Hi dwferg,
-Chris
Yes there is a really good reason. If you have ever looked at the eye pattern from one of these, the CD one is a complete mess with high noise, distorted signal and a pretty closed eye. All this means a ton of digital errors that the (cheap) DEP has to deal with. Junk in, junk out. What consumers don't know will hurt them.
-Chris
Hi Salar,
Excellent! So Kyocera was the OEM, I thought it was NEC. Their OMS-3/4 transport is Kyocera as well as they have the same transport as a Kyocera model I serviced once.
Wouldn't it be great if they had heads, belts and disc motors? A new clamper wouldn't hurt either! Kyocera used ceramic slide rods that were problematic in the OMS-3/4 mechanism instead of steel. I think they wore out the head bearings too quickly (new head time). Great find!
-Chris
Excellent! So Kyocera was the OEM, I thought it was NEC. Their OMS-3/4 transport is Kyocera as well as they have the same transport as a Kyocera model I serviced once.
Wouldn't it be great if they had heads, belts and disc motors? A new clamper wouldn't hurt either! Kyocera used ceramic slide rods that were problematic in the OMS-3/4 mechanism instead of steel. I think they wore out the head bearings too quickly (new head time). Great find!
-Chris
Hi dwferg,
Without actually checking one yourself, there is no way for you to have known.
If you have an oscilloscope, check the eye pattern in various machines. Then check a good, old machine that still works. You'll be amazed at the difference. Now, imagine feeding the new electronics with a clean eye pattern as to what it might sound like.
-Chris
Without actually checking one yourself, there is no way for you to have known.
If you have an oscilloscope, check the eye pattern in various machines. Then check a good, old machine that still works. You'll be amazed at the difference. Now, imagine feeding the new electronics with a clean eye pattern as to what it might sound like.
-Chris
Hi Mark,
The eye pattern is everything! I use it much like a blood test that you might have, except I order all the tests on the sheet.
Every signal, every bit of information is highly dependent on the eye pattern. Even in RF communications, the eye pattern is the go-to test if you have a problem, especially a quality problem.
The eye pattern is the RF analog waveform that is how the data comes off the disc to begin with. The way the digital information is created comes from whether the waveform is high or low (actually the transitions between high and low). Those transitions are also used to generate ... wait for it ... the clock. This is compared to the oscillator in the CD player and the difference is used to control the ... disc speed! If you have ever heard a disc spinning out of lock, it is because the eye pattern is too degraded to lock to. So now we have a clock and the disc spinning (in lock) at a constant linear velocity (that's why "Stable Disc" products are harmful, and belt drive is so stupid). The clock determines when the eye pattern (that again ??!) is read to determine when transitions occur, but they are read from the short term stable highs and lows of the waveform. If the eye pattern is noisy or even closed, the decisions of high and low will be incorrect. The disc may very well be mis-tracking with high digital distortion (if it hasn't muted yet). Once we get to the EFM pattern you are ready for the DSP to do it's thing. However, when you have errors, it sets the C1 flag (internal these days) and an attempt to recreate the proper data is made from the information in adjacent. If that fails, then the C2 flag is set and the data is (hopefully) discarded. Then the DSP will try to mute, reinsert previous values or interpolate the values as a best guess as to what the value should have been.
That occurred for one frame. What happens when adjacent frames are also corrupted? No values to interpolate with. Eventually it will mute the audio.
So from the eye pattern, we can find out about tracking, E-F balance (critical), focus, sled operation, laser power (derived from amplitude) and disc motor + disc motor servo. Just be glad you don't have to adjust the diffraction grating or slice level any more.
The eye pattern is the root of everything in a disc player. If you aren't checking it, you are basically working blind in every sense of the word. Just like a doctor who never does blood tests on his live patients. Mark, the eye pattern is your go-to test. You can learn so much from an eye pattern that you might be able to diagnose a problem without going any further.
I haven't heard the term E22-E32 before. Care to share?
-Chris
The eye pattern is everything! I use it much like a blood test that you might have, except I order all the tests on the sheet.
Every signal, every bit of information is highly dependent on the eye pattern. Even in RF communications, the eye pattern is the go-to test if you have a problem, especially a quality problem.
