alandamp said:
The magazine seems to be quoting Hollman(and in my first version of this reply, I actually believed that to be the case). But this is hard to believe. Some serious corrections are in order. I think that probably, the article author is ignorant but is pretending he knows something, and the random quotes by Hollman make it appear that the source of all information in the article is Hollman. The number of errors in the just the first part of the article were very high.
Traditional CD sampling is 44.1kHz, with a 16-bit word length, but this leads to a wave form that is severely staircased. Heard like this, CD audio would be terrible. But by using a technology called dithering, this staircase is smoothed out and the audio is cleaner and more enjoyable.
Anti-alias filters remove the 'staircase' appearance of the waveform. This 'staircase' is nothing more than the sample frequency itself. It is removed easily with modern anti-alias filters.
The digital filters which apply this dither smear the sound.
Dithering does not address[
or have anything to do with] the 'staircasing'. Dither is used when encoding the analog to digital data[
or when converting between bit depths] in order to increase the signal to noise ratio[
thus eliminating quantization noise that would otherwise occur in low level signals in PCM digital audio].
But when you increase the sampling rate you can apply a far less severe filter, with less smear. Greater audio clarity is achieved.
It is alarming to see these erroneous conclusions made by such a noted audio expert[if this information is from Hollman], or even by a writer of an audio magazine[whom should do some fact checks before publishing]. I would challenge him to show a credible perceptual study showing greater 'clarity' by using a less 'severe' filter.
It's a good thing that most audio components that claim to deliver 24-bit dynamics can't manage it. 'A very good 20-bit system will always sound better than an inferior 24-bit one,' concludes Holman.
Actually, no analog audio electronic components, anywhere, will achieve 24 bit audio performance. So none would able to manage it. Even if they could, it would be impossible for any speaker[except for the insane set up; imagine a horn loaded line array] to actually use 24 bits. 20 bits is about the maximum achievable in the best analog audio electronics, under the best conditions, and this far exceeds the capability of nearly every audiophile speaker set-up in existance.
It's worth remembering that 24kHz is the upper limit of human hearing.
Actually, it's usually 20kHz. 24kHz may be the upper limit of some very exceptional adults.
Most of us can't hear more than 20kHz
If he is referring to adults, most of us can't hear more than 16-18Khz.
'And for 24kHz you only need to sample at 60kHz.' So would a 60kHz/20-bit sound system have been the better option for DVD-Audio?
Actually, for 24kHz you only need to sample at 48kHz. Nyquist theorem, anyone?
I am far from an expert on digital audio, but even I could spot these false statements from 2 miles away....
Is this magazine writer just spouting his ignorance, or was this info actually given by Hollman?
-Chris
