Articles

Oversampling

Many audio engineers feel that a higher sampling rate would improve the sound, but at a great expense to the convenient size of current discs (5 inches, and 3 inches) which can easily be used not only in the home, but in automobiles, and portable units. By Nyquist's theorem, it was felt that you only need two samples to decode the maximum frequency in a waveform, and therefore, since the maximum audible frequency is 20 kHz, the sampling rate could be around 40 kHz. It ended up being 44.1 kHz, but now this number is considered too low. DVD-A remedies this problem because it can hold much more data. "Oversampling" is a process whereby the player reads two samples, and, using DSP, additional values are subsequently "interpolated" in between those two. Thus, if the CD player uses 8X oversampling, then the player creates 7 additional values in between each of the actual single samples on the disc (7 + 1 = 8, for 8X oversampling). Oversampling is for the purpose of removing artifacts when the signal is passed through an analog filter to the output. The DAC performance probably does improve. However, there is controversy over this procedure.