Voice isolator

deepfilternet3

DeepFilterNet 3. Same extra knobs as DeepFilterNet 2.

Native waveform: 48 kHz, keeps mono or stereo. Live price and availability are on GET /v1/models/deepfilternet3.

Parameters

FieldTypeDescription
params.attenuation_limit_db
Default model default
number0 to 100. 0 means no attenuation.
params.post_filter
Default model default
booleantrue applies the post filter.
params.lsnr_min
Default -15
number-40 to 10.
params.lsnr_max
Default 35
number0 to 80.
params.output_format
Default mp3
stringwav, flac, mp3, or model.
params.output_sample_rate
Default source
string or integersource, model, or one of 8000, 16000, 22050, 24000, 32000, 44100, 48000.
params.output_channels
Default source
stringsource, model, mono, or stereo.

Example

curl https://api.pascalai.co/v1/voice-isolator \
  -H "Authorization: Bearer $PASCAL_API_KEY" \
  -H "Idempotency-Key: $(uuidgen)" \
  -F model=deepfilternet3 \
  -F params='{"attenuation_limit_db":80,"post_filter":false}' \
  -F file=@speech.wav

curl https://api.pascalai.co/v1/voice-isolator/$JOB_ID \
  -H "Authorization: Bearer $PASCAL_API_KEY"