The eye pattern is the RF analog waveform that is how the data comes off the disc to begin with. The way the digital information is created comes from whether the waveform is high or low (actually the transitions between high and low). Those transitions are also used to generate ... wait for it ... the clock. This is compared to the oscillator in the CD player and the difference is used to control the ... disc speed! If you have ever heard a disc spinning out of lock, it is because the eye pattern is too degraded to lock to. So now we have a clock and the disc spinning (in lock) at a constant linear velocity (that's why "Stable Disc" products are harmful, and belt drive is so stupid). The clock determines when the eye pattern (that again ??!) is read to determine when transitions occur, but they are read from the short term stable highs and lows of the waveform. If the eye pattern is noisy or even closed, the decisions of high and low will be incorrect. The disc may very well be mis-tracking with high digital distortion (if it hasn't muted yet). Once we get to the EFM pattern you are ready for the DSP to do it's thing. However, when you have errors, it sets the C1 flag (internal these days) and an attempt to recreate the proper data is made from the information in adjacent. If that fails, then the C2 flag is set and the data is (hopefully) discarded. Then the DSP will try to mute, reinsert previous values or interpolate the values as a best guess as to what the value should have been.
That occurred for one frame. What happens when adjacent frames are also corrupted? No values to interpolate with. Eventually it will mute the audio.
So from the eye pattern, we can find out about tracking, E-F balance (critical), focus, sled operation, laser power (derived from amplitude) and disc motor + disc motor servo. Just be glad you don't have to adjust the diffraction grating or slice level any more.
The eye pattern is the root of everything in a disc player. If you aren't checking it, you are basically working blind in every sense of the word. Just like a doctor who never does blood tests on his live patients. Mark, the eye pattern is your go-to test. You can learn so much from an eye pattern that you might be able to diagnose a problem without going any further.
I haven't heard the term E22-E32 before. Care to share?
-Chris
The eye pattern is processed into digital data, followed by error correction. There are two parts to this error correction.
E1.1 E2.1 E3.1 ,also called BLER and should be <220/second, indicates the amount of error correction of C1.
E1.2 E2.2 E3.2 indicates the amount of error correction in C2. As long as E3.2 is zero then all the read errors were perfectly corrected.
The eye pattern is so far removed from the output signal, it only becomes relevant when it is so bad it causes errors and then only if uncorrectable, which is indicated by E3.2.
It is bit like me correcting this sentence and you receiving it correctly.
E1.1 E2.1 E3.1 ,also called BLER and should be <220/second, indicates the amount of error correction of C1.
E1.2 E2.2 E3.2 indicates the amount of error correction in C2. As long as E3.2 is zero then all the read errors were perfectly corrected.
The eye pattern is so far removed from the output signal, it only becomes relevant when it is so bad it causes errors and then only if uncorrectable, which is indicated by E3.2.
It is bit like me correcting this sentence and you receiving it correctly.
Hi Mark,
Okay, so the E designation is simply what we called the C1 and C2 error flags back when CDs first hit the market. I agree with your points until you say that the C2 error is the only one that consumers need worry about. I'm guessing that the Ex.1 and Ex.2 stand for intermediate steps in the DSP. Subdividing the C1 and C2 flags into smaller steps.
Our BLER rate definition was per second, but I can't remember what the actual limit was. The point was that you wanted a minimal error rate as more would accumulate over time as the disc surface became "not perfect". However, the point where the data is bad and unrecoverable is not the E3.2, it is the E1.2 flag.
The one thing I hope you will reconsider is the fact that everything begins with a good eye pattern. Here is where we disagree somewhat. For an E1.2, E2.2 and E3.2, the data is already flagged as bad and unrecoverable, and the resulting action is replacing the bad data with something that hopefully is close to what the data should have been. Basically, once the errors have mounted to a certain level, even the best DSP just packs up and goes home <mute> which I guess is the E3.2 action.
Outside of the CD/DVD business, we also use eye patterns. This is with signal generation and degraded signal through the medium used for transmission. I can assure you that the eye pattern is the critical test for quality. Problems are identified and the eye pattern is used as a quality indicator for the channel in use. All long distance telephone traffic and cell traffic is digital, and the quality is checked via the eye pattern and constellation patterns. Just for an example.
I think we can both agree that the first place you will see problems developing will be in the eye pattern. It's important to catch them there before your C2 flags become active. If all you are looking at are the C1 flags, you aren't getting nearly as much qualitative information. If you are only monitoring the C2 flags, I guess you want to watch E3.2, you've completely missed the boat once that flag goes active.
For years, repairing a CD player and properly adjusting it has been noticed by some owners. They report that the sound quality is improved even though the C2 error levels hadn't been reached yet. The same thing was noticed in the 8 track DAT service. We tuned the unit watching the C1 flags (tune for minimum smoke). The eye pattern was still the best indicator of quality.
-Chris
Okay, so the E designation is simply what we called the C1 and C2 error flags back when CDs first hit the market. I agree with your points until you say that the C2 error is the only one that consumers need worry about. I'm guessing that the Ex.1 and Ex.2 stand for intermediate steps in the DSP. Subdividing the C1 and C2 flags into smaller steps.
