Probably a touchy subject, but WHY?? (1 Viewer)

[Lewis Besze]

You do know that Stereophile measures every CD player they test for jitter right? Just because you don't understand something, doesn't make it a "crock". One more time for those who don't read:

Yes they do and many others are measuring this anomoly,however,to my knowledge there werent any study that proved,that jitter is readily audible to anyone,or that that's the big reason for CD players for "sounding" different.
If there was I'd like to read about it.
------------------
"You Hungarians always disagree"

 

[Saurav]

You could hold half the CD in memory, and jitter could not be an issue.

You would need to read the CD at a higher speed than 1x though. Reason is, you need to fill this memory buffer at least 50% before you can start moving data out of it. This is to ensure that the buffer never gets full or empty (i.e., you never get a buffer overrun or underrun error). So... at normal CD reading speeds, 1 CD is 650 MB which is 74 minutes, so to fill 128 MB (1/2 of 256) would take about 14 minutes. Every time you skipped a track or picked a different CD on a changer, you'd have to wait 14 minutes before you heard any music.
Anyway... I'm not going to go into any technicalities here, and I assume your statement was in jest, because if it were really that simple to design this stuff, then you're right, someone would have already done it.

 

[Saurav]

Yes they do and many others are measuring this anomoly,however,to my knowledge there werent any study that proved,that jitter is readily audible to anyone,or that that's the big reason for CD players for "sounding" different.
If there was I'd like to read about it.

There are informal accounts available that show that sufficient amounts of jitter are audible. However, AFAIK there isn't much research data available on this, especially, what the thresholds are. One can obviously set up a system with an extreme amount of jitter, and that would certainly be audible. However, we're usually talking milli or microseconds (nanoseconds?) with normal equipment, and again AFAIK no one is sure if we can actually hear jitter in that range. I think the S/PDIF and AES/ABU specifications have some reference to jitter, in terms of the maximum allowed level or something like that.

 

[Lewis Besze]

Saurav,
most cd dvd player's jitter measures in pico an nano seconds,which is why it's very debateable,if it's audiable at all.
Many in the audiophile community claims this as one of the reasons why,even extremely overengineered cd players in the market exhibit "different" sound among each other.
However this remains a theory,and once again comes from the crowed,that "hears music,not numbers".

------------------
"You Hungarians always disagree"

 

[John Morris]

Well, I hear music and not 1s and 0s, and I don't believe that I've ever heard jitter? Then again, I have never invested in an outboard DAC, so maybe it has always been so minute that it was never apparent to my ears.
What does "jitter" sound like?
If no one can answer that question, then I guess that no one else has ever heard it either.???

------------------
Take Care,
merc
----------------
Link Removed
God Bless America!!!

 

[Saurav]

If no one can answer that question, then I guess that no one else has ever heard it either.???

Very true. You'll never know until you've actually heard something different. That's usually how the human mind works - for instance, if all you've heard is a grainy amplifier, you will never realize that it's grainy and you'll just accept that that's how amplifiers are supposed to sound, until you hear a not-so-grainy amplifier.
That's part of the reason why these things are so heatedly debated - sometimes, you just need a well trained ear to hear some things. Any musician will know what I'm talking about. Let's say you play bass guitar, and your band is trying to cover another band's song - when you're listening to the song, you need to shut out the rest of the music and zero in on just the bass line. Most people cannot do that, they hear all of the music. This gets even tougher if there are two guitars playing and you're trying to figure out what one of them is doing, or with multiple part vocal harmonies, and so on.

 

[Lewis Besze]

John, It was an "insider" joke.Of course noone hears "numbers".I meant to show,that while "audiophiles" like to bash science and EE,time to time they like to "borrow" from them to make a point.
As for what jitter sounds,well the above mentioned group,like to treat this like "distorsion" in the signal,so the adjectives they use is similar,as to what they use with it's "analog counterpart".
Namely, "loss of detail,shrill top end,looser bass[or ill defined],narrower [less enveloping]soundstage,less sweet[or warm] midrange,etc......"
You got the picture,right?
------------------
"You Hungarians always disagree"

 

[Lewis Besze]

That's usually how the human mind works - for instance, if all you've heard is a grainy amplifier, you will never realize that it's grainy and you'll just accept that that's how amplifiers are supposed to sound, until you hear a not-so-grainy amplifier.

