I have to revise initial obsevations about copied CD's. After quite a few experiments with reading and writing speed variations, and using Nero's "jitter correction" (but haven't yet tried normalization or different media) it seems that no one conclusion is always valid.
In general the copy is less detailed than the original, proportional to the speed of both read and write. Not clear that "jitter correction" does anything.
In a very good recording this is a bad thing, it gets to sounding quite thin and the soundstage suffers - the point sources of speakers are much more obvious.
In harsh, digital noise recordings (like Lovett's Pontiac), it can sound better, as it seems to filter out more noise than music. So it can sound more musical and pleasant even though there is less detail, wierd.
Who knows if this is system specific, probably, but the differences are very noticeable. btw the cd is ROtel RCD bx955 with slight power supply cap mods.
In general the copy is less detailed than the original, proportional to the speed of both read and write. Not clear that "jitter correction" does anything.
In a very good recording this is a bad thing, it gets to sounding quite thin and the soundstage suffers - the point sources of speakers are much more obvious.
In harsh, digital noise recordings (like Lovett's Pontiac), it can sound better, as it seems to filter out more noise than music. So it can sound more musical and pleasant even though there is less detail, wierd.
Who knows if this is system specific, probably, but the differences are very noticeable. btw the cd is ROtel RCD bx955 with slight power supply cap mods.
You see Guido and others... Unless jitter is now defined as "anything within the read window other than the original bit", then jitter is not the issue.
If it was that definition, than any system that has that much information within the read window will sound bad. And contrary to Aghead's comments my experience has been that on my particular drive (TDK) the CDs always sounds better and more imaged.
Hence my ongoing dilemma.
We need to have a common ground though. What is "jitter" nowadays? Back in the day when I had to align hard drives, jitter was eliminated or minimized by some adjustments in the drive speed, the read circuit clock, the read window width (in milliseconds. Old systems. Slower) and so on. Those things went out of alignment on discrete circuits because of heat differences.
Nowadays with all that junk on a single chip, and automatic PLL type clock synchronizations etc., jitter is all but a historical footnote. In hard drives, anyway.
In any event, jitter was the back and forth movement within the read window of a bit, due to the hard drive platter's rotation inconsistencies. If jitter was too much, it would occasionally cause the bit to move outside the read window. Data errors ensued.
So you see my problem with the ubiquitous and to me gratuitous use of the term "jitter", it may have lost its true meaning, or taken on a new one. I read an article a few weeks back I think someone here posted a link to and I couldn't help but exclaim "BULL$H!T!". The person did not really have a full understanding of digital technology, as far as I could tell.
I mean, it was a real problem with linear data storage. But... CD data is not linearly stored. It is stored in an encoded format. Therefore, if there were some bytes with some bits of incorrect data, the BIOS coded program would decode the data stream and correct for any errors. This being after the data has been read into memory, thereby eliminating any "jitter" induced errors to begin with. Oh, any jitter induced errors would most likely be corrected for with CRC checking even before being decoded and rearranged for DAC processing.
I mean, if jitter was bad enough on either the disc or the drive, the freakin thing would not read very well, and would really go nuts if both were bad. Guess what? Time for a new CD player!
Or... maybe things have changed that much in the 23 years I have been in the business?
Naaaaaahhhhhhh.
Gabe
If it was that definition, than any system that has that much information within the read window will sound bad. And contrary to Aghead's comments my experience has been that on my particular drive (TDK) the CDs always sounds better and more imaged.
Hence my ongoing dilemma.
We need to have a common ground though. What is "jitter" nowadays? Back in the day when I had to align hard drives, jitter was eliminated or minimized by some adjustments in the drive speed, the read circuit clock, the read window width (in milliseconds. Old systems. Slower) and so on. Those things went out of alignment on discrete circuits because of heat differences.
Nowadays with all that junk on a single chip, and automatic PLL type clock synchronizations etc., jitter is all but a historical footnote. In hard drives, anyway.
In any event, jitter was the back and forth movement within the read window of a bit, due to the hard drive platter's rotation inconsistencies. If jitter was too much, it would occasionally cause the bit to move outside the read window. Data errors ensued.
