(Finally) Getting the Highs Right on CDs

David Matz -- Wed, 10/20/2010 - 21:35

I have had the pleasure to hear some very good cd players recently.  One common denominator is that the highs sound so much better than on the older players.  They are smooth and liquid, not brittle or hard.  Highs on a good CD recording sound like it is an SACD.  Does anyone know what is going on?  What are these companies doing to finally get it so much better?

johnny p. -- Fri, 10/22/2010 - 20:53

All a CD player can do is reveal what's on the recording. So, if you're (finally) hearing smooth and extended highs, it's because CD players are at last doing their job. 
 
The CD standard is high-resolution - naysayers be damned. It covers the full range of human hearing and covers more dynamic range than a symphony orchestra produces.......

Robert Harley -- Fri, 10/22/2010 - 21:07

The treble from CDs does sound noticably better in the past couple of years. I attribute it to better digital filters.
 
The CD standard was set in the late 1970s, and is definitely not "high resolution." 44.1kHz sampling and 16-bit quantization were compromises dictated by the technology of 1978 when the standard was set. It's taken 30 years for CD sound to mature, as curiousmind has noted.

David Matz -- Sat, 10/23/2010 - 10:07

Thanks for your reply. Do you know if the better treble is primarily due to removing the pre-ringing, or are the filters removing the pre-ringing + doing other stuff?

Also, do you think upsampling the sound to DSD (or twice the DSD) has an effect on this also? Or are these completely different stages of sound processing?

Additionally, (I'm not sure this is an area of your expertise) does a better treble result in a better overall sound of the cd psycho-acoustically, kind of like absorbing the bass waves in your small room results in a better overall sound along the entire frequency spectrum?

Thanks

johnny p. -- Fri, 10/22/2010 - 23:25

CD, as I hear it, is Hi-Rez. This, because neither SACD (or downloaded high-bit files) sound better than the best CD systems. One combo that works great together is the Soolos music server feeding the new Meridian 808.3.
 
How is a coding system that goes beyond human hearing *and* and symphony orchestra a "compromise" ?

Josh Hill -- Thu, 10/28/2010 - 19:08

44.1/16 can't accommodate the dynamic range of live music. See the excellent "Dynamic Range Issues in the Modern Digital Audio Environment" by Louis Fielder of Dolby Laboratories:

http://www.zainea.com/Dynamic%20range.htm

Arguably, 44.1 kHz is sufficient to accommodate the frequency response requirements of the human ear. However, to do so requries transparent "brick wall" anti-aliasing and reconstruction filters. It's difficult or impossible to implement such filters, although filters have gotten better over the years and the sonic gap between 44.1 and hi res recordings has been closing as a result.

johnny p. -- Thu, 10/28/2010 - 22:54

I'm assuming you've read the link you provided. So, can you point out why (exactly) these people don't think 16 bits is enough (on disc).

There's no doubt that 16/44 *recording* wasn't enough - at least for wide dynamic-range music. This, because of the intermediate calculations that are done on the signal, in post production. There was (some) loss of resolution. That's why audiophile-type labels started recording at 20 bits/88kHz in the early 1990's - to absorb these losses.

How could 115db of dynamic range (CD with noise shaping) and sampling that covers the full-range of human hearing not be enough ?

Josh Hill -- Fri, 10/29/2010 - 09:21

The author of the Dolby Labs paper determine the noise detection threshold and the maximum peak levels of music performances, both acoustical and electronically augmented, and conclude "Therefore music reproduction, limited only by sound levels and human auditory capabilities, requires a dynamic range of 122 dB for monophonic, 122 dB for stereophonic, and 124 dB for five-channel reproduction circumstances." So even noise shaped 16-bit recordings will have audible noise. Arguably, you need even more than this to allow for level-setting discrepancies. Certainly in the recording format, though bit depth can be reduced for release.

I suspect that the reason we seldom hear the noise is that most commercial recordings are compressed and that music is typically played back at lower than natural levels. In fact, when the author measured loudspeakers, he found that many consumer units couldn't reproduce recordings of un-amplified acoustical music at natural levels. See Fig. 2 in the Dolby paper, peak levels of music, and also Fig. 8, maximum peak output level of 23 consumer loudspeakers. It's easy to spot the Quad, LOL.

I don't know whether today's best converters can record and reproduce 44.1 kHz recordings with open high end of higher sampling rates. Critics seem to differ on the issue and I haven't tried any comparisons myself. One study found the sampling rates indistinguishable; unfortunately, their methodology was shoddy, e.g., they used recordings that were upsampled from 44.1(!).

johnny p. -- Fri, 10/29/2010 - 11:53

First of all, I noticed the references for this report end in 1992 - indicating it was written that year (or right after). If this is the case, then it makes sense - Dolby wasn't the only one saying "we need more" back then. Other researchers, inc. Keith Johnson (of Spectral Labs) said the same. Johnson allegedly found that higher sampling rates are needed "to better capture transients". Possibly - but there has been little proof of this in the real world.....

The next thing is that electonics have *nowhere near* the capability to reproduce 122db's of dynamic range !! I wish they did - but it will never happen.