You open up a whole new can of worms here,I think.
1 How you define grainy?
2 Do you think that listening to the same amp,everybody will conclude that it's "grainy"?
3 Can we determine the sound to be "grainy",by other means,other then listening? If no then ,see #2.
So my points are that hearing is very personal,and any differences between any components,can be traced to individual perception,therefore cannot be argued,on academic level,and all personal preferences should be respected as such,IMO.
------------------
"You Hungarians always disagree"

 

[Saurav]

Lewis,
Very good points, all of them. Which reminds me of something which I'd read a long time ago - all equipment reviews are pretty much meaningless unless they answer the question "when compared to what?". Or, in other words, you always need a frame of reference in order to meaningfully describe anything. I believe Einstien said something like that too

 

[Neil Weinstock]

Earlier in this thread, Saurav wrote:
quote: However, your input waveform isn't just the voltage levels - a waveform consists of specific voltage levels at specific instants of time. If you get the levels right but mess up the timing, your reproduced waveform will not be the same as the original.[/quote]
Just wanted to re-emphasize that, since it is such a succinct summary of the problem. Digital audio out = Data PLUS timing. Getting the data right is no problem. Timing is another matter.
Anyway, regarding the notion that a cheap transport plus a memory buffer should solve the jitter problem: I tend to agree that this oughta be the answer. It has been done in car audio players, if my memory serves, which have a buffer for shock protection (not expensive!), in addition to the high-end CD players that do this exclusively for sound purposes.
Regarding buffering, Saurav wrote:
quote:
You would need to read the CD at a higher speed than 1x though. Reason is, you need to fill this memory buffer at least 50% before you can start moving data out of it. This is to ensure that the buffer never gets full or empty (i.e., you never get a buffer overrun or underrun error). So... at normal CD reading speeds, 1 CD is 650 MB which is 74 minutes, so to fill 128 MB (1/2 of 256) would take about 14 minutes. Every time you skipped a track or picked a different CD on a changer, you'd have to wait 14 minutes before you heard any music.[/quote]
Yeah, but the buffer doesn't need to be so big. Please poke holes in this argument:
1) Assume the maximum oscillator drift between the output stage and the CD mastering device is 200 PPM. I pulled this number out of the nearest available orifice, but it seems reasonable (conservative, even). That is the same as 1 part in 5000.
2) On a 70 minute CD, there are 4200 seconds. Therefore, the maximum drift over an entire CD is less than 1 second. Therefore, you'd only need to pre-buffer about 1 second to guarantee no buffer underruns over the course of a CD. I one-second delay in startup would be hardly noticeable, since there's always a bit of delay anyway.
3) So, a CD player that only reads at 1x would only need a 1 or 2 second buffer, which is nothing.
4) A CD player that reads at >1x has no problem at all.
5) Shock-protected car CD players prove that the technology to do this is routine.
I must be missing something, because not all CD players are made this way. I feel like if I knew why, I'd be a step closer to understanding the universe...
[Edited last by Neil Weinstock on November 11, 2001 at 11:58 PM]

 

[Neil Weinstock]

Just had another thought. Has anyone ever built a jitter demonstrator? It should be possible to build a CD player with a jitter level knob, for the ultimate refutation to the notion that "bits are bits". It would be most interesting to play with such a device...

 

[John Morris]

Lewis: ROFL! I love it. The last time I actually heard numbers when listening to music, it went something like this: "a one, a two, a one, two, three, four." Still, I didn't hear a single zero... damn!
Anyway, all I wanted to say is that for me so far... jitter has been a mythical figure spoken about mostly by high end dealers who wanted me to spend four times as much on their high end products. Even so, I've never really seen(or heard) that mythical figure demonstrated in person for me. I've listened to probably a hundred CD players in my life and many have sounded crappy. Still, I guess what you are saying is that they may have sounded crappy due to "jitter" and not to the fact that they were just crappy CD players with possibly crappy DACs or crappy transports.
Unless someone can convince me, "Jitter" just seems like another word that supposed audiophile dealers use to scare us mid-audiophiles into spending way too much on price controlled inflated audio hardware...???
------------------
Take Care,
merc
----------------
Link Removed
God Bless America!!!