Our BLER rate definition was per second, but I can't remember what the actual limit was. The point was that you wanted a minimal error rate as more would accumulate over time as the disc surface became "not perfect". However, the point where the data is bad and unrecoverable is not the E3.2, it is the E1.2 flag.
The one thing I hope you will reconsider is the fact that everything begins with a good eye pattern. Here is where we disagree somewhat. For an E1.2, E2.2 and E3.2, the data is already flagged as bad and unrecoverable, and the resulting action is replacing the bad data with something that hopefully is close to what the data should have been. Basically, once the errors have mounted to a certain level, even the best DSP just packs up and goes home <mute> which I guess is the E3.2 action.
Outside of the CD/DVD business, we also use eye patterns. This is with signal generation and degraded signal through the medium used for transmission. I can assure you that the eye pattern is the critical test for quality. Problems are identified and the eye pattern is used as a quality indicator for the channel in use. All long distance telephone traffic and cell traffic is digital, and the quality is checked via the eye pattern and constellation patterns. Just for an example.
I think we can both agree that the first place you will see problems developing will be in the eye pattern. It's important to catch them there before your C2 flags become active. If all you are looking at are the C1 flags, you aren't getting nearly as much qualitative information. If you are only monitoring the C2 flags, I guess you want to watch E3.2, you've completely missed the boat once that flag goes active.
For years, repairing a CD player and properly adjusting it has been noticed by some owners. They report that the sound quality is improved even though the C2 error levels hadn't been reached yet. The same thing was noticed in the 8 track DAT service. We tuned the unit watching the C1 flags (tune for minimum smoke). The eye pattern was still the best indicator of quality.
-Chris
Your interpretation of E1.2 and E2.2 is incorrect and the correction is perfect. Only E3.2 indicates uncorrectable errors, after which there is interpretation style of correction taking place.
Your eye pattern theory is not supported by the CD technology used for playback. The main objective was to design a system that's not sensitive to read errors.
Writing Quality - Printer Friendly version
Your eye pattern theory is not supported by the CD technology used for playback. The main objective was to design a system that's not sensitive to read errors.
Writing Quality - Printer Friendly version
Last edited:
Hi Mark,
Okay, I'll bite. Please go through the steps for E1.2~E3.1, then from E1.2 ~E3.2.
When we were taught CD service by several companies, they went into great depth so that we even understood the Reed-Solomon error correction. This is still the way CD error correction works, it's locked in to the standard. Those companies included Philips Teac / Tascam Revox and I'll assume that Yamaha was also correct since they agreed with everyone else. Denon was probably from Sony's technical services.
Where you might be getting confused is by considering the DSP output as a faithful decoding from the material encoded on the CD, but it isn't. Let me expand on that.
If your E1.2 and E2.2 is considered referenced to whether the data is valid or not, that would explain it. When the level 2 error detection is flagged, that means that level 1 failed, and that means the data is unrecoverable - period. Now, in the level 2 (C2in my world), the original data is discarded, make no mistake, it's gone. What happens in the C2 next is that, depending on the DSP programming, a previous data value is substituted, or an interpolation is made considering the previous values and the next valid value. Whichever action is taken, what comes out of the DSP chip is valid data, but it ain't the value that was stored on the CD unless by chance it was. Now, if the preceding data is bad, and the next data is also invalid, there is no choice but to mute the output in the digital domain. I think you are calling this step E3.2. But make no mistake, if the level 1 error correction fails, level 2 is concealment. It's not some miracle of technology that pulls the corrupted data out of thin air. The level 1 corrector already did the Reed-Solomon thing and it didn't work. The distributed data was corrupted as well (at least the CRC was).
The fact remains, the instant the C2 flag is raised (any of your Ex.2 stages), the original data has been marked as bad, it's over for that packet. The C2 "corrector" now does it's best to send data packets out that have a good CRC so the DAC can make an intelligent output level. But the data is no longer what was on the CD. All it has going for it is a data packet with a matching (good) CRC bit. The data is valid, not what was on the disc.
The "bit perfect" folks get this bit wrong every time. But I can see where the confusion comes from.
-Chris
Okay, I'll bite. Please go through the steps for E1.2~E3.1, then from E1.2 ~E3.2.
When we were taught CD service by several companies, they went into great depth so that we even understood the Reed-Solomon error correction. This is still the way CD error correction works, it's locked in to the standard. Those companies included Philips Teac / Tascam Revox and I'll assume that Yamaha was also correct since they agreed with everyone else. Denon was probably from Sony's technical services.
