cjl
Science Advisor
- 2,001
- 612
It's worth noting that as with many audiophile beliefs, there's perhaps a grain of truth to this one. CDs only sample at 44.1kHz, and DVDs at 48, and in practice, it's at least non-trivial to make a brickwall filter that is down to effectively zero at 22kHz but has no phase or magnitude impact on 18-20kHz. There's a real argument to be made that sampling at perhaps 60kHz would make it far less likely for devices to have audible filter artifacts because then you'd have the entire range from 20-30kHz to implement your low pass filter.boneh3ad said:While we are at it, I should also mention that audiophiles in general have effectively zero grasp on signals processing, specifically the Nyquist frequency and its implications. All of this high-resolution audio BS pilfered by people like Neil Young is pure snake oil.
Similarly, 16 bit is only 96dB of dynamic range, which means that you have to be a bit more careful with your mastering because if you have a mean level of say -30dB or so, it's possible you could start to hear some of the noise from dithering or quantization (if you have one of the awful "audiophile" DACs that doesn't dither) during quiet sections. This also applies to using digital volume control (say, the windows volume control on your computer) - if you're outputing a digital signal at 16 bits, and you turn the digital volume down 20-30dB, you could start to hear dithering noise during quiet periods (since you have to have more analog gain in your amp to compensate for the low digital volume level). 24 bit solves both of these issues - you could have a 24 bit encoded signal at a mean level of -50dB and the dithering/quantization noise will still be completely inaudible.
Of course, that's not the argument they're using, and with a properly implemented filter (which is definitely possible), 44.1 is definitely sufficient for audibly perfect results.