Convert audio sample rate without changing the programme speed

State what must stay the same

When an owned recording needs another sample rate, the usual delivery goal is the same programme at the same speed and pitch. The output contains a different number of samples representing the same elapsed time. That goal is different from deliberately speeding up the performance or changing pitch.

Write the destination sample rate and channel requirement beside the source rate. Keep these fields separate from the audio codec and bitrate. A change in sample rate does not by itself specify the compressed file size or the complete quality of the result.

Use a duration calculation as a check

For a hypothetical ten-second mono recording at 48000 samples per second, there are 480000 samples. At 44100 samples per second, a duration-preserving output represents those ten seconds with 441000 samples. This arithmetic is an expectation for a simple example, not a demand that every container's reported duration match to an arbitrary decimal place.

FFmpeg documents resampling separately from setting a rate without changing the PCM data. The latter can change speed and pitch. Choose the operation that matches the delivery goal, and confirm its behaviour in the installed version rather than treating similarly named controls as interchangeable.

Keep conversion out of an unrelated diagnosis

A silent track, a wrong language and a rate mismatch are different problems. Identify the actual rejection or production requirement before resampling. Repeatedly converting a file because it sounds unfamiliar can make the original problem harder to inspect and create unnecessary generations of processing.

If audio must accompany picture, compare a recognisable event near the beginning and later in the export. A candidate that begins correctly but drifts over the programme has failed the practical task. Preserve the original rate and timing information until the result has been accepted.

Inspect the actual deliverable

Listen for unintended changes in voice pitch, clicks at boundaries and an altered ending. Confirm the output's rate, channel inventory and intended duration, then test it in the destination that requested the conversion. A successful processing command is only one part of that check.

Digital audio concepts by Mozilla Contributors provides background on rates and samples under CC-BY-SA 2.5 or later. The arithmetic and production review here are original. Keep the conversion recipe and source identity so the output can be recreated, without claiming that a higher numeric rate makes an existing recording contain more captured detail.

Related reading

• Media formats: codecs, containers and decoding
• Follow the audio from the programme to the speaker

Questions answered

Sources & further reading

Official references for the subjects discussed. Service terms and device software can change.

• https://developer.mozilla.org/en-US/docs/Web/Media/Guides/Formats/Audio_concepts
• https://ffmpeg.org/ffmpeg-filters.html