The attributes of Jitter. Can we hear it or not?

This could be another good explanation of why turning off the digital circuitry on many players delivers better sound (take note you low-cost Pioneer player owners!).

 

And also the 'power supply regulation overkill' of units like the Naim CD players, single or two-box (in this case the second box *is* the multiple psu, not the decoder...)
with each active circuit having it's own separate regulated rails (they don't meet until back at the star earth in the PSU..)

 

You forgot listener bias, power of suggestion, placebo effect, etc.

 

Give it up, Chris. You have way too many people here who HEAR it. Be glad you don't. It;s a curse.

 

I'm jumping back into this a little late, but I had a sort of similar problem with my CD player a few months ago (tho' it was a bit more like crackling than a helicopter sound) and it was due to a problem with the transformer on my surround amp decoder/amplifier. It wasn't consistently sending quite the right voltage and, so the tech support people at Acoustic Energy said, digitial is a lot more picky about that than analog. So when I sent anything over the digital outs from any source, problems. Over the analog outs, no problem. Does this tie in with some of the power issues discussed over the last couple of pages? Not necessarily in terms of the amp per se, but in terms of voltage irregularities from whatever source causing digital "noise"?

Well, no. Not really. Ever tried looking at an x-ray? Granted if we didn't in general agree on most of the basic aspects of ordinary life, there's be trouble. But once you start looking at anything like "complicated" cases of perception, things do break down rather quickly. I'm sure that I and my doctor, or I and my deep space astronomer, or I and my crime scene investigator don't agree on what we see (and that's probably a good thing, 'cause I don't see squat).

Geoff

 

Thank you Geoff!

Since I have people starting to back me, and all of us who hear and know there's a difference up, I think i'm going to sit the rest of it out.

BTW, the reason all the players I have are hooked up digitally is because one is part of a HT setup, and I am using my very best analog cables for other things in the system. I have been using a digital connection for the last couple of years on my setup. I just can't recall if the same thing happened with the analog connection. The reason I haven't tried it out yet is because I really don't feel like getting back there and disconnecting a lot of stuff right now.

 

There's this book called 'Extraordinary Popular Delusions and the Madness of Crowds'. Interesting read. Written well over a hundred years ago.

 

True, but more: when you burn different media in different drives at different speeds, you will find different levels of C1 and maybe C2 error. Whether the 'largest' is audibly any different from the 'smallest' is another issue.

 

<100 ps is usually what's aimed for, though the frequency spectrum of the jitter and of the recording, are going to matter. It's isn't necessarily true that 200 ps will be audible, much less
an order of magnitude more audible than 10 ps.

Anywah, here's an informative and amusing thread about certain wild claims about CDRs, with some sidetracks into the jitter issue.


http://www.hydrogenaudio.org/forums/index.php?showtopic=6907&hl=audible jitter&st=0

 

If two doctors could never agree on what's on an X-ray -- including *details* -- hospitals would grind to a halt. Moreover, it is NOT the case that when people 'see' different things, they are necessarily all 'true' in their own way. It is possible to be plain *wrong*. It is possible to 'hear' and 'see' something that doesn't really exist. The mind has extraordinary powers to convince itself of things, by taking shortcuts around reason. Scientific work always has to be on guard against 'wishful' results based on that phenomenon.

I suspect few here have bothered to verify whether what they think they hear, is really 'there' at all -- whether it holds up to any more careful scrutiny. Everyone's quick to assume that whatever they perceive, must really *be*, therefore let's just figure out where all this jitter is coming from, OK?

 

Let's start with the observation that CD jitter is neccessarily completely irrelevant when playing back from another media, e.g. hard drive. Since the CD transport isn't being used at the time the music is played, it can't - even in theory - have any effect (assuming the bits are correct, which they usually are).

