What's new

Will digital CD audio differ from player to player? (1 Viewer)

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
Various blind tests have in fact been done in the audio industry. Unsurprisingly (to someone who knows the technical details), it turns out that people who think differences from transports, cables, etc. are "blatantly obvious" in non-blind listening, are often completely unable to pick out those same differences in blind testing.
:rolleyes

Ian,

This is just your opinion, and may be based partly on DBTs that others have conducted, but there are very real differences in transports and many other things including digital cable which also exhibits jitter.

I am no engineer but I have engineered or assistant produced over 14 albums, mostly for audiophile and mainstream labels. We have noted and adjusted for transport differences in our playback chain in most of these albums, several of which won industry awards for their clean sound.

(This also addresses Lance's comments: )

Jitter can be detected as smearing the music down to 20 Picoseconds. Recently, I have been doing orchestral and chamber music ensembles where we record in hi-rez PCM on Alessis Masterlink hard drives at 88.2khz. To get better sound at a recent concert we added a Lucid Audio GenX 96 master clock that drove down jitter from 200ps to 20ps. The difference was more detail, more "air" around the instruments, and an overall improvement in focus.

It was suprising to me initially but I have heard similar sonic improvements as we lowered jitter in the recording and playback chains at Chesky Records. The reading of 20 picoseconds is extremely low in technical terms but one has to keep in mind that the human brain is an amazing processing engine capable of hearing very subtle nuances through its neural network retrieval and analysis.

I am recording my music samples at 200 and 20 ps for a test CD so if you want to listen what each sounds like, I would be happy to send you a copy for evaluation.

:)
 

ChrisAG

Supporting Actor
Joined
Aug 26, 2001
Messages
503
If you did a blind test you'd quickly realize there was no discernable difference. The mind can be extremely deceptive about sound quality when it knows what it is listening to and when.
Ian,

I used to have the same belief as you, thinking that as transports, CD players cannot have different sound characteristics, since they are only passing digital data.

That was before I started shopping around for a new player.

I had three different players hooked up to my system, using the same cable (coax), speakers, and receiver. After several months of getting used to my room mates Sony player, it was removed from my system under a divorce settlement. The next player I tried was my cheap Toshiba DVD player. It sounded very lifeless in comparison. Thinking that it was too much to expect of a cheapo DVD player to do double duty, I decided to get a dedicated CD changer, the Marantz CC3000. This player looked great, matched my Marantz receiver, and had a super-fast and quiet changer. In short, it was the perfect player for me.

Until I listened to it.

It didn't sound any better than the Toshiba - lifeless and dull compared to the older Sony. This was all the more disappointing to me due to the fact that I WANTED to like this player, so if anything, the psychological bias should have made me enjoy the sound. The fact of the matter was, it simply did not sound as good as the Sony (and I was not a Sony loyalist, having had several of their products fall apart on me in the past).

Long story short, I ended up getting another Sony, and I love the sound.

Perhaps many "double-blind" listening tests involve subjects that can't tell the difference between a Real Audio file and a well-recorded CD. Or, maybe it takes a great familiarity with the sonic signature of specific equipment (your own speakers and amp, etc.) before the average person can tell a difference when switching out a piece of the chain.

Whatever the reason, I now have less faith in these so-called unbiased sound tests. Not all CD players sound the same, and that's a fact.
 

Ian Montgomerie

Stunt Coordinator
Joined
Feb 2, 2002
Messages
112
This is just your opinion
No, it's a cold hard fact. The sort of thing with which you've displayed little familiarity in the past. As evidenced by your repeated inability to understand the extreme and fundamental difference between clock jitter in the ADC/DAC, and clock jitter in the digital transmission chain, despite it having been explained to you over and over. Basically you are almost entirely ignorant about how digital playback equipment works, and are unwilling or unable to learn about how it actually works, but remain willing to hold forth on the subject at great length and with great frequency. I am beginning to think responding to you is a complete waste of time.
 

