Question regarding digital signal transfer (long)

Chrisssssssss

Ok, I have a question here but a thread sparked a question that I have been always wondering and thinking about, but in my mind I see it this way (but I am sure that there are other opinions...) but:

-I am wondering about digital signal transfer. I can see how a different DAC can make a huge world of difference because each one had a different algorhythm of taking digital information and converting it to the analog waveform that will ultimately be passed through the various stages of electronics. Ultimately this waveform will be ending up at the speaker who's job it is to take that electrical analog signal and convert it into sound pressure waves that end up at our ears. Fine. I understand that an analog signal ends up being altered with certain signatures of each analog stage it passes through just due to the design characteristics of the electronics. Everytime an analog signal passes through any sort of electronic component, it is altered somewhat, this is why for a truer (for lack of a better word) signal, the shorter, most direct path will have less effect on the signal, therefore being closer (actually less altered from) the original signal. Agreed? Did I make any sence? Capice? Bueller? Bueller?

Well, here's what I don't get...

1) There was mention that the transport (assuming the definition of transport is the team of laser, motor, platform, clamping, etc... The process of reading the disc...) makes an audible difference on the overall sound that is observed in final playback.

and

2) The digital signal transfer method (Toslink-optical, Coax-non optical, and their related connectors and cables that relay this digital data signal between stages) also makes an audible difference in the resulting soundwave.

The point here I am getting to (finally I know... I just wanted to clarify what I am trying to say...hang on!) In my thought and the way I see it is a digital signal no matter where it comes from or how it gets there should be an EXACT bit for bit relay of the data from the media, to the stage at which this data is processed into an analog signal. A "1" is a "1" and a "0" is a "0". There are no "half states" meaning a "1" is not a "1.2" or a "0.9" and a "0" is not a "0.1". The Bit is either "on/1" or "off/0". That is the only possibility. There are only two possible choices here and if this information is transferred bit by bit identically from a-b then there should be no resulting distortion or difference to this signal. So, there should be no resulting difference to this data provided that this data is accurately transferred through whatever means it goes through, provided that it is uncompressed (meaning mathematically unmodified in any way) and clocked (you know what I mean...) the same.

So... How does the media, cable, transport, or a different transfer technology (optical/coax), make a difference to the data being transferred, provided that each stage does it's job with bit accuracy? A "1" at the source end (a) should still be a "1" when it get's to it's final destination before it get's converted to an analog signal. Using the idea of simple math there should be no bering on the data regardless of the method used as a carrier.

Sorry to ramble on here, I just always wondered that and had to get it off my chest. I don't want to start a great debate here but... I just don't understand as the signal is transferred through it's stages (digitally) how it would make any difference at all to the overall sound as long at the signal is bit for bit identical from beginning (media) to end (DAC)? Please explain! Am I completely off my rocker here? Am I the only one that has thought about this and thinks this way?

Take for example data on a computer. Whatever the data file may be, no matter what it is, where it's stored, copied to, transferred to, should be identical bit for bit. If it wasn't, then audio would be the least of our worries! Can you imagine if any digital data file gets altered somehow between source and destination and what effects that may have on anything? Internet?

A digital copy does not degrade (provided of course the stream is not corupted or altered by error or compression technics) no matter how many times it is duplicated or whatever "generation" it is. Like I said. A "1" is a "1" and a "0" is a "0".

Ok, it's late and my eyes and head hurt. I am going to stop typing now. Thank you for bearing with me.

Chris

steveinaz

Jitter, IMO is probably the main varible that can affect how digital media sounds. Jitter is timing errors that smear the signal (like an out-of-focus camera) causing the glare in midrange/treble and affecting the soundstage that many complain about with digital music.

Better equipment usually implements better jitter reduction, and many things can cause it. Cable topology, power supply quality, digital output circuit quality, the list goes on and on. Get your read on....

For generations who grew-up on CD's, it's not as apparent to those of us you started with vinyl/analog music. Digital glare and flat soundstage was very apparent in some recordings to the latter group who had alot of exposure to vinyl.

The system has to be relatively high resolution for most to hear differences in transports, cables, and dac's. There is still a lot about digital sound reproduction that engineers don't understand...

