Those eye-patterns are the representation of the data from the pits coming off the disc before the DACs or any other digital processing. That is what the 1s & 0s look like. (actually i think the diagram shows a sequence ...101010....)
dave
Hi Dave,
Couldn't resist popping in after your note.
All,
Your comments so far have been pretty close to the truth. The one error I see is that the frequencies of the RF pattern are both variable (depending on the information encoded) and really around the AM band of frequencies. When I have ever connected my frequency counter to the signal of the RF test point, I get variable readings around 450 KHz. I should allow it to run for a minute and average the reading. I am using an HP 5335A that calculates frequency via reciprocal counting, HP's method. I may have the nomenclature a little mixed up there, but that's because I don't sell the stuff, I just use it.
So, the encoded information comes off the CD with varying frequencies. That is a fact. Being an analog signal, it may suffer from level changes due to variable reflection / transmission (right through the CD and gone for good), surface defects causing focus failures (drop outs), pin holes that do the same thing, eccentricity that causes major jitter (timing errors), warping that may cause focus problems (variable signal strength) and "dished" discs, causing similar things and possible transport contact (gotta love that!). Consider also that all these effects may occur in combination. Add to that some dust, material on the disc table (causing a wobble), lens actuator movements in combination that are simply impossible to expect, mechanical misalignment (almost 100% of transports out there) and servo mis-alignment. I'm amazed that CDs play at all. They are an engineering feat that we've managed to make error tolerant (not for sound quality, but "playability"). Most transports today are complete junk, even the expensive ones.
BTW, a belt drive CD disc assembly? Come on people! It is not an turntable that runs at a constant angular velocity! The CD spins the disc to attempt a constant linear velocity. Near the center it spins shy of 500 RPM and near the outside it spins closer to 200 RPM. Just watch one. Okay, so the belt drive may work? Naw, it may function, but not very well. Remember the eccentric CDs? Yeah, those need to be sped up and slowed down on a per revolution basis. In this situation, high rotating mass and a compliant coupling are not your friends - but this works great on my Thorens turntable.
Alright, so now we have a signal coming from two pairs of photo-diodes that we are attempting to retrieve data from in the analog domain. Photo-diodes are current operated devices (little voltage variation across them). We convert this to a varying voltage and balance the two sets of diodes to obtain the focus information. It would be really great if they were matched (they are close on the better heads). By combining both sets of signals, we work with a signal with less noise. This is amplified (analog still) and filtered in an effort to reduce out of band noise, then a "slice level" is used to determine whether there is a pit or land passing by the head assembly. The information is critical on the pit edges - which leads us to the other problem with CDs. They are pressed just like records were, finer detail. So all types of pressing defect may occur (nice eh?), then they are coated with a sputtered aluminum layer. It's pretty thin, but the first CDs had a thick coating that would reflect most of that infrared beam mentioned earlier. Then, cost cutting measures reduced the thickness of the reflective layer, this allows some percentage of the beam to pass through, never to return - ever. No, it does not bounce around and come back to pollute other data. In fact, ambient light or LEDs or markers will not affect anything unless they are placed to block the beam. Of course, play with a light source of the same wavelength and you may be able to make things worse.
Who still feels good about the idea of perfect data coming back off the CD?
Recognizing that many data errors would definitely occur in reading, the original standard included an encoding method where by data was repeated in tracks close by the track being read. The Reed-Solomon (I think) code that we were forced to learn. The upshot of this is that small, localized data errors can be completely corrected. Good thing, because in a perfect system and perfect media, you have an error rate that is unavoidable. It's called the BLER rate, or Block Error Rate. It's counted on a frame by frame basis. So the one thing you must accept is that life isn't perfect -AT BEST!
In early CD players, there were adjustments galore. There were also test points galore to complement all those adjustments. Two in particular bear directly on this discussion. They are called the C1 and C2 error signals. They each signal the result of error correction DSPs. As you might expect, the C1 point follows the first data correction block and it signals an error that couldn't be corrected. In other words, bad data folks. That's pretty cool though, all is not lost. Good because this is a very active test point. The flagged data is then processed by the next DSP and correction is attempted using surrounding data frames. Then we come to the C2 signal. Unfortunately, this one indicates that the data is bad, invalid, unable to be fixed and whatever else you want to call it. The next operation depends greatly on how much money was spent on the DSP section. In really old cheap machines, the bad data was passed along as is (Eeeek!
