The Official (and only) Optical vs. Coaxial Thread

[I 8 Piglet]

I have a question, is digital coax diffrent form the regular RCA coax cable, or will a simple video cable give me digital audio when run from the digital out on a DVD player to the digital in on a reciver?

[1138]

Originally posted by I 8 Piglet
I have a question, is digital coax diffrent form the regular RCA coax cable, or will a simple video cable give me digital audio when run from the digital out on a DVD player to the digital in on a reciver?

Yes, it will. I've used a thin little RCA mono cable as a digital cable just fine.

[Reznor007]

Originally posted by I 8 Piglet
I have a question, is digital coax diffrent form the regular RCA coax cable, or will a simple video cable give me digital audio when run from the digital out on a DVD player to the digital in on a reciver?

No, RCA wire is pretty much all the same. It will work fine.

[1138]

Originally posted by Tyler Durden
I will not apologise as I have nothing to apologise for but IMO 1138 is talking nonsense - as I'm certain he thinks I am too!

No, I don't, and I wouldn't say so. Much of what you have said is in fact true. It just doesn't apply to the current discussion.

Originally posted by Tyler Durden
As for being 'informational' - can I refer you to your own links at the top of this thread? I think you'll find more relevant information there to the topic at hand than anywhere else...

But clearly you didn't read the previous monstrosity very well. Could you point out precisely which part of the thread was not about digital?? I have just re-read it and found only a very small amount of information that related to analogue interconnects. A VAST chunk of it deals quite obviously and specifically with the digital medium. Can't be bothered to cut and paste which bits as there's far too much info and it will clutter up this thread still further...

The problem is that, in both the huge post you refer to and what you've said here, all the talk is about analog to digital conversion and digital to analog conversion. And again, those don't apply here.

The huge post in the other thread spends the first half talking about cable qualities. It even says that impedance mismatching is a digital problem. It is in fact just a general transmission line problem. But it's a big problem for analog because in analog form, the exact level of the voltage is used to create the sound/video/information. But when we enocde that infromation in digital, there's only two voltage levels we care about. Those two voltage levels correspond to 1 and 0. As long as we're able to tell the difference between the two, we can piece back together the information. If we can't tell the two apart, then we get nothing (we're either seeing all 1's or all 0's).

Then there's the second part of the huge post. It goes into great detail about the problems with converting analog audio information into digital form. Great, it goes into very in depth and thoorough information and looks good to me. The problem is that all that infomation is not germane to this discussion. All of that analog to digtal conversion is done by the time we're popping in a DVD/CD into our players and reading information off of it to send down our digital connections. Let me repeat that because it apparently needs repeating: all of that talk is moot by the time we get to what we're talking about. There is no analog to digtal (or vise versa) conversion going on between reading the information with a laser off of the disc and your receiver getting it over the digital connection.

Again, I'll point out here that most of what you mention in this thread Tyler also has nothing to do with this discussion. You're either talking about A/D or D/A conversions or some disc mastering problem. We're talking digital connections and you're bringing up analog aspects. They don't come into play here. Sorry for repeating that so often in this post, but it doesn't seem to be getting through.

Anyway, give me some time and I'll attempt to comeup with some basic explanations of A/D, transmission line qualities, digital signals, etc.

[Wilson]

I don't believe I just read all of those threads...

I'll start by saying I don't know anything about the subject but I must say that overall 1138 has made the best case. I have an understanding of the 0 and 1's of the digital world as they apply to computers but don't fully understand the process of converting a digital stream into an analog signal.

I can totally understand the idea that the cable is transmitting 1 and 0's and just as long as they are in the same order at the receiving end and they all show up their should be no difference between cables. Other people here are saying that this just isn't the case but none have dumbed it down enough for me so I understand what is going on. If the cables are in fact causing some sort of signal loss or degradation how can someone say it sounds brighter or bassier etc. Shouldn't the dropped bits be random and just generally cause the sound to be crappy or non-existent? What is happening that makes it sound brighter etc. 1138 is saying 10001001 is coming out as 10001001 at the other end no matter what cable you use but no one who supports the other side of the argument has said what it's turning into that makes it sound different.

Without a doubt I'm confused.