 

[PatrickM]

The affect of Jitter on audio isn't as mythical as you think. In something like Voice over IP telephony, which is coming to your home soon, one of the key factors in voice quality is jitter and for the best sound quality you need to have a QoS (quality of service) specified that will keep the jitter below a certain level.
I know VoIP is not CD data but it is similar.
Patrick
------------------
Evidence of my obsession.

 

[Nick G]

>
OR, prove the reverse. I think it is a great idea for a test. The thing that strikes me is if jitter in automobile CD players is easily fixed with a cheap buffer (as a previous poster suggested) it seems like a no brainer to do the same for home CD players. Unless jitter in home CD players is one of those phantom problems that cannot be clearly identified or quantified and which may, or may not really exist.
Nick

 

[Saurav]

OK, there's one thing that a lot of people seem to be confusing - the buffer in a car or portable CD player is for error correction, not jitter correction. That buffer comes into play when the CD player takes a shock which cases it to skip - i.e., the laser loses its tracking. This is a gross error which will definitely cause the 1s and 0s to go wrong, and that's what a shock protection buffer does, prevent data error. While I too believe that the way to address jitter is with buffering, it's not as simple as using a portable CD player, or a cheap player with a stick of RAM.
quote: Yeah, but the buffer doesn't need to be so big. Please poke holes in this argument:[/quote]
There are no holes in your argument. Numbers I've seen are in the range of a few seconds, based on oscillator difference measurements with average consumer-grade equipment.
I think this is the classic issue of marketing vs. engineering - you're right, the technology to implement this is trivial, but someone probably decided that a CD player which paused 2-3 seconds whenever someone changed tracks would not sell, even if it sounded better than other CD players. And think about it - how many people in your CD player's target market even care about jitter? When you go to high-end players, they do address this issue, but those players are expensive, for other reasons - parts and engineering quality, as well as brand name markup, of course. However, IMO the added price does bring with it superior engineering, and it's not all markup. For instance, Linn has a transport/DAC combo, where in addition to the regular digital data cable (optical or coax) between the transport and the DAC, there's a second cable going between them too, which carries a clock signal. And get this, it's the DAC which feeds the clock back to the transport, not the other way round. This eliminates jitter a lot, because the master clock driving the digital system (and specially the DAC) is now derived purely from an oscillator chip, and not recovered from an incoming bitstream using VCOs and PLLs. As you can see, this results in much better jitter performance.
quote: I one-second delay in startup would be hardly noticeable, since there's always a bit of delay anyway.[/quote]
You must remember that this isn't just at startup, it would pause when switching CD's, or jumping tracks, or even cueing within a track. In short, anything which required reading something other than the next immediate bit of data.
quote: Has anyone ever built a jitter demonstrator? It should be possible to build a CD player with a jitter level knob, for the ultimate refutation to the notion that "bits are bits". It would be most interesting to play with such a device...[/quote]
I believe such devices exist. It should be fairly easy to add controlled amounts of jitter into a digital signal.
Also, there's a DAC called the Genesis Digital Lens which does actually read the data into a buffer and then clock it out using an independent clock. AFAIK, the LCD display on this device shows the level of jitter in the incoming signal, I believe in ppm (parts per million).
quote: Unless someone can convince me, "Jitter" just seems like another word that supposed audiophile dealers use to scare us mid-audiophiles into spending way too much on price controlled inflated audio hardware...???[/quote]
Like I said - go to http://www.aes.org No offense to anyone, but it's always amusing to read posts which go "I've worked with computers all my life and I know everything there is to know about digital electronics"... ummm... right.
There, I just demonstrated the size of my ego. Anyway...
quote: Unless jitter in home CD players is one of those phantom problems that cannot be clearly identified or quantified and which may, or may not really exist.[/quote]
Well, it exists, whether you believe it or not It can and has been clearly identified and quantified. It can actually be measured, and yes, on home audio equipment.
Whether it can be heard or not, or rather, how much can be heard, is still being debated.
[Edited last by Saurav on November 12, 2001 at 03:07 AM]

 

[Lewis Besze]

It can and has been clearly identified and quantified. It can actually be measured, and yes, on home audio equipment.Whether it can be heard or not, or rather, how much can be heard, is still being debated.

