Resample to 32 kHz
Drop an audio file and export it at 32 kHz as a WAV. The rate carries everything up to 16 kHz, which covers almost all of what most listeners can hear while storing a third less data than 48 kHz.
A broadcast rate
32 kHz was the rate used for FM studio-to-transmitter links and for digital radio, where bandwidth was expensive and a 16 kHz ceiling was more than the transmission path could deliver anyway. DAT machines offered it as a long-play mode, and it still turns up in voice recorders and camcorder audio.
The 16 kHz ceiling
Half the sample rate sets the limit. Hearing above 16 kHz is uncommon past early adulthood, so on speech the reduction is close to inaudible and on music it mainly costs cymbal shimmer and the sense of air at the top.
Content above the ceiling is filtered out before the rate changes, which stops it folding back into the audible band as aliasing.
A friendlier ratio than 44.1
48000 / 32000 reduces to 3 / 2. The two sample grids align every few samples, so the resampler has far less interpolating to do than it does converting 48 kHz to 44.1 kHz, where the ratio is 160 / 147. The arithmetic is simpler and the result slightly cleaner.
When to pick it
- Matching an archive or a broadcast chain already running at this rate.
- Voice recording where 16 kHz of bandwidth is generous and file size matters.
- Halving data against 48 kHz stereo with a smaller audible cost than dropping to 16 kHz.
Other delivery rates
Music wants 44.1 kHz, video wants 48 kHz, and speech models want 16 kHz. For a mono downmix or another output format, use the full Sample Rate Converter.