). Yup, right out to the DAC where that part attempted to output the analog value the bad data represents. Nice eh? Better DSPs had some other options available. They could (in order of "niceness") interpolate the value by looking at data points surrounding the bad data, they could simply reuse the previous data point or they could mute that value (analog zero value). As the errors become worse, the options are limited until you reach the only thing that makes sense, and that is to mute the signal. All of these last options will result in a data stream that is valid. In other words, the better DSPs will output a data stream that is not in violation of CRC checks and such. Looks great, but you have to understand that while the data is good, the information it conveys is not accurate. This is a very important point to understand, and is probably where the idea of "bit perfect" data comes from. It ain't bit perfect and you can't really tell unless you have the original digital data available before encoding. It's a pretty safe bet that this isn't going to happen. BTW, all those digital tapes suffer from similar problems. In fact, in servicing digital recorders, you adjust the reading heads for minimum C1 or C2 error counts and the best "eye pattern". Notice I didn't say "for no error flags". That is important to remember, and the same applies to CD players.
The concept of an "eye pattern" is pretty wide spread these days. All digital communication is transmitted and the quality is checked by monitoring the "eye pattern". Here you have better quality signals, and the equipment used has a signal mask that represents an error zone, a do not enter zone. This waveform shapes levels where either an amplitude or timing error will probably cause that data bit to be interpreted incorrectly. They even have expensive testers for optical communications. BTW, those optical cables used in data transfer in CD to DAC connections have imprecise mating areas with the connectors, and because the cables are handled, many defects that also reduce amplitudes and cause reflections in the cable. Not a pretty picture, I'll use the copper connection, thank you very much. There is no chassis ground connection because it's transformer coupled. There goes another myth I guess.
-Chris
Couldn't resist popping in after your note.
All,
Your comments so far have been pretty close to the truth. The one error I see is that the frequencies of the RF pattern are both variable (depending on the information encoded) and really around the AM band of frequencies. When I have ever connected my frequency counter to the signal of the RF test point, I get variable readings around 450 KHz. I should allow it to run for a minute and average the reading. I am using an HP 5335A that calculates frequency via reciprocal counting, HP's method. I may have the nomenclature a little mixed up there, but that's because I don't sell the stuff, I just use it.
So, the encoded information comes off the CD with varying frequencies. That is a fact. Being an analog signal, it may suffer from level changes due to variable reflection / transmission (right through the CD and gone for good), surface defects causing focus failures (drop outs), pin holes that do the same thing, eccentricity that causes major jitter (timing errors), warping that may cause focus problems (variable signal strength) and "dished" discs, causing similar things and possible transport contact (gotta love that!). Consider also that all these effects may occur in combination. Add to that some dust, material on the disc table (causing a wobble), lens actuator movements in combination that are simply impossible to expect, mechanical misalignment (almost 100% of transports out there) and servo mis-alignment. I'm amazed that CDs play at all. They are an engineering feat that we've managed to make error tolerant (not for sound quality, but "playability"). Most transports today are complete junk, even the expensive ones.
BTW, a belt drive CD disc assembly? Come on people! It is not an turntable that runs at a constant angular velocity! The CD spins the disc to attempt a constant linear velocity. Near the center it spins shy of 500 RPM and near the outside it spins closer to 200 RPM. Just watch one. Okay, so the belt drive may work? Naw, it may function, but not very well. Remember the eccentric CDs? Yeah, those need to be sped up and slowed down on a per revolution basis. In this situation, high rotating mass and a compliant coupling are not your friends - but this works great on my Thorens turntable.
Alright, so now we have a signal coming from two pairs of photo-diodes that we are attempting to retrieve data from in the analog domain. Photo-diodes are current operated devices (little voltage variation across them). We convert this to a varying voltage and balance the two sets of diodes to obtain the focus information. It would be really great if they were matched (they are close on the better heads). By combining both sets of signals, we work with a signal with less noise. This is amplified (analog still) and filtered in an effort to reduce out of band noise, then a "slice level" is used to determine whether there is a pit or land passing by the head assembly. The information is critical on the pit edges - which leads us to the other problem with CDs. They are pressed just like records were, finer detail. So all types of pressing defect may occur (nice eh?), then they are coated with a sputtered aluminum layer. It's pretty thin, but the first CDs had a thick coating that would reflect most of that infrared beam mentioned earlier. Then, cost cutting measures reduced the thickness of the reflective layer, this allows some percentage of the beam to pass through, never to return - ever. No, it does not bounce around and come back to pollute other data. In fact, ambient light or LEDs or markers will not affect anything unless they are placed to block the beam. Of course, play with a light source of the same wavelength and you may be able to make things worse.
