Storage Options

[rowuk]

Here is a real heretical thought: jitter is insignificant as CD playback is NOT in real time, it is buffered. It is treated like wow and flutter, but that shows the big misunderstanding. That jittery digital signal path feeds the bits to a buffer before decoding. If the buffer runs dry, we have a problem. Therefore it has to be at least big enough for the error correction to be able to compute missing values and putting the stereo back together as it is stored sequentially not in parallel on the disc.

Can anybody tell me what artifacts show up in the audio signal when jitter is introduced? Can we see or measure them?

Please note that I am not saying that low litter playback chains do not sound different. I am just questioning if jitter is really the poison.

[jkeny]

Jan 16, 2010 23:00:17 GMT rowuk said:

Here is a real heretical thought: jitter is insignificant as CD playback is NOT in real time, it is buffered.

I've seen this argument before - i.e. buffering avoids jitter! BUT (it's a big but) ask yourself, is the whole song buffered before playback starts? If this doesn't happen then there is CD reading happening while simultaneously bits are being read out of the buffer. I believe the CD reading causes all sorts of power supply noise which is one of the main ways for the introduction of jitter!

It is treated like wow and flutter, but that shows the big misunderstanding. That jittery digital signal path feeds the bits to a buffer before decoding. If the buffer runs dry, we have a problem. Therefore it has to be at least big enough for the error correction to be able to compute missing values and putting the stereo back together as it is stored sequentially not in parallel on the disc.

It's assumed that any decent buffer operates properly in this regard

Can anybody tell me what artifacts show up in the audio signal when jitter is introduced? Can we see or measure them?

I can hear how lower jitter sounds - more like very good analogue

Please note that I am not saying that low litter playback chains do not sound different. I am just questioning if jitter is really the poison.

It's probably just one of the poisons of digital.

[rowuk]

Jan 16, 2010 23:19:40 GMT jkeny said:

Jan 16, 2010 23:00:17 GMT rowuk said:

Here is a real heretical thought: jitter is insignificant as CD playback is NOT in real time, it is buffered.

I've seen this argument before - i.e. buffering avoids jitter! BUT (it's a big but) ask yourself, is the whole song buffered before playback starts? If this doesn't happen then there is CD reading happening while simultaneously bits are being read out of the buffer. I believe the CD reading causes all sorts of power supply noise which is one of the main ways for the introduction of jitter!

It's assumed that any decent buffer operates properly in this regard

I can hear how lower jitter sounds - more like very good analogue

Please note that I am not saying that low litter playback chains do not sound different. I am just questioning if jitter is really the poison.

It's probably just one of the poisons of digital.

This kind of avoided all of my "questions". If the problem is power supply, then we just need more power supply isolation between stages. Not that technically difficult.

I am still interested in what "artifacts" jitter leaves in the analogue output. If I have a bucket of water with a 1mm hole, the water leaks at a constant rate regardless if the bucket is completely or half full or tanked back up in 44.1khz bursts. I think a decent buffer behaves the same way. I think jitter is just something that is easy to measure and quote and therefore an easy target to market. A better buffer would probably accomplish more. Caching an entire CD to a fast RAM Disc is not expensive or difficult. Then jitter should be a non-issue, even in lower cost systems. I just have not read any compelling research on why jitter should change the sound if the rest of the system works properly.

[jkeny]

Jan 16, 2010 23:40:14 GMT rowuk said:

This kind of avoided all of my "questions". If the problem is power supply, then we just need more power supply isolation between stages. Not that technically difficult.

Again you are quoting a stock answer - more ps isolation - not difficult! Oh really, so easy? - then why is there so few very low jitter CD players?

I am still interested in what "artifacts" jitter leaves in the analogue output. If I have a bucket of water with a 1mm hole, the water leaks at a constant rate regardless if the bucket is completely or half full or tanked back up in 44.1khz bursts. I think a decent buffer behaves the same way. I think jitter is just something that is easy to measure and quote and therefore an easy target to market. A better buffer would probably accomplish more. Caching an entire CD to a fast RAM Disc is not expensive or difficult. Then jitter should be a non-issue, even in lower cost systems. I just have not read any compelling research on why jitter should change the sound if the rest of the system works properly.