Now, what about playing the CD in real time, on a normal CD player? It *should* be irrelevant there too, but it's theoretically possible - say if the DAC's power supply is polluted by the CD transport in a one-box player. Simple solution is to use an external DAC. The problem can't be solved at the CD end, so that's an exercise in futility.

 

Yup. I've assumed people are 'hearing' this 'jitter' when playing actual discs back on actual transports. If not, I'll need something more powerful than a rolls-eyes emoticon to convey my
incredulity.

And, in case people aren't following along closely, running your player's digital out to your preamp/receiver digital in -- quite common these days -- is an example of using an external DAC.

 

Going back to the original questions:

Yes. Jitter in the AD/DA processes introduces audible distortion that manifests itself as lack of image stability and frequency and amplitude modulation of the signal. Just listen to a solo soprano or a piano (recorded through a jittery ADC/DAC) and you will immediately be able to tell that something is wrong.

This, I feel, is where the discussion here is getting derailed. People are confusing sampling jitter with disc jitter. All optical media are designed to cope with high jitter levels. In fact, the jitter specification for CD is 35 nanoseconds. While that may seem like a small number, it can be shown that sampling jitter affects the signal-to-noise ratio as a function of frequency, with the highest frequencies most affected. To guarantee 16-bit resolution at 20 kHz, you need a sampling jitter value less than 120 picoseconds. So, how is it that the disc specification is so much larger? Because a WELL-designed optical pick-up will remove ALL of the disc jitter. The only thing that then remains is the sampling jitter of the ADC (which should be very small in any high-quality professional converter) and the jitter in the DAC which is a function of the quality of the crystal oscillator in the player (or outboard DAC).

So, does burning at higher speed cause more jitter? Yes and no. No, it does not change the sampling jitter. Yes, it may increase (or decrease) the disc jitter. As several have pointed out, different drives behave better or worse at different speeds, and with different discs. However, provided the burned disc stays within the jitter specification, it should make no difference (in theory). Why only in theory? Because there are many more factors that affect a player's ability to properly read and decode the data on a disc.

It may come as a surprise to some that all pits/lands on a CD are not created equal. As a result of the EFM coding used, the most sensitive pits/lands are the ones of length 3T (the pits and lands can only have lengths of 3T to 11T). The 35 ns jitter specification is based on measurement of the 3T pits/lands. Whether the pits/lands are created by a professional laser-beam recorder or your home player, inaccuracies in the lengths of and gaps between the pits/lands give rise to disc jitter. However, the jitter can also be increased by other factors such as tilt, disc eccentricity, disc flatness, etc. As a general rule, as the jitter level goes up, so the C1/C2 error rates also increase. C1 errors (also called BLER) are common, even on factory pressed discs, and should be less than 220 per second. There should be zero or very, very few C2 errors. It goes without saying that there should be zero CU (uncorrectable) errors!

To compound matters, the readout spot of the laser is rather broad (to help deal with scratches/debris on the disc surface). This leads to intersymbol interference (ISI) from the pits/lands along a track, and cross-talk from longer pits/lands in adjacent tracks, both effects increasing the level of disc jitter. The ISI is worse at CD's lowest recording velocity of 1.2 m/s, since the 3T pits are then at their shortest, and closest together. Much of the readout problems can be overcome by employing HF equalization in the optical pick-up, but, since this is not mandatory, the jitter specification has to be met for pick-ups without EQ. All high-quality transports/players likely employ HF EQ (if they don't, they should!)!

Now, let's consider what happens as you increase the burn speed, the laser power has to go up accordingly. This can lead to heating of the disc, which can result in poor pit formation. The laser power also needs to be kept at a level that leads to a balance between pits and lands. If the power is too high or too low either the pits or lands will become more pronounced. This balance between the pits and lands is called the Beta value, and it can vary across the disc surface. Ideally, there should be no difference, and Beta should be 0.0 across the disc. This will result in better readability (the pick-up will not need to change laser power continuously as it tracks). The better the readability, the fewer focussing and tracking errors there will be, which again contributes to lower error rates.