Who still feels good about the idea of perfect data coming back off the CD?
Recognizing that many data errors would definitely occur in reading, the original standard included an encoding method where by data was repeated in tracks close by the track being read. The Reed-Solomon (I think) code that we were forced to learn. The upshot of this is that small, localized data errors can be completely corrected. Good thing, because in a perfect system and perfect media, you have an error rate that is unavoidable. It's called the BLER rate, or Block Error Rate. It's counted on a frame by frame basis. So the one thing you must accept is that life isn't perfect -AT BEST!
In early CD players, there were adjustments galore. There were also test points galore to complement all those adjustments. Two in particular bear directly on this discussion. They are called the C1 and C2 error signals. They each signal the result of error correction DSPs. As you might expect, the C1 point follows the first data correction block and it signals an error that couldn't be corrected. In other words, bad data folks. That's pretty cool though, all is not lost. Good because this is a very active test point. The flagged data is then processed by the next DSP and correction is attempted using surrounding data frames. Then we come to the C2 signal. Unfortunately, this one indicates that the data is bad, invalid, unable to be fixed and whatever else you want to call it. The next operation depends greatly on how much money was spent on the DSP section. In really old cheap machines, the bad data was passed along as is (Eeeek!
). Yup, right out to the DAC where that part attempted to output the analog value the bad data represents. Nice eh? Better DSPs had some other options available. They could (in order of "niceness") interpolate the value by looking at data points surrounding the bad data, they could simply reuse the previous data point or they could mute that value (analog zero value). As the errors become worse, the options are limited until you reach the only thing that makes sense, and that is to mute the signal. All of these last options will result in a data stream that is valid. In other words, the better DSPs will output a data stream that is not in violation of CRC checks and such. Looks great, but you have to understand that while the data is good, the information it conveys is not accurate. This is a very important point to understand, and is probably where the idea of "bit perfect" data comes from. It ain't bit perfect and you can't really tell unless you have the original digital data available before encoding. It's a pretty safe bet that this isn't going to happen. BTW, all those digital tapes suffer from similar problems. In fact, in servicing digital recorders, you adjust the reading heads for minimum C1 or C2 error counts and the best "eye pattern". Notice I didn't say "for no error flags". That is important to remember, and the same applies to CD players.The concept of an "eye pattern" is pretty wide spread these days. All digital communication is transmitted and the quality is checked by monitoring the "eye pattern". Here you have better quality signals, and the equipment used has a signal mask that represents an error zone, a do not enter zone. This waveform shapes levels where either an amplitude or timing error will probably cause that data bit to be interpreted incorrectly. They even have expensive testers for optical communications. BTW, those optical cables used in data transfer in CD to DAC connections have imprecise mating areas with the connectors, and because the cables are handled, many defects that also reduce amplitudes and cause reflections in the cable. Not a pretty picture, I'll use the copper connection, thank you very much. There is no chassis ground connection because it's transformer coupled. There goes another myth I guess.
-Chris
I've got 20 years of ink formulation experience and several more patents. It's also nonsense. Organic dyes (nonmagnetic) and metal oxides (usually nonmagnetic, always non-ferromagnetic) are the pigments. Get a jar of silk screen ink and hold a magnet near it. No effect. Then ponder the fact that the CD label has about 1/100,000 of that amount of ink on it...
To continue ....
The information coming off the CD is loaded into a FIFO (First In, First Out) type memory. The disc motor roughly controls the timing of the data going in, and it is clocked out referenced to either the system clock, or a clock source in the DSP/DAC/ reclocker circuits. the source of the speed correction signal is the difference error between the data coming off the disc and it's regenerated clock, and whatever clock source is used for the rest. As you can imagine, there are great differences between the reference clock and the data rate. We are not talking about phase here, we are talking about buffer under-run or over-run. It's a variable mechanical system that will not even frequency lock. All we can do is hope that the buffer remains partially full at all times. This is an extremely loose coupling between the internal data-rate and the rate that the data enters the buffer. I can't stress this enough. It even works most of the time.
