Dacs and Transports

[George]

<font color='#000000'>Hi there,
I currently use a Marantz DV 8300 for DVD as well as DVD Audio and SACD. For two channel music(stereo), I have read that it may not be as good as a purely dedicated cd player(although I am currently using it for stereo).
This then led me onto reading more and finding out about high end cd players(mind you with what I layed out for the Marantz, I assumed it was also high end).
In reading, I came across the use of transports, DACS and things like jitter etc.
Could someone explain in fairly simple terms what difference all these make to sound quality and why there is such a huge difference in prices of ccd players.
At the moment I am contemplating getting a Proceed CDD transport to be coupled to my Denon 5803, to be used for stereo.
Again, I have the option of coupling it directly or using a Proceed DAP before inputting it into the Denon.If I input direct then I would be using the Dacs in the Denon.
Is it correct that analogue inputs into receivers are first converted into digital and then back again into analogue before being sent to the speakers? If so, why? With the various conversions would there not be some degradation in the signal?
Thanks
George.</font>

 

[av_phile]

<font color='#000000'>Many exotic or designer gears use consumer electronics dressed-up with arguably better parts and housed in better-built chassis and then priced more than 10 times. &nbsp;I guess anything new in the market that is not priced high wouldn't get notice. &nbsp;To me. if there is any sonic improvement, such an improvement doesn't justify the cost makr-up. &nbsp;Understandably, these high end brands have to market their products much higher as they produce only so few units. So the appeal of having a rare breed of gear is also there.

Personally, I'd get those high ends not because of any sonic superiority over a flagship consumer product, but becasue thei are better built and can impart an aura of &quot;rarity&quot; and &quot;class&quot; to my audio gears. &nbsp;

Jitter and a host of other data corruptuon are being heralded to justify sonic differences between digital players and digital interconnects. &nbsp;Many internet sites consider these claims snake oils as there is no sceintific date to make such observations repeatable. &nbsp;Personally, I don't subscribe to them. &nbsp;Digital players retrieve 0s and 1s or they don't. &nbsp;Error correction enters the picture to make sure the 0s and 1s are complete prior to analog transcirption. &nbsp;If the correction fails, you would hear pops or skips in the material that last no more than a few milliseconds or seconds at the most, not in the imaging or clairities one ascribes to ENTIRE musical passages or CDs. &nbsp;Difference in sonic character &nbsp;between digital players, such as better imaging, more defined highs, punchier bass, etc are more attributable to their post-DAC circuits.

Some recievers will convert the analog signals from a DAC/Player to digital especially if such receivers use digital volume controls or when using DSP and prologic. &nbsp;The better ones allow analog signals to by-pass thsse conversion circuits when playing in STEREO mode. &nbsp;Some would, by using the digital output of players, eliminiate 1 or 2 conversion steps. &nbsp;They let the receiver do the digital to analog conversion rather than have the player do it only to be converted back to digital by the receiver. &nbsp;In general, audiphiles frown at so many circuits thorugh which a signal passes through. The simpler the better. &nbsp;I would opine this is true for analog signals since the more cuircits or stages are used, the more phase inversions and added harmonic distortions are introduced. &nbsp;In digital, however, the process is a lot cleaner. &nbsp;But still, I'd prefer lesser conversion processes just the same. &nbsp;

For whatever my 2 cents is worth.</font>

 

[abe]

<font color='#000000'>av_phile is right on the mark. &nbsp;I know this is very controversal but digital is digital. &nbsp;The [/U] conversion [/U] part is the key to the sonic quality, either digital-to-analog (DAC) or electrical-to-mechanical (speakers). I learned my lesson and sold my high-end digital gears. &nbsp;


abe</font>

 

[Guest]

<font color='#000000'>Actually I think you have simplified a few things and quite possibly ignored a few others. Timing Jitter is only a problem when you convert from either analog to digital or digital to analog. The way jitter effects the analog audio typically shows up under test for dynamic range. The dynamic range test was rarely done for consumer audio products so nothing really came of it.
One of the major problems with many DAC's was that they formed a pretty hefty ground loop when they were connected to most pre amps. The other major problem is that you will need a very low noise floor in the equipment downstream in order to begin to hear what a good DAC is capable of.
I realize the above is rather simplified but if you have any more specific questions; feel free.
         d.b.</font>

 

[Rip Van Woofer]

<font color='#000000'>Hear, hear, av-phile! Well said!