Jagan Seshadri

Supporting Actor
Joined
Nov 5, 2001
Messages
528
Ah yes, double-blind tests (DBTs). Isn't it the Audio Asylum that has non-DBT zones? That's like having a "No Scientific Methods" zone. Now, that deserves the roll-eyes.

Whatever floats your boats.

-JNS
 

Ian Montgomerie

Stunt Coordinator
Joined
Feb 2, 2002
Messages
112
I used to have the same belief as you, thinking that as transports, CD players cannot have different sound characteristics, since they are only passing digital data.

That was before I started shopping around for a new player.
As I said, your mind is deceiving you. It is very likely that on a blind test these differences would disappear. And to be frank, you obviously never believed the same thing I do, because you don't have the experience of listening to something and being confident you could hear a difference, then actually subjecting yourself to blind listening and having it completely disappear. Thinking that people under blind tests must simply have inferior listening ability is a classic, arrogant self-delusionary rationalization.
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
Ian,

I have worked with Bob Katz, arguably (with Ed Meitner) one of the world's leading authorities on the subject, on several albums and several projects related to jitter mitigation. On the contrary it is you that simply cannot fathom how certain audiophile phenomena are in fact real.

Perhaps you should read Jon Risch's summary at:

http://www.geocities.com/jonrisch/jitter.htm

Basically you are almost entirely ignorant about how digital playback equipment works, and are unwilling or unable to learn about how it actually works, but remain willing to hold forth on the subject at great length and with great frequency. I am beginning to think responding to you is a complete waste of time.
This kind of attack and deep slandering is not productive here Ian. If you have tried to communicate an idea that you have failed to get across perhaps it is your communication skills that are lacking. Also, you have never addressed the very real phenomena of replication induced jitter in any of your posts.

Your understanding and work in hardware is a valuable asset but it does not extend on down the line to deep knowledge of the recording chain.
 

Jun-Dai Bates

Stunt Coordinator
Joined
Aug 16, 1999
Messages
148
If jitter is a problem, then why don't they put the data in memory and then read from the memory? It wouldn't be expensive, once the processor and board were designed for the task. It would require such a small amount of energy and processing that it shouldn't affect any of the other functions in the DAC. The same computer could be used for pretty much any DAC, and it would eliminate any time delays that developed between the source and the processing unit.

Of course any digital signal that contains oversampling and error correction will have to be processed before being converted to analogue, and should therefore lose any timing inconsistencies between that processing and the source. (correct me if I'm wrong and you have a convincing argument).

My suspicions lie on Ian's side of the fence, but I'm open to being persuaded that I'm wrong. In general I am skeptical about what people claim they perceive, especially when the results are in opposition to the results of experimentation.

First of all, the cited article by Jon Risch does bring up the possibility of buffering the data. He concedes that this seems like a likely solution, but argues that it doesn't prevent the problems of fluctuation in the power source to the DAC itself. If what he claims is true, buffering the data would still seem to solve every problem but the power source, and that problem isn't going to be solved by cables or transports, but only by the internal design of the DAC and its power supply.

But Jon Risch's article is so poorly written that it isn't worth much consideration. Anyone that takes himself as seriously as that should consider proofreading his articles before self-publishing them. Even on the internet, writing in all caps is a no-no for technical articles (it should be a no-no pretty much everywhere, but alas). Typos and grammatical errors abound, and his line of reasoning in that article is very poor.

Anyways, I think that buffering should be employed in the DACs and digital signals should all be passed through Cat-5e cables, because I've got lots.
 

Chriss M

Second Unit
Joined
Dec 14, 2001
Messages
320
Last night, my roommate told me that she would only drink whole milk, and that any other type of milk was gross. I bet her that she could not tell the difference between whole milk and 2% milk in a blind test. We tried it, and i was right. She couldn't tell with any degree of accuracy. Does this mean that 2% milk and whole milk taste the same?

I don't know if different transports will sound different but i also don't think that failing a DBT is any kind of conclusive proof that they don't.
 

Todd Schnell

Second Unit
Joined
May 21, 2001
Messages
255


Absolutely!
The more revealing the system, the more obvious these differences will be.