My simplistic understanding of jitter is as a timing problem! When the signal went through A/D the samples were taken at a particular time. When going through D/A the exact same time for these samples needs to be used! Imagine a sine wave is the signal - what happens if some samples are translated back to analogue at the wrong time? You work it out!
The ps is critical to all this - if it's not perfectly clean it modulates the timing & causes jitter.

[Deleted]

This kind of avoided all of my "questions". If the problem is power supply, then we just need more power supply isolation between stages. Not that technically difficult.

Robin
You have to be joking !
It would be a nightmare trying to correct all the ills caused by the SMPS in a typical CD/DVD player of today. You would ideally need to fit a whole new linear supply with multi regulated sections, and perhaps a special transformer with several different secondary windings. It is just too hard to improve the output sections of an SMPS as most of the derived voltage supplies are "tuned" Even replacing a capacitor in this area with an incorrect type is fraught with danger.
Alex

[Deleted]

Jan 16, 2010 23:53:49 GMT jkeny said:

Jan 16, 2010 23:40:14 GMT rowuk said:

This kind of avoided all of my "questions". If the problem is power supply, then we just need more power supply isolation between stages. Not that technically difficult.

Again you are quoting a stock answer - more ps isolation - not difficult! Oh really, so easy? - then why is there so few very low jitter CD players?

I am still interested in what "artifacts" jitter leaves in the analogue output. If I have a bucket of water with a 1mm hole, the water leaks at a constant rate regardless if the bucket is completely or half full or tanked back up in 44.1khz bursts. I think a decent buffer behaves the same way. I think jitter is just something that is easy to measure and quote and therefore an easy target to market. A better buffer would probably accomplish more. Caching an entire CD to a fast RAM Disc is not expensive or difficult. Then jitter should be a non-issue, even in lower cost systems. I just have not read any compelling research on why jitter should change the sound if the rest of the system works properly.

My simplistic understanding of jitter is as a timing problem! When the signal went through A/D the samples were taken at a particular time. When going through D/A the exact same time for these samples needs to be used! Imagine a sine wave is the signal - what happens if some samples are translated back to analogue at the wrong time? You work it out!
The ps is critical to all this - if it's not perfectly clean it modulates the timing & causes jitter.

Robin
I am 100% with John on this one !
BTW, have you ever tried modifying or improving CD/DVD players, even those older ones with linear PSUs ? HUGE improvements can be made by fitting additional regulators,better decoupling etc., but try doing this with a recent player.
Alex

[rowuk]

Jan 17, 2010 0:21:03 GMT @sandyk said:

Jan 16, 2010 23:53:49 GMT jkeny said:

Again you are quoting a stock answer - more ps isolation - not difficult! Oh really, so easy? - then why is there so few very low jitter CD players?

My simplistic understanding of jitter is as a timing problem! When the signal went through A/D the samples were taken at a particular time. When going through D/A the exact same time for these samples needs to be used! Imagine a sine wave is the signal - what happens if some samples are translated back to analogue at the wrong time? You work it out!
The ps is critical to all this - if it's not perfectly clean it modulates the timing & causes jitter.

Robin
I am 100% with John on this one !
BTW, have you ever tried modifying or improving CD/DVD players, even those older ones with linear PSUs ? HUGE improvements can be made by fitting additional regulators,better decoupling etc., but try doing this with a recent player.
Alex

Maybe I am just too stupid to understand the process. Numbers are moved from the disc into a buffer. The path between the two has timing problems, but the numbers do not change. All the pieces are stored in the buffer until the word is complete, then that word is turned into part of an audio signal. If the buffer works right (and it does), no amount of jitter changes the characteristics of each word or the timing of the words in relation to one another. My research into CD players shows that the buffers work even in cheap players. It also shows big differences in sound, that cannot be accounted for if we understand the real jitter process. The words are not subject to any serious amount of jitter and looking at the reconstructed audio signal at the RCA jack does not show a "wavering" time base proportional to jitter.

