Head-Fi.org › Forums › Equipment Forums › Sound Science › Why does the transport matter?
New Posts  All Forums:Forum Nav:

Why does the transport matter? - Page 4

post #46 of 85

Apology accepted. normal_smile%20.gif

 

Thanks for the links.

 

I'm aware of Nick's work:

http://www.head-fi.org/t/486598/testing-audiophile-claims-and-myths/2100#post_9676361

 

To add to the list:

 

Paper on jitter simulation by M.O.J. Hawksford:

http://www.essex.ac.uk/csee/research/audio_lab/malcolmspubdocs/C134%20Paper%20121st%20convention%20(corrected).pdf

 

Part of a relevant thread on another board:

http://forums.stevehoffman.tv/threads/cd-identical-data-different-sound-the-argument.305408/page-4#post-8415713

post #47 of 85

Just a quick note, the CDROM jitter is not the same as play clock jitter. As we have discussed in another thread the DAC's play clock is totally independent and self generated (Async mode), it is totally insulated from transport jitter. This includes USB, CD transport and WiFi.

post #48 of 85

Even CD players with 1x speed produced in the 80s had such buffer memory, which is not only used to decouple but also required to control the reading speed of the disc.


Edited by xnor - 8/14/13 at 10:45am
post #49 of 85

There is also some signal processing before they hit the DAC. They use Reed=Soloman coding to CORRECT the error. A trivial, R-S coding is also the reason why you want to clean the CD with perpendicular instead of circular motion.

post #50 of 85

The argument with CD players was that noise elsewhere in the CD player, such as noise on the power and ground busses caused by the servo used to keep the laser on track, might differ for different CDs due to manufacturing tolerances, and this noise might cause jitter on the DAC output clock or be injected directly into the analog output. The same argument is applied to DACs, in that jitter on the input bitstream or noise on the ground line might equally affect the output.

post #51 of 85
Quote:
Originally Posted by dvw View Post

There is also some signal processing before they hit the DAC. They use Reed=Soloman coding to CORRECT the error. A trivial, R-S coding is also the reason why you want to clean the CD with perpendicular instead of circular motion.


You clean your CDs?!?
 I suppose you can scrap off large pieces of gunk, but I never clean my CDs because it makes them sound too sterile and clinical. I much prefer the flavorful sound of the CDs left in their natural condition. 

 

It's the same reason why you don't scrub your grill or a wok or a pizza stone --- it's the natural deliciousness that makes your food worth eating --- same thing with the CDs! Sometimes after a long session of Licensed to Ill, I swap it out of the CD player for a little bit of Ludwig Van's #9. I think some of the Beastie Boy flavor gets on the Beethoven disc, and I swear there's a little bit more syncopation in the 2nd movement, and about 3/4 of the way through the 3rd movement there's a little bit of brass monkey. Seriously, who doesn't want a hint of girls during the Ode an die Freude finale?

 

Also, be careful you don't accidently clean off the green sharpie markings on your CD that make it play better. If you do, you'll have to redraw them. Don't use a regular black sharpie either---it has to be green for best results.

 

This is a joke. Please use common sense! It's pretty difficult to get digital audio to sound only slightly different by random flipping of a bit stream. The analog section is the only real place where things can get slightly botched and any competently designed section its effectively transparent :)

Cheers!

post #52 of 85

Don Hills-Your reference to power supply noise affecting     the CD player is correct my Cyrus XTSE+  Transport all built and designed in house by Cyrus has TWO internal transformers to counteract just that which you state each has its own function to separate the motor from the digital circuit. . But even then I also have a large PS same size as the transport that operates via an umbilical cord to the transport to reduce-noise/hum/power distortion even more--overkill maybe but I get a clear totally open sound that goes down to great depth. 

post #53 of 85
Quote:
Originally Posted by Don Hills View Post

The argument with CD players was that noise elsewhere in the CD player, such as noise on the power and ground busses caused by the servo used to keep the laser on track, might differ for different CDs due to manufacturing tolerances, and this noise might cause jitter on the DAC output clock or be injected directly into the analog output. The same argument is applied to DACs, in that jitter on the input bitstream or noise on the ground line might equally affect the output.


 here is an example of an SPDIF -> analog DAC. Please help me understand where nanosecond-scale jitter on the digital transmission can find it's way into the analog output signal?

 

here's the data sheet for the SPDIF reciever CS8414

here's the data sheet for the DAC converter CS4334

 

Given this simple DAC, please help me understand how an advanced DAC on any competently designed device since 1995 can have jitter influence the output.

 

Cheers

post #54 of 85
Quote:
Originally Posted by Don Hills View Post

The argument with CD players was that noise elsewhere in the CD player, such as noise on the power and ground busses caused by the servo used to keep the laser on track, might differ for different CDs due to manufacturing tolerances, and this noise might cause jitter on the DAC output clock or be injected directly into the analog output. The same argument is applied to DACs, in that jitter on the input bitstream or noise on the ground line might equally affect the output.

I feel terribly for noise. They' always get blamed for everything and now for jitter. Well, there might be some truth to this, but it may not be the case in CD player. There are two assertions here; the noise is injecting jitter in the output stage (clock + analog output) and jitter is in the input stream due to noise.

 

1. If noise is the cause for all these troubles, then we can take the CD player out of the equation. With this theory, any CD player with a standalone DAC will not have this issue since noise generated by the CD player will not propagate to the DAC (separate power and ground). This should answer the OP's question. Now, with P/G(power ground) noise issue, the problem is easily mitigated with proper PCB design. Power and ground are easily separated. Most chips provide analog P/G, digital P/G and even separate transmit/receive P/G.

