Convert a voice recording without damaging it
Published · Updated 2 min read
A recorded voice, for example a lecture, a meeting or a message, handles compression well: it contains fewer frequencies than music, and the loss is hardly noticed. But each format has a cost, and a setting such as normalization changes the shape of the file in a way you do not always expect.
Formats for a voice
Take a 15-second voice recorded as WAV at 22 kHz, weighing 669 KB. The weight depends on the output format.
| Format | Weight |
|---|---|
| MP3 | 365 KB |
| AAC | 255 KB |
| OGG | 212 KB |
| FLAC | 353 KB, lossless |
What this means
FLAC loses nothing of the sound it receives, for about half the weight of WAV: it is the right choice for archiving. MP3, AAC and OGG weigh less and are enough for listening or sharing. One detail may surprise you: all four outputs set the file to 48 kHz, so the sampling rate goes up without the sound gaining finesse, since the starting recording did not contain those frequencies.
Normalize the volume
Normalizing brings the overall loudness to a reference value, -16 LUFS, used by listening platforms. On a voice recorded too low, it raises the level and reduces the gaps between a loud passage and a quiet one. The peak, meaning the loudest sound in the file, goes down at the same time to avoid clipping. Watch the weight: the result is a 48 kHz stereo WAV, which makes the file more than four times heavier than the 669 KB source, about 2.8 MB. Convert it to MP3 afterwards to lighten it.
Do not recompress
Going several times through a lossy format, for example an MP3 converted to AAC and then to MP3 again, gradually damages the sound, like a photocopy of a photocopy. Always start from the original recording, and keep the compressed versions only as sharing copies.
For which recordings
A lecture, a meeting, a voice message or an interview: a voice handles compression well, which lets you lighten the recording before sending it or transcribing it.
Check the result
- Listen to the start, the middle and the end of the file you get.
- Compare the level before and after.
- Keep the original recording.
To act only on the sound level, see Normalize audio volume.
Frequently asked questions
- Which format should you choose for a voice?
- MP3, AAC or OGG are enough for listening or sharing. FLAC keeps everything for about half the weight of WAV, the choice for archiving.
- What is volume normalization?
- It brings the overall level to a reference value, -16 LUFS. It raises a voice recorded too low and reduces the gaps between loud and quiet passages.
- Why is the normalized file heavier?
- The result is a 48 kHz stereo WAV, more than four times heavier than the source in the example. Convert it to MP3 afterwards to lighten it.
- Can you convert a file several times without loss?
- Not with lossy formats. Each pass damages the sound a little. Always start from the original.