
* Lower order Ambisonics.
Clarity Challenge
---
RT @clarityprojuk
More details on the new eval set for signal processing grand challenge. Including a set of example recordings. Release date 1 Feb 2023.
claritychallenge.org/docs/icas
twitter.com/clarityprojuk/stat

This dataset for ICASSP2023 @clarityprojuk is for this research Q: "Can speech enhancement systems trained on simulated data generalize to more ecologically-valid measurement data?" We're measuring with an ambisonic mic and live actors.
---
RT @clarityprojuk
We're setting up ready to record our evaluation set for the @ieeeICASSP signal processing challenge on for hearing aids. St…
twitter.com/clarityprojuk/stat

if you enjoyed this exploration of pleasant sounds, @BBCRadio4 has a program at 11.30 today all about the sounds of the sea bbc.co.uk/programmes/m00193ng
---
RT @trevor_cox
We have a winner "Waves on a Beach" 🏆 🌊 🏖
twitter.com/trevor_cox/status/

final will be on Monday. Waves on a Beach vs Leaves Rustling in the Trees. After all the votes, we've come down to two sounds with really similar audio qualities!

Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.