I recently added a Sony C555ES to my system.
In comparing red-book audio on it to my old Sony 200 disc changer the differences are very obvious.
The 555 is far more detailed, crisp, & clean sounding.
Whether using as just a transport or internal DACs, the difference is quite obvious.
I'll quote my wife here "Wow that new cd player really does sound better!". :D
Indeed it does!

Todd
 

ChrisAG

Supporting Actor
Joined
Aug 26, 2001
Messages
503
As I said, your mind is deceiving you...
And to be frank, you obviously never believed the same thing I do, because you don't have the experience of listening to something and being confident you could hear a difference, then actually subjecting yourself to blind listening and having it completely disappear. Thinking that people under blind tests must simply have inferior listening ability is a classic, arrogant self-delusionary rationalization.
You're rather sure of yourself, hmmm?

When referring to double blind tests, I was proposing a possible explanation for the results of some test, rather than claiming my listening ability was better than others (though it may be when it comes to sensing changes in my own equipment chain). Would you not agree that some testing is fallible? Sometimes the testing method (or subjects used) ensures a certain result.

I approached this argument in good spirits, but since you are feeling somewhat vitriolic towards anyone who disagrees with you, I'll peruse the lounge for a while until you get it out of your system.
 

Ron Economos

Auditioning
Joined
Jan 25, 2003
Messages
11
If jitter is a problem, then why don't they put the
data in memory and then read from the memory?
Doesn't help for uni-directional interfaces like S/PDIF.
We still have to recover the sampling clock somehow.
The main source of jitter is the PLL based clock
recovery circuit that is driven by the data stream
itself.

However, if you have an intelligent bi-directional interface
that can also handle control messages (IEEE1394), then
a receiver can rate control the delivery of the data
stream from the source. Now we can use a memory buffer
and asynchronously re-clock the receivers DAC's with
a nice inherently low-jitter crystal based clock circuit.
The rate control protocol keeps the receiver memory
buffer approx. half-full so that no audio samples
are lost to under or overflow.

Ron
 

DaveLenhert

Stunt Coordinator
Joined
Nov 1, 2001
Messages
79
Lee-

I'm very skeptical that picosecond fluctuations in a clock signal can be audible. The reason for this skepticism is that we are talking PICO scales, ie 10^-12 seconds! To put that into perspective, light travels at 186,000 miles/sec, that is approximately 1 foot per nanosecond (10^-9 seconds), but you are talking scales that are 1000 times smaller. Even current state-of-the-art oscilloscopes are only capable of bandwidths of approximately 50 GHz, that is a resolution of 20 picoseconds. So, what you are saying is the human ear can hear the differences between 200 & 20 picoseconds. This resolution is barely possible with state of the art oscilloscopes, so I HIGHLY doubt that the human ear can discern such subtle changes in time.

Furthermore, your idea to make a CD to show the differences between 20 and 200 picosecond jitter has a few problems. First, you are making the assumption that your master clock is accurate to such time scales, ie, did your hardware come certified to those time scales. Second, even if your clock was accurate, you cannot simply play this CD in our systems to show the differences between 20 & 200 picoseconds. The reason is that our systems also add jitter to your CD so it is not possible to quantify how much jitter is added to the final audible signal. Finally, there can be multiple sources for the addition of jitter. As stated before by Ron, “The main source of jitter is the PLL based clock recovery circuit that is driven by the data stream itself.” Now, I don’t know if this is indeed true, but his point is valid that you have assumed that the major source of jitter is a result of the master clock on the CDP.

On a side note, being an experimentalist researcher, I’m a born skeptic. BUT, I would be more then happy to change my options that jitter is such a massive problem and very audible if you would kindly show me a paper that was published in a PEER REVIEWED journal, not some corporate white paper or a “Audiophile” magazine.

-Dave

PS: No hostility or personal attack is intended, just scientific doubt.
 