[Wirehed]

If it was nothing but a perfect order of 0 and 1 going accross the cables we wouldn't need error correction.

Error corection can only do so much.

Both DD and DTS use not so perfect lossy compression (and so does 1138s Satelite TV set up).

The better the cable, the better the 0s and 1s make it to the error correction process and the less work it has to do.

If the cable was perfect, we wouldn't need error correction.


that should some up my side of the argument.

[Reznor007]

This just popped into mind, and I'm not sure if it's right or not, but don't DD/DTS use a checksum for every packet?

[Mr. Furious]

Seamonkey -

In the computer world there are numerous ways to handle bad packets - discard and resend is fairly common. Sometimes you can also in interpolate the data (for example RAID 5) based on parity bits. I imagine that it would be most efficient for digital transfer between our devices if they were to buffer the data stream and resend when there are errors. Keep in mind we are talking very short runs of wire with minimal chance of picking up a ton of interference. I don't find it very likely that the data stream is constantly being corrupted to the point where we have a lot packets being lost.

[Wirehed]

I've never heard of a receiver that buffers data. And in audio/video I would doubt there is enough time to resend packets.

[Mr. Furious]

Seamonkey -

You've heard of skip protection right (on portable CD players, car decks, etc)? It is simply buffering the data by reading ahead. When the device can't read or gets bad data, it resends the data. All that is transparent to the user because the data has been buffered, and there isn't any interruption in play. The same can be done with video, or any combination of the two.

My guess is that the receiver works in combination with the DVD player in order to achieve this - the big trick would be timing.

[Wirehed]

Oh, I think I mistook what you siad. You are just saying it's possible, I thought you were saying current systems ARE doing this.

Interestingly enough, I always turn the skip protection off on my portible CD player because it degrades the sound quality.

[Enzian]

Originally posted by 1138
The problem is that, in both the huge post you refer to and what you've said here, all the talk is about analog to digital conversion and digital to analog conversion. And again, those don't apply here.

We have a winnah!!! DACs are not the issue. A DAC with a higher resolution will undoubtedly sound a million times better than a cheap DAC with sub-phone quality. That is what you've paid for, Tyler.

There is error-correction. Any digital signal that is Tx'd and Rx'd has some level of error correction - I'm 99% sure. Do a search for the "OSI Model"... see what you pull.

A DVD player costs more than others based on features. What features would run a player into the $1000s... I have no idea.

DACs are found in the reciever. That is where loot can be lost.

Tyler, the equipment you're using is so good (or it must be) I've never heard of it. You have to admit - unless you blow a wad on speakers, it's no use to buy a $4000US unit. "You're only as strong as your weakest unit"... so to speak.

And, while Tyler's comments weren't purely "constructive", I don't think they deserved a lecture.

[Matlock]

Hi! Not a frequent poster here I must say.

You can explain the difference heard by I don't remember who when using different players.

Many things come into play. First is the quality of the support (CD, DVD, whatever). These expensive audiophile CD are not made of gold just for the cool factor. Gold is more durable, first. It also reflects light better, therefore the laser beam is more certain if it read a hole or a bump.

Speaking of that laser, in a good player il will be more "precise" and the "reflexion receiver" (don't know how to call it) will be error-proof. That's for the player itself.

Now the connexion. I don't know much about optical cables, but in the coaxial one, digital signal is not sent as "some current means 1, no current means 0". Instead, there is a certain defined potential that means: above it 1, below it 0. Of course, we try to send potential values far away from this limit to avoid any confusion, but it is still error-prone, especially if it's not shielded properly or else.

As for what I know of optical cables, they are very fragile physically. If they are bent too much the signal can be lost, and if the internal structure is not perfect, the light beam is seriously wrecked when it cross the imperfection. There are probably also some problems due to uncertainty of the signal like in the coaxial case.

Despite all this, I must say that the laser beam of the player is in my opinion the thing that really matter. The cable is so short compared to all the path the signal went through inside the player, its effect is negligible. And for what few glitches that might appear once in a while, error correction should do the job.

Thanks for listening

[Wirehed]

DACs are important, no doubt about that at all. And Analog is a totally seperate issue and not the point of this thread.


