Analog-to-Digital and Digital-to-Analog Conversion
Sampling and reconstruction are the two gates that decide what digital audio keeps.
Related
The question
A converter turns voltage into numbers and back again. Change the converter and the air, the soft notes and the tails change with it.
How it works
Sampling needs a low-pass filter first, or content above half the rate folds back into the audible band. Each sample is then quantized, and the remainder becomes noise. Playback rebuilds a staircase and smooths it with a reconstruction filter. Real converters add filter transition bands, quantization noise and clock jitter. Oversampling spreads that noise wider before filtering it back, easing the analog stage.
Settings and practice
44.1 and 48 kHz cover most work; 96 kHz suits heavy pitch shifting. Track with 6 to 12 dB of headroom and one word clock. Linear-phase filters keep transients but pre-ring; minimum-phase filters delay less. Avoid repeated rate conversions.
What to listen for
On a sustained organ chord, hear whether overtones thin. In the soft passages of Clair de lune, hear whether the lightest touch survives. In quiet Satie, listen to the noise floor. Dull highs suggest a steep anti-alias filter; smeared soft notes suggest too few bits.
Common misconceptions
A higher rate is not automatically better, since 44.1 kHz covers the audible range. Bit depth is not just loudness; it decides whether quiet detail clears the noise. A DAC is not passive wiring. And 96 kHz captures content only up to 48 kHz.
Listen in the library
All tracks from midicn-lib — inline badge is the license tier (C1 commercial / C2 non-commercial / C3 study only).
Sources
- Pohlmann, Principles of Digital Audio, sampling and conversion chapters — background checked, not quoted
- 转换器滤波模式与时钟的差异,依厂商公开技术文档核查,未抄文本
Used for fact-checking only; all prose is original.