Finally, the recordings. You mention that many are "compressed". As an audiophile, I play only natural recordings that aren't. Recordings that use the full potential of the Red Book CD standard. And listen as I do, I hear *no noise* on my recordings. Actually, I doubt any audiophile (who listens to acoustic music) would say they hear "noise" on their digitally-recorded music.

This is why I hear no improvement of HRx downloads, over CD. At least not with the system I mention above. I specifically asked Robert Harley what he thinks of the Sooloos-Meridian combo vs. HRx - and he declined to answer. Anyone wondering why ?

Josh Hill -- Fri, 10/29/2010 - 13:43

To quote again from the Dolby paper, "Power amplifier equipment noise is limited in dynamic range due to the fact that amplifier designers have considered the l10—120-dB dynamic range to be completely adequate for all applications. Fortunately the design of a power amplifier with wider dynamic ranges is possible, given the need to do so."

This sort of chicken-and-egg scenario has discouraged advances in the reproduction of dynamic range. We have the technology now to record and reproduce music in the home at natural levels, uncompressed, and with little if any audible noise, but we haven't availed ourselves of it, because the record companies put out compressed recordings (don't want to overload the speakers), the speakers can't play at natural levels (recordings are compressed, why go to the expense?), etc.

The thing is, as Robert Harley pointed out, I have a feeling you'd think differently if you could hear the original mic feed, or the studio master. Compression in particular does terrible violence to the music. What wide dynamic range recordings aren't compressed for release? Almost all of them are.

By the way, in one of the 44.1/hi def comparisons in that AES Journal article, they *were* able to hear the difference in the noise floor between 16 and 24 bit, but only on one recording, which they played back at a level of 115 dB SPL. Whether the 16 bit version was noise shaped, I don't know. But I think it's a given that only the very loudest, wide dynamic range music is going to challenge today's 16 bit technology. You won't hear it on most pop (quiet passages not quiet enough), and you won't hear it on most classical works (maximum levels not high enough). The idea here is to accurately reproduce the most dynamically varied material at natural levels, and I think it's a sound one. Bits are cheap, why should they stand in the way of transparent reproduction?

johnny p. -- Fri, 10/29/2010 - 14:40

Bits are cheap - that's why they use 20-24 of them when they record. But subsequently, many recordings are *not* compressed - as far as reducing the dynamic range of the program material. Chesky, Reference, Hyperion, Harmonia Mundi are a few labels that come to mind.

If "Hi-Rez" is better than CD on playback, it hasn't manifested yet. It has not surpassed the best Red Book-only players, to my ears. The same thing happened to SACD 10 years ago - not better than the best CD machines (like the Linn Sondek CD12).

Since RH said HRx files "sounded like a live mic feed", he's gonna have to explain (at some point) how Red Book could do the same !!

Josh Hill -- Fri, 10/29/2010 - 18:04

I don't know. As usual, there are conflicting reports of whether you can or can't hear the difference, with no way to determine which are right. There was certainly a time when you could hear the difference -- I heard A/B comparisons back when SACD was being developed -- but today? I do think it's probably easier and cheaper to design a good sounding converter for hi res material. But I haven't done anything but cursory A/B'ing in recent years.

OTOH, there is proof that you can hear the added bit depth, both theoretical as in the Dolby paper, and from double-blind testing. Whether this matters to most people is another question. I do know a few people online who like to listen at natural levels, and have gone out of their way to select some of the relatively few high end speakers that can do it. For them, it could potentially matter.

Robert Harley -- Sat, 10/23/2010 - 22:04

Compare 44.1kHz/16-bit digital audio to a mike feed, to an LP, or to Reference Recordings HRx and you  might reach a different conclusion.
 
The Sooloos feeding a Meridian 808.3 via Ethernet is indeed wonderful. That's what I'm listening to now, as a matter of fact.

David Matz -- Sun, 10/24/2010 - 17:59

You were very enamored with the 808.2 last year, to say the least. Is the 808.3 much better or just different?

johnny p. -- Sun, 10/24/2010 - 20:55

I don't think we'll hear back from RH. At least I won't - my question was (probably) too much for him to handle. He would have to admit that the HRx files sound *no better* than CD - via the Sooloos/808.3 system.........

David Matz -- Sun, 10/24/2010 - 21:14

I just left that post a few hours. And it is Sunday. Let's give the man a chance.

johnny p. -- Sun, 10/24/2010 - 10:00

Glad to see you're listening to a *great* Red Book combination.
 
If you don't mind, I'd like to ask how close (you think) the Sooloos-808.3 is to LP or HRx files.........

JA FANT -- Sun, 10/24/2010 - 22:19

I, too, love a liquid and full-body sound!

Ded Frag -- Tue, 12/14/2010 - 20:47

 I don't understand the technology behind my CD players' vastly improved (standard Red Book) treble but the Meridian 808.2 is the first player I've owned that almost equals the best analogue in higher frequencies. Apparently Ayre and Moon have achieved an equal, if not better, outcome. Can anyone explain what's so different about these new players that no longer allows CD to sandpaper our eardrums?
There's another issue here that the audio magazines always avoid. How accurate were all the reviews we've read over the years that passed judgement on amplification and speakers when they were being fed by what we now see as an inferior and lower resolution source?

All content, design, and layout are Copyright © 1999 - 2011 NextScreen. All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.