Viewing profile — Sirsho29

The "cassette" framing is a nice reversal of what most voice-memo apps do -- treating the recording as the primary object instead of immediately reducing everything to text is a re…

The vocabulary-learning angle is the right problem to chase -- most ASR failures I see aren't model quality, they're proper nouns. Curious how the personal dictionary handles near-…

The real-time question is the interesting one -- most enhancement models are tuned for offline processing where you can see the whole clip. Have you benchmarked latency on live aud…