This particular line of hooey can mostly be laid at the feet of Stereophool mag. Jitter, especially, has been misinterpreted to justify all manner of BS.

Actually, using an outboard DAC increases the likelihood of jitter errors, I've read. Something to do with the clocks not synching or something.

The subtle degradations of sound that we 'philes are so keenly (obsessively?) attuned to is an &quot;analog thing&quot;. As pointed out, when digital goes bad it does so unsubtly and unmistakably, by clicking, skipping, or simply not playing at all.

To make an analogy with computers (think CD player) and software (the CD): when Word, say, gets corrupted, you don't notice subtle degradations like the text getting a little fuzzy on the screen! Instead, gross errors like a major slowness in opening files occurs. Or your formatting goes kerflooey. Or it crashes.

As The Audio Critic mag put it: unlike with an analog signlal, you can't make 1's and 0's any more or less &quot;1-ish&quot; and &quot;0-ish&quot;.</font>

 

[Yamahaluver]

<font color='#0000FF'>Stereophool

&nbsp;

&nbsp;

Fantastic Rip Van Woofer, I couldnt have made that better, they do make good TP.</font>

 

[Rip Van Woofer]

<font color='#000000'>Thanks Yamahaluver, but I can't take credit! Saw it somewhere else.</font>

 

[av_phile]

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>

Guest : <font color='#000000'>Actually I think you have simplified a few things and quite possibly ignored a few others. Timing Jitter is only a problem when you convert from either analog to digital or digital to analog. The way jitter effects the analog audio typically shows up under test for dynamic range. The dynamic range test was rarely done for consumer audio products so nothing really came of it.
One of the major problems with many DAC's was that they formed a pretty hefty ground loop when they were connected to most pre amps. The other major problem is that you will need a very low noise floor in the equipment downstream in order to begin to hear what a good DAC is capable of.
I realize the above is rather simplified but if you have any more specific questions; feel free.
         d.b.</font>

<font color='#000000'>Hi there db,

I am curious how a timing eror in the digital domain translates into the analog. &nbsp;Correct me if wrong, but a well designed DAC is supposed to retime or re-sync the digital signals and buffered so that they are complete and corrected prior to any analog conversion. &nbsp;

So a low noise floor is necessary to hear the difference between DACs. &nbsp;How low should it be? &nbsp;Thanks</font>

 

[Dan Banquer]

<font color='#000000'>It's been a few years since I designed and built and outboard DAC but I will attempt to explain. I observed that timing jitter will effect low level linearity in a multibit DAC. Before I start talking directly about this I will fill you in some basics that you may be lacking. For 16 bit audio every bit represents 6 db of dynamic range, so for a 16 bit system we have a 96 db dynamic range. (6 x 16 = 96). When you have a signal that is -60 db below full scale you are now just using the bottom six bits. This is where both Burr-Brown and Analog Devices concluded that dynamic range should be measured. The THD + N at this level should be -36 db ( 8X oversampling will give you an extra 1.5 db to -37.5db, but I don't think anyone has actually achieved that). The dynamic range measurement just adds 60 db to what ever you get for the THD + N at - 60 db below full scale.
So we now know that a near ideal Multi Bit DAC should have a dynamic range of 96 db and the THD + N at -60  db should be -36 db THD + N.
I observed the following. When I was using a Yamaha DIR (Digital interface receiver) my THD + N at -60 db was approximately -32 db THD + N. When that was replaced with a Crystal DIR the measuremant was repeated and the THD +N went to - 36 db THD + N. This is pretty close to ideal.
The major reason for this was the Crystal DIR had less jitter than the Yamaha DIR.
Most DAC's have some kind of jitter attenuation, the real question is how much. One bit DAC's are much more sensitive to timing jitter than multi bit DAC's so it's really all about applications here.
Now in order to be able to appreciate an extra 4 db of dynamic range and less overall lower level distortion you will  need a quiet system, a quiet, and well acoustically treated  room and speakers that have some decent low level linearity.
The above can be pretty rare in consumer audio.
I hope the explanation helps. It's the best I can do at the moment.  

