44.1 kHz
44.1 kHz is a popular sample rate, for legacy reasons.
Digital audio processors
Digital audio processors generate a lot of digital data. Back in the 1980s, it was simply too much to fit on any household medium, with the exception of videotape. So that's what they recorded to.
| Standard | Framerate (contains two interlaced fields) | Visible lines | Lines used by digital audio processor | Total lines |
|---|---|---|---|---|
| PAL | 25 Hz[1] | 576 | 588 | 625[1] |
| B&W NTSC | 30 Hz[1] | 480 | 490 | 525[1] |
| Colour NTSC | ≈29.97 Hz[1] | 480 | 490 | 525[1] |
By multiplying the framerate (frames per second) by the number of lines used per frame, we get the number of lines used per second:
| Standard | Frames per second | Lines used per frame | Lines used per second |
|---|---|---|---|
| PAL | 25 Hz | 588 | 14,700 |
| B&W NTSC | 30 Hz | 490 | 14,700 |
| Colour NTSC | ≈29.97 Hz | 490 | ≈14,685.3 |
Each line can store 3 stereo pairs of 16-bit samples. By multiplying the lines used per second by the pairs of samples per line, we get the sample rate per second:
| Standard | Lines used per second | Pairs of samples | Sample rate |
|---|---|---|---|
| PAL | 14,700 | 3 | 44,100 Hz |
| B&W NTSC | 14,700 | 3 | 44,100 Hz |
| Colour NTSC | ≈14,685.3 | 3 | 44,056 Hz |
Each line stores 3 pairs of 16-bit samples, so 3 × 2 × 16 = 96 bits per line.
For example, Sony's PCM-F1 could record onto NTSC videotapes at 44.056 kHz or PAL/SECAM videotapes at 44.1 kHz, depending on whether you bought the Japanese/American version or the European version. (Theoretically, B&W NTSC could support 44.1 kHz, but there was no such thing as B&W NTSC consumer videotapes.)
CDs
The PAL sample rate was adopted by the global CD format. CDs were originally made using the output of a professional digital audio processor, such as Sony's PCM-1600.
Curiously, while CDs are split into sectors and frames (analogous to fields and lines respectively), they're different to any of the above videotape field and line lengths, at 75 fields per second, 98 lines per field, and 6 stereo pairs of 16-bit samples per line.[2]
Incidentally, a full 75 Hz sector is the shortest addressable point on a CD, just as a single field is the shortest addressable point on a videotape. This explains some tracks — such as Nine Inch Nails's "A Warm Place" — starting with a very brief snippet of the previous track. Perhaps they should have addressed the next sector instead...
While I'm not clear on how CDs were able to be so much smaller than LaserDiscs, it presumably helped that they have no colour signal, and that the luminance signal can be entirely Boolean, as that's all a PCM decoder needs.
CDs became wildly popular, as the first pre-recorded digital format available to consumers. As a result, the sample rate of 44.1 kHz became even more popular, spreading to home computers and their beige successors, and subsequent formats such as DAT and MiniDisc.
References
- "ITU-R BT.470-6" International Telecommunication Union, p. 2
- "CD-Recordable FAQ: What's a Frame? CIRC Encoding? How Does ECC Work?" Andy McFadden
Sample rates: 8.363 kHz | 31.25 kHz | 32 kHz | 44.056 kHz | 44.1 kHz | 48 kHz