Where you might be getting confused is by considering the DSP output as a faithful decoding from the material encoded on the CD, but it isn't. Let me expand on that.
If your E1.2 and E2.2 is considered referenced to whether the data is valid or not, that would explain it. When the level 2 error detection is flagged, that means that level 1 failed, and that means the data is unrecoverable - period. Now, in the level 2 (C2in my world), the original data is discarded, make no mistake, it's gone. What happens in the C2 next is that, depending on the DSP programming, a previous data value is substituted, or an interpolation is made considering the previous values and the next valid value. Whichever action is taken, what comes out of the DSP chip is valid data, but it ain't the value that was stored on the CD unless by chance it was. Now, if the preceding data is bad, and the next data is also invalid, there is no choice but to mute the output in the digital domain. I think you are calling this step E3.2. But make no mistake, if the level 1 error correction fails, level 2 is concealment. It's not some miracle of technology that pulls the corrupted data out of thin air. The level 1 corrector already did the Reed-Solomon thing and it didn't work. The distributed data was corrupted as well (at least the CRC was).
The fact remains, the instant the C2 flag is raised (any of your Ex.2 stages), the original data has been marked as bad, it's over for that packet. The C2 "corrector" now does it's best to send data packets out that have a good CRC so the DAC can make an intelligent output level. But the data is no longer what was on the CD. All it has going for it is a data packet with a matching (good) CRC bit. The data is valid, not what was on the disc.
The "bit perfect" folks get this bit wrong every time. But I can see where the confusion comes from.
-Chris
When the level 2 error detection is flagged (C2), that means that level 2 failed, not level 1.
There are three error count ( E12, E22, and E32 ) at the C2 decoder.
"E12 count indicates the frequency of occurrence of a single symbol (correctable) error in the C2 decoder. A high E12 is not problematic because one E31 error can generate up to 28 E12 errors due to interleaving. Each C1 byte is sent in a different C2 frame, therefore never affecting more than one byte in any C2 frame
E22 count indicates the frequency of double symbol (correctable) error in the C2 decoder. E22 errors are the worst correctable errors.
E32 count indicates triple-symbol, or more, (un-correctable) errors in the C2 decoder. E32 should never occur on a disc."
Here you can clearly see what is correctable, everything but E32, no confusion. The problem is that it is possible for the eye pattern to look good and the playback to be bad with an excessive amount of errors or visa versa. BLER, E22 and E32 give you a real indication of the quality of playback, the actual result of the system.
There are three error count ( E12, E22, and E32 ) at the C2 decoder.
"E12 count indicates the frequency of occurrence of a single symbol (correctable) error in the C2 decoder. A high E12 is not problematic because one E31 error can generate up to 28 E12 errors due to interleaving. Each C1 byte is sent in a different C2 frame, therefore never affecting more than one byte in any C2 frame
E22 count indicates the frequency of double symbol (correctable) error in the C2 decoder. E22 errors are the worst correctable errors.
E32 count indicates triple-symbol, or more, (un-correctable) errors in the C2 decoder. E32 should never occur on a disc."
Here you can clearly see what is correctable, everything but E32, no confusion. The problem is that it is possible for the eye pattern to look good and the playback to be bad with an excessive amount of errors or visa versa. BLER, E22 and E32 give you a real indication of the quality of playback, the actual result of the system.
Last edited:
Hi Mark,
I'll leave it at this. The C2 comes after C1 and does indicate an unrecoverable error. The C1 errors are completely corrected in an error-free manner. So C2 anything means that data is toast.
A clean eye pattern is the best indicator of health as it clearly shows errors and problems long before it shows up in the C1 corrector. Even with more robust error tolerant codes, the eye pattern is the critical signal to assess. You will see phase errors clearly, and jitter is highly visible. Unless you inject errors in the coding, a clean, stable eye pattern will result in the lowest C1 flag count.
If everything has gone to hell and the eye pattern is trash, the only signals you can watch to try and improve the situation will be the C1 and C2 flags. However, if this is your only reasonable signal, the eye pattern is already unacceptably compromised. The Philips VAM series of transports would be an excellent example of a poor, distressed, jittery and closed eye signal I see on a regular basis. When you get one that has a better eye pattern, the machine will work right away and sound good. Now, if you feed the electronics a clean eye pattern, the number of C1 flags drop considerably and the C2 flags disappear. The sound quality is improved.
