Time to discuss: the Jitter Thread

Hi folks, there're many tests (at least 20) about jitter on the testers such as AP2722, those manufacturers won't tell us about these "small" things. There's a guy who once worked in a Brit pro DAC company told me, THD ain't important (you can treat), THD+n is important (you can't treat), but more important is those measurements about jitter, but outsiders (we) know nothing about these jitter measurements before purchasing, somebody has to test DACs in detail before buying, that's a paradox.

 

You should probably just go here: http://www.superbestaudiofriends.or...olutions-cd-and-sd-players-daps-and-more.212/
Read about jitter measurements here: http://archimago.blogspot.com
And look at the last page on any of these reviews (the "measurements" page): http://www.stereophile.com/category/computer-audio-reviews

Audibility of jitter is the important part. Frankly, most modern DACs measure superbly in jitter. You can check the Archimago and the Stereophile links.

 

What does jitter sound like?

When people outside the world of deeeeep digital processing talk abut one piece of kit having more or less jitter than another, on the basis of listening tests, I have to admit that I am sceptical. We have analog problems with names like wow and flutter, and they are not so hard to detect and identify, and they are not so hard to identify with their mechanical causes.

Jitter, though, is something else. The closest we get to jitter is some helpful diagrams explaining, in simplistic terms for us non engineers, why it happens, but they do not tell us how (or even if) it sounds. At extreme levels (google for example files) the word does suggest the effect, as in our old-fashioned analog distortions, but anything that sounds like that is never have going to have made it onto the market.

My suspicion is that it is often imagined, or used as a convenient word to describe some perceived audible difference which may have nothing to do with it at all.

 

Jitter doesn't sound like anything. From my own experiences, Jitter is more about a problem (or problems) in the translation of digital audio data to analog signals, where the improved (jitter "free") signal just sounds a little bit better; separation of instruments can be placed more easily and depending on what is being played, how well it's been recorded, and how good your hearing is, you may actually experience things like "more depth". I wish I could say it was BS, because I'm one of those that approaches non-obvious things with skepticism and doesn't like to spend more money than they have to, but I'm sorry to say it's not bullshit.

First, I have to say this (before anyone asks me to produce scientific data): I work in audio and have about 20 years of experience working with digital playback systems. Like a lot of people that work in audio, we get paid to do work and listen to all kinds of music and that eventually leads to having experienced hearing; a lot of us are not scientists, we simply buy tools that do the job. Often times we try things out and if they're worth their keep, we go on with them (or we don't and move on). I really, really wish Jitter was bullshit. But it's not. I honestly would have loved to save money in investing in things like an external word clock.

Jitter doesn't affect FR or dynamic range, which is probably why it's so difficult to make out on less than pristine playback systems (did I already mention that you need pretty good hearing too?) I first heard it when I got a "Pro-sumer" recording interface years ago, that has a notoriously bad word clock, the Digidesign 002. It's not the best ADC/DAC but it is for sure "decent" even by today's standards. I had a friend of mine bring in an Apogee Big Ben (word clock) after I complained that I wasn't really happy with what I was hearing after doing some recordings and he said "It's the shitty clock; with a better clock, that thing sounds fantastic."

So, he left it for a week. There's no clock output on the 002, so I clocked to the Big Ben via SPDIF and immediately, I heard what my friend was referring to. Wait, no, I didn't "hear" the clock - I just heard clearer audio because I had "cleaned" the 002's bad clocking issue by using the Big Ben. That opened up a huge can of worms for me and the search began for a better ADC that was "jitter immune" (ended up buying Crane Song converters for my main rig).

I still use the 002 to this day when doing stuff on Pro Tools (I refuse to give them more money after their decision to make that system "subscription based" but that's another topic) but I clock it to a Lucid 88192; I can't tell the difference between clocking to that and the Big Ben, but I simply won't use the 002 without an external clock (and believe me, I wish I could because every time I power cycle, I have to tell the 002 to see the external clock).

 

Jitter definitely does sound like something, if it is above audible levels for the current listener.

 

+1 for dedicated word clocks. Wish more consumer gear had it. It does make a difference.

 

I believe in an ideal case you want the clock as close as possible to the dac (ie: inside the case with it) and you'd slave external components to the dac clock. But yeah the Digidesign 002 is notorious for having a bad clock, and was probably one of the devices that really sparked off the external clocking debate online.

For the typical home hifi user though, they aren't going to have a ton of processing gear and whatnot that needs to be synced up, so for the cost of an external clock it may just be better to buy a better dac in the first place.

 

Funny this thread pops up, today i'm receiving a Marian seraph d4 which also has wordclock input. And i'm already thinking of getting a Mutec MC-3+ non usb to give it a good clock signal and then reclocking it before it goes to my dac. Total price would be a bit more expensive then a MC-3+ usb without the headaches that usb can cause.

About jitter i always thought it sounded a bit fuzzy/dirty in certain frequencies and perhaps a bit crunchy sounding. Most noticeable on optical. On coax i hear it a bit on my Oppo blu ray player. But if you can't really hear jitter then what am i hearing?

 

Hahaha, now THAT is what I can refer to as "Jitter BS".