So you see my problem with the ubiquitous and to me gratuitous use of the term "jitter", it may have lost its true meaning, or taken on a new one. I read an article a few weeks back I think someone here posted a link to and I couldn't help but exclaim "BULL$H!T!". The person did not really have a full understanding of digital technology, as far as I could tell.
I mean, it was a real problem with linear data storage. But... CD data is not linearly stored. It is stored in an encoded format. Therefore, if there were some bytes with some bits of incorrect data, the BIOS coded program would decode the data stream and correct for any errors. This being after the data has been read into memory, thereby eliminating any "jitter" induced errors to begin with. Oh, any jitter induced errors would most likely be corrected for with CRC checking even before being decoded and rearranged for DAC processing.
I mean, if jitter was bad enough on either the disc or the drive, the freakin thing would not read very well, and would really go nuts if both were bad. Guess what? Time for a new CD player!
Or... maybe things have changed that much in the 23 years I have been in the business?
Naaaaaahhhhhhh.
Gabe
I think nowadays the issue may be present only in case of
(a) really unstable internal cdp clock (throw the cheap junk away)
(b) external dac over spdif link with unstable transmitter clock and receiver that follows too closely this clock (get a decent external dac, the all-on-one-chip that has spdif on the left and analog out on the right costs about 5 bucks, so it's not rocket science to build one)
The rest consists of audiophile horror tales in which any perceived sound problem from bad recording to poor speakers / placement is blindly blamed on jitter.
(a) really unstable internal cdp clock (throw the cheap junk away)
(b) external dac over spdif link with unstable transmitter clock and receiver that follows too closely this clock (get a decent external dac, the all-on-one-chip that has spdif on the left and analog out on the right costs about 5 bucks, so it's not rocket science to build one)
The rest consists of audiophile horror tales in which any perceived sound problem from bad recording to poor speakers / placement is blindly blamed on jitter.
I have a CD player that uses a Philips CDM4/!9 and when playing original CDs there is a chirping noise from the mechanism which is disc speed related when I play copied CDs the noise is considerably reduced or not present at all. This suggests that there is a difference in the amount of work the servo does with different discs. I have not yet had a chance to investigate further
Stuart
Stuart
Gabevee said:
Gabe
All proper CD plaback systems retrieve data from the disc without errors, that is NOT the issue (it has been shown by measurements many times). Discs however contain jitter, and that amount differs disc by disc (Not only among various CDr discs but also among commercially made discs !).
That jitter ripples through and affects playback quality. Crosstalk of jitter takes place through power supplies and layout errors.
Finally the playback is affected once the conversion clock is affected as well (Note, I mean the clock quality ON-CHIP). At that moment a time error is converted in an amplitude error. That is why we hear differences.
I agree if you don't agree but would expect another explanation rather than stating "it is not the jitter".
best
I just wanted to put my two cents in, so here goes. A CD has two sources of jitter, as most of you know. There is inherent jitter in creating the original analog-digital transfer. Let's call that JO (original jitter). Next, when the transfer is burned to a CD, there is some additional jitter, due to any imprecision in the pit spacing/depth, etc. Let's call that JT (transfer jitter). So, when you play that disc on a standard CD (without buffering or anything fancy), the playback is subject to those two sources of jitter. Now, suppose instead that you rip the CD to your computer. Well, you have suddenly gotten rid of JT (not JO, of course). So, when you burn this copy to a new disc, you will create a new source of jitter. Let's call it JB (burn jitter). Clearly, if JB<JT, then this new disc may sound better (if you believe jitter is audible) than the orginal CD on the same cd player (or through the same DAC). If JB>JT, the CD could sound worse. It simply depends on whether the burner imparts less jitter than the original CD transfer had.
ezkcdude said:
Hi,
The JO does not expose as jitter but as errors present in the data due to jitter present in the conversion process. You will never get rid of that, never, so make sure your AD has lowest jitter possible (among other things).
Newly written discs may have lower pit jitter indeed, depends on the disc and writer. It is good to know that original discs differ in jitter too as the mould ages.....
In addition, other parameters affect read-out jitter like a-centricity and reflections from the upper surface.
best
Just to add oil to the fire...
