Page 21 of 33

PostPosted: Wed Jun 18, 2003 10:49 pm
by dodecahedron
KCK wrote:BTW, I've been wondering why this thread has not achieved a "sticky" status so far. :roll: :P
me too! (for quite some time actually...)

PostPosted: Wed Jun 18, 2003 10:50 pm
by Ian
No reason really. It was always near the top anyway.

PostPosted: Wed Jun 18, 2003 10:52 pm
by dodecahedron
along the same lines, the CD Doctor thread might as well be "unstickyfied" now... :o

PostPosted: Fri Jun 20, 2003 5:28 am
by MediumRare
Matt- I think that would be a super idea and hope that Karr Wang is willing to work with you/us. I'll certainly contribute anything I can (although I'm trying to limit my forum time :o ). A separate forum would help to organize the information too- instead of having one record-breaking thread.

Ian/dodecahedron: having this thread "sticky" makes it a bit easier to find :D . I think the CD-Doctor thread should stay sticky, though, since it's the canonical reference for this program (at least in the English speaking world).

dolphinius_rex wrote:heh, I never considered the format of the disc to affect the read speed!

That really surprised me too when I ran across it. See CD Speed with video CDs (one of my first contributions) for more info. I don't know if the older LiteOn's do this too.

G

PostPosted: Fri Jun 20, 2003 5:28 pm
by Inertia
MediumRare wrote:
dolphinius_rex wrote:heh, I never considered the format of the disc to affect the read speed!

That really surprised me too when I ran across it. See CD Speed with video CDs (one of my first contributions) for more info. I don't know if the older LiteOn's do this too.G


(S)VCD discs do normally read slower than data discs. See my explanation for this behavior which I have added to your CD Speed with video CDs thread. :)

Strange difference between KProbe/CDSpeed testing results

PostPosted: Sat Jun 21, 2003 11:48 am
by abgm
I'm testing one old RW disc - Princo 2x.
And when I pass three different testing programs, I get very strange results.
1. Nero CD Speed 1.02e
Image
2. CDDoctor 1.04
Image
3. KProbe 1.1.14
Image

As you can see, C2 level in KProbe/CDDoctor graph are similar, but very differ from CDSpeed results. In my opinion, "errors" in CDSpeed is C2 errors. But why CDSpeed's C2 - 5877 and KProbe's C2 - only 1?!

PostPosted: Sat Jun 21, 2003 11:58 am
by rdgrimes
abgm
Your results are not really inconsistant. CDSpeed is a very different program with different drive controls and reporting. If you run the same disc through the same program several times you will see similar variations. In KProbe, you need to slow the read speed to get improved C2 reporting, as it skips a lot of the errors at high speed.

PostPosted: Sun Jun 22, 2003 10:13 pm
by abgm
to rdgrimes
In KProbe, you need to slow the read speed to get improved C2 reporting, as it skips a lot of the errors at high speed

As you can see, C2 errors in CDSpeed graph appear in the beginning of disc, where linear speed is only 20x, and when I testing with KProbe/CDDoctor the speed is the same.
CDSpeed is a very different program with different drive controls and reporting

According Nero DAE Quality Test, Lite-ON-based drives have 100% accuracy of C2-results, so, in my opinion, the same 100% accuracy can have CDSpeed results.
If you run the same disc through the same program several times you will see similar variations

Yes, certainly, I passes this tests many times :wink:. And results of passes has same variations... But sizes of results are kept.

I think, that it does not speak about discrepancy of the program KProbe, but maybe "vendor unique command" gives out other values then "standart MMC command for C2", that CDSpeed used?

P.S. Sorry, my English is too poor :(

Re: Strange difference between KProbe/CDSpeed testing result

PostPosted: Sun Jun 22, 2003 10:46 pm
by Inertia
abgm wrote:As you can see, C2 level in KProbe/CDDoctor graph are similar, but very differ from CDSpeed results. In my opinion, "errors" in CDSpeed is C2 errors. But why CDSpeed's C2 - 5877 and KProbe's C2 - only 1?!