I have plugged a multitude of DVD players into my system, currently I use two on my main system. A Marantz I paid $1400 for and a Pioneer 919 combo player.

I auditioned countless players before I settled on the Marantz. I was not interested in features at all. I can tell you with out a doubt EVERY machine I plugged into my system looked and sounded ENTIRELY different. The Marantz sounded MUCH better than any other model I looked at, that's why I bought it.

Again, I tested these units by plugging them into my own system, using the same cables.

To suggest that they all were exactly the same and my wife, family and friends that helped audition them are all completely nuts is simply asinine.

[1138]

Here's a healthy chunk of information on the digital audio conenction of HT, SPDIF:

http://www.epanorama.net/documents/audio/spdif.html

[1138]

Originally posted by Matlock
Now the connexion. I don't know much about optical cables, but in the coaxial one, digital signal is not sent as "some current means 1, no current means 0". Instead, there is a certain defined potential that means: above it 1, below it 0. Of course, we try to send potential values far away from this limit to avoid any confusion, but it is still error-prone, especially if it's not shielded properly or else.

Are you sure? Because, based on that link I provided, it sounds like logical "1" and "0" are two different voltage levels (.2V - .6V for the coax connection. On - Off for the optical connection).

And besides, even if we are having trouble telling 1 and 0 apart, then we'll have trouble figuring out any info since that's all we are reading. In other words, if all we're doing is looking for 1's and 0's, and we can't tell them apart, then we won't be seeing much.

Despite all this, I must say that the laser beam of the player is in my opinion the thing that really matter. The cable is so short compared to all the path the signal went through inside the player, its effect is negligible.

Isn't the digital cable usually longer than any path inside the player?

These expensive audiophile CD are not made of gold just for the cool factor. Gold is more durable, first. It also reflects light better, therefore the laser beam is more certain if it read a hole or a bump.

You're mixing singal types here. Gold has nothing to do with light (though I've had a good laugh when I've seen gold plated optical cables in a hi-fi store!). Gold is better for transmitting based on electric reasons. It doesn't cororde as easily, it's a better conductor, it's more maluable, etc.

And for what few glitches that might appear once in a while, error correction should do the job.

I completely agree. (Sorry if I'm rough on what you've said, especially since you don't post much) It's not that a digital signal is totally imune from degredation. It's that the point of transmitting in digital is because we can reconstruct on the receiving end exactly what we sent. That's why we've converted all of this info into a world of 0/1, black/white, off/on. We're either getting it or we're not.

As the section of that link about SPDIF expalins, when there are errors, you'll certainly notice them because there will be big effects along with any minor ones.

[Matlock]

Originally posted by 1138

Are you sure? Because, based on that link I provided, it sounds like logical "1" and "0" are two different voltage levels (.2V - .6V for the coax connection. On - Off for the optical connection).

Yeah, they try to send a precise potential just to be sure, but in the end they check if it's above or below a limit because for sure it won't be exactly .2 and .6

And besides, even if we are having trouble telling 1 and 0 apart, then we'll have trouble figuring out any info since that's all we are reading. In other words, if all we're doing is looking for 1's and 0's, and we can't tell them apart, then we won't be seeing much.

Sure, most of the time the signal is clear, but as I said many things can happen that blur the distincition. That's when errors appear

Isn't the digital cable usually longer than any path inside the player?

Well, I would think that the path inside error-correction chips, and othes,r before getting to the DAC is relatively long, but I admit I don't know the exact numbers

You're mixing singal types here. Gold has nothing to do with light (though I've had a good laugh when I've seen gold plated optical cables in a hi-fi store!). Gold is better for transmitting based on electric reasons. It doesn't cororde as easily, it's a better conductor, it's more maluable, etc.

Sorry, it's you that is mixing stuff.
The gold I was talking about is on the disc itself, therefore it does relate to light.

(Sorry if I'm rough on what you've said, especially since you don't post much)

It's alright, all done in a friendly and constructive manner!

Thanks!

(I'd never think my English writing skills would be so rusty...)