Straight from the EAC homepage
Known Bugs:
Some older CD-ROMs cause extreme jitter errors to occur on some CD-ROM drives causing the extraction to slow down to an unacceptable rate (e.g. Mitsumi FX240)
Straight from the EAC homepage
Known Bugs:
Some older CD-ROMs cause extreme jitter errors to occur on some CD-ROM drives causing the extraction to slow down to an unacceptable rate (e.g. Mitsumi FX240)
Data isn't just poured from pits and lands to the dac. You must decode it (from efm to octets) and run it through error detection and correction (rs codes). Not sure you can do all this with absolutely no buffering. So much about this kind of jitter.
---
When you rip using DAE the jitter it another kind of jitter. Audio sectors do not include address information (header) so you cannot be sure you got the right sector when you start a read command. You mai get data from the previous sector for example. This used to happen on old cdrom units but it is no longer the case with modern ones (don't ask me how they do it).
This does not apper at audio playback because you just play continuosly, not do read-sector-save-to-hdd loop.
---
So, here we are stuck mostly with recording jitter from studio ADC conversion. This one remains constant no matter what you do. Hopefully they have better gear than we do.
---
When you rip using DAE the jitter it another kind of jitter. Audio sectors do not include address information (header) so you cannot be sure you got the right sector when you start a read command. You mai get data from the previous sector for example. This used to happen on old cdrom units but it is no longer the case with modern ones (don't ask me how they do it).
This does not apper at audio playback because you just play continuosly, not do read-sector-save-to-hdd loop.
---
So, here we are stuck mostly with recording jitter from studio ADC conversion. This one remains constant no matter what you do. Hopefully they have better gear than we do.
I am not sure that this jitter discussion really helps correlate what I'm hearing.
Now that I have done a few tests, the results are getting predictable.
As in, give me two recordings - original and copy, and I will tell you which is which. Same with 1st copy and second copy (of the 1st copy)
Also, if you have two otherwise identical hard drive images, I can tell apart two burnings made at different write speeds and identify the higher and lower.
This is true of both CD burners I've tried, with totally different burners, hardware, software, and locations.
That kind of empiric reproducibility must have a testable basis - like compare the digital signals being read on the player to maybe discover whether this discussion on jitter really means anything- is that possible?
Now that I have done a few tests, the results are getting predictable.
As in, give me two recordings - original and copy, and I will tell you which is which. Same with 1st copy and second copy (of the 1st copy)
Also, if you have two otherwise identical hard drive images, I can tell apart two burnings made at different write speeds and identify the higher and lower.
This is true of both CD burners I've tried, with totally different burners, hardware, software, and locations.
That kind of empiric reproducibility must have a testable basis - like compare the digital signals being read on the player to maybe discover whether this discussion on jitter really means anything- is that possible?
Can you do this in a _blind_ test using a new cdrw unit and a recent (and not the cheapeste available) standalone player?
Why the question... for one older or used cdrw units may experience problems at audio extraction. Than it is not uncommon for old standalone players to not like at all burned cd's (they have lower reflectivity than pressed ones) and experience serious read errors.
Ripping can be done without errors, once I ripped 3 times one audio cd and the images were bit-identical. Burning not an issue at 4x or 8x if the blank is not the cheapest one. Then the only thing that can make a difference is the player and if in such conditions it experiences audible read errors, well, it has a problem.
Why the blind test... human brain is _very_ good at auto-suggestion. If it sees the copy it will know for sure that it must sound worse.
Why the question... for one older or used cdrw units may experience problems at audio extraction. Than it is not uncommon for old standalone players to not like at all burned cd's (they have lower reflectivity than pressed ones) and experience serious read errors.
Ripping can be done without errors, once I ripped 3 times one audio cd and the images were bit-identical. Burning not an issue at 4x or 8x if the blank is not the cheapest one. Then the only thing that can make a difference is the player and if in such conditions it experiences audible read errors, well, it has a problem.
Why the blind test... human brain is _very_ good at auto-suggestion. If it sees the copy it will know for sure that it must sound worse.