Obviously there must be something else that accounts for what I hear. We all accept that the power supply plays an incredible role in all audio devices, digital, analog or otherwise. There are also well accepted topologies for "best" reproduction. We have no issue here.

I have some suspicions where the real audible differences come from (in addition to power supply issues common to digital and audio) and have been able for instance to measure RF in the audio signal at the RCA output as well as a very unnatural noise floor (natural noise is harmonically coherent, that means multiples of some fundemental frequencies). Error correction is also a "possible" culprit.

Higher end playback gear does lower jitter, but also uses other approaches to filtering and isolation. I simply have found no compelling technical argument that shows how jitter can get into or change the audio signal or change the characteristics or timing of the "words". The problem must lie somewhere else. I am looking for help where that could be.

Jitter is just an easy sell to someone not familiar with the actual process and trying to place the blame on something for they really do hear. Just like comparing a Mercedes and a Fiat, the differences cannot be accounted for in numbers of cylinders, horsepower, gas consumption or weight. The differences are the sum of many things. Picking out jitter is like picking out power windows as the source of problems.

We have been taken for a ride by the audiophile press in the past. I think jitter is another issue that has a bit of truth, but has been blown WAY OUT OF PROPORTION by the marketing people. It exists, resembles at least to the "casual" observer wow and flutter and the rest is journalistic privilege.

What does Leo say?

[jkeny]

Jan 17, 2010 18:23:38 GMT rowuk said:

Maybe I am just too stupid to understand the process. Numbers are moved from the disc into a buffer. The path between the two has timing problems, but the numbers do not change. All the pieces are stored in the buffer until the word is complete, then that word is turned into part of an audio signal. If the buffer works right (and it does), no amount of jitter changes the characteristics of each word or the timing of the words in relation to one another. My research into CD players shows that the buffers work even in cheap players. It also shows big differences in sound, that cannot be accounted for if we understand the real jitter process. The words are not subject to any serious amount of jitter and looking at the reconstructed audio signal at the RCA jack does not show a "wavering" time base proportional to jitter.

We are all trying to understand what is going on - so take my comments as my best guess so far - I may & probably definitely WILL change my mind in the future about all of thi - that's the nature of investigation. Anyway, I have no disagreement with you in what you said - but what happens when the data is taken out of the buffer to be transmitted over SPDIF or I2S some translation has to happen within electronics from PCM to SPDIF or I2S. My contention is that when this is happening if power fluctuations are happening then jitter will be introduced into the signal. These PS fluctuations are more likely to arise within CD mechanisms with mechanical servos drawing current peaks then in solid state storage. So buffering only gets you some of the way down the tortuous path of clean audio data transmission.

Obviously there must be something else that accounts for what I hear. We all accept that the power supply plays an incredible role in all audio devices, digital, analog or otherwise. There are also well accepted topologies for "best" reproduction. We have no issue here.

I have recently modded a HiFace USB transport feeding I2S to a DAC & when I changed the ps to the clock to a clean battery supply the sound quality jumped a couple of notches. Nothing else changed & the original ps was still in there so in this case I believe the jitter was the main culprit of les than optimal sound.

I have some suspicions where the real audible differences come from (in addition to power supply issues common to digital and audio) and have been able for instance to measure RF in the audio signal at the RCA output as well as a very unnatural noise floor (natural noise is harmonically coherent, that means multiples of some fundemental frequencies). Error correction is also a "possible" culprit.

Higher end playback gear does lower jitter, but also uses other approaches to filtering and isolation. I simply have found no compelling technical argument that shows how jitter can get into or change the audio signal or change the characteristics or timing of the "words". The problem must lie somewhere else. I am looking for help where that could be.