In spite of your opinion the 5,877 "errors" shown in the CD Speed: CD Qualilty Check are hardly likely to be C2 errors. KProbe and CD Doctor do measure C2 errors, and their close correlation gives reassurance of a reasonably accurate measurement.

Even the CD Speed help file states that the ScanDisc test is better suited for error checking with a CD recorder. Have you tested the same disc with CD Speed ScanDisc? I think that you might find all green blocks except for one or two yellow blocks at the beginning indicating where the one or two C2 errors are located.

Please check the disc with CD Speed ScanDisc and let us know the results.

PostPosted: Sun Jun 22, 2003 10:50 pm
by rdgrimes
As you can see, C2 errors in CDSpeed graph appear in the beginning of disc,

CDSpeed does not force a constant speed through the scan, KProbe and CDDr do. In Kprobe, when the drive is unable to read a sector, it skips ahead, thereby missing the erorr, and any more errors that might fall into that sector. The speed is not the issue so much as the skipping, the higher the read-speed, the more skipping occurs. CDSpeed allows the drive to re-read, and slow if needed, to read the sector, but will still report the error. However doing that also changes the error count on the sector in comparison to the others.
Lowering the read speed in Kprobe eliminates most of the skipping that occurs, but also means fewer errors reported over-all.
They are different tools, CDSpeed more closely duplicates the drive's behaviour in actual use, where KProbe is a pure testing tool that eliminates most of those variables. The relative speed at the start of the discs compared to the end is not so much the issue as is the different speeds in any part of the disc when you compare, say 52x to 8x. I use them according to those differences, depending on the nature of the disc I'm testing. These differences in the testing tools account for most of the differences in error counts. The actual number is not significant except when compared to the number on a different disc or read speed, or burn speed.....whatever.
BTW, many, if not all, RW disc show a spike of errors at the start of the disc.

PostPosted: Mon Jun 23, 2003 2:40 am
by abgm
Thank big for your answers!

Inertia wrote:CD Qualilty Check are hardly likely to be C2 errors
But what it then :( ?
ErikDeppe wrote:The yellow lines in the CD Quality check represent the number of C2 errors per second

Inertia wrote:and their close correlation gives reassurance of a reasonably accurate measurement

IMHO, "gives reassurance" that both programs (KProbe and CD Doctor) use the identical mechanism of reception of C2-errors :wink:, distinguished from the program CDSpeed.
Inertia wrote:Please check the disc with CD Speed ScanDisc and let us know the results
OK. I shall try.

rdgrimes wrote:In Kprobe, when the drive is unable to read a sector, it skips ahead
In KProbe - yes, but, AFAIK, in CDDoctor No. This program try reread such sector again, slowdown reading speed, etc. Besides, where on diagrams there are unreadable sectors :) ?
rdgrimes wrote:They are different tools, CDSpeed more closely duplicates the drive's behaviour in actual use, where KProbe is a pure testing tool that eliminates most of those variables
Interesting, you want to tell, what their results basically are non-comparable?
rdgrimes wrote:BTW, many, if not all, RW disc show a spike of errors at the start of the disc.
I know :D , but I'm interesting why these errors are visible on CDSpeed's diagrams and would be absent on KProbe/CDDoctor diagrams? Only because of different speeds of reading?

PostPosted: Mon Jun 23, 2003 3:41 am
by Inertia
abgm,

If you want a useful comparison of the results provided by the three programs, you should use a CD-R and not a CD-RW disc. CD-RW discs can give spurious results, particularly with CD Speed.

PostPosted: Mon Jun 23, 2003 3:53 am
by abgm
Thank for the help, Inertia! But it would be desirable to understand a physical explanation of this phenomenon :wink: .

PostPosted: Mon Jun 23, 2003 10:39 am
by abgm
Well, next diagrams:) !

RW Princo 2x. Write: Sony185@LTR-40125S-XSU1 2x, read: Sony210@LTR-48246S-SS0E max (40x).

1. Nero CD Speed | CD Quality Test:
Image

2. Nero CD Speed | ScanDisc | Surface Scan:
Image

3. CDDoctor 1.04:
Image

3. KProbe 1.1.14:
Image

Any comments?