Lastly, there are no significant magnetic effects that occur, and demagnetizing a CD does not affect it in any way. If it did then larger de-magnetizers would cause damage and possible de-lamination if this was to be effective in doing anything. The greater threat comes from the sun. Recordable CDs are very susceptible to sun damage and the stamped variety may suffer from the aluminum layer decomposing. Some earlier CDs may even allow oxygen to come in contact with the aluminum layer (reactive metal - remember chemistry class?). This is the type of deferred damage that those truing up tools may cause. It may take a couple years, but if oxygen is allowed into the information layer, the CD will be damaged beyond repair.
Dave (planet10) also brought up a valuable point earlier on. the interaction between the servo section and both the DSP sections and analog processing sections. This can be a real concern.
Modern CD players often have two individual transformers and at least three supplies to service these three circuit blocks. Some CD players may simply use different windings and very good regulation, and this works just as well. Really, it does work every bit as well as different transformers. In this application, I would rather use EI or split "C" cores, EI being my preference. Toroid type transformers are inherently wide band devices. That means that line noise is coupled more readily into the circuitry through the transformer. No matter what arguments are placed to counter my statement, practically speaking, toroid transformer = more noise. This doesn't mean all is lost. It only means that the following power supplies need to address more problems than the EI (for example) types. Another outcome of this is that noise generated in other sections will couple to other windings more easily than with EI types, with the same remedy. A more carefully designed power supply section.
The problem I have with this is that the use of a toroid transformer has been sold as a good thing over an EI type transformer. Also, many newer - younger (?) designers are not addressing power supply designs very well. Not all, but certainly a small, "high end" manufacturer will generally not be up to the job. Of course, the low priced branding does not consider these things as important. Not even on their radar, sorry but it's true.
In a modern, well designed and built CD player, interactions between various circuits is not a problem. Unfortunately, cost will not guarantee a well designed machine. This probably will guarantee that the popular capacitors and chips have been used. There may even be very questionable methods of converting a D/A chips current output into a voltage signal.
What does all this mean to the average person? It means that the earlier advice to buy a cheap transport and use an expensive D to A solution (system or however they describe it) was completely wrong. Anyone who suggested such an approach does not understand the physics and signal processing issues involved here. Audible differences between transports are really showing which one(s) do the least wrong. Very importantly, it has to be realized that no two playing cycles are the same as far as the errors that occur retrieving the data off the CD, except for those large defects that you can see.
To do the least damage to the data, a well designed transport should be used. Examples of "less good" transports are those that increase the rotating mass of the CD - clamp/disc table - motor assembly. This includes "stable platters" and any compliant coupling between the motor and CD - clamp/disc table at all. Belt drive is an example of "unclear on the concept", unless the concept is to sell expensive .... <no comment>. I think CEC has a model like this, and the member I spoke to on this idea at the first Burning Amp was extremely upset with me over my comments. Sorry, it's physics. Rapid changes in rotational speed means that a low mass system should be used.
The information coming off the CD is loaded into a FIFO (First In, First Out) type memory. The disc motor roughly controls the timing of the data going in, and it is clocked out referenced to either the system clock, or a clock source in the DSP/DAC/ reclocker circuits. the source of the speed correction signal is the difference error between the data coming off the disc and it's regenerated clock, and whatever clock source is used for the rest. As you can imagine, there are great differences between the reference clock and the data rate. We are not talking about phase here, we are talking about buffer under-run or over-run. It's a variable mechanical system that will not even frequency lock. All we can do is hope that the buffer remains partially full at all times. This is an extremely loose coupling between the internal data-rate and the rate that the data enters the buffer. I can't stress this enough. It even works most of the time.
Lastly, there are no significant magnetic effects that occur, and demagnetizing a CD does not affect it in any way. If it did then larger de-magnetizers would cause damage and possible de-lamination if this was to be effective in doing anything. The greater threat comes from the sun. Recordable CDs are very susceptible to sun damage and the stamped variety may suffer from the aluminum layer decomposing. Some earlier CDs may even allow oxygen to come in contact with the aluminum layer (reactive metal - remember chemistry class?). This is the type of deferred damage that those truing up tools may cause. It may take a couple years, but if oxygen is allowed into the information layer, the CD will be damaged beyond repair.
Dave (planet10) also brought up a valuable point earlier on. the interaction between the servo section and both the DSP sections and analog processing sections. This can be a real concern.