Jun-Dai Bates

Stunt Coordinator
Joined
Aug 16, 1999
Messages
148
A few points:

1. What Ian has stated is not a matter of opinion, however much you might disagree with him. His points are either factual or wrong (his main points, anyways), but not matters of opinion.

2. Blind testing, when done properly, is about as reliable as placebo studies--which is to say, more reliable for testing the capabilities of human perception and differentiation than reports from people that feel they can hear the difference in quality between two machines.

3. Don't underestimate the powers that your mind has to trick you. When you make associations with something, it affects your perception of it deeply. When you take away your ability to make those associations (as with blind testing), you will often lose your ability to differentiate that sound. Your mind, wanting to maintain those associations, may convince you that you can correctly perceive the differences, but if the test is performed correctly and the differences are indeed beyond human perception (or more importantly, your perception), then the results will be random.

4. If S/PDIF is unidirectional, and the digital information sent over it is sent raw (i.e. not in frames or packets), then that means that error correction is impossible (and therefore removed from the CD data before being sent through the digital output), which seems like it would be an incredibly poor design. If, on the other hand, the data is error corrected, then the data would have to be processed, in which case the 1s and 0s would be interpreted and then fed to the DAC at a rate and with a consistency dictated by the processor. Am I wrong? (if so, how?)

And if the data stream is raw, how does the DAC know what sampling rate the data stream is encoded in? In terms of 1s and 0s, the same number of bits per second would be sent in a data stream at 24bit*48Hz vs. 12bit*96Hz. Without this information, the data stream can't be processed correctly. But if the DAC does have this information, why can't it retime the 1s and 0s to a correct signal. No control signal is needed for any of this, just an understanding of the sampling rate of the data stream. Something here doesn't make sense.

5. The argument that one would have a motivation to hear the sound one way as opposed to another, and therefore the fact that one heard it the other way is proof that the judgement is correct/valid is worthless. It doesn't even qualify as poor evidence. If you take two identical sounds, your perception of which sounds better is based on the very complicated workings of your mind, and it is not necessarily going to have a direct correlation to which sound you feel you have a reason to prefer.

6. It is possible that you might be able to perceive small differences in sounds that are beyond simple scientific measurement (measurement of your perception, that is, not measurement of the differences). It is also possible that you might over time develop associations with that sound that you weren't able to make during the blind testing, during which you were only just becoming acquainted with that sound quality and not able to distinguish it from another because of all the other confusing, imperceptible things affecting that sound (moving air molecules, etc.). If this is true, then using the exact same setup (amplifier, speakers, cables, room, furnishings, etc), you ought to be able to pass a blind test in that setting that you weren't necessarily able to pass in a laboratory. If this is also true, then you have a good (
 

John Royster

Screenwriter
Joined
Oct 14, 2001
Messages
1,088
In my experience I was able to have my wife switch sources between a sony ES and an adcom cd player, both of which were hooked up via coaxial cable.

I would leave the room while she switches sources or not switched source.

8/8 trials I got right being able to tell her which player was playing (same CD was playing simultaneously in both). I would consider that a decent test and proves to me how transports can differ.
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
If S/PDIF is unidirectional, and the digital information sent over it is sent raw (i.e. not in frames or packets), then that means that error correction is impossible (and therefore removed from the CD data before being sent through the digital output), which seems like it would be an incredibly poor design.
The answer is that jitter can occur in uni-directional junctions even while error correction remains in tact because there are in fact two different data streams. It has been shown, and not refuted, that jitter can be introduced by using certain connections. Optical cables surprisingly are prone to jitter due to lack of quality manufacturing at the ends. This is why many high end reviewers shy away from using them.
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
Here is some more research and corroboration on the jitter issues provided by JVC in its XRCD mastering process...

http://www.xrcd.net/Shopping/process.asp

If you click on the PDF file and read the 7th paragraph, it states

The K2 Laser interface is used to reshape the EFM signal just before going to the laser beam of the glass cutter. This last stage is the same circuit used for the K2 Super Coding interface, which eliminates any time-based jitter that may be present in the data stream.
So what we have is JVC, one of the world's largest replicators of discs, admitting to jitter in the glass master process (disc replication) and then having corrected it with similar jitter reduction circuits used earlier in the chain.