[1138]

Originally posted by Matlock

Yeah, they try to send a precise potential just to be sure, but in the end they check if it's above or below a limit because for sure it won't be exactly .2 and .6

Sounds like your describing voltage trigger levels, In that casem there would be a level for each logical state. I agree, it won't be exactly .2 or .6. But if the signal crosses, say .45 for example, then it registers as .6. .35 would trigger a .2 reading.

Just checking if it's above or below a single voltage level isn't a very good system. What if it's at that voltage level? What do you do then?

Originally posted by Matlock
Sure, most of the time the signal is clear, but as I said many things can happen that blur the distincition. That's when errors appear

But what I'm saying is that, when the cable can't deliver a signal that distinguishes from "1" or "0", then you're not going to read anything because that's all your trying to read, 1's and 0's.

Originally posted by Matlock
Well, I would think that the path inside error-correction chips, and othes,r before getting to the DAC is relatively long, but I admit I don't know the exact numbers

Either do I, but the difference in the player is that going through all of those chips and what not, the singal is rejuvinated by each chip. In fact, one way to fight singal attenuation is to run it through a powered chip because the powered chip will spit it back out on the others side at full voltage.

Originally posted by Matlock
Sorry, it's you that is mixing stuff.
The gold I was talking about is on the disc itself, therefore it does relate to light.

OK, I didn't realize you weren't talking about the players or the cables, but the actual media.

Originally posted by Matlock
It's alright, all done in a friendly and constructive manner!

Yes, I appreciate it.

[Wirehed]

Originally posted by 1138
Yes, I appreciate it.

Group hug.

[Enzian]

There's no way in hell I'm gonna argue with someone who's done a side-by-side comparison of DVD players - just not gonna do it. It only makes sense that a more expensive plkayer produces better viseo/sound - but I can't explain it. I think there's a difference in players - but I don't think it has anything to do with weather the optical or coax cables were used - I could be wrong on that, too.

I used to work as an EE... you guys are on-target, but your terminology is making me cringe. I'm doing my best to limit my anal impulses from breaking off a page-long breakdown of strandard CMOS voltage transitions.

BTW - I forget the actual voltage values, but it's something like 2v high, -1.5 low is called ECL. I garauntee those are NOT the exact voltages, but it's something like that.

Before those impulses keep me typing....


...must... stop....

[Matlock]

Originally posted by 1138

Sounds like your describing voltage trigger levels, In that casem there would be a level for each logical state. I agree, it won't be exactly .2 or .6. But if the signal crosses, say .45 for example, then it registers as .6. .35 would trigger a .2 reading.

Just checking if it's above or below a single voltage level isn't a very good system. What if it's at that voltage level? What do you do then?

You get an error (well, half-chances to be right)?

But what I'm saying is that, when the cable can't deliver a signal that distinguishes from "1" or "0", then you're not going to read anything because that's all your trying to read, 1's and 0's.

Like I just said, it is the cause of error. How could there be any errors if you were always sure of what you received?

Either do I, but the difference in the player is that going through all of those chips and what not, the singal is rejuvinated by each chip. In fact, one way to fight singal attenuation is to run it through a powered chip because the powered chip will spit it back out on the others side at full voltage.

Hmmm, good point here. Unless you have bad chips (there you have a difference between cheap/quality players). And unless your cable is broken or absolutely not shielded, nothing worse than slight attenuation can happen, therefore chips in your amp shouldn't have any problem.


Oh, and sorry for our lack of correct EE terminology. I'm just a physicist, my only Signal Processing class was, let's just say it, crap. And I won't go into what we, physicist, think of engineers (just kidding). But thanks for letting us know we're not that far from the Truth (with a capital T).

Have a nice day!

(I should be studying for my exam tomorrow instead of posting here...)

[Enzian]

X-axis = time
Y-axis = voltage

This is a representation of a signal starting low, going high, and returning low.

If any of the above needs further explaination, please stop reading.

[Matlock]

I think you forgot to attach the graph, which I'm sure is great!

Now I go get some sleep...

[Enzian]

I have the pic... I just decided to spare everyone.

Must... resist...

[Wirehed]

Enzian told me to stop reading.

[1138]

And if Enzian told you to jump off a cliff, would you?

[Wirehed]

Originally posted by 1138
And if Enzian told you to jump off a cliff, would you?