Modern CD players often have two individual transformers and at least three supplies to service these three circuit blocks. Some CD players may simply use different windings and very good regulation, and this works just as well. Really, it does work every bit as well as different transformers. In this application, I would rather use EI or split "C" cores, EI being my preference. Toroid type transformers are inherently wide band devices. That means that line noise is coupled more readily into the circuitry through the transformer. No matter what arguments are placed to counter my statement, practically speaking, toroid transformer = more noise. This doesn't mean all is lost. It only means that the following power supplies need to address more problems than the EI (for example) types. Another outcome of this is that noise generated in other sections will couple to other windings more easily than with EI types, with the same remedy. A more carefully designed power supply section.
The problem I have with this is that the use of a toroid transformer has been sold as a good thing over an EI type transformer. Also, many newer - younger (?) designers are not addressing power supply designs very well. Not all, but certainly a small, "high end" manufacturer will generally not be up to the job. Of course, the low priced branding does not consider these things as important. Not even on their radar, sorry but it's true.
In a modern, well designed and built CD player, interactions between various circuits is not a problem. Unfortunately, cost will not guarantee a well designed machine. This probably will guarantee that the popular capacitors and chips have been used. There may even be very questionable methods of converting a D/A chips current output into a voltage signal.
What does all this mean to the average person? It means that the earlier advice to buy a cheap transport and use an expensive D to A solution (system or however they describe it) was completely wrong. Anyone who suggested such an approach does not understand the physics and signal processing issues involved here. Audible differences between transports are really showing which one(s) do the least wrong. Very importantly, it has to be realized that no two playing cycles are the same as far as the errors that occur retrieving the data off the CD, except for those large defects that you can see.
To do the least damage to the data, a well designed transport should be used. Examples of "less good" transports are those that increase the rotating mass of the CD - clamp/disc table - motor assembly. This includes "stable platters" and any compliant coupling between the motor and CD - clamp/disc table at all. Belt drive is an example of "unclear on the concept", unless the concept is to sell expensive .... <no comment>. I think CEC has a model like this, and the member I spoke to on this idea at the first Burning Amp was extremely upset with me over my comments. Sorry, it's physics. Rapid changes in rotational speed means that a low mass system should be used.
Hi Dave,
SY is right. There is nothing significant in the way of mass that can be magnetized in a normal CD. Watch someone bring out a metal based CD next.
Food for thought. The average hard drive uses (or used) an aluminum platter onto which the magnetizable surface for information storage was applied. Given that the heads are reading a magnetically generated signal and the heads are skimming on air just above the head, wouldn't magnetization issues with the platter substrate be a problem?
-Static doesn't matter much (within reason!).
-Magnetization is not an issue, and the best clamping systems are magnetic.
-Extraneous light has little to no effect (again, within reason!).
You will have far greater problems with RFI and magnetic fields from external sources than anything generated from a spinning CD and it's magnetic clamp system.
Things that can cause problems:
-Vibration from loud music
-Heavy foot falls or movement due to machinery or heavy traffic (including aircraft)
-Condensation
-Power line conditioners interacting with internal power supplies.
-Alignment, mechanical and servo.
-Whether the DAC is internal or external (internal is much, much better!).
-The condition of the CD and it's information layer.
-Temperature
-Air borne particles or anything that can be deposited (cooking oil, smoke).
-Bad modifications and bad workmanship.
There are probably more issues, but you can see there are no shortage of possible problems. These should all be common sense issues that are only listed to show how completely insignificant magnetization really is.
Oh, I almost forgot another rampant concern. I have no idea where this originated, but it has been postulated that a rotating magnetic disc table and clamper could interact with the focus and tracking positioning coils in the laser head assembly at the beginning of a disc in play mode. You know what? Unless the CD player is very cheaply made, I can't see this as a real issue. What is an issue are the very poor optics and alignment in cheap laser heads. The Philips VAM1202 and brother heads are an excellent example of this. The quality is extremely variable in these, as in the process is completely out of control and they are shipping heads that fail to operate properly.
-Chris
SY is right. There is nothing significant in the way of mass that can be magnetized in a normal CD. Watch someone bring out a metal based CD next.
Food for thought. The average hard drive uses (or used) an aluminum platter onto which the magnetizable surface for information storage was applied. Given that the heads are reading a magnetically generated signal and the heads are skimming on air just above the head, wouldn't magnetization issues with the platter substrate be a problem?
-Static doesn't matter much (within reason!).
-Magnetization is not an issue, and the best clamping systems are magnetic.
-Extraneous light has little to no effect (again, within reason!).
You will have far greater problems with RFI and magnetic fields from external sources than anything generated from a spinning CD and it's magnetic clamp system.