Also, please note the effect of power supply stability mentioned in the PDF. This also coincides with Bob Katz work and that of Jon Risch on his website.

I do agree that Risch's grammar & text formatting could use some work, so we are in agreement on that. I am no English major either. ;)
 

Jun-Dai Bates

Stunt Coordinator
Joined
Aug 16, 1999
Messages
148
more points:

1.
Agreed which is why I don't place much value in disproving statements by saying that "since no DBT exists therefore it can not be true."
I don't believe that anyone here has made such a statement. Ian claims that there are reliable Blind Tests on the subject and that they back his argument. I claim that I don't know of any Blind Tests, but I'm skeptical of the claim that the quality of a CD player matters when you are using the digital output, because it goes against what seems like common sense to me, and I haven't seen any convincing science to back it up (corporate white papers about corporate products aimed at audiophiles--such as the JVC document--does not qualify as science).

3. Also, I don't understand the comment about error correction. I've never heard of error correction being sent as a separate data stream (I didn't even realize that S/PDIF cables were even capable of transmitting multiple data streams), but even so, if errors are actually checked for (and corrected if found), then the data stream has to be digitally processed and then sent on to the DAC with altogether new timing. Or am I missing something?

4. As for Risch's writing: the problem is more than just grammar, but more importantly, his writing problems, while acceptable in online forums (most people don't bother to proofread their posts before submitting them), reflect very poorly on the writer for a full-length article. If he wants anyone to take that article seriously (except for people that already believe what he's saying before they read it), he should rewrite it. The misuse of English will turn off most of his readers, and his poor reasoning will turn off the rest. (If Risch is reading this, I don't mean to insult his intelligence, but he's surely capable of writing better, and even a quick proofread would improve the document's readability. It really needs to be rewritten, though)

I would like to fish around for more opinions. Are there any electrical engineers here?

Jun-Dai
 

John Royster

Screenwriter
Joined
Oct 14, 2001
Messages
1,088
Are there any electrical engineers here?
Yes. Graduated from purdue '93.

I used to have a quote something along the lines of "I'm a audiophile and a EE...I argue with myself everyday." :)

It is my understanding of SPDIF that it is not framed or packetized. The clock is contained within the stream but I don't know if it is with a pre-amble or not. Doubtful because a pre-amble would mean it is framed.

But I will say that jitter is always measurable and crucial to keep within tolerance/spec for digital transmission. Ask any of the old bellcore guys who wrote all the standards.

So this kind of discussion leads to "is jitter audible?".

I don't know but in my home I can reliably pickout between two players used as transports only. And that's good enough for me.;)
 

ChrisWiggles

Senior HTF Member
Joined
Aug 19, 2002
Messages
4,791
OK! Enough about jitter. Sheesh. Why does everyone just assume that jitter is the only thing going on here? I personally don't know one way or the other whether jitter causes differences between transports. Obviously, the analog sections of different players will have an impact, but you can't compare a DVD player running as a transport to a reference levinson or krell or whatever. There is more going on that just jitter. The actualy transport and laser mechanism in even an entry-level but very good NAD cd player will be more stable, have better tracking, and overall pickup the data off the cd with more accuracy than your discman does as you are out jogging. So what if there is a little jitter that may or may not be audible, the accuracy is probably way more based on the actual transport mechanism itself, not just the digital output timing and buffering or whatnot. I think vibration would be 1000 times more significant in a cd player, and I am very skeptical that that is very audible at all. The transport mechanism is key here, and not all are created equally.

You guys are doing the equivalent of being down on your hands and knees recording to the nano-meter the tread patterns on car tires trying to find out why the BMW drives so much better than the GEO Metro...

And I'd be curious to hear this jitter cd too. I'm always open for a chance to test new things.
 

Users who are viewing this thread

Forum statistics

Threads
356,810
Messages
5,123,599
Members
144,184
Latest member
H-508
Recent bookmarks
1
Top