EDIT: If it is "above audible levels for the current listener", guess what - it's another way of saying it's inaudible (meaning, you can't hear it). Dislike me all you want, it still doesn't change the definition of "inaudible".

 

Truth. A lot of today's DACs are very well implemented, and many don't have the same jitter error problems that ADC/DAC units had even 10 years ago. Good point!

 

Perhaps I haven't been following close enough, but it seems you guys have dug yourselves into a semantic hole. Assuming that jitter is the actual issue here, it's a form of distortion that is obviously unlike a noise floor or hum when you talk about it being "audible". Rather it is an effect that can color the signal in a more subtle way, not unlike harmonic distortion. As others have said, even low levels of jitter produce a slight haze or lack of absolute clarity. It does seem that relatively high amounts of jitter (as found in certain USB interfaces, transports, DACs, etc.) can produce an even more offensive sibilance, grain or stridency in the treble. Or at least that is what members of our community have correlated with excess jitter via years now of listening tests.

 

Its not obvious in measurements but somehow even low levels (-110db) can make a dac sound like mud. Also imo, consumer gear shouldn't have word clock, there is simply no need for extra paths in pcb, increased cost, increase the number of signal paths in the pcb for possibly negligible benefit aside from causing increase audiophile nervosa. Its justifiable with pro gear needing it with large number of dacs/adcs that require slaving to a single clock. Consumer gear should just imo come packaged with the clocks as it is status quo.

 

For an $100 DAC sure, but for anything that sells itself as high-end it makes sense to sync your source's word clock to that of the DAC's. If audiophiles are going to worry about resistor ladder accuracy, digital filters, ultra-high performance analog stages, 15 sections of independent power regulation, then adding mechanism to keep the receiver/transmitter in lockstep doesn't seem like a bridge too far.

 

I know I've seen jitter measured (on maaany AES papers and graphs I've seen by several AEs who are way more knowledgeable than I on this subject) but as far as "hearing" the effects of jitter, that's where I disagree - because frankly, I have tried. many. times.

In audio, we often do null tests to be able to amplify, then hear artifacts and I've done that with all sorts of gear (mic pres, amps, effects processors, consoles, shit, even cables one time when someone came to one of the studios I've worked at trying to sell us some). I've never heard of a "null test" for Jitter (just a ton of science and data about the subject that demonstrates jittery data vs data that isn't). I'm someone who runs away from science in audio; once I start seeing an oscilloscope and numbers, I'm out.

THD can be heard AND measured, to many varying degrees. Some effects processors are coveted for their harmonic distortion (and even with that, you have some exact units that vary slightly in their level of distortion). Harmonic distortion can be obvious AND subtle and all sorts of values of "in between".

Jitter (in my book) is simply an error in the translation of digital data into an electrical signal that our brains then translate as "audio".

One of the biggest revelations of a "jittery" converter vs. a "jitter-free" converter I have ever heard was on a 24/48 recording of a drum kit, and the cymbals simply sounded slightly "veiled" on the jittery converter as opposed to the converter that was clocked properly (no difference in frequency response nor dynamics). It wasn't a "colored" signal, it wasn't at all like harmonic distortion in the ways I would recognize them; it was that I could hear nuances of the decaying cymbals more clearly. It's not a subtle effect but rather a subtle difference. I wouldn't know how to "capture" the difference between the poorly clocked signal vs. the one that was more revealing.

 

LIke I said, a silly semantic argument

For those of us operating solely on the playback side, the distinction is lost. But yes, I am not aware of any method extract the "sound" of jitter so I understand your approach to the question

 

So feeding say for example the Multibit Gungnir with a USB source that has measurably less jitter than an SPDIF source and comparing the two isn't "hearing the jitter"? I was under the assumption that you can hear jitter and it sounds as people described above. Lack of resolution/clarity/focus are some of the ways I would describe the comparison.

Plug the Chromecast Audio into Gungnir Multibit and compare to a USB source. That's the sound of jitter.

 

Something forgot to mention: r2r dacs ain't so sensitive to jitter. So when we talk about jitter, we mainly talk about delta-s dacs.

According to my little research, there are some basic test about jitter, subjects included: jitter rejection slope, jitter attenuation, jitter rejection corner frequency, intrinsic jitter. Still don't fully understand these testing subjects, if any insider could privode some information, that would be a big help.

 

Schiit, that's a good idea!

 

Especially since we tend to categorize as "jitter issue" everything related to digital transport.
There are other causes of sound deterioration in digital transmission. Jitter is one of them. Electrical noise is another.

Electrical noise has various origins as well:
- Noise during transmission (RF, EMI)
- Noise generated by the receiver when recovering the signal (which in turns affects other elements in the DAC, unless the receiver is isolated from them - total isolation is rarely achievable).

A jittery signal will cause the receiver to generate more noise (or at least, a different profile of noise), which in turn might affect the other components differently -> "Cables sound different".

Stereophile has some nice articles about all this. "Bits is Bits?" is one of them. Can't recall the name of the one that compared digital transport, which seem to correlate sound differences between transports with their jitter profile (not just the integration in a single ppm value).