Things that can cause problems:
-Vibration from loud music
-Heavy foot falls or movement due to machinery or heavy traffic (including aircraft)
-Condensation
-Power line conditioners interacting with internal power supplies.
-Alignment, mechanical and servo.
-Whether the DAC is internal or external (internal is much, much better!).
-The condition of the CD and it's information layer.
-Temperature
-Air borne particles or anything that can be deposited (cooking oil, smoke).
-Bad modifications and bad workmanship.
There are probably more issues, but you can see there are no shortage of possible problems. These should all be common sense issues that are only listed to show how completely insignificant magnetization really is.
Oh, I almost forgot another rampant concern. I have no idea where this originated, but it has been postulated that a rotating magnetic disc table and clamper could interact with the focus and tracking positioning coils in the laser head assembly at the beginning of a disc in play mode. You know what? Unless the CD player is very cheaply made, I can't see this as a real issue. What is an issue are the very poor optics and alignment in cheap laser heads. The Philips VAM1202 and brother heads are an excellent example of this. The quality is extremely variable in these, as in the process is completely out of control and they are shipping heads that fail to operate properly.
-Chris
Or....
Rip with EAC.
Foobar kernel-stream data to envy24 chipset which does not change 44.1/48/44.1mhz.
SPDIF TO DAC.
Deal with jitter on SPDIF until better connections become cheaper or I can actually hear a difference, whichever comes first.
Put cd player in closet and cds on the high shelves as decorations.
Works for me until I can acquire better speakers, when I might consider upgrades. Hopefully additional digital outs to implement a digital x-over. I doubt any cd player will be an upgrade. All this set up really needs is DACs designed with significant buffers and their own clocks.
Rip with EAC.
Foobar kernel-stream data to envy24 chipset which does not change 44.1/48/44.1mhz.
SPDIF TO DAC.
Deal with jitter on SPDIF until better connections become cheaper or I can actually hear a difference, whichever comes first.
Put cd player in closet and cds on the high shelves as decorations.
Works for me until I can acquire better speakers, when I might consider upgrades. Hopefully additional digital outs to implement a digital x-over. I doubt any cd player will be an upgrade. All this set up really needs is DACs designed with significant buffers and their own clocks.
What about them? Are they ferromagnetic? Calcuim? Lithium? Sodium? Reality check- take the strongest magnet you have in house and hold it next to some aluminum foil.
I did read one hilarious document where a guy sawed up a CD and found that some of the sawdust bits were attracted to a magnet. He triumphantly proclaimed the inherent magnetism of CDs and used his result to justify the demagnetizer snake oil. I leave it as an exercise for the reader to figure out why this was a stunningly stupid experiment and an even stupider conclusion.
What I mean is that simply because the logic of how something is not clear to you is no reason to discount it and bash others for asking about it. How long did humanity believe the earth was flat? The universe revolves around the earth? These ridiculous audiophile toys can't make a difference, but they do. Why? I don't know and don't plan on trying to explain them.
I have tried in the past and spent far to many hours thinking about it. In the end, I can't explain it. What I can tell you is that a tape demagnatizer changes the sound of a cd slightly. I have one I will lend out if anyone wants to borrow it. It is not some high end yada yada smoke and mirrors audiophile gadget. It is a Radio Shack bulk tape demagnetizer. In other words, it is a big coil with a handle and a power switch. Don't ask me why it changes things, but it does. It doesn't really make things sound better in my opinion, but it changes the sound. I used for about a year before deciding that my time was better spent, well, doing pretty much anything other than demaging disks.
Hi DaveM,
I used the same demagnetizer when I had my shop. We have played and joked around with it, and it didn't make any difference at all. Nor should it.
I'm not discounting what you are saying out of hand, but I'm sure that one hears what they expect to hear. Consider that this is an optical system. Magnetized or not, the light doesn't care at this scale. What I can tell you is that no two plays will be identical, and that the RF patterns look identical. I can't see individual bit changes, but I can see the overall quality of the signal. Another telling experiment we tried (okay, we were playing), the C1 error counts did not change with demagnetizing or "the green marker". Not outside the varying numbers we were getting. So at best, your actions changed something that made a smaller difference than the other totally random effects.
You know what does make a big difference that can be audible? Correcting the alignment/adjustment on a CD player that is out or directly from the factory. Most of the time anyway.
Always remember. There were people who heard enhanced bass after using the green marker edge treatment, and I know a guy who had green LEDs installed in his player. He was adamant there was a positive difference. He doesn't like being reminded, so you will no be the first, or the last on things like this.