Jitter is just an easy sell to someone not familiar with the actual process and trying to place the blame on something for they really do hear. Just like comparing a Mercedes and a Fiat, the differences cannot be accounted for in numbers of cylinders, horsepower, gas consumption or weight. The differences are the sum of many things. Picking out jitter is like picking out power windows as the source of problems.

We have been taken for a ride by the audiophile press in the past. I think jitter is another issue that has a bit of truth, but has been blown WAY OUT OF PROPORTION by the marketing people. It exists, resembles at least to the "casual" observer wow and flutter and the rest is journalistic privilege.

What does Leo say?

I do believe you are correct that a lot of issues get lumped into the category of jitter but that isn't to say that "real" jitter isn't important

[Spirit]

Dec 22, 2009 1:36:27 GMT @sandyk said:

You are still sticking with laptops and USB

Alex, from all I read last night, the HiFace is an audibly better transport than your Xonar...

Can't comment on the ref1, or the ROC, but there's nothing wrong with usb as implemented in the hiface...
Jkeny of course has gone further, supplying it with cleaner power, but even stock, it's better than any soundcard it was compared to in the headfi thread...

[Deleted]

Mar 13, 2010 2:09:43 GMT Spirit said:

Dec 22, 2009 1:36:27 GMT @sandyk said:

You are still sticking with laptops and USB

Alex, from all I read last night, the HiFace is an audibly better transport than your Xonar...

Can't comment on the ref1, or the ROC, but there's nothing wrong with usb as implemented in the hiface...
Jkeny of course has gone further, supplying it with cleaner power, but even stock, it's better than any soundcard it was compared to in the headfi thread...

Don't believe everything you read in Head Fi !

[Spirit]

Mar 13, 2010 3:11:20 GMT @sandyk said:

Don't believe everything you read in Head Fi !

I don't

Seriously... what kind of a reply is that?

[Deleted]

Mar 13, 2010 3:45:03 GMT Spirit said:

Mar 13, 2010 3:11:20 GMT @sandyk said:

Don't believe everything you read in Head Fi !

I don't

Seriously... what kind of a reply is that?

Perhaps it's time to upload a few high resolution .wav files to see if your setup is capable of resolving what Leo's setup can ?

[Spirit]

Mar 13, 2010 3:47:02 GMT @sandyk said:

Perhaps it's time to upload a few high resolution .wav files to see if your setup is capable of resolving what Leo's setup can ?

I can tell you as it stands currently, there's no way in hell it will.

I'm not familiar with what soundcard, if any, Leo has.... perhaps we should ask him for his opinion on how the hiface sounds relative to a reasonable (xonar?) one...

[Deleted]

Mar 13, 2010 3:50:36 GMT Spirit said:

Mar 13, 2010 3:47:02 GMT @sandyk said:

Perhaps it's time to upload a few high resolution .wav files to see if your setup is capable of resolving what Leo's setup can ?

I can tell you as it stands currently, there's no way in hell it will.

I'm not familiar with what soundcard, if any, Leo has.... perhaps we should ask him for his opinion on how the hiface sounds relative to a reasonable (xonar?) one...

Phil
Don't try and drag Leo into your present beef with me, which appears to have been due to my stirring you up about Lego blocks ?
As I have stated previously, I would love to try a Buffalo DAC with Paul Hynes regulators, based mainly on Leo's reports. However the simple fact is that I can no longer afford to do so. I look forward to seeing your reports about your setup based on good quality source material, into a decent headphone amplifier or Power amplifier and good speakers.
That is why I offered to upload some high resolution .wav files now that your listening gear has been substantilly upgraded.

Alex

P.S.

The point is that you were equating the rubbish sound from a usb dac, to what usb transport is capable of, which imho, is silly

I didn't say it was rubbbish, you did. I said that it wasn't as good as my modified X-DAC V3 via the Oppo and SPDIF. I was equating the performance of the Benchmark USB DAC via USB with it's performance from the same PC via SPDIF.
The Benchmark USB DAC isn't a normal USB DAC because of the way it is implemented. The manufacturer claims anti jitter properties via USB
YMMV with different Desktops and Laptops.
I am out of this discussion, as you appear to be spoiling for a fight.