HBombToo

Agreed Steve but IMHO there is far more to the story.

1. What was the source and how was it recorded?
2. What type ADC was used?
3. What is the BW of all the inter connects?
4. Is there EMI impacting the digital stream?
5. Is the pressing of high quality?

the list can go on but I stopped here.

In a perfect world bits are bits; however, we don't live in a perfect world.

HBomb

steveinaz

Yep, there are alot of varibles. In audio 2+2 doesn't always = 4.

heiney9

Well it just happens I came across this post in the DIY forum the other day. I'm not endorsing the views of this post, nor am I denying the views. Just a very interesting read to make one think. These individuals aren't what I'd call experts but many of them have practical experience and some may just be blowing smoke. Personally I can see where post #1 is coming from right up until he states Jitter is necessary. One thing is for sure digital codeing and decoding is a very complicated process and has almost an infinite amount of variables to contend with. I'll be posting more if I find more interesting stuff. Don't take it as the absolute truth just a spring board for discussion.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=66839&highlight=

H9

heiney9

Another linky-technical but a great read even if you don't understand the minutiae.

http://www.diyaudio.com/forums/showthread.php?s=&threadid=56954&highlight=

H9

heiney9

http://www.diyaudio.com/forums/showthread.php?s=&threadid=23566&highlight=

Some good food for thought.....some good stimulation for the brain. Jitter free stimulation .

H9

mldennison

dont really have the experience to comment on the whole idea here but i did want to make one comment.

i see alot of people making the comparison between audio data and computer file transfer. this is not really a good comparison. computer data files, when transferred have redundant checking systems in place since the data MUST be correct. the more complex the transfer, the more systems in place to make sure the data is corrent. with audio it is different. for regular CD audio, you must deliver 44,100 samples every second, any data that is late is useless. there is no time to correct or resend the data. there is some error checking involved, but the "correct" data is not always provided. so what you are getting at the preamp or at the DAC is not always going to be exactly what was on the disc.

my guess is that this is why things like jitter are still a big problem.

h9 - that is an interesting concept, i find it hard to believe that different dithering algorithms would be audible but it definately is interesting.

heiney9

mldennison wrote:

dont really have the experience to comment on the whole idea here but i did want to make one comment.

i see alot of people making the comparison between audio data and computer file transfer. this is not really a good comparison. computer data files, when transferred have redundant checking systems in place since the data MUST be correct. the more complex the transfer, the more systems in place to make sure the data is corrent. with audio it is different. for regular CD audio, you must deliver 44,100 samples every second, any data that is late is useless. there is no time to correct or resend the data. there is some error checking involved, but the "correct" data is not always provided. so what you are getting at the preamp or at the DAC is not always going to be exactly what was on the disc.

my guess is that this is why things like jitter are still a big problem.

h9 - that is an interesting concept, i find it hard to believe that different dithering algorithms would be audible but it definately is interesting.

I think the big misconception between computer data and audio data is that with computers there is no analog element. The problem, knock, concern with digital audio is the mis-representation of the original analog wave on both ends A->D and D->A. As you know comparing apples to oranges. The conversion to and from analog to digital and back again is where the contraversy is.

H9

mldennison

yeah, i totally agree, that is where the big sonic impact is. quanitzation error is definately the biggest problem with digital audio reproduction. i think there is still alot of work to be done in this area. that is what really intrigues me about SACD, it is not the same old PCM, it is a totally different way of looking at a->d and d->a, i am hopeful more companies look at this instead of just settling for the same old PCM techniques.

steveinaz

mldennison wrote:

i see alot of people making the comparison between audio data and computer file transfer. this is not really a good comparison.

Ed Zachry.

steveinaz

mldennison wrote:

i think there is still alot of work to be done in this area. that is what really intrigues me about SACD, it is not the same old PCM, it is a totally different way of looking at a->d and d->a, i am hopeful more companies look at this instead of just settling for the same old PCM techniques.

My thought is, we need to get PCM right before we move onto bigger and better things; otherwise the same old weak-link will exist.

Redbook/PCM can sound very good. It's getting people in the recording studio who know what the hell they're doing, and cost cutting that ruins the format. IMHO.

mldennison

steveinaz wrote:

Ed Zachry.