Hi malcontent,
If you are actually saying that you can accept things the way they sound at the moment, that's fine and perfectly acceptable. If you intend to say that you are getting accurate data with your process, we disagree. The data loss/errors occur before the information leaves the DSP chip, long before you have any access to it. No matter what you throw at it by this point in time, all you can do is mess it up some more. There are no improvements to be had after the signal has been spit out unless you have a good up-sampler. Then you are only calculating intermediate samples from the stuff that was sent out your way. Bad data will create more bad data.
The coaxial digital out has a lower data rate, as long as the cable is in good condition. Try using a video patch cable if you want.
Just sit back, listen and enjoy. Not much you can do beyond having your machine set up properly.
-Chris
I used the same demagnetizer when I had my shop. We have played and joked around with it, and it didn't make any difference at all. Nor should it.
I'm not discounting what you are saying out of hand, but I'm sure that one hears what they expect to hear. Consider that this is an optical system. Magnetized or not, the light doesn't care at this scale. What I can tell you is that no two plays will be identical, and that the RF patterns look identical. I can't see individual bit changes, but I can see the overall quality of the signal. Another telling experiment we tried (okay, we were playing), the C1 error counts did not change with demagnetizing or "the green marker". Not outside the varying numbers we were getting. So at best, your actions changed something that made a smaller difference than the other totally random effects.
You know what does make a big difference that can be audible? Correcting the alignment/adjustment on a CD player that is out or directly from the factory. Most of the time anyway.
Always remember. There were people who heard enhanced bass after using the green marker edge treatment, and I know a guy who had green LEDs installed in his player. He was adamant there was a positive difference. He doesn't like being reminded, so you will no be the first, or the last on things like this.
Hi malcontent,
If you are actually saying that you can accept things the way they sound at the moment, that's fine and perfectly acceptable. If you intend to say that you are getting accurate data with your process, we disagree. The data loss/errors occur before the information leaves the DSP chip, long before you have any access to it. No matter what you throw at it by this point in time, all you can do is mess it up some more. There are no improvements to be had after the signal has been spit out unless you have a good up-sampler. Then you are only calculating intermediate samples from the stuff that was sent out your way. Bad data will create more bad data.
The coaxial digital out has a lower data rate, as long as the cable is in good condition. Try using a video patch cable if you want.
Jitter in this leg of the journey is an entirely different thing. You missed the point completely. Of course, you may hear a difference - or not. Also, re-sampling the data (not changing the sample rate, like 44.1 KHz to 48 KHz) shouldn't have any effect as long as you change by an integer amount. For example, 44.1 KHz to 88.2 KHz will give you interpolated samples between each original sample, and that can sound better. Just by virtue that the audio filtering can be cut off at a higher frequency, a lower order filter. Going from 44.1 KHz to 96 KHz could be interesting. I haven't even thought about that, but there is probably heavy signal processing going on. That makes it a quality thing.
Just sit back, listen and enjoy. Not much you can do beyond having your machine set up properly.
-Chris
Do you understand how digital works? (no, really understand, or did you just gleam over it looking for ways to discredit it)I am not saying that the amount of magnetism will affect anything, just pointing out that the blanket statement that a CD has nothing magnetic in it is a false statement. SY's analysis indicates that it is probably not large enuff to have any affect (but not a proof).
I am trying to stimulate discussion of possible mechanisms, and why they will or will not affect things. I do not have a position on these devices but i would like to see a clear & intelligent discussion of why or why-not.
For instance, we can dismiss the bits-is-bits, because we are dealing with an analog representation that could be misinterpreted, Here is an image of idea, practical ideal, and something more typical in a cheap CD player (maybe they are even worse in really cheap players -- anyone have some real world eye-pattern captures they can post?)
dave
All the reader has to do is detect if the signal is over half way, which would be no problem for any of those waveforms. The timing signals come from a clock which tells the reader how fast to spin.(no jitter of the disc) Boy the designers of CD players must be idiots if these kinda things make any difference.
How do you demagnitize something that was never magnetized, or even magnetic? You haven't changed the disc, so you havent changed the sound.
We are not bashing for asking, its for believing. Believing something that makes no sense, has no proof. Its like saying demagnetizing my camera lens has improved my pictures. Would you believe that, if so I have products to sell you.
Hello anatech,
I am not claiming to get perfect sound. I do claim to have bit perfect output though.