Bingo!

------------------
"You Hungarians always disagree"

 

[Zbigniew]

Hm...
There are few things we need to keep in mind.
Jitter is defined as variation of delay on arriving datastream. Constant delay is not a problem - there is lot of gear which will nicely remove echo (G.164 standard for telco is a good reference point). For a music or TV it is even easier - as long as it is constant, you will not notice it with music, and unless you will get sound out of synch with video, it does not create a problem.
If your data is arriving with varying speeds (gradual, or just one packet faster, one slower) you will get into troubles. Unless you use external reference clock, and start timing whole equipment together, you will have troubles. Timing synchronization is one of major issues for telco; lack of proper timing was/is a limiting factor in implemenation of different VoiceOver{IP/ATM} implementations.
From what I understand, PCM encoding used for CD does not include timing - each device uses either it's own internal clock (and pray to god that it is stable in short and long terms); or it can retrieve clocking from circuit (Tosling or coax). If you retrieve from circuit, you are using god knows what quality circuit on other end + errors (slips, variations) introduced by circuit.
A good, stable reference clock can be retrieved from GPS devices, or specially build devices - anywhere from a few hundreds to few thousands $$. Of course, you can use nework based reference clocks - but I am still looking for a Ethernet equipeed HTV system.
Bottom line - even if we get all bits in proper sequence, timing of bits is critical for some protocols. Sadly, Toslink was not designed to be a very reliable protocol.
We can today send at speed up to 10GB over 4000 KM hops; but the interfaces used for it costs much more than top of line HT.
Would I use a separate jitter removal device ? When I will upgrade my processor, I will look at how potential unit deals with retiming, reclocking of signals. There are some units which will do just that - so using cheaper CD-changer + jitter corrector might be a good idea...
_zjt

 

[RicP]

quote: Unless someone can convince me, "Jitter" just seems like another word that supposed audiophile dealers use to scare us mid-audiophiles into spending way too much on price controlled inflated audio hardware[/quote]You know, sometimes I wonder why I bother. Does anybody actually read the links that are posted? All the information you need to understand was posted and is readily available.
As a self-described "mid-audiophile", you are willing to admit that there are things that you do not fully understand right? Not everything that you don't understand is "mythical".
You are free of course to believe what you wish, but when it flies in the face of scientific documents that have been posted, well...that's a little silly.
------------------

Ric Perrott - My DVD's
[Edited last by RicP on November 12, 2001 at 10:24 AM]

 

[PaulKH]

I don't know why you bother either, RicP.

I did read Bob Katz' article, and I thought it was an even bigger CROCK. His only 'evidence' that 'jitter' is a problem is his own listening tests. No oscilliscopes, no re-digitizing at higher rates those 'jittery' sources that lack soundstage and whatever other piffle he had to say about them. Sounded more like a wine tasting description... yes, that CD player wasn't woody enough.
Someone back there even said PCs can only play audio CDs at 1x because the data stream is so special. Another CROCK. If that were true, it would take 74 minutes to make a copy of a CD.
CD audio is two 16 bit samples at a piddling 44,100 times a second. Period. No ifs, ands, or buts. Reading that data rate and delivering it reliably out of a digital cable is TRIVIAL. You'd have to have one shitty circuit to mess up that data rate. Something like a pea shooter maybe.
This discussion sounds EXACTLY like those about cables and how those $1000 cables make a difference. If you believe it does, more power to you.
Yeah, I know this is 'stirring it up' but what the hey...
[Edited last by Paul Higginbottom on November 13, 2001 at 09:05 PM]

 

[Todd Hochard]

I think over the years, I've convinced myself that high dollar transports sound better, simply because they are more expensive. The WOW factor in your head turns on, preventing you from clearly judging the material.
One of these days, I need to sit down with someone who CAN hear the difference, and see if they can illustrate the "where" and "why" to me. Virtually all of the articles I've read on the subject invariably fall back on the "golden ear" reason, so I simply can't get any value from such statements. But perhaps I can sit down with a golden ear, and be shown how. I'm genuinely curious.
Todd
P.S. I'm not being facetious, either. I am curious.