?? - Pardon my Ignorance.

steveinaz

Ed Zachry = Ex-actly. I agree.

heiney9

I've just begun to learn about the Delta Sigma modulators( 1 bit converters) and PDM (Pulse Density Modulation) conversion. Regular 16 bit uses PCM (Pulse Code Modulation).

Just for informational purposes here are a couple links explaining 1 bit conversion which is used for SACD.

http://www.howstuffworks.com/question620.htm

http://en.wikipedia.org/wiki/Pulse-density_modulation

FWIW, I agree with Steve we need to get PCM right, or atleast affordable for the mainstream before we move on to something else. I am also completely unconvinced that 1 bit Delta Sigma conversion (PDM) is any better than correctly done 16 bit PCM conversion.

This whole batch of info in this thread is why a $100 cdp and $1000 cdp sound different. It's all in the selection and matching of components and design. The problem is the general public doesn't understand, doesn't care to understand how and why there are differences. So they would rather pay $100 for a cdp which is very mediocre compared to what is actually possible than pay $1000 and perhaps not care or be able to tell the difference.

I have had several cdp's come thru my system and a few DAC's and they all sound markedly different. The better sounding ones were always those that took on the better design philosopy and used precision grade expensive parts.

You don't have to buy into spending big bucks on the source for your system but just understand that there are differences and most of the time they aren't subtle. It depends on what you are willing to settle for and far you want to go in this hobby.

Steve, I agree you are one picky SOB, but then so am I . I don't have an ultra high end system, but what I do have sounds damn fine and I've taken it as far as it can go right now. Using a high quality sepreate DAC has really opened everything up, beyond even what I thought was capable when I started to research units a few years ago.

Anyway thats my $3 worth

H9

heiney9

Just picking thru my sources and posting as I run across them. Here's another discussion article from a post I made a while back. I just snipped the relevant info from the article for informational purposes.

http://www.polkaudio.com/forums/showpost.php?p=409423&postcount=9

This is a very technical process and is not for the faint of heart. I understand the basics of conversion and the inherent issues as they relate to each other in the whole picture. But that is just scratching the surface of the conversion process and to go much beyond this you'd need an audio engineering degree to sort things out. I just know when things sound good or bad an am just now understanding some of the reasons why and how it relates to the conversion process.

H9

heiney9

nm........

danger boy

hasen't this been discussed at length before? someone always brings this topic up.. who really cares about 1's and 0's anyway? I don't. that's it. i'm outta here

madmax

Just play vinyl and forget about these two digits. Life is too short to worry about 1's and 0's... I'd even go as far as to suggest we start at 2 and go up from there.
madmax

kingsqueak

danger boy wrote:

hasen't this been discussed at length before? someone always brings this topic up.. who really cares about 1's and 0's anyway? I don't. that's it. i'm outta here

I care, simply because I've yet to see anyone explain what the original poster is asking. I'm with the first guy, I don't get it and until I've seen someone show me that somehow in a bit stream there could be a difference between a 10110 over one cable vs the same 10110 over another...I will remain a skeptic.

Yes, I understand that if you have a bad transport it might not read data or drop bits that are there, but if you have a K-Mart transport and one of the $9k CD deck transports, sending the same bits, bits is bits. If you wire up that $9k reference CD reader with a $14 RCA optical cable, it will sound the same as with a $120 Monster optical cable. It's a beam of light....and bits is bits.

If someone can show me something that would explain otherwise I might change my mind, but nothing I've seen explains how 10110 from one cable is any different than 10110 from another, or the same for 10110 read by one transport or another.

10110 will sound like 10110 no matter how it is delivered, as long as it gets there.

My favorite dunsel quote I've seen yet was I think from Stereo Review, it was a review of some insane $9k CD player, Arcam maybe...the reviewer actually said that it sounded better to him while suspended in a basket, compared to it sitting on a shelf. These are the only people I've seen insisting that there are audible differences in optical/digital cables.

I just want to add, since people never read carefully when this topic comes up.

If you are dealing with analog signals or changing the actual signal path, all bets are off, this isn't what I'm talking about.

heiney9

kingsqueak wrote:

I care, simply because I've yet to see anyone explain what the original poster is asking. I'm with the first guy, I don't get it and until I've seen someone show me that somehow in a bit stream there could be a difference between a 10110 over one cable vs the same 10110 over another...I will remain a skeptic.