If you check the reviews at a site like www.cdrinfo.com you will see measures of the change in jitter, C1/C2 errors, Beta function, Focus Error and Tracking Error with recording speed and from disc manufacturer to disc manufacturer. You will also see that some discs burned at the highest speeds exceed the stated specifications for audio CDs. My assumption here is that the added error correction for data CDs means that the drive manufacturer can burn at <high number>X, compare the data on disc to the original data file, see that they are the same, and so claim a fanciful write speed. As the articles on www.cdrinfo.com point out, audio CD-Rs burned at these speeds may still produce music, but may not pass muster with serious listeners.

So, clearly, it is possible to create CD-Rs that will measure appreciably different, but are those differences audible?

Yes and no! Re-iterating what has been posted here already, a poorly designed player that does not de-couple supply ripple/RFI/etc. generated by the optical pick-up/servo/error correction logic from the DAC stage will be far more susceptible to revealing audible differences between discs burned at different speeds (or pressed CDs made at different plants). However, assuming the disc has jitter, Beta and C1/C2 errors within reasonable bounds, a well-designed player can eliminate all of the problems, resulting in the same data being delivered to the final FIFO buffer (ahead of the DAC) irrespective of which disc is played. Then, assuming no issues such as thermal changes during warm-up, the audio out of the DAC will be the same in all cases. A good transport/player will behave just like computer memory, and it is accepted that rips from two different discs that result in bit identical audio files MUST sound identical when played from RAM.

Reference has been made in this thread to work done by Prism on "identical" CDs (I use quotes here since they admit that they used different EFM codes, which shouldn't have any influence, but does mean that CDs with the same data on them are not necessarily identical at the pit/land level!). Despite the assertion that there are audible differences between data-identical discs, their conclusion after blind testing of expert listeners was: "Listening tests have so far failed to produce convincing evidence for consistent sonic differences among the TD-2 disc sets."

So, burn speed certainly has an influence on a disc's measurable performance, but whether that results in audible differences will depend on how far you've pushed the parameters, and how good your player is.

Tools like Plextor's PlexTools Professional will meaure all these parameters for those that really want to determine the best combination of drive, disc and burn speed!

 

Black Elk, I have to jump back in and say that your post above makes a LOT od sendse, and helps explain things a bit better. Where were you several pages ago?

 

Sorry, I was trying to find the time to write it all out and post, and wanted to have read all the other posts before stepping in (just in case someone else had done the honors). I've also been following some of the related posts like the needle-drop to CD-R one. So many threads, so little time!

 

Okay I'll answer the question, Yes.

 

Black Elk, thanks for the post. If I'm reading this quote well I have to arrive at the conclusion that burning at higher speeds will also "burn" the burner faster because the laser is being powered high in order to achieve its purpose.

 

Well, the laser power is proportionately higher, but then it's on for proportionately less time because the burn is completed faster, so I really don't know what, if any, effect this will have on burner longevity. If you had two burners designed to work at 1x and 100x and you spent 8 hours a day burning on them, then I imagine the 100x burner is going to "age" much more quickly unless its laser and motor system have been designed to cope with such high levels of continuous use. If there are any people involved in PC burner design on the board I'd love to hear their views.

 

Resurrecting this part of the thread. Here's the article, from December 1994(!)

http://stereophile.com/features/55/

it's by Bob Katz, who has been mentioned here before. Conclusion: bitwise identical but audibly different.

 

He provides no details of any listening tests, admits that any disc jitter has been correctly eliminated when the samples were compared in a DAW, does not test whether the files played from the DAW sound identical, and provides no measured data on any of the discs to support his hypothesis. As I mentioned earlier in this thread, people keep referring to the Prism paper and their conclusion was that expert listeners in blind test could not reliably tell the difference even after a training period where they admitted that they could recognize the "characterisitcs" of each pressing.