             d.b.</font>

 

[goodman]

<font color='#000000'>George is looking to get the best CD redbook sound possible and is considering trading up from his Marantz DV 8300 to get it. &nbsp;What if, instead of buying an unobtanium CD player, he connected his DV 8300 to his AVR 5803 with a digital connection? &nbsp;Wouldn't the DACs of the DV8300 then be bypassed, so that the AVR 5803 would then convert the digital output of the DV8300 to analog? &nbsp;Gene DellaSala, in his review of the AVR 5803 says it has just about the best DACs around, and he likes that they operate in dual differential mode, whatever that means. &nbsp;Then George would have great DACs without spending another dime.
I need some of you bright engineers to tell me whether I have this right or wrong. &nbsp;Thanks.</font>

 

[goodman]

<font color='#000000'>George, I was right.  You don't need to buy some fancy stand-alone DAC. If you connect your Marantz to your AVR 5803 with a digital connection, the internal DACs in the Marantz become irrelevant, because you will be using the terrific DACs in the Denon AVR 5803.  Good listening.</font>

 

[av_phile]

<font color='#000000'>That is applicable only for playng CDs. &nbsp;Not for high reoslution DVD-A and SACD which will still require the internal DACs of the marantz player. &nbsp; Unless the Denon AVR has its wn DVD-A and SACD processor, does it?</font>

 

[av_phile]

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>

Dan Banquer : <font color='#000000'>It's been a few years since I designed and built and outboard DAC but I will attempt to explain. I observed that timing jitter will effect low level linearity in a multibit DAC. Before I start talking directly about this I will fill you in some basics that you may be lacking. For 16 bit audio every bit represents 6 db of dynamic range, so for a 16 bit system we have a 96 db dynamic range. (6 x 16 = 96). When you have a signal that is -60 db below full scale you are now just using the bottom six bits. This is where both Burr-Brown and Analog Devices concluded that dynamic range should be measured. The THD + N at this level should be -36 db ( 8X oversampling will give you an extra 1.5 db to -37.5db, but I don't think anyone has actually achieved that). The dynamic range measurement just adds 60 db to what ever you get for the THD + N at - 60 db below full scale.
So we now know that a near ideal Multi Bit DAC should have a dynamic range of 96 db and the THD + N at -60  db should be -36 db THD + N.
I observed the following. When I was using a Yamaha DIR (Digital interface receiver) my THD + N at -60 db was approximately -32 db THD + N. When that was replaced with a Crystal DIR the measuremant was repeated and the THD +N went to - 36 db THD + N. This is pretty close to ideal.
The major reason for this was the Crystal DIR had less jitter than the Yamaha DIR.
Most DAC's have some kind of jitter attenuation, the real question is how much. One bit DAC's are much more sensitive to timing jitter than multi bit DAC's so it's really all about applications here.
Now in order to be able to appreciate an extra 4 db of dynamic range and less overall lower level distortion you will  need a quiet system, a quiet, and well acoustically treated  room and speakers that have some decent low level linearity.
The above can be pretty rare in consumer audio.
I hope the explanation helps. It's the best I can do at the moment.  

             d.b.</font>

<font color='#000000'>Thanks DB for enlightening me on the analog effects of digital jitter. &nbsp;I never had you experienc in builing an outboard DAC, but based on my scant technical knowledge, I suspected that jitter does affect dynamics and will really not be that audible on most home systems. &nbsp;And if at all, the difference would be subtle. &nbsp;And not the mushy bass, lost details and poor imaging that some say they hear. &nbsp;

So based on your explanation, a single-bit DAC oir 1-bit DAC generates more noise than multi-bit DACs. &nbsp;Does that mean that SACD's 1-bit sampling gets more noisy than a DVD-A's 24-bit DAC?