2. Does noise really affect jitter? If so, how? The noise influence is real. When the noise is coupled to the power, it affects the accuracy of the VCO. With today's silicon, this is a critical factor. However, the noise that actually compromise the jitter performance is largely generated by the switching logic on chip. The combination of small geometry and low voltage made the design a real challenge.There are a lot of design "tricks" that can be used to isolate the analog portion from digital noise in a mix signal design. The second part I don't understand. How does noise cause jitter in the output buffer?

3.Noise coupling into the data stream at the output DAC. This is not an issue because the data and the clock line are now decoupled. The clock is no longer embedded in the data. The data is always synchronous because it's clocked by the play clock from the read buffer.

 

Summary for the people that don't want to read my mumbo jumbo:

1. CD transport is not critical

2. Noise:induced jitter can be mitigated with competent designs

3. To my knowledge, I don't know how noise induce jitter at the output stage of a DAC. IMHO, I don't think it matters.

post #55 of 85
Quote:
Originally Posted by dvw View Post

3. To my knowledge, I don't know how noise induce jitter at the output stage of a DAC. IMHO, I don't think it matters.

 

As dvw described, the various digital parts of the signal chain---by design---operate independently of one another as long as the data transfer between the different sections is within the design specification. When data buffered and reclocked, whatever jitter might have existed in the previous component is irrelevant. If clock jitter is causing bit-errors between components, then something is broken.

 

The only place jitter can enter the audio signal is in the digital to analog conversion process, and that process's jitter is entirely determined by the quality of the DAC's clock.  If the DAC reclocks its data with its own dedicated high performance clock---like just about every modern design---then the jitter you end up with is exactly the jitter of the DAC's onboard high quality clock. 

 

If the DAC's power supply is susceptible to noise, and the design is such that the jitter of its clock is adversely screwed up by that noise, then you have yourself a badly designed piece of equipment---send it back or throw it away!

 

 

Cheers


Edited by ab initio - 8/15/13 at 12:33am
post #56 of 85

Amazing ! Cyrus electrical Engineers --digital engineers --analog electronic engineers . Design engineers -all with a whole string of  letters after their names are WRONG??? All qualified at major UK universities. Must write to Cyrus and tell them they are "not up to scratch" That will be a revelation to the MD.

post #57 of 85
Quote:
Originally Posted by duncan1 View Post

Amazing ! Cyrus electrical Engineers --digital engineers --analog electronic engineers . Design engineers -all with a whole string of  letters after their names are WRONG??? All qualified at major UK universities. Must write to Cyrus and tell them they are "not up to scratch" That will be a revelation to the MD.

Who? Regarding what exactly?

 

edit: total correlated jitter of their XT SE2 is about 115 ps. This player costs what, over € 1,000?


Edited by xnor - 8/15/13 at 8:00am
post #58 of 85

Title ?

 

Quote:

Originally Posted by Don Hills View Post

Apology accepted. normal_smile%20.gif

 

Thanks for the links.

 

I'm aware of Nick's work:

http://www.head-fi.org/t/486598/testing-audiophile-claims-and-myths/2100#post_9676361

 

Fame at last ! wink.gif

 

To add to the list:

 

Paper on jitter simulation by M.O.J. Hawksford:

http://www.essex.ac.uk/csee/research/audio_lab/malcolmspubdocs/C134%20Paper%20121st%20convention%20(corrected).pdf

 

Nobody (nobody rational anyway) disputes the damaging effects of jitter, the extent to which it introduces distortion or even the effective bit-depth to which even small amounts of jitter can degrade a signal of a given bit-depth, none of that is in dispute. The issue is whether it is actually audible. Both Dunn  and Hawksford created models of what should be theoretically audible but they never took the next logical step to test these empirically.

 

As an aside some of the assumptions of Dunn's model are shall we say extreme, for instance they talk about audibility of jitter in a musical signal that is 120db above the noise floor, unless you are in a room decorated with egg cartons this would give you peaks of er, well very loud.

 

To date those few (very few) who have tested jitter audibility empirically have not found these models and theoretical thresholds to be upheld.

 

 

 

Part of a relevant thread on another board:

http://forums.stevehoffman.tv/threads/cd-identical-data-different-sound-the-argument.305408/page-4#post-8415713

 

 

 


Edited by nick_charles - 8/15/13 at 9:45am
post #59 of 85
Quote:
Originally Posted by duncan1 View Post

Amazing ! Cyrus electrical Engineers --digital engineers --analog electronic engineers . Design engineers -all with a whole string of  letters after their names are WRONG???

Oh no. They aren't wrong. They fixed something that didn't need fixing and used it to get you to buy their product. If I was them, I would consider that 100% correct .

There can be quality differences between low end and high end equipment. It's just that the difference isn't audible.
post #60 of 85
Quote:
Originally Posted by bigshot View Post


Oh no. They aren't wrong. They fixed something that didn't need fixing and used it to get you to buy their product. If I was them, I would consider that 100% correct .

There can be quality differences between low end and high end equipment. It's just that the difference isn't audible.

 

 

In some pathological cases the differences can be audible. I have a pair of WD HDTV music streamers which fed an Entech 203.2 and one day listening to a pretty saturated MP3 track (Crowded House) I heard an annoying distortion that was not apparent on the source CD, nor was it readily apparent on the MP3 played back via a computer. I took recordings from the digital outputs and found they were different. On an already saturated track the WD was actually increasing the digital signal output and causing clipping which was visible when captured and analyzed in Audacity, the digital output from the PC did not clip though it was hitting maximum a lot. I am at a loss to explain this but this was what I measured. My only wild guess is that the WD native sample rate was 48K and so the 44.1 track already right on the limit was being resampled badly and thus clipping. In context this was a $130 movie played that also did music after a fashion. I then did some DBTs of the wav file vs the MP3 and found I actually could hear the difference, the mp3 was slightly distorted.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Why does the transport matter?