Downsample 32 kHz audio to 16 kHz at the same pitch
Admin ·
Lower the audio sample rate to fit a destination that expects fewer samples per second. Resampling filters the high frequencies which the new rate cannot represent and keeps the selected passage at its original speed.
Make a one-second, 16 kHz stereo WAV
Open the noisy stereo WAV sample. Select Start 0.75, End 1.75, an output rate of 16,000 Hz and WAV. The original interval contains 32,000 frames per channel. The exported interval contains 16,000 frames per channel and still lasts one second.
Check that both the 440 Hz left tone and the 630 Hz right tone are present at the same pitch. The resampler attenuates the original 10 kHz and 11 kHz components above the new 8 kHz Nyquist frequency, preventing strong spurious 6 kHz and 5 kHz tones from folding into the result.
| One-second stereo selection | Frames per channel | 16-bit PCM audio bytes |
|---|---|---|
| 32,000 Hz sample | 32,000 | 128,000 |
| 16,000 Hz WAV export | 16,000 | 64,000 |
Download the noisy stereo WAV sampleFFmpeg resampler options and cutoff
Choose the sample rate for the recipient
A 16 kHz file can suit a speech pipeline that requests it. For music, test a higher supported target if cymbals or other bright sounds lose detail. Keep the source as an editing master before downsampling further.
Read the receiving app's expected sample rate and channel count before exporting. This tool keeps the existing channels. A different step is needed if the recipient expects mono.
Separate audio bytes from total file size
The rate choices estimate the number of 16-bit PCM bytes for the selected length. A WAV adds header bytes, while FLAC stores the processed PCM with lossless compression. For a size limit, compare the complete saved files.
Select MP3 for a smaller mono or stereo file. For the same sample use 16,000 Hz and MP3, then verify 16,000 decoded sample frames in each channel at 96 kbps. Its container may report slightly more than a second because of MP3 encoder padding.
The browser reads the file, resamples the selection and builds the output locally. The clipping guard reacts to peaks that can occur during resampling. Review high-level passages in the output before replacing a working copy.
IETF RFC 9639, Free Lossless Audio CodecFFmpeg, LAME MP3 encoderFFmpeg lookahead audio limiter
Common questions
Will converting back up recover the high frequencies?
Upsampling creates more samples of the downsampled signal. High-frequency information filtered out during the reduction remains unavailable.
Which input rates allow a 16 kHz export?
Choose a supported source rate greater than 16,000 Hz. The workspace shows only output rates below the source.