Published by PitchChanger · Updated · How these guides are made
How to use this workflow
Read the source and destination rates
Use hertz, not a bitrate in kilobits per second. 48 kHz means 48,000 samples per second per channel.
Divide by two
Compute the theoretical boundary independently for each sample rate.
Choose a delivery rate
Use the destination’s requirement. Upsampling can satisfy a workflow requirement but does not create genuine captured detail.
A boundary table for the converter’s rates
44,100 / 2 = 22,050 Hz; 48,000 / 2 = 24,000 Hz; 96,000 / 2 = 48,000 Hz. These values describe sampling limits rather than the frequency response of a microphone, speaker or codec.
For an illustrative unfiltered 48 kHz sampler, a 30 kHz sinusoid aliases to 18 kHz because |30,000 − 48,000| = 18,000. This is a mathematical example, not a claim that the converter produces that alias: resamplers use filtering.
Swipe or scroll sideways to read all columns.
| Sample rate | Nyquist boundary | One sample interval |
|---|---|---|
| 44,100 Hz | 22,050 Hz | 22.676 microseconds |
| 48,000 Hz | 24,000 Hz | 20.833 microseconds |
| 96,000 Hz | 48,000 Hz | 10.417 microseconds |
Why the exact edge is not a useful audio target
A sine at exactly half the sample rate can produce all-zero samples for one phase alignment. The useful band must therefore stay below the theoretical boundary, with transition space for a real filter.
Audio Converter offers 44.1, 48 and 96 kHz for supported lossless outputs, with narrower choices for MP3/AAC. A 44.1 kHz source converted to 96 kHz still carries the source’s original bandwidth limitation. Match the receiving project rather than choosing the largest number automatically.