Yes, I understand that if you have a bad transport it might not read data or drop bits that are there, but if you have a K-Mart transport and one of the $9k CD deck transports, sending the same bits, bits is bits. If you wire up that $9k reference CD reader with a $14 RCA optical cable, it will sound the same as with a $120 Monster optical cable. It's a beam of light....and bits is bits.

If someone can show me something that would explain otherwise I might change my mind, but nothing I've seen explains how 10110 from one cable is any different than 10110 from another, or the same for 10110 read by one transport or another.

10110 will sound like 10110 no matter how it is delivered, as long as it gets there.

My favorite dunsel quote I've seen yet was I think from Stereo Review, it was a review of some insane $9k CD player, Arcam maybe...the reviewer actually said that it sounded better to him while suspended in a basket, compared to it sitting on a shelf. These are the only people I've seen insisting that there are audible differences in optical/digital cables.

I just want to add, since people never read carefully when this topic comes up.

If you are dealing with analog signals or changing the actual signal path, all bets are off, this isn't what I'm talking about.

Well the cable end of it is easy (atleast on the analog end) cables do sound different with analog signals so it's reasonable to question whether the same can be said of digital bits. No, I don't have an answer because I haven't been able to test the cable that just transmits the digital bits (by itself). Also the nature at which digital info is sent between types of connections is completely different and it seems reasonable again to question if there can be sublte differences between 75 ohm coax, optical toslink, AES/EBU connections. Again there's lots of info out there to Google. It may just come down to the design of said output and how it's incorporated in the DAC. Again I don't have an answer either way. I've choosen 75 ohm coax for the main system and Toslink for my 2nd system (mainly becasue the sound card from the computer only allows Toslink as its Digital output)

As far as the physical properties of the transport material's and isolation process, that's a debate for the ages. I have read lots of posts from Steveinaz and he's a pretty articulate guy who seems to be very knowledgeable and expericened in many things audio, so I believe him when he states he feels he can hear a difference between his transports.

Personally with my gear I can't say I hear any difference between a very good transport and very poor transport. I've never tried a side-by-side comparo of Toslink vs. Coax

I will say that all the info discussed in previous posts has far a greater effect on the end result (the sound) than the transport and the cable used for the digital side of the transfer. All the improvements can and should be made on the analog side of the conversion. That's not getting into what engineers can do when mastering music. We all know that Jitter artifacts can be added during the mastering process and if that happens there is no possible way to correct it.

H9

steveinaz

kingsqueak wrote:

Yes, I understand that if you have a bad transport it might not read data or drop bits that are there, but if you have a K-Mart transport and one of the $9k CD deck transports, sending the same bits, bits is bits. If you wire up that $9k reference CD reader with a $14 RCA optical cable, it will sound the same as with a $120 Monster optical cable. It's a beam of light....and bits is bits.

If someone can show me something that would explain otherwise I might change my mind, but nothing I've seen explains how 10110 from one cable is any different than 10110 from another, or the same for 10110 read by one transport or another.

The point is, it's not about "dropping bits" or error correction at all. For one, error correction is either grossly audible (a drop-out in the sound or skip) or it's completely inaudible (corrected by the circuit).

What can potentially make different transports and cables sound different is jitter, or timing errors. Think of it like "wow & flutter" of the old days of cassette decks (and turntables).

It's an expensive proposition to reduce jitter, as it requires top notch componentry, good design layout, and better power supplies, therefore, your better transports are more expensive. Do you need to spend 10k on a transport?...hell no, most of Musical Fidelity's CDP's are very low in jitter, as are models from Cambridge Audio (especially the older D500SE). Both of these can be found under a grand.

Nothing will ever convince you---if your interested enough, you'll experiment and see & hear the differences; provided the system has the resolution to display the anamolies.

Do a search for jitter and read up on its effects on digital audio transmission, it'll make sense once you understand what jitter does. I can tell you from first hand experience that I readily heard a difference between a Denon DVD-2910 and my CEC CD-3300 when A/B'd as transports thru an external DAC. And I assure you, I'm no "golden ear."

