Compact Disc Sampling Rate of 44,100 Hz*

Folks, As I said before let your ears make the decision. IMHO, higher sampling rates are the key to better digital sound, but please remember there are people in this field who would disagree with me. My listening experience is that 24/96 sounds better than redbook CD and SACD using DSD (especially if the music was recorded in analog and the converted to DSD) sounds almost as good as all analog. Remember DSD and DVD-A both use exactly the same number of bits, they just chose to use them differently. DSD chooses to do very high sampling rates, a delta sigma digitization and noise shaping to get rid of any quantization noise from the low bit sampling. If you do not understand what this means do not worry about it. The two key things that high sampling allows you to do is use anti-aliasing filters that are at very high frequency cut-offs, so they little to no effect on the audio frequencies and capture ultrasound frequencies so that the leading edges of impulse signals (drums, guitars, cymbals, etc.) sound right.

 

Yes. I also remember Steve Hoffman and others saying that trailing edges of waveforms are reproduced with the proper decay profile with the high-res formats and that you can hear that in the more realistic ambience and so forth.

 

Except with CD recording, the "wheel rolling backwards" would occur at recorded frequencies over 22.05khz (half of the 44.1khz "sampling rate"). I wish I had saved the article (if I recall, it might have been printed in Absolute Sound or some other audio publication around the mid 90s - can't say for sure). What they mention is that any recorded frequency over 22.05khz will be read wrong (therefore during mastering, the recording's frequencies needs to be cut off at 22.05khz). In other words, a 22.1 frequency may be mis-read as a 22.00, or 22.2khz as 21.9khz, or a 23.0khz being mis-read as 21.1khz (and so on - much like the wheels spinning backwards in a "24 frame per second" video of a car going at 55mph). To visualize the clock rate of the frequency and to understand in layman's terms, I remember the magazine article explaining it like this: Get a battery operated quartz movement clock with an HOUR, MINUTE, AND SECOND (sweep) hand. Remove the HOUR and MINUTE hands from the clock, reinstalling only the sweep hand. Photograph the clock every 15 seconds, then playback all the photos in sequence ......It'll appear the hand is moving forward 15 seconds. NOW, take photos every 45 seconds and do the same thing! The sweep hand actually appears to go BACKWARDS by 15 seconds (even though the photographer would have intended for it to appear to be moving forward 45 seconds). So in other words, the frequency response (from what I understand) is actually half of what the sampling rate is. In which case, the 96khz sampling rate is obviously superior. Hope I explained it correctly (I did the best I could).

 

Folks, I really do not want to try to teach a course on Integral Transforms here on the forum because it should be about music. Let's make life simple since I am drinking wine right now. Let's say our sampling frequency is 40 KHz and thus we should have an anti-aliasing filter set at 20 KHz. Now let's say that the filter craps out and our incoming signal has frequencies up to 21 KHz. The 21 KHz signal will alias as a 19 KHz signal. Imagine that 20 KHz is a mirror and that any thing above it reflects back, so 21 KHz becomes 19 KHz, 22 KHz becomes 18 KHz and so on. I hope this helps. sorry I could not do the decimals.

 

And the argument always digresses into source preparation, hardware, software, processing, etc. Conducting an actual test where the field is completely level is not really feasible with a mass marketed piece of media. Defenders of the Hi-Rez formats always use the "it sounds more "analogy" which in essence means natural or lifelike. It's really difficult to nail with quantifiable measurements.

A simple home needle drop using two different specifications actually "proves" it to me...one sounds more like the record I was just listening to than the other one does.

 

This is not correct. DSD (or SACD in format terms) is straightforward. It uses a 1-bit data stream at a sampling frequency of 2.8224 MHz (64 x 44.1 kHz) giving 2.8224 megabits per second per channel (and this holds for all channels in stereo and M-ch). DVD-A has no fixed PCM definition. Word lengths can be 16-, 20- or 24-bits, and sample rates can be 44.1 kHz, 48 kHz, 88.2 kHz, 96 kHz, 176.4 kHz or 192 kHz, the latter two rates restricted to stereo due to data rate limitations in the DVD specification. Moreover, DVD-A allows different configurations for different channels (e.g., 24/96 for left/right front; combined with 24/48 for the rest in M-ch). The common rates of 24/96 and 24/192 equate to data rates of 2.304 megabits per second per channel and 4.608 megabits per second per channel. A PCM configuration of 16-bit/176.4 kHz uses the same amount of data per channel as DSD.

 

BE, Sorry for the mistake, for some reason I had it stuck in my head that 24/96 was the same number of bits as DSD but it is slightly less.

 

Tony, I think the point you missed is, as BE says, DVD-A "has no fixed PCM definition".