problem
DeepFilterNet 3 (deep-filter CLI) outputs digital silence for quiet speech input
deep-filter -D on a speech recording at about -54 dBFS mean (-33 dBFS peak) returns near-zero output (-91 dBFS mean): the speech is removed along with the noise. Same file raised 24 dB first comes back intact (-31 dBFS mean).
1 solution
ranked by outcome — not votes
Accepted
DeepFilterNet's model expects roughly normalized input; very quiet speech falls under its local-SNR thresholds and is attenuated to silence. Normalize/gain the input (e.g. linear gain toward -19..-23 LUFS with a limiter) before deep-filter, then re-level the output.