CD/DVD-A/SACD Dynamic Comparisons

[gene]

This article, courtesy of Christine Tham illustrates some very interesting differences between CD, DVD-A and SACD in terms of spectral content, and dynamics.

CD, DVD-A, SACD Dynamic Comparisions

 

[O'Shag]

DSD Vs PCM

I wanted to put in my 'two cents worth' regarding DSD vs PCM, which is brought to mind by the article written by Christine Tham.


Gene et al are rightly focused on measured results as a means to determine performance for a given product or technology. In the quest to better understand the technical characteristics of PCM vs DSD, or DVD-audio vs SACD, articles such as Christine's are really informative. Could it be that the graphs which show spiking or a higher noise floor for DSD are signals representing ambient or spatial noise information that is missing in the PCM formats?

My experience so far is this - I have listened to state of the art PCM/CD reproduction systems, either with DAC and transport combined in one case as in most CD players, or systems where DAC and transport are seperated. I have owned several CD players myself, including the very good Sony SCD-XA777ES such as Christine's. I have two speaker systems; the $7,500.00 Polk audio RT5000 Cinema system, and flagship KEF Reference 107s for serious audio listening.

I first experienced SACD for the first time about two and a half years ago.
For me, it was nothing short of a revalation. Articles have discussed how some listeners are not so impressed, perhaps expecting 'huge' differences rather than subtle ones. This puzzles me. I'm darned, but if you ask me, you'd have to be slightly hearing impaired not to hear the very noticable difference between the DSD and PCM technologies. DSD technology, with its far higher bit rate transfer allows a great deal more information to be recorded on disc and replayed by a compatible player. Overall the result is much more dynamic, a much fuller more rounded presentation of instruments, voice, and soundstage. Subtle but important ambient information that was previously unable to be recorded due to bandwidth limitations in in PCM technology is captured using DSD, with the result of producing a more faithful rendition of the original recording. I still love my 'redbook' CD collection and my CD player, And with some material I just like the way my CD player sounds. But SACD produces a more accurate representation of the music - there's no doubt about it.

I can say with confidence that I am not just hearing things, because my fiancee had a similar reaction to myself, and she did not know she was listening to SACD. I sat her down on the couch, and put on some music. She is not an audiophile per say but truly loves music. Her first reaction was, "my God this sounds so real, as if the performers are right there in front of you..."

It seems to me that there is a natural reluctance for many audiophiles to let go of older, more familiar technologies, or to admit that their 20K CD player can be bettered by a 2K SACD player. I have some such friends who've not even tried SACD yet - is this living in denial or what!

I would guess that any flaws people may hear are probably attributable to the current capability of lower grade consumer players rather than the core technology itself. You will probably remember the same comments during the first few years of PCM/CD technology when many players seemed to be excessively bright in the mid-range, and somewhat lacking in spatial soundstage as compare to the best turntables.

Another big roadblock to DSD becoming established as the new benchmark standard has to be the lack of available software to consumers. They've got to get more recordings out there before people are comfortable with a complete transition.

If you want to get a taste of how good SACD really is, try listening to BB King's album - Reflections......

Best,

O'Shag

 

[docferdie]

A major flaw of this "scientific" article is that the soundcard used does A/D conversion using PCM so is it surprising that the PCM formats (DVD/A and CD) are doing so well?
These comparisons are flawed from the outset as there is no device used that does Analog to DSD conversion. Another way of looking at it is if your measuring device only does PCM then of course the PCM formats are going to look good.
The only way to scientificially compare the three formats is if there were a universally agreed upon gold standard of analog to digital conversion and then do analog to digital conversion using PCM, DSD and comparing these to the bitstream of the gold standard process.

 

[Rob Babcock]

I agree with respects to SACD vs 16 Bit PCM. But I personally feel 24 Bit (via DVD-A) sounds superior to SACD to my ears. Granted, I've only got maybe fifty or sixty hi rez discs, counting both formats. But given the same albums in both formats on several different players, I'll give the edge to DVD-A.

Both are great, though.

 

[Rip Van Woofer]

Somehow, I'm not sure that Docferdie's argument holds water for the parameters being tested even though it has a certain logic but I'm not techy enough to say quite why...I doubt that frequency response, noise, and such have anything to do with the soundcard being PCM...but it seems that Ms. Tham's results square reasonably well with what I've read about the three formats elsewhere (and not in Stereophool). Any digital experts to set us straight out there?