"Ripping can be done without errors, once I ripped 3 times one audio cd and the images were bit-identical. Burning not an issue at 4x or 8x if the blank is not the cheapest one. Then the only thing that can make a difference is the player and if in such conditions it experiences audible read errors, well, it has a problem."
Is this assessment done with the computer's cd drive?
And yes, my player is quite old. Perhaps it is all due to "the pits".
"Why the blind test... human brain is _very_ good at auto-suggestion. If it sees the copy it will know for sure that it must sound worse."
I didn't say it sounds worse, and often times it sounds better. The thing is, the chararcter of the sound changes is a particularly reproducable way. It is always less detailed, but can be cleaner, and this can sometimes mean that instruments are easier to distinguish, improved bass distinction etc... I'm repeating gabavee but I agree...
As for "suggestion", I'm the world's worst critic around objectivity, but I've found that the advice "don't believe anything but your ears" is the only way to listen to mods and to get anywhere in audio.
Is this assessment done with the computer's cd drive?
And yes, my player is quite old. Perhaps it is all due to "the pits".
"Why the blind test... human brain is _very_ good at auto-suggestion. If it sees the copy it will know for sure that it must sound worse."
I didn't say it sounds worse, and often times it sounds better. The thing is, the chararcter of the sound changes is a particularly reproducable way. It is always less detailed, but can be cleaner, and this can sometimes mean that instruments are easier to distinguish, improved bass distinction etc... I'm repeating gabavee but I agree...
As for "suggestion", I'm the world's worst critic around objectivity, but I've found that the advice "don't believe anything but your ears" is the only way to listen to mods and to get anywhere in audio.
Agreed, but you must do it under controlled conditions, otherwise you are listening with your eyes, ears, brain and so on. Psycho-acoustics is a powerful force. Which brings us to DBT (Double Blind Testing).
ezkcdude summing up
best [/QUOTE]
IMHO, this is a really fair synthesis and modelization of all this thread.
Thank you,
Stefano
...ezkcdude said:
best [/QUOTE]
IMHO, this is a really fair synthesis and modelization of all this thread.
Thank you,
Stefano
Dumb question, but don't most modern CD players use an output buffer for anti-skip protection or to allow for error correction? Most don't even advertize it any more as far as I know, but wouldn't this render most of the discussion about "read-jitter" moot?
Just as an aside, [rhetorical question] if I were to copy my wordprocessor documents over to another CD, would the text somehow change if the jitter was different on the burned copy? [/rhetorical question] I guess I don't see the difference for audio CD's. Assuming computer encoding and decoding, I am not sure how the bits change from one disk to another.
I personally have never noticed a difference, but like I have said in other threads, my gear is probably not up to what some other's standards! 😀
Just as an aside, [rhetorical question] if I were to copy my wordprocessor documents over to another CD, would the text somehow change if the jitter was different on the burned copy? [/rhetorical question] I guess I don't see the difference for audio CD's. Assuming computer encoding and decoding, I am not sure how the bits change from one disk to another.
I personally have never noticed a difference, but like I have said in other threads, my gear is probably not up to what some other's standards! 😀
dfdye said:
every player has a small buffer, but a buffer is a buffer, not a jitter eliminator. And if it where, other mechanisms are still present to assure pit jitter riples through here and there.......
cheers
Data mode has an additional ECC step than audio mode (out of 2352 octets/sector only 2048 are used for data, rest is header and ecc), so data mode is more reliable at the expense of smaller capacity.
But this does not prevent a modern transport to read an audio cd without errors (just try ripping a cd a couple of times and compare the images)
But this does not prevent a modern transport to read an audio cd without errors (just try ripping a cd a couple of times and compare the images)
burning cd's
i have burnt many, many, many cd's... best luck i've had is to set your computer writer to 1x (single speed) even though they can burn faster and to use black blank cdr's. can be hard to find the black ones but memorex and a couple others make them in 50 pack spindles... i use a top quality yamaha cd burner and a teac cd-rom. try to avoid the cheap cd-burners.
as for audio cd's they are the same but the extra money you pay for the disks go to musical artists for copying their music...
i have burnt many, many, many cd's... best luck i've had is to set your computer writer to 1x (single speed) even though they can burn faster and to use black blank cdr's. can be hard to find the black ones but memorex and a couple others make them in 50 pack spindles... i use a top quality yamaha cd burner and a teac cd-rom. try to avoid the cheap cd-burners.
as for audio cd's they are the same but the extra money you pay for the disks go to musical artists for copying their music...
