Definition of sample rate
Sample rate is the number of signal samples per second in digital audio, written in hertz. A 48 kHz stream represents 48,000 sample instants per second per channel. Under ideal sampling theory, the Nyquist frequency is half the sample rate, so 48 kHz corresponds to a theoretical 24 kHz boundary.
That boundary does not state how much useful content a recording contains, how an anti-alias filter behaves, or whether a listener prefers it. Sample rate is not kbps, and it is not an overall quality score.
Definition of bitrate
Bitrate is the amount of encoded data used per second, commonly shown as kbps for delivery audio. In a fixed-bitrate compressed model, duration × bitrate provides the payload estimate. A codec uses those bits according to its design; the same number across different codecs or encoders does not guarantee an identical result.
Average VBR bitrate can be used in the same size formula with uncertainty. CBR vs VBR Audio focuses on that distinction.
How sample rate and bitrate affect file size
For fixed-bitrate lossy audio, do not multiply by sample rate again. The selected bitrate already expresses encoded bits per second, so multiplying it by sample rate would count data twice. Sample rate can still affect encoder behavior and bandwidth, but it is not another direct multiplier in this model.
For uncompressed PCM, sample rate, bit depth, and channels are all direct multipliers. Doubling channels doubles the PCM data. Doubling sample rate also doubles it, assuming duration and bit depth stay fixed.
Compressed audio vs PCM
The interactive switch changes both the formula and the list of direct inputs. Compressed mode uses duration and bitrate. PCM/WAV mode uses durationSeconds × sampleRate × bitDepth × channels ÷ 8, with a minimal header allowance. That contrast is the main reason a 16-bit stereo WAV is much larger than a modest delivery MP3.
Lossless FLAC is compressed but does not use a fixed lossy bitrate formula. Its ratio depends on content, so Lossy vs Lossless Audio shows a range instead of a fake exact number.
44.1 kHz, 48 kHz and speech rates
44.1 kHz is common in music delivery histories, while 48 kHz is common in video and production workflows. Neither is simply “good” and the other “bad.” Choose a rate compatible with the actual source and destination to avoid needless resampling. Speech-only workflows may use lower rates when the intended bandwidth and codec support them.
A rate choice should reflect production and delivery requirements, not a belief that a larger number restores content. Use Sample Rate Converter only when a deliberate destination rate is required.
Common misconceptions
- Sample rate is measured in hertz, not kbps.
- Fixed-bitrate file size is primarily duration × bitrate; sample rate is not multiplied again.
- PCM size directly includes sample rate, bit depth, and channel count.
- Upsampling cannot recreate frequencies or detail absent from the source.
- Raising a lossy output bitrate cannot restore information removed by an earlier lossy encode.
- 44.1 kHz and 48 kHz are workflow choices, not a universal quality ranking.
Interactive worked examples
| Model | Inputs | What the result means |
|---|
| Compressed | 10 min · 128 kbps · 48 kHz displayed | About 9.6 MB payload; 48 kHz is not multiplied |
| PCM/WAV | 10 min · 48 kHz · 16-bit · stereo | About 115.2 MB plus header |
| Compressed mono | 60 min · 64 kbps | About 28.8 MB before overhead |
| PCM speech | 60 min · 16 kHz · 16-bit · mono | About 115.2 MB plus header |
Actual compressed outputs vary with container, metadata, and encoder mode; PCM data is deterministic for the listed parameters.