One thing I wonder about is whether the differences in compression were entirely due to the inherent properties of each format, or whether the CD version was compressed more in post-production for "mass" tastes and car players, and the DVD/SACD releases were left relatively uncompressed for the presumed "discriminating" listener and home systems. Yes, I know this begs the question re: the difference between the DVD and SACD. Properly recorded 16 bit PCM can have full dynamic range if the recording engineer is very careful with setting levels (trouble is, they often aren't and there's little room for error). 20 bit PCM has much more room for error and can, according to some experts, equal DVD-A and SACD even when released as a standard 16 bit CD. There is some reason to be skeptical of the need for an entirely new/different format such as SACD.

And there are good reasons to suspect that the impetus behind DSD/SACD is not technical or musical but mostly business-driven. This has been discussed before here at Audioholics; do a search if you're curious.

 

[docferdie]

Let's take a page from digital video. The old standard was 1 CCD. Now 3 CCD has come out. But if you take the output of a 3 CCD camera and then digitize it using 1 CCD then how will you be able to evaluate the differences between the 2 technologies. DSD was developed precisely because there are people who did not believe the PCM system to be the best method of analog to digital conversion. Let's take it to the extreme. Let's say the output was recorded onto a cassette tape and then analyzed then all three devices would have a frequency cutoff of around 10 Khz which would render the comparisons meaningless. This article relies on the assumption that 96/24 PCM is enough to resolve the sonic differences between the three formats. This is obviously not true as there are DVD/A's which are encoded at 192 Khz. Just that alone should be enough to show that the "measuring device" is not up to the task.

 

[Dan Banquer]

Cd/dvd/sacd

To all; I recommend that you read this article posted here at audioholics.
http://www.audioholics.com/techtips/specsformats/CurrentFormatTrends.php
To Doc Ferdie: there is a lot of debate on higher sampling rates on one of the pro audio forums I am on. Recently Dan Lavry of Lavry engineering posted up a very interesting white paper on why 192 kHz is total and utter overkill. It's worth a read.
d.b.

 

[docferdie]

Dan,
I agree that 192 Khz may be overkill but I just want people to be aware that just because a paper touts measurements, it isn't necessarily valid. One always has to be wary of the precision of the measuring devices.
I'm sure Joseph Lister would have thought the electron microscope to be overkill but can you imagine the many advances that would have been lost without all that resolving power.
I still have yet to see a convincing technological article comparing DSD and PCM head to head. The soundcard in the article simply does not have the resolving power to make any meaningful comparison. Another analogy would be, it would be pretty difficult to compare a 1080i broadcast and a 720p broadcast if you're watching on a 480i screen.
To the engineers on the forum, is there a measuring device can definitively answer which format does a more accurate analog to digital representation?

 

[Dan Banquer]

Dvd/sacd/cd

The Audio Precision One is very good piece of test gear.
d.b.

 

[av_phile]

This article, courtesy of Christine Tham illustrates some very interesting differences between CD, DVD-A and SACD in terms of spectral content, and dynamics.

CD, DVD-A, SACD Dynamic Comparisions

I am not so sure if the conclusion on the above test matters to me. One thing for sure is that the spectral analysis reveal that there is content energy above the 20Khz for which CDs have a brickwall filter. Now I am sure at my age, I can no longer hear above 17K, confirmed when my 11-yr old son can hear 18K to 20k on the Sheffield test disc, and I can't. But I often wonder if the 40k and 80k bandwidth of DVD-A and SACD respectively matters at all in terms of perceiving depth, detail, accuracy, airiness, etc.

 

[Rip Van Woofer]

Okay, getting down to brass tacks, the main advantages to various hi-rez formats including 20+ bit PCM seem to come down to the following in varying degrees depending on the format:

Increased dynamic range (always a good thing). That alone, IMO, accounts for a lot of the "air", "detail", "resolution", "microdynamics", etc. in 'phile-speak.

More tolerance of less-than-perfect 0db settings by the recording engineer, resulting (we hope) in fewer disks exhibiting clipping. Cynical rejoinders to this point are entirely understandable!

Increased bandwidth which, if I understand things correctly, can be "traded" for increased dynamic range (kinda hazy on that - further enlightenment apreciated). BTW, the noise shaping in DSD/SACD that puts all that noise just above 22kHz seems to render the increased bandwidth somewhat moot there, hmmm?

Increased data density on the disk so you have room for things like multichannel, "enhanced" features like multimedia presentations you can play back on a video or PC system, etc. For example, the Harmonia Mundi DVD-A release of the Bach Mass in B Minor (and others) contains multimedia extras on the life of Bach, analysis of the score, etc. Hint: that would make a princely birthday present for the Bach lover. (My birthday is Sept. 19 )

 

[gene]

And here is part II

Part II of CD vs DVD-A vs SACD

Two more parts to go, stay tuned......