No way man, I'm not falling for that twice! I'm still mad at you for that one 1138 and as soon as I get this cast off, you better watch out!

[brian]

I'm wondering how an optical to coaxial converter would affect some people's perception of the sound...

http://minidisco.com/minispecs/maudioco2.html

[Enzian]

I think our network uses those as regens... Perhaps Cisco should invest.

Clever. I wonder what they say about the differences in signal quality. Apparently, from reading their blurb, coax has difficulties with long transmissions. Makes sense to me.

No one but Tyler has said anything I can't agree with. [checks logic of statement ] I'm almost begging Tyler to tell me more about how he thinks one is better than the other.

Most people don't know that the signal in coax is transmitted along the outer braid (ring) - not the solid center (tip). At high frequencies, the sold center becomes the "ground", cancelling noise. Just thought I'd share.

[AirSickMoth]

I was just thinking....

In the good old days of CD's the CD player would error check the data it is reading off the disc. It would check it in case there was a small scratch on the disc. If data was missing due to a scratch, the CD player could use redundancy bits to recreate the missing info. Then it would convert it to analog, and send it to your amps.

Now in the world of DVD....

The DVD player does the same thing when it is sending out info through the analog conections.

But when the DVD player is sending through the digital connections, how much error checking does it do? I imagine that it still does the checking for the errors in reading off the disk. Corrects any errors. Then it sends the correct bit stream to the reciever. The reciver would then check the bit stream it gets for transmision errors.

Now, there might be a small difference between DVD players in this error checking. But error checking is only as good as the encoded material is prepared for it. So, in the end there is probably no difference.

So, I agree completly with 1138.

What? Do, more expensive DVD players add extra bits to the bit stream? Do they replace bad sounding bits with better ones?

Well, it is imposible for the answer to be yes.

How is it possible for it to sound better? If there is digital info missing, you don't get a fuzzy sound or a quiet sound, you get no sound or small pops.

1138, I really enjoyed your modem comment. Just think of all the MB's that go through your cable modem every day.... with no errors. Well, no errors after it goes through error checking.

Just think, when you request to view a page (say 5kB big) the server splits it up into hundreds of packets. These packets get sent through hundreds of different paths. They travel across thousands of miles of Fiber optic cable. They all end up at your computer, almost always in the improper order. Your computer checks them out, fixes any errors (sometimes asks for packets to be resent), combines the packets, and displays the page.

Do people really think that if they have better equipment the page will be a little bit better? Do people think that spelling mistakes are due to crappy computers? No, stuff like that does not happen. If there was an error with the data you recieved, the page would not show, or more likely, your computer would crash.

Millions of pages viewed every day by people all over the world and there are only a hand full of fatal errors every day.

[AirSickMoth]

Originally posted by Enzian
Most people don't know that the signal in coax is transmitted along the outer braid (ring) - not the solid center (tip). At high frequencies, the sold center becomes the "ground", cancelling noise. Just thought I'd share.

Well, I knew that

But I was hoping you could help me out on a few other things. I've only taken a few Phys classes at University a while ago, so you are probably fresher with it.

What happens when you change the compacitance of the wire? I've heard that this is what Monster and other high-end cable companies do to their Speaker Wires. I've heard it creates a high frequecy wave in the wire. This wave alters the sound created by speakers, making it more pleasent for many people. So, when someone says it sounds better, for them it does. It is just not an accurate sound they like.

Can this wave interfere with a digital signal?

Now, when we are dealing with voltages in a wire, how can you interfere with it?

In a coax cable the 1's and 0's correspond to certain voltages. So to get interference that would mess this system up, the voltages would need to be messed up.

Now, to do something like this don't you need to subject the wire to a large fluctuating electromagnetic field. But where can you find a field large enough, in the 3 feet between your DVD player and reciever, to alter the voltages by more than a volt?

[1138]

One small quible Moth: I don't think there's the equivalent of packet resending in HT.

But then, the stream isn't being split up and sent out nilly willy throughout the net either. Just one, solid, direct connection.


"But where can you find a field large enough, in the 3 feet between your DVD player and reciever, to alter the voltages by more than a volt?'

In a shielded cable no less.

"At high frequencies, the sold center becomes the "ground", cancelling noise."

Not to mention the skin effects that would happen on the solid core at high frequencies.

[Enzian]

That skin effect is exactly why the signal travels along the outer braid.

And YOU: AirSickMoth!!!! You're gonna make me delve into a quick rundown of CMOS transitions. It's got pretty much the same characteristics found anywhere when dealing with signal transmission.

Fresher? Hmm... I'd been working as a test engineer in a microwave lab before I made the leap to Optical Networking. Two totally different worlds - that are worlds apart from standard analog/digital electronics. I'm about as fresh as Lane Staley. Oh, that's not cool.

...kinda funny, though.

[AirSickMoth]

Originally posted by 1138
One small quible Moth: I don't think there's the equivalent of packet resending in HT.

Well, I know that. I was using it as an extreamly complex example of error checking. Just think, All that internet tech (packets, error checking) is based on tech that is 20 years old. And it works perfectly. You think that DVD tech, less than 5 years old, would be even better.

[1138]

True. Once you learn about packet transmission and path determination that make up the net, you start to be amazed how well it does work most of the time.

[Enzian]

Packet Tx is (usually ) indicative of ATM (Asynchronous Transfer Mode) which uses cell headers to route packets on-the-fly to their destination. While ATM is widespread (I maintain an optical and an ATM network) the bulk of "The Net" is made of an optical transport network. There is no "scattering to the wind" of packets in the optical domain. Optical networks usually rely on rings. Data gets sent in both directions around the ring to their destination. Even if one span breaks, the data will get to it's destination via the other route.

Example:
Your modem (IAD or CPE) Tx your data. It gets picked up by an optical transport network which carries your data 99% of it's final journey. Once it reaches it's optical destination, it enters a (typically) small ATM network. The small ATM network takes your packets and routes them to other devces, like an inter-net-work dvice, a voice switch or a tie-line (think red phone hotline) device. It's only in the small ATM network where the packets are free to roam, so to speak.

[Chromy]

Tyler, just a few questions on one small section of one of your statements above about using a digital cable to copy PCM digital audio using a home CD recorder (that has been hit on by other posters in this thread, but I'll expand).

Just to make sure that I am clear what your "test case" would be...

Two CD players... one a player with a digital output, the other a recorder that has a digital input that records PCM audio to the disc with no conversion (some CD recorders convert away from PCM before re-encoding back to digital audio, for the purposes of this test these would not work as it throws an 'analog' stage in an otherwise all digital path, it makes a bit for bit copy impossible with any cable).

Label the CD's with the corresponding cable that made them.

You then record the same track repeatedly to multiple cd's with varying qualities of cable and then AB the CD's to see which cable resulted in the most accurate recording. The 'best' sounding recording would be made by the corresponding cable.


This is the end of your test case


Would you agree that this slight modification fo the test would yield the same result of a listening test, but take out the subjectivity that listening lends to the above test case.

Take the two CD players, identical to the setup above and make several copies of the master CD using several different cables. In this case we will record the same song multiple times to each disc from the master.

Label the CD's with the corresponding cable that made them.

Now, instead of listening to the result with our fallable ears We load the master CD into our computer and read out the entire bitstream from the master track. We then compare the bitstream that came through the cable onto the burned CD and compare each of the tracks that we made... counting the deviations from the master (errors) and average them across the multiple track recordings that we made on the CD. We do this for each cables CD. The cable that has the fewest number of average errors is the best!

You can then make this even more interesting by seeing which cable has the best noise rejection by replicating this test but introducing sources of RF noise close to the cable during the recording (trying to introduce more errors into the recording)

Are there any problems with this slight modification?

Chromy

[Chromy]

Anyone?

[Wilson]

Originally posted by Chromy
Anyone?

I suggested doing a bit by bit comparison in the other thread twice and didn't really get any response either. I can't see any reason why you would rely on human ears when this should clearly show which cable is reproducing the source with more accuracy. Maybe there is some type of flaw that we can't see in this kind of test.

[Chromy]

Exactly, Wilson. Just wanted to put up a specific proposed test case to see if I was missing anything (and to make sure that I was getting the jist of your earlier posts)

Chromy