01
How much should be spent on sound?
More than most people assume, because it is nearly free. Sound is typically a small single digit percentage of a video file, so halving it barely changes the total while directly degrading the one thing the words on screen are built from. There is no useful saving here.
02
What audio actually costs in a video file
Halving the audio on a typical file saves under one percent of its size and directly degrades the thing the words on screen are built from.
| Video bitrate | Audio at 128 kbps | Audio share of the file |
|---|---|---|
| 5 Mbps | 128 kbps | About 2.5 percent |
| 10 Mbps | 128 kbps | About 1.3 percent |
| 20 Mbps | 128 kbps | Under 1 percent |
| 50 Mbps | 128 kbps | Around a quarter of one percent |
03
Problems a higher setting cannot fix
Echo from a bare room, a microphone placed too far away, and the hum of an air conditioner are all baked into the recording. Raising the audio setting on export preserves them more faithfully; it does not remove them. Most bad sound in talking video is a recording problem, not a compression problem.
The cheapest fixes happen before you press record. Move closer to the microphone, add soft furnishings to a hard room, switch off fans and fridges, and record a few seconds of silence at the start so any cleanup tool knows what the background sounds like.
A small clip-on microphone costs less than most cables and usually improves a talking clip more than any other purchase, because it keeps the voice loud and close while leaving the room quiet in comparison.
04
Listening the way viewers listen
Most people hear short video through a phone speaker or cheap earbuds. Both roll off low frequencies and exaggerate harshness in the upper middle, where sibilant sounds like s and t sit. A voice that sounds warm on studio headphones can sound thin and spitty on a phone.
Before publishing, play the clip on a phone speaker at half volume in a quiet room and then in a noisy one. If the words are hard to follow in either, the fix is usually a louder, more even voice level rather than any change in bitrate.
05
Keeping levels consistent across clips
Viewers scroll from one clip to the next without touching the volume. If your clip is noticeably quieter than the one before, many will scroll on before they hear the first sentence. Aim for a consistent loudness across everything you post, so your clips never feel like a sudden drop.
Loud peaks are the other problem. A laugh or a raised voice that clips into distortion sounds broken on every device. Leave a little headroom when recording, and let a limiter catch the occasional spike rather than recording everything close to the maximum.
06
Protect the part the words come from
- Leave audio settings alone when compressing. The saving is negligible and the cost is not.
- Cut video data instead. That is where the actual size lives.
- Check the recording, not the export. Poor sound at the source cannot be improved by any setting later.
Send a recording with good sound and hear the difference.
Questions people ask
- Does audio quality affect the words on screen?
- Directly. The text comes from what was said, so clearer sound produces more accurate text. It is the single biggest lever there is.
- Is speech easier to compress than music?
- Yes, considerably. A voice occupies a narrower range than a full arrangement, which is why speech survives heavier compression.
- Should I record in a lossless format?
- It is unnecessary for this purpose. A well recorded compressed track is fine; a badly recorded lossless one is not.
- What matters more than the setting?
- Microphone distance. Close and quiet beats every export choice combined.
- Does the platform compress the sound too?
- Yes, along with the picture. Sending something clean is what limits how much that shows.