 

[docferdie]

av_phile
It's true that most adults don't perceive frequencies above 18000 Hz. Just because you don't perceive it doesn't mean your brain doesn't hear it. A test called the brainstem auditory evoked response, measures electrical signals in the brain when the ear is exposed to various sonic signals. The clinical significance of eliciting a response to something like 24000 Hz is unclear but it is present. Like speakers the human ear does not have a brick wall filter that prevents frequencies above 20000 Hz from getting through.

 

[Audioholics]

Part 3 is now up which goes even further into the compression and dynamics of the various formats. Each article refines the process of discovery further and Part 3 eliminates the argument that the player might have some effect on the recorded dynamics. In this article Christine uses a Denon DVD-2200 Universal player to source all of the tracks at constant gain. Very cool reading.

Check out Part 3...

 

[gene]

Wow a real active thread It's good to see people take an interest in what really matters in audio

Here is Part IV which compares LP to CD, DVD-A and SACD

 

[Rip Van Woofer]

I had already skipped ahead on Ms. Tham's site to the vinyl vs. digital article and on first reading all I can give is a somewhat skeptical "hmmm..." Might have more to say on a more thorough re-reading. But I still say when all factors are considered, digital rules.

The whole schmear did make the Denon 2200 look pretty good to me, though!

 

[jeffsg4mac]

Hmmm, kind of lost me on the vinyl thing too. Been there done that. Vinyl lost and it is a dead format, gone the way of the 8-track. Have I heard some excellent vinyl recordings? you bet. Do I want a turn table? Nope, I sold all my records 20 years ago when I bought my first cd player. I do not ever intend to buy another one.

 

[av_phile]

If I recall right LPs have a theoretical dynamic range limit of around 80db while CDs have around 96db. Those Hi res formats, even more. Perhaps the recorded specimens considered in the study have been mixed to compensate for the media's inherent limitations. Boosting the dynamics in LPs while retaining the dynamics of the master recordings in CDs. I often wonder how similar tests will look like with the origianl multi-track open reel masters.

Curiously, I recall the Telarc 1812 Orerture engineers had to to tone down on the dynamics becasue the canon shots were so powerful to allow the lathe cutter to work properly. No such limitations with the digital media.

Regarding the noise floor data, I wonder if the CD player used has the same hi-end attributes as the hi-end turntable used. What is so good about the digital medium is that the typical Joe can experience excellent sonics usually reserved for expensive set-ups using esoteric turntables and esoteric gears. With motor rumble noise, wow and flutter, tangential tracking errors, tracking force errors, etc, etc, I wonder how a commercial turntable system would stack up against even a discman in terms of clean and accurate sonics. Whatever, after getting rid of my Technics turntable in 1987, I really doubt if I'd go back even if turntables were on sale for a dollar. Unless they can discover a material for LPs that are impervious to grime and dust-causing tics and pops, static crackles, warpage, surface noise and other vagaries.

 

[WmAx]

I question the value of these comparisons of different masters/mixes of titles against 'each other' on different formats.

Additinoally, the statement from the author quoting "given the surface noise on LP (which is quite audible on my system). " and then following later with different S:N ratios for specific bands, making the performance seem 'very good' is perplexing. If an actual 80dB(as asserted theoretical S:N by av_phile), for example, was possible under realistic conditions, the noise woudl literally be non audible on speakers and headphones[1] under ideal circumstances.

[1] Signal-To-Noise Ratio Requirement for Digital Transmission Systems, Spikofski, Gerhard, AES preprint 2196, Convention 77(1985)

It seems to me that the limited good performance of LP on noise in certain bands is not signficant considering......

The inevitable/eventual random surface noise pops/clicks etc. are unavoidable, unless of course you are using your LP player in a 'controlled clean room', the low frequency response dynamic range is limited mechanically as Telarc displayed( the needle could supposedly unseat itslef on many record players on the 1812 overture). However, properly dithered 16 bit audio files can manage a reliably inaudible noisefloor(assuming the recording's noisefloor is not the limiting factor) throughout the entire passband. Of course, we are only discussing S:N. THe LP process introduces single digit THD values, that are known to be within human thresholds of audibility[2]

[2] Just Detectable Distortion Levels, James Moir, Wireless World, February 1981

LP, being a self-destructive format under proper use, is another serious issue. Of course, they do have that LP laser player avaialbe for a few thousand to remedy this problem....

How do we fairly compare these different mixes/masters on different formats? I think the anser is pretty simple: we can't. Stereophile recently released a SACD, CD and LP set that is claimed to be created from the same masters. I recommend performing the comparison using these unique examples. Though, conclusions of audibility are still independant on what may be measured in this test, since any shortcomings on any medium in this set may be only a result of the recording/mastering.

-Chris

 

[docferdie]

Again I question the validity of using the audiotrak 7.1 as a measuring device to carry out the desired comparisons in this article. CD is PCM and SACD is DSD and yet they are both being evaluated by a card that can only work on PCM.