Another question. &nbsp; I saw an add for an ultra-expensive CD transport-only device that claims its machine can extract more bits than most ordinary CD players. &nbsp;How true is this?</font>

 

[Rip Van Woofer]

<font color='#000000'>A transport only spins the CD, moves the laser pickup to read the CD, and shuttles it in and out. It's pretty much a mechanical device (well, it does send a signal to the DAC).

Neither it nor any other device can &quot;extract more bits&quot; than those that are already burned into the CD. You can pretty safely assume that all modern CP players can already do so. All CD players from portables to megabuck have more than adequate upsampling (aka oversampling) to give you all the bits that can be gotten.

It's pretty much the quality of the DAC stage that makes or breaks a CD player. And even the ones on a $50 portable are pretty good these days. Plug one into your system just for fun and hear for yourself! Ain't technology grand?</font>

 

[Guest]

<font color='#000000'>To AV phile: SACD uses a noise shift out to about 100 Khz. There has been great debate in the pro audio area as how this effects tweeters, and the preceeding circuitry. Being the old fart that I am I am generally of the opinion that if it doesn't belong there, get rid of it.
As far as the transport and retrieving more bits, this sounds like they are using a FIFO memory set up to temporarily store all of the bits coming off the CD and then reclocking the bits out. If this is implemented properly then less timing jitter will be the result. For more information I would recommend www.digido.com. This is the Bob Katz web site who is a rather well known recording and mastering engineer, and there is a boatload of relevant information on this subject. There is also some possibly useful information in the article &quot;Current Trends in the Recording Industry&quot; which is on the Audioholics web sitre. Enjoy.</font>

 

[Dan Banquer]

<font color='#000000'>The post above is from Dan Banquer, It's early in the morning and I haven't had enough coffee to kick start me yet.
&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; d.b.</font>

 

[Rip Van Woofer]

<font color='#000000'>Dan -- that Digital Domain site looks good on first perusal. Being a non-techy trying to understand and assimilate this stuff, it might help a lot. So there can be a bit more to a CD transport than I thought...thanks for the correction.

Maybe you can confirm/clarify something I've read (a bit OT for this thread); that one of the main reasons SACD and DVD-A offer better performance is because they allow more room for error in recording, specifically in setting the 0dB level. In other words, you can get the same quality from CD, but only if the recording engineers are really on the ball and Murphy's Law doesn't rear its head. True?

Have you heard of the so-called giant memory, or FeeFIFOFum?

(sorry!)</font>

 

[Dan Banquer]

<font color='#000000'>One of the problems they found in the overloading of CD's was that the digital peak meters would only register on 3 consecutive samples. (Note: This was a mangement decision overiding an engineering fact from what I understand) This would allow &quot;clipping&quot; to get through because the peak meter needed to register below 3 consecutive samples.
At present SACD does not have the software or hardware to the best of my knowledge to do things like hyper compression. This does not mean it can't be developed. As for DVD, it's on the way; I have already warned GDS to keep an ear peeled for this.
If you are noting a touch of cynicism you are correct.
               d.b.

P.S. I was an audiophile until I found out what they were doing in the recording studio.  

</font>

 

[Guest]

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>

Dan Banquer : <font color='#000000'>One of the problems they found in the overloading of CD's was that the digital peak meters would only register on 3 consecutive samples. (Note: This was a mangement decision overiding an engineering fact from what I understand) This would allow &quot;clipping&quot; to get through because the peak meter needed to register below 3 consecutive samples.</font>

<font color='#000000'>To expand further on DB's comments, digital peak meters register an 'over' with three consecutive samples at either +full scale or -full scale. &nbsp;I fail to see this as managers overiding engineers. &nbsp;One sample at +/-FS does not indicate clipping. &nbsp;Two samples at +/-FS is also fully recoverable as no information has been lost. &nbsp;When three or greater samples are at either +/-FS, then the waveform cannot be accurately reconstructed to mimic the original because information has been lost. &nbsp;This, then, is indeed clipping.</font>

 

[Dan Banquer]

<font color='#000000'>There appears to be quite a debate on this issue. This topic is certainly not my area of semi expertise, however I would like to send you two pdf files that concern this very subject. Can you provide an e mail address?
                d.b.</font>