I have followed this thread.
Let's get one thing straight. A CD is nothing like a vinyl album. The information on a CD is interleaved. The laser reads the information in chunks and not in a constant stream like a needle. A CDM functions roughly like a hard drive. The speed isn't constant. It neither can nor need to be. The FIFO devise and buffer take care of the chunks of data. Based on that I guess it's time to chalk the 47 Laboratory PiTracer, Philips CD-Pro2, Teac VRDS, and the various belt-drive CDMs out there up to snake oil. The belt-drive CDM is especially stupid. The belt-drive is extremely poor at the rapid speed changes. Ask Meridian what kind of CDM drives they use. The price of ignorance is murder.
Junji Kimura of 47 Laboratory has said that CDs are so poorly made they shouldn't even be played. CDs are, after all, manufactured by the lowest bidder. But the interesting question is: Is it the fault of the PiTracer that it refuses to play about 1 out of 10 CDs, or is it the fault of the CDs?
Now I cannot speak for the difference between original CDs and copies. But if we try to make a deduction based on what we know, it would seem that the difference is those worthless original CDs. After all, the information is exactly the same on both.
While it would seem there's no such thing as "pre FIFO jitter," asynchronous playback seems to be the only hope for digital formats. Everything else is damaged goods.
Let's get one thing straight. A CD is nothing like a vinyl album. The information on a CD is interleaved. The laser reads the information in chunks and not in a constant stream like a needle. A CDM functions roughly like a hard drive. The speed isn't constant. It neither can nor need to be. The FIFO devise and buffer take care of the chunks of data. Based on that I guess it's time to chalk the 47 Laboratory PiTracer, Philips CD-Pro2, Teac VRDS, and the various belt-drive CDMs out there up to snake oil. The belt-drive CDM is especially stupid. The belt-drive is extremely poor at the rapid speed changes. Ask Meridian what kind of CDM drives they use. The price of ignorance is murder.
Junji Kimura of 47 Laboratory has said that CDs are so poorly made they shouldn't even be played. CDs are, after all, manufactured by the lowest bidder. But the interesting question is: Is it the fault of the PiTracer that it refuses to play about 1 out of 10 CDs, or is it the fault of the CDs?
Now I cannot speak for the difference between original CDs and copies. But if we try to make a deduction based on what we know, it would seem that the difference is those worthless original CDs. After all, the information is exactly the same on both.
While it would seem there's no such thing as "pre FIFO jitter," asynchronous playback seems to be the only hope for digital formats. Everything else is damaged goods.
As a pro, regarding digital signal conversion, I have some comment to this discusion about copying music to CD /HDD
When playing a CD on a normal player, there are some bit errors, making necessary for re-read the disc (read: causing delays in conversion), that the player correct by interpolation each time there is a read error, this is a natural part of playing a CD, however when once recorded on a new CD media or HDD, the bits of the original material have been corrected, and all read errors have been corrected when stored on a new media, when re-playing from the new media CD/HDD, there are no errors, and no need for re-read, causing no delays, and lower jitter. Of course a copied CD will by time also get scratches, and need re-reads, here the Hard drive wins.
I am working on a project (professionally) with a high-end Integrated amplifier with built-in Harddisk, for High End music reproduction.
When playing a CD on a normal player, there are some bit errors, making necessary for re-read the disc (read: causing delays in conversion), that the player correct by interpolation each time there is a read error, this is a natural part of playing a CD, however when once recorded on a new CD media or HDD, the bits of the original material have been corrected, and all read errors have been corrected when stored on a new media, when re-playing from the new media CD/HDD, there are no errors, and no need for re-read, causing no delays, and lower jitter. Of course a copied CD will by time also get scratches, and need re-reads, here the Hard drive wins.
I am working on a project (professionally) with a high-end Integrated amplifier with built-in Harddisk, for High End music reproduction.
- Status
- Not open for further replies.