For voice recordings I was given the advice to use this parameter:
-cutoff 15000
The 15 kHz cutoff matches human voice fundamentals (up to ~8-10 kHz), cutting high-frequency waste while preserving clarity. Human speech rarely exceeds 8-10 kHz (sibilants like "s" top out there), so 22.05 kHz (Nyquist limit ~11 kHz) captures everything needed with zero excess bandwidth. Pairing it with your -cutoff 15000 prevents aliasing perfectly.
Is your feature request related to a problem? Please describe.
I can put -cutoff 15000 in ffmpeg, but not in ffmpeg-normalize, which yields an error
Describe the solution you'd like
Accept and pass the -cutoff option to ffmpeg
Describe alternatives you've considered
Use other -prf filters as mentioned in the manual, but they are very different.
For voice recordings I was given the advice to use this parameter:
-cutoff 15000
The 15 kHz cutoff matches human voice fundamentals (up to ~8-10 kHz), cutting high-frequency waste while preserving clarity. Human speech rarely exceeds 8-10 kHz (sibilants like "s" top out there), so 22.05 kHz (Nyquist limit ~11 kHz) captures everything needed with zero excess bandwidth. Pairing it with your -cutoff 15000 prevents aliasing perfectly.
Is your feature request related to a problem? Please describe.
I can put -cutoff 15000 in ffmpeg, but not in ffmpeg-normalize, which yields an error
Describe the solution you'd like
Accept and pass the -cutoff option to ffmpeg
Describe alternatives you've considered
Use other -prf filters as mentioned in the manual, but they are very different.