[Spirit]

Mar 13, 2010 4:32:43 GMT @sandyk said:

Phil
Don't try and drag Leo into your present beef with me, which appears to have been due to my stirring you up about Lego blocks ?
As I have stated previously, I would love to try a Buffalo DAC with Paul Hynes regulators, based mainly on Leo's reports. However the simple fact is that I can no longer afford to do so. I look forward to seeing your reports about your setup based on good quality source material, into a decent headphone amplifier or Power amplifier and good speakers.
That is why I offered to upload some high resolution .wav files now that your listening gear has been substantilly upgraded.

Alex

What?

You were saying usb audio was rubbish... and that Valter couldn't possibly hope to hear the differences in your rips with such substandard gear. I was merely making the point that while most usb dacs use rubbish usb transports (ala benchmark), that not all usb TRANSPORTS are rubbish.
(and seemingly, the hiface is a superior transport to your xonar...)

WTF does that have to do with PH or Buffalo or anything else...

I mentioned the Buffalo/hiface so you could hear the hiface (and compare to your xonar)

(none of which I actually have yet ) BII ships next week, and I'll pick up a hiface when I've had a good play with the BII.


I'm not allowed to believe what I read on headfi, and the only person I know of with a highface whose opinion you trust (who might also have a decent soundcard), is Leo
Hence mentioning him ;D

[Deleted]

Robin is RIGHT on the subject of jitter (my tech opinion of course, my ears are crappy).

a lot of things are blamed on jitter when it has probably more to do with the proper decoupling of power lines, as Alex already has mentioned is of importance in HIGH speed circuits (CD players, PC, DAC, SS amps with huge bandwiths)

Should someone feel the need for extra info on jitter there are lots of tutorials on this on the web or P.M. me.

Mind you... It is hard to explain technical shit into everyday words/analogies for non-technical persons.

[Deleted]

a lot of things are blamed on jitter when it has probably more to do with the proper decoupling of power lines, as Alex already has mentioned is of importance in HIGH speed circuits (CD players, PC, DAC, SS amps with huge bandwiths)

Frans
That is another reason why I do not believe that you can say that a particular brand of soundcard will perform exactly the same in every PC or laptop. Radiated crap will also vary from PC to PC , as will the quality and "stiffness"of it's SMPS. I have also been caught out several times by those crappy Molex connectors . They often go high resistance due undoubtedly to the cheap metal used in their contacts.I mentioned that previously in a post about fan problems.Ideally, you would also have separate power feeds, without piggybacking connectors. The more upmarket SMPS often have more individual isolated feeds.
Alex

[rowuk]

Mar 13, 2010 7:36:07 GMT @sandyk said:

a lot of things are blamed on jitter when it has probably more to do with the proper decoupling of power lines, as Alex already has mentioned is of importance in HIGH speed circuits (CD players, PC, DAC, SS amps with huge bandwiths)

Frans
That is another reason why I do not believe that you can say that a particular brand of soundcard will perform exactly the same in every PC or laptop. Radiated crap will also vary from PC to PC , as will the quality and "stiffness"of it's SMPS. I have also been caught out several times by those crappy Molex connectors . They often go high resistance due undoubtedly to the cheap metal used in their contacts.I mentioned that previously in a post about fan problems.Ideally, you would also have separate power feeds, without piggybacking connectors. The more upmarket SMPS often have more individual isolated feeds.
Alex

This sounds like a real case for an external DAC with its own power supply............ Or a properly designed internal DAC with power supply isolation/stabilization! Maybe we need a thread on PC internal PS stabilization? It would not be difficult to have 3x molex connectors on the input for redundancy and 5 or 6 quality, filtered molex outputs.

Cutting the power supply traces on a DAC board is not that tough.