Prompted by discussion over in the Blowtorch thead over error correction in CD players, I thought a little experiment was called for. I have a pair of
official CD test discs, one is bit perfect (to Red Book standard), the other has manufactured defects present. The content of each is identical. So the
idea for a little test presented itself, one that you guys that like looking at and comparing files might enjoy.
This shows the disc and the track chosen is the one at greatest interuption to the recorded layer.
There is a gap of 1000um. According to Philips, any player should cope with up to 400u. So this is pushing things. The servo will be working overtime and
so will the error correction.
The track is such that the defect begins 3 seconds into the track and finishes at 2 minutes 51 seconds in. Those timings are only exact on the ripped version, the recorded tracks have a little leeway due to starting the recording manually.
Attached are two zipped folders, one called 'Ripped WAV' which is a rip of the two tracks made on a laptop using Windows Media Player and the other called
'Real Time Playback' which is a recording made from a very late 1980's early 90's Sony silver spinner, a midrange CDP790.
Here are the files,
Ripped WAV
Real Time Playback
official CD test discs, one is bit perfect (to Red Book standard), the other has manufactured defects present. The content of each is identical. So the
idea for a little test presented itself, one that you guys that like looking at and comparing files might enjoy.
This shows the disc and the track chosen is the one at greatest interuption to the recorded layer.
There is a gap of 1000um. According to Philips, any player should cope with up to 400u. So this is pushing things. The servo will be working overtime and
so will the error correction.
The track is such that the defect begins 3 seconds into the track and finishes at 2 minutes 51 seconds in. Those timings are only exact on the ripped version, the recorded tracks have a little leeway due to starting the recording manually.
Attached are two zipped folders, one called 'Ripped WAV' which is a rip of the two tracks made on a laptop using Windows Media Player and the other called
'Real Time Playback' which is a recording made from a very late 1980's early 90's Sony silver spinner, a midrange CDP790.
Here are the files,
Ripped WAV
Real Time Playback
How have you determined that any differences are error correction rather than error concealment? E32 monitor?
I don't know that there are any differences, but I know a lot here like analysing these kind of things, so I was rather hoping they could tell me 😀
Good question though, I've looked at the chipset but couldn't find any error flag points that were accessible.
Good question though, I've looked at the chipset but couldn't find any error flag points that were accessible.
Should just mention that I do know which track is which (the track with errors vs the clean) but I thought it would be fun to see that if there were any differences when analysed, whether those differences could be pinned to a particular track by analysis (and of course listening) alone.
You need to determine this at the time of playback. I strongly suspect that you're not testing what you think you're testing.
You also might want to go back and search some of the literature from 30+ years ago when all this stuff was determined.
A CD track really is merely just one number, albeit a very big one. Successful EC recovers that number exactly, and then in principle errors aren't audible. However, if the servo is working hard then PSU feedthrough to analog conversion or clock gen etc might influence audibility I suppose, but that might well depend on the transport and circuit and not be repeatable between players I think.
If EC doesn't recover number exactly then it generally doesn't just miss, the result is a junk sector, and so interpolation algorithms are used to try to disguise the effect. But such things one might reasonably expect to be audible, especially if they are periodic ie once per rev.
If EC doesn't recover number exactly then it generally doesn't just miss, the result is a junk sector, and so interpolation algorithms are used to try to disguise the effect. But such things one might reasonably expect to be audible, especially if they are periodic ie once per rev.
FWIW: Ripped files - no difference between the two. Zero amplitude, straight line.
Downloading real time ones right now.
Last edited:
Hi,
Drilling a 2 mm hole in a CD should not affect playback.
Most errors from CD are caused by track skipping issues.
rgds, sreten.
Drilling a 2 mm hole in a CD should not affect playback.
Most errors from CD are caused by track skipping issues.
rgds, sreten.
Part 2
BIG difference between Y.wav and Z.wav files!
First, despite the same lenght, the two do not align. I've searched for the peak at the very beginning of the files and, by cutting some samples, was able to align the two, BUT the difference is not the straight line 🙁
What capture software are you using, Mooly?
BIG difference between Y.wav and Z.wav files!
First, despite the same lenght, the two do not align. I've searched for the peak at the very beginning of the files and, by cutting some samples, was able to align the two, BUT the difference is not the straight line 🙁
What capture software are you using, Mooly?
Part 3
What is also interesting is the amplitude differences at the below mentioned peak location.
Something is rotten in Denmark...
P.S. Before anyone jumps to any conclusions - first you have to make sure that your playback and capture devices are 'bit transparent'. And that means throughout testing of every single device, and every new software version!
What is also interesting is the amplitude differences at the below mentioned peak location.
Something is rotten in Denmark...
P.S. Before anyone jumps to any conclusions - first you have to make sure that your playback and capture devices are 'bit transparent'. And that means throughout testing of every single device, and every new software version!
Attachments
Last edited:
I used Audacity. The two files were made consecutively, only difference is swapping one disc for the other.
Perhaps as a test of the real time method two identical tracks could be compared i.e. recording the same track twice.
Could clock drift as the player warms account for slight misalignment ?
The amplitudes between ripped and recorded wasn't matched. That's not easy to do without manipulating files and I felt that whether the difference is 0.1db or 10db it would always be picked up on. It was really just whether differences between the two were apparent.
As far as I can tell its incorporated directly into the reflective layer. The label side has heavy printing so its impossible to 'look through'. The dots and simulated fingerprint are added onto the surface and cast shadows on the reflective layer when viewed with a torch but not the gaps.
I do tests like this to see how well a disc will stand-up to abuse. It also gives some idea of how good the different production processes are. Not once did I consider that this was audible up to the point of dropouts.
No, if you've captured CD player's digital output. You did transfer over SPDIF, right?
Yes, the amplitudes shouldn't be matched. What surprises me is that there are amplitude differences between the two captured files and between those two and computer based rip.
Why?
Error correction happens inside the decoder IC (CXD2500 in your case), there's no need to go thru several analog stages and another sampling process which might mask the effect you're trying to uncover.
Audioquest LaserGuide comes to mind..
All the talk of freezing CD's was the next one..I'm not saying its right or wrong..the idea was the stress in the Cd material..
So yes I put them in the freezer for a laugh..then thought it gets colder outside in the winter so I put them outside in a plastic bag..😀
Thought well it can't hurt because they probably get this in transit..
My friend hadn't been told what I had done because I thought he would laugh..and he borrowed a few..then liked the tracks so he bought them..he said he couldn't understand why his copys were different and asked to look at the disks again..LMAO
More of the snake me thinks..however it drove him nuts he bought another copy so I told him what I had done...he didn't believe me..I don't either it must be my disks were older production.
He had a full Linn system..probably why I don't see how an optical link could sound different..
Regards
M. Gregg
All the talk of freezing CD's was the next one..I'm not saying its right or wrong..the idea was the stress in the Cd material..
So yes I put them in the freezer for a laugh..then thought it gets colder outside in the winter so I put them outside in a plastic bag..😀
Thought well it can't hurt because they probably get this in transit..
My friend hadn't been told what I had done because I thought he would laugh..and he borrowed a few..then liked the tracks so he bought them..he said he couldn't understand why his copys were different and asked to look at the disks again..LMAO
More of the snake me thinks..however it drove him nuts he bought another copy so I told him what I had done...he didn't believe me..I don't either it must be my disks were older production.
He had a full Linn system..probably why I don't see how an optical link could sound different..
Regards
M. Gregg
Last edited:
- Status
- Not open for further replies.