LuSh

Since I yammer on about how important the source is, and have stuck with my statements dating back a couple years ago long before most on this forum started using DAC's, I'll give it a shot and try to explain to you how and why digital screws up simple 1's and 0's.

First you must understand that the 1's and 0's are represented in digital streams as Full voltage or No voltage. Ideally then the digital signal would be a perfect square wave. In the real world however, the perfect "square wave" is actually a crooked slopped wave. This is caused both because of a circuits inability to rise infinitely fast enough to full voltage and by the inability of the cable to carry an infinite frequency range.

There is a short but finite time between the initial rise of the signal voltage and its arrival at full voltage. The receiver circuit at the front of the converter must interpret the signal and decide when it will consider that a "1" (full voltage) has arrived. This present's a potential for timing errors. A simple 1 might be recognized as a 0.

Worse yet, because neither the circuits nor the connectors nor the wire are perfectly matched for impedance, some energy arriving at the converter will actually bounce back toward the transport. This signal is then again bounced back to the DAC as an echo. Which is then sent again as a signal to the DAC. Some voltage signals can have as many as 3-5 reflections before they decay!

For a more in-depth understanding pickup UHF NO. 74 off their website. It goes into greater detail and gives examples of how a 1.5 meter digital connection will sound better then a 1 meter. Crazy?! I know!

Different cables introduce different impedance curves which can affect reflections. Remember, I've never pretended to be a techo wizzard but I trust my ears and it's nice to understand how and why you're hearing a difference after of course you've in fact heard a difference.

steveinaz

LuSh wrote:

Different cables introduce different impedance curves which can affect reflections. Remember, I've never pretended to be a techo wizzard but I trust my ears and it's nice to understand how and why you're hearing a difference after of course you've in fact heard a difference.

indeed, known as "cable induced jitter." One argument in the "pro" column for Toslink.

kingsqueak

Ah o.k. fair enough, finally some info that seems rational. I've never had an explanation of physically how the cable could have an effect.

Now, for maybe some more clarification if you gents know.

My understanding is, if a bit is dropped, you'll have a 'skip' in the audio...right? Or is there some sort of delay/interpolation done at the receiving end. If the latter is true then I can understand even more how a cable could effect the signal. My impression had been either the signal is there or it isn't.

heiney9

Good point Lush, one I had forgotten to mention. The digital transfer is voltage driven, hence it has it's limitaions and tendancies regardless of what the voltage is carrying (analog info or digital info).

H9

heiney9

kingsqueak wrote:

My understanding is, if a bit is dropped, you'll have a 'skip' in the audio...right? Or is there some sort of delay/interpolation done at the receiving end. If the latter is true then I can understand even more how a cable could effect the signal. My impression had been either the signal is there or it isn't.

Now you are getting into error correction. Essentially it's replacing missing or damaged bits of info. CIRC is the standard error correction scheme/circuit. It's either on or off. It works basically by interpolating previous info and filling in the damaged or missing bits. If too many bits are missing then you hear an audible click or in worst case senarios the disc will simply stop playing or fail to spin up at the damaged spot. There is no audible/measureable degradation as the error correction works until it reaches it's limit and starts clicking. Most times EC is working and you don't even know it.

H9

heiney9

CIRC explained a bit better (pun intended)

http://en.wikipedia.org/wiki/Cross-Interleaved_Reed-Solomon_Coding

H9

kingsqueak

O.k.

So, in filling in the missing bits...the hole is filled with what, a duplicate of preceeding bits already in the buffer?

Square wave vs sloped was the start, that was a great explanation, I was thinking too linearly.

If missing bits are interpolated, then it's totally clear as that would certainly reduce the quality of the end product.

The answer is YES...

Your Wiki link crossed over this when I clicked send, thanks.

I'm a believer now and I can rest easy :-) Thank you.

Chrisssssssss

ok,

I can see the point of reflections and impedance induced echoes. That actually seems to make some sence! I guess that there are all sorts of real world factors that can attribute to signal differences. I never looked at it that way. Hmmm... That helps alot, well at least it changes my thought process on the whole theory.

Thanks Lush.

Also thanks to Kingsqueak for understanding my question and for helping me convey my point. I also will be sleeping better with this knowledge.