Not sure what you mean by this:
Where does the data get lost? What DSP?
Exactly. And most sound card chipsets do change it from 44.1 to 48mhz. Even if they convert it back before they spit it out. That's why I specified a chipset that does not and kernel streaming past my operating system.
When I (hopefully) implement digital X-overs, then I will be using DSP. Should compare very favourably to a passive one though.
I am not claiming to get perfect sound. I do claim to have bit perfect output though.
Not sure what you mean by this:
Where does the data get lost? What DSP?
Exactly. And most sound card chipsets do change it from 44.1 to 48mhz. Even if they convert it back before they spit it out. That's why I specified a chipset that does not and kernel streaming past my operating system.
When I (hopefully) implement digital X-overs, then I will be using DSP. Should compare very favourably to a passive one though.
Hi cbdb,
The waveforms do not look like that on a CD player. A 50% slice level will not extract the data from the RF pattern, it's certainly not that simple (I wish). Early Yamaha machines are an example that had adjustable slice level and also RF amplitude.
Hi malcontent,
It sounds to me as if you have read things on the internet without actually getting any true research done. In fact, if you read my previous posts I made today, you should find the answers you are looking for. I did go through several areas where the audio data can be corrupted. Try reading that stuff again tomorrow when you aren't tired. I know I'm hitting the sack now.
-Chris
The waveforms do not look like that on a CD player. A 50% slice level will not extract the data from the RF pattern, it's certainly not that simple (I wish). Early Yamaha machines are an example that had adjustable slice level and also RF amplitude.
Hi malcontent,
But, you are not. The data stream is valid, but it's not always correct. See the difference?
Any errors in decoding the RF eye pattern are processed in the digital signal processor chip(s). The RF amplifier deals with the signal in the analog domain and outputs the analog information to the DSP chip in what is called an EFM pattern. The DSP takes this, errors and all, and processes it in an effort to correct any errors. EFM stands for Eight to Fourteen Modulation, involved in the Reed - Solomon code, the DSP also decodes this to give us a serial data stream. All this takes place well before you have access to the signal. The DSP chip also sends the audio data out it's serial data pin, if so equipped. Most early ones do not provide this output.
Not true. Soundblaster used to do that, but starting with the X-Fi series, this no longer occurs.
It sounds to me as if you have read things on the internet without actually getting any true research done. In fact, if you read my previous posts I made today, you should find the answers you are looking for. I did go through several areas where the audio data can be corrupted. Try reading that stuff again tomorrow when you aren't tired. I know I'm hitting the sack now.
-Chris
Thanx for your posts Chris. They are really good.
So from this discussion so far, the olny possibility of an improvement is if the cutter gives better balance to the CD leading to less work by the servo... BUT, that make let O2 in ruining your CD. Besides, for the $1k you pay for the device you could do a whole lot to the power supply and make it moot?
What about light scattering affecting the read (You hinted at this, but didn't cover it fully)?
dave
So from this discussion so far, the olny possibility of an improvement is if the cutter gives better balance to the CD leading to less work by the servo... BUT, that make let O2 in ruining your CD. Besides, for the $1k you pay for the device you could do a whole lot to the power supply and make it moot?
What about light scattering affecting the read (You hinted at this, but didn't cover it fully)?
dave
Yes. Unfortunately I made a random apples to oranges comment about my system. I now think anatach and I are talking about two different things.
@ anatech -
I think we are talking about two different things here. I was referring to data ripped from a disc to hard drive using brute force error correction and data caching turned off.
Yes, I am taking the manufacturers word for what it does with the signal after it is in their chipset. I have no clue how accurately it is clocked or how accurately my DAC interprets the signal. But, I am confident the actual data is correct. All the rest of the data on my computers seems just fine. Applications run, pictures are all there and it doesn't crash, so I'm guessing the music is intact. (Music copied from another hard drive. 7 weeks of music in less than two hours. No errors.)
I guess the point of my original post was that disc players seem to be made from design concepts from 1978. Any device that attempts real-time playback, error correction and clocking based on it ability to read the disc, is fundamentally flawed.
My system is, I'm sure, highly flawed. But only downstream of the accurate data it produces. AFAIK PC clocks are dithered to reduce RF emission, so that can't be good. Not sure if the card attempts to re-clock or not. But, that is jitter, not data.
Eric
Not a valid comparison. CD-ROMs have a different encoding than Audio CDs. Multiple reads & no time restrictions give you a much better chance thou.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.