Two numbers describe how audio is digitised. Both are widely misunderstood, and one of them matters much more than the other.
Sample rate is how many times per second the audio is measured. 44,100 times per second is 44.1kHz.
The rule that decides this is the Nyquist theorem: to represent a frequency accurately, you must sample at more than twice that frequency. Human hearing tops out around 20kHz, so 44.1kHz (capturing up to ~22kHz) covers it.
So why go higher? Two genuine reasons:
- Lower latency. At 96kHz, a 128-sample buffer is half the time it is at 48kHz. Useful when tracking.
- Nonlinear processing. Distortion and saturation generate harmonics above the original frequencies. At a higher sample rate there is more room before those fold back as aliasing. (Though most good plugins now oversample internally, which solves this without a project-wide rate change.)
Why not always go higher? Double the rate is double the file size, double the CPU, double the disk throughput — for a difference almost nobody can hear in a blind test.
Practical answer: 48kHz. It's the video standard, it gives slight headroom over 44.1kHz, and it's what most people work at.
Bit depth is how precisely each sample is measured. This one matters more.
- 16-bit = 65,536 levels ≈ 96dB of dynamic range
- 24-bit = 16.7 million levels ≈ 144dB of dynamic range
- 32-bit float = effectively unlimited practical headroom
The practical consequence is headroom. At 24-bit you can record with your peaks at -18dBFS and still have vastly more resolution than a 16-bit recording peaking at -1dBFS. That means you never need to record hot, which means you never clip.
32-bit float recorders take this further: they cannot clip in any practical sense, because the format's range exceeds anything a converter produces. If your interface or field recorder supports it, it removes gain-staging anxiety from tracking entirely.
| Stage | Setting |
|---|
| Recording | 48kHz, 24-bit (or 32-bit float if available) |
| Mixing | Same as the project — never change mid-project |
| Master delivery | 48kHz or 44.1kHz, 24-bit WAV |
| CD delivery | 44.1kHz, 16-bit with dither |
| Streaming upload | 24-bit WAV; the platform re-encodes |
Never change sample rate mid-project. Resampling introduces artifacts and pitch problems.
Only dither once, at the very end, when reducing bit depth for final delivery. Dithering repeatedly accumulates noise.
Related reading: audio file formats explained, gain staging explained, and buffer size and latency.
Start making music in Veena
Free, browser-based, no downloads required.
Try Veena Free