I think what you need to do is to sit down at a bench and play with some working transports using your BLER rate counters and a good oscilloscope. If you have Philips 5A test disc, try that out on a poor transport and a good one that is in good adjustment. You will witness that the clean eye pattern will handle the defects much better than the one with a crappy eye pattern. If you can find a Pierre Verany (? sp) discs, they have some defects that go beyond the specifications. With a good disc and Nakamichi OMS-5 or 7 adjusted properly, the player will track through everything and might even give you a constant 500 Hz (I think) tone with no chirps. Other machines will typically throw in the towel and either stop playing or they will lose lock and run the head to the outside stop.
Always remember that a good data frame with valid CRC does not mean it is the correct data that was on the CD. Only if there are no C2 error flags will it be correct data that came off the disc. Time to roll up your sleeves and check things in the trenches.
-Chris
I'll leave it at this. The C2 comes after C1 and does indicate an unrecoverable error. The C1 errors are completely corrected in an error-free manner. So C2 anything means that data is toast.
A clean eye pattern is the best indicator of health as it clearly shows errors and problems long before it shows up in the C1 corrector. Even with more robust error tolerant codes, the eye pattern is the critical signal to assess. You will see phase errors clearly, and jitter is highly visible. Unless you inject errors in the coding, a clean, stable eye pattern will result in the lowest C1 flag count.
If everything has gone to hell and the eye pattern is trash, the only signals you can watch to try and improve the situation will be the C1 and C2 flags. However, if this is your only reasonable signal, the eye pattern is already unacceptably compromised. The Philips VAM series of transports would be an excellent example of a poor, distressed, jittery and closed eye signal I see on a regular basis. When you get one that has a better eye pattern, the machine will work right away and sound good. Now, if you feed the electronics a clean eye pattern, the number of C1 flags drop considerably and the C2 flags disappear. The sound quality is improved.
I think what you need to do is to sit down at a bench and play with some working transports using your BLER rate counters and a good oscilloscope. If you have Philips 5A test disc, try that out on a poor transport and a good one that is in good adjustment. You will witness that the clean eye pattern will handle the defects much better than the one with a crappy eye pattern. If you can find a Pierre Verany (? sp) discs, they have some defects that go beyond the specifications. With a good disc and Nakamichi OMS-5 or 7 adjusted properly, the player will track through everything and might even give you a constant 500 Hz (I think) tone with no chirps. Other machines will typically throw in the towel and either stop playing or they will lose lock and run the head to the outside stop.
Always remember that a good data frame with valid CRC does not mean it is the correct data that was on the CD. Only if there are no C2 error flags will it be correct data that came off the disc. Time to roll up your sleeves and check things in the trenches.
-Chris
Hi stocktrader200,
Well, you just proved that you have never worked with any of this stuff.
Data CDs are completely different from music CDs. The error correction used in data discs are far more robust than your music CD. Your data stream for whatever you plan to use still depends on an antenna or cable, both of which can corrupt data (send it again), and any physical media depends critically on the mechanics. They have become much better at hiding problems these days, but the errors still exist lurking just below the surface.
You have got to love blind faith and the wonderful job advertising is doing!
-Chris
Well, you just proved that you have never worked with any of this stuff.
Data CDs are completely different from music CDs. The error correction used in data discs are far more robust than your music CD. Your data stream for whatever you plan to use still depends on an antenna or cable, both of which can corrupt data (send it again), and any physical media depends critically on the mechanics. They have become much better at hiding problems these days, but the errors still exist lurking just below the surface.
You have got to love blind faith and the wonderful job advertising is doing!
Maybe for phones. Not for good stationary equipment.
-Chris
Anatech - then your computer should be the best CD reader of all
digital data is digital data. it either works or it doesn't.
As a Test, you can open a CD player and *stop* the disc with your finger and the music will play on for several seconds. This is because the error correction requires that data from the disc is stored in memory so that when an error is detected the player has sufficient time to re read the data from the disc.
Iphones are the almost perfect audiophile source as they are powered by battery (no hum , signal isolation) the dac is integrated ( low latency) and there are no switching frequencies anywhere near the audio band ( Easy to filter out). there are more advantages, NAND produces no current spikes on read and very low read errors requiring correction
digital data is digital data. it either works or it doesn't.
As a Test, you can open a CD player and *stop* the disc with your finger and the music will play on for several seconds. This is because the error correction requires that data from the disc is stored in memory so that when an error is detected the player has sufficient time to re read the data from the disc.
Iphones are the almost perfect audiophile source as they are powered by battery (no hum , signal isolation) the dac is integrated ( low latency) and there are no switching frequencies anywhere near the audio band ( Easy to filter out). there are more advantages, NAND produces no current spikes on read and very low read errors requiring correction
Last edited: