Its all about the error correction. Theres always a percentage of pit misreads off any optical disc, but the data can come off perfect every time (or not at all). When was the last time you pulled a word document off a CD and a few words changed? If you convert your music to wave files and store them on CD, they will come back bit perfect (even spinning 48x or vibrating the computer). The error correction on audio CDs isnt as good as data CDs but as I understand it the only time bits are "lost" is when there are C2 errors. Does anyone know how often these occur in a average CD player?
I avoid Word as much as possible. It is a terrible piece of software.
But i get what you mean. You do have to get it off the audio CD to start with thou.
dave
Chris deserves a big thank you for taking the time and trouble to help explain the actual processes at work here... and to keep reminding folk that we are dealing with what is an analogue signal off the disc. Discs don't have 1' and 0's, pits and lands are not directly representing 1's and 0's. The change from 1 to 0 and vice versa occurs whenever there is a transistion from pit to land and vice versa. This is also intimately related to the wavelength of the laser "light"... the difference in "height" between a "pit" and a "land" is the difference between "light is reflected back" or "no light is reflected back". That microscopic difference in "distance" the light travels allows for "cancellation" or not of the reflected signal.
CDR's... any recordable disc is fundamentally "different" here as there are no pit's and lands, just reflective material, or non reflective (where the state has been altered by burning). The "edge" definition of these transistions is far less well defined than a pressed disc.
As an aside to this, and I know Chris will have found this many times when looking at RF patterns, that if you just lightly "hold" the leads to a pickup the RF deteriorates so badly with noise/jitter... you name it, yet the player goes on regardless.
I think I read in there "you hear what you want to hear" or something along those lines.
I think that sums up most "tweaky" arguments here, not just this thread but cables, burn in of IC's etc etc.
Good on you Chris
CDR's... any recordable disc is fundamentally "different" here as there are no pit's and lands, just reflective material, or non reflective (where the state has been altered by burning). The "edge" definition of these transistions is far less well defined than a pressed disc.
As an aside to this, and I know Chris will have found this many times when looking at RF patterns, that if you just lightly "hold" the leads to a pickup the RF deteriorates so badly with noise/jitter... you name it, yet the player goes on regardless.
I think I read in there "you hear what you want to hear" or something along those lines.
I think that sums up most "tweaky" arguments here, not just this thread but cables, burn in of IC's etc etc.
Good on you Chris
Sorry to bring you bad news Bas, but you're just plain deaf, like the rest of us whom can't hear the difference
Which would lead to the conclusion, that SY is stone deaf!
Magura
The good news is there's often a BIOS setting that allows you to switch that off.
IIRC the magic word is "spread spectrum" (for my Gigabyte motherboard anyway). Turn that off to get a constant clock speed.
I found out about that in context of the effect it might have on system stability when overclocking.
It didn't occur to me till now that it could also jitter the audio. Duh!
Thanks for the heads-up.
Posted by SY -
I still have one of those in my closet. Oddly enough, even though it is ancient and crude, it had a wonderful sound to it. For the longest time, I was not satisfied by any player I attempted to replace it with. Guess it was well broken in.
I keep meaning to listen to it now, after all these years to see how bad it sucks, now that it is no longer "broken in".
I still have one of those in my closet. Oddly enough, even though it is ancient and crude, it had a wonderful sound to it. For the longest time, I was not satisfied by any player I attempted to replace it with. Guess it was well broken in.
I keep meaning to listen to it now, after all these years to see how bad it sucks, now that it is no longer "broken in".
Religion and faith are always controversial.
Sy,
I think you pegged it. Religion and faith... I used to be a believer. Then I woke up. I still use the green pen as I still have a green pen. When it dries out, will I replace it, or will I buy some better parts for my next project. Hmmm.... One is a likely subjective gain, while the other is measurable. I will go measurable. In my experience all of these little tweaks make some minute change in the sound that as someone else said, were likely no greater than the random changes in read.
It didn't occur to me till now that it could also jitter the audio. Duh!
Thanks for the heads-up.
Your sound card has its own local clock. IMO the CPU clock dither should only improve things, after all it is there to reduce EMI.
I would think the data would make it intact. SPDIF has no error detection, but AFAIK, it's problems are mostly jitter related, which wouldn't show up in your difference measurement. (At least I don't see how it would.)
Eric
No, I'm talking about the data from the CD making it without errors to the second PC via SPDIF. I suppose you could use a CD player with SPDIF out then rip the CD and compare the two files. I can't count how many threads have passed full of wild speculation where this simple test could put this to rest and everyone could go back to talking about jitter. Even stranger I'll bet you could find CD's that are not bit perfect in this context.
Anytime.
re:C2 errors
So there are very few non correctable errors being read off music CDs. So what are these devices fixing? How do you fix errors that dont exist? It would seem that modifying anything in a CD player before the DAC (other than the clock) including the actual CD will not improve the sound. If there is a difference you are probably increaseing the C2 error count and making it sound worse.
many posts since!
Hi cbdb,
The problem is that this process is entirely hidden from your view. There is no way to detect any problems until everything goes right to pot! There are generally a lot of C2 errors, but the better DSP chips these days hide them very well. All you can do is hope that the attempts at correction are close to the actual lost information. They generally are.
Hi Scott,
Thank you.
All data transmission is encapsulated and has error correction built in, so the SPDIF signal is more reliable. To the best of my knowledge though, the data gets spit out and is not repeated. The same situation as in reading the data the first time from the CD. The medium is far more robust though! Same for the optical link, but this has issues with the quality of the connectors and jacks. Other mating connector standards have been tried, but Toslink is too entrenched to dislodge. By far the best connection would be the AES-EBU link using an XLR. The structure of the signal is better as well.
"Wild speculation" is all most people have. The technical details have not been stated very often, and it's not to the manufacturers best interests to clear any of this up. The same holds true for the horrible manufacturing quality of the software. Now there is a crime!
Sorry Scott, I know you probably understand this better than I might. Just agreeing with explanation for some who haven't studied this stuff.
Hi Fran,
Ahhh, no. The Earth's magnetic field is far stronger than anything you could possibly imagine with a CD. It would easily swamp out effects involving CDs. It's entirely possible to take an argument concerning improbable possibilities to ridiculous lengths, and it would never ever finish. With those limits, you could drag up counter arguments forever, complete waste of time.
It's romantic to think a person can detect minuscule differences between various effects, and that's how the snake oil people nail victims. There are more real effects in electronics than the average consumer could imagine, but the stuff out on the 'net is mostly false or wishful thinking. For instance, just ask Scott Wurcer how many things vary inside transistors and integrated circuits. The odd cosmic ray that will change a data bit in memory. I tell you, this stuff that is real will keep you up at night. Instead, debate over imaginary issues will continue to occupy people's fancy.
Like many other things Fran, the average person does not want to know the truth. It doesn't even involve aliens!
I think the major problem is that most people don't understand how things work well enough to put things into perspective. An effect may or may not occur, but if other things affect the system more, those little issues are totally swamped out. Unnoticeable due to the "noise" of other issues that matter a great deal more.
-Chris
Hi cbdb,
The problem is that this process is entirely hidden from your view. There is no way to detect any problems until everything goes right to pot! There are generally a lot of C2 errors, but the better DSP chips these days hide them very well. All you can do is hope that the attempts at correction are close to the actual lost information. They generally are.
Hi Scott,
Thank you.
All data transmission is encapsulated and has error correction built in, so the SPDIF signal is more reliable. To the best of my knowledge though, the data gets spit out and is not repeated. The same situation as in reading the data the first time from the CD. The medium is far more robust though! Same for the optical link, but this has issues with the quality of the connectors and jacks. Other mating connector standards have been tried, but Toslink is too entrenched to dislodge. By far the best connection would be the AES-EBU link using an XLR. The structure of the signal is better as well.
"Wild speculation" is all most people have. The technical details have not been stated very often, and it's not to the manufacturers best interests to clear any of this up. The same holds true for the horrible manufacturing quality of the software. Now there is a crime!
I have to say that I have little motivation for that test. The encoding is different between the data as it is used inside the CD player, and what is spit out to the external device. For one thing, the clock signals are combined with the actual data. The end device must reconstruct the various clocks (got one, got them all), then decode the data using the reconstructed clock signal and sync to the packet beginning. Within the CD player, the DAC and digital filters are presented with a clean rate clock, clean word clock and L-R clock (for a single DAC chip). So you can see that internal filters and DACs have quite an advantage over anything that is external. To accomplish this test, one would probably be further ahead to allow the audio signal to be reconstructed and compared in the analog domain. I'm more than sure you could figure out how to do that! Once you have things synced up, an instrumentation op amp with the audio to each input should do the trick. Just monitor the output (probably extremely noisy due to pulse length differences). I am suggesting that the audio signal be compared before the reconstruction filter, using the same DACs and filters between them. Run each with the same clock as well, which would eliminate problems due to clock regeneration issues if they exist.
Absolutely! Most subsystems in a computer have their own master clock and use FIFO buffers to exchange data. Commonly referred to as memory-on-board. Video cards are another example of this.
Sorry Scott, I know you probably understand this better than I might. Just agreeing with explanation for some who haven't studied this stuff.
Hi Fran,
You can often hear differences between a CD player in adjustment and not. Over the years, many customers were wondering what I did to improve the sound of their machine. Unexpected comments, I didn't think they would notice, but some do!
Ahhh, no. The Earth's magnetic field is far stronger than anything you could possibly imagine with a CD. It would easily swamp out effects involving CDs. It's entirely possible to take an argument concerning improbable possibilities to ridiculous lengths, and it would never ever finish. With those limits, you could drag up counter arguments forever, complete waste of time.
It's romantic to think a person can detect minuscule differences between various effects, and that's how the snake oil people nail victims. There are more real effects in electronics than the average consumer could imagine, but the stuff out on the 'net is mostly false or wishful thinking. For instance, just ask Scott Wurcer how many things vary inside transistors and integrated circuits. The odd cosmic ray that will change a data bit in memory. I tell you, this stuff that is real will keep you up at night. Instead, debate over imaginary issues will continue to occupy people's fancy.
Like many other things Fran, the average person does not want to know the truth. It doesn't even involve aliens!
I think the major problem is that most people don't understand how things work well enough to put things into perspective. An effect may or may not occur, but if other things affect the system more, those little issues are totally swamped out. Unnoticeable due to the "noise" of other issues that matter a great deal more.
-Chris
Hi Dave, Mooley SY,
Thank you sirs.
Hi Dave,
So, we have a wavelength (colour) issue to look at. Then, there is the fact that the head is peering at a tiny spot on the CD with a narrow field of view. Really narrow since it is focused at one point! Any light with a different angle of incidence will simply not make it through the optics in order to strike the pick up diodes. Refracted light will not exit the CD surface perpendicularly in order to enter the optics directly. A surface defect might refract the light for an instant, but then the information couldn't be read anyway and the servos would have entered a "mute" mode where the lens movement was halted. This is to prevent wild movements in the lens assembly (a good thing).
Hi SY,
It's amazing to see how much ego can be riding on an idea where a person has little to no training to back it up. When I'm wrong, I learn (and thank the person correcting the idea). There is no shame in learning.
Once we get to non-mechanical media for information storage, there will be an entirely new crop of concerns to keep everyone busy and a little freaked out.
-Chris
Thank you sirs.
Hi Dave,
Yes, great points. Audio CDs and data CDs use completely different error correction methods. This is probably causing a great deal of confusion as the same media is used for both. They should have used different colours or something.
This should be easy to understand really. Firstly, the wavelength used is in fact "blind" to other wavelengths. However, I'm sure you could set up a situation where enough light enters the optics in the head and causes trouble. It would be an extreme situation though.
So, we have a wavelength (colour) issue to look at. Then, there is the fact that the head is peering at a tiny spot on the CD with a narrow field of view. Really narrow since it is focused at one point! Any light with a different angle of incidence will simply not make it through the optics in order to strike the pick up diodes. Refracted light will not exit the CD surface perpendicularly in order to enter the optics directly. A surface defect might refract the light for an instant, but then the information couldn't be read anyway and the servos would have entered a "mute" mode where the lens movement was halted. This is to prevent wild movements in the lens assembly (a good thing).
That would be my major concern also. The CD is hopefully sealed and removing some of the edge stands a high chance of allowing atmospheric gases to make contact with the aluminum layer. The other big concern is that the CD is a rotating mass, it should be dynamically balanced. What the answer to that is, I don't know. The CD is defective from my point of view.
Hi SY,
So true!!!
It's amazing to see how much ego can be riding on an idea where a person has little to no training to back it up. When I'm wrong, I learn (and thank the person correcting the idea). There is no shame in learning.
Once we get to non-mechanical media for information storage, there will be an entirely new crop of concerns to keep everyone busy and a little freaked out.
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.