AI article

I trained on two more kinds of damage. The kind I held out moved 0.2 dB.

Community description: I wrote a declicker, benchmarked it against two established tools, and it won by a lot. Then I noticed I had written the test set too.

Dev.to | Sep 12, 2026 | Christo

Automated excerpt

So I broadened the training damage model and measured again. So version 1.1 put both families in the mixer. Writing more kinds of damage into the training set bought competence at those kinds of damage. The model genuinely learned something about damage quieter than the music, transferred from being shown destructive damage that was louder. About 1.5 dB on percussive material, and roughly 4 dB of headroom on the faintest ticks: the default now finds harvested clicks down to about 7 dB over the local music level where the old weights reached about 4 dB.

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.

Read the original article

More AI news