PostPosted: Fri Jun 27, 2003 12:12 am
by abgm
All thank for the help! I have calmed down a little, after reading ErikDeppe's explanations http://www.cdrlabs.com/phpBB/viewtopic.php?t=9237. And while the questions I take off :) .

PostPosted: Mon Jul 07, 2003 5:20 pm
by KCK
Apparently setting the speed in KProbe v1.1.14 doesn't work as expected.

For my LTR-48125W VS0D, when the reading speed is set to Max, everything looks fine: an almost full 80 min disc is read in 2:15 min (I start timing after seeing the "Testing..." message).

However, it takes 4:56 min for the reading speed of 24x, and 19:43 min for the reading speed of 8x (as if the reading speed were 4x).

This disc is not perfect, but no C2 errors are detected, and I can't hear any slowdowns in reading.

I would appreciate if other users reported similar timings for their drives.

BTW, since the forum was down, I posted a similar message at CDFreaks

http://club.cdfreaks.com/showthread.php ... post440797

BoSkin confirmed that 8x works like 4x, but he didn't specify his drive.

PostPosted: Mon Jul 07, 2003 7:27 pm
by dolphinius_rex
I odn't have any testing speed numbers, but I too have noticed something not quite right with the length of time it takes to do the testing. I couldn't put my finger on it though, so I ignored it until now... I'll pay a little more attention next time!

PostPosted: Mon Jul 07, 2003 8:34 pm
by KCK
dolphinius_rex:

For me, KProbe is mostly of academic interest, and I had always run it at max speed until I saw some recent results of Halcyon, which got me thinking about reading speeds. Thus I find it hard to believe that KProbe's running at about half of the desired speed has gone unnoticed so far.

Before complaining to Karr Wang, we should determine whether this is a real bug which affects most burners.

Your results should be comparable to mine, since similar burners are involved.

Owners of other drives (e.g., LTR-52246S), wake up! :roll: :D

PostPosted: Mon Jul 07, 2003 11:39 pm
by dolphinius_rex
Okay, I tested 1 CD-R at every possible speed (and a few impossible speeds it seems :wink: ), here are the results:

Code: Select all
Disc Length: 79m 31s 31f

Speed Tested  |  Time Taken
@Max(48x)     |     2:23
@    48x      |     2:43
@    40x      |     3:26
@    32x      |     4:27
@    24x      |     4:57
@    16x      |     6:36
@    12x      |     9:55
@     8x      |     19:51
@     4x      |     19:50
@     1x      |     19:50

PostPosted: Tue Jul 08, 2003 12:13 am
by KCK
Are you also using VS0D on your LTR-40125S?

Your results agree with mine, and indicate bugs. We should expect something close to CLV for 24x, and definitely CLV for 16x. Now, for CLV, the time should be 3:20 for 24x (like your 40x), 5:00 for 16x (like your 24x), 6:40 for 12x (like your 16x), 10:00 for 8x (like your 12x), and 20:00 for 4x (like your 8x). The tail may mean that 4x is the slowest speed, but otherwise there is a "one-notch-below" pattern.

Definitely something Karr Wang should look into.

PostPosted: Tue Jul 08, 2003 12:18 am
by dolphinius_rex
yup, VS0D on my 40125S :D

I agree, Karr Wang should be informed. Since you were the one to catch this big, why don't you do the honors?

PostPosted: Tue Jul 08, 2003 9:33 am
by rdgrimes
Confirm that 8x read is taking 20 min.

PostPosted: Tue Jul 08, 2003 2:10 pm
by dolphinius_rex
By the way, the 48125W doesn't support CLV for 16x writing, only 12x and below. The funny thing is that 24x writing starts at 16x, so 16x CLV *IS* possible with the LiteON 48125W, but not implemented :cry:

I wonder why...

PostPosted: Tue Jul 08, 2003 2:27 pm
by dodecahedron
it's to keep the rotation speed down as much as possible.

this is seen in other drives too.

PostPosted: Tue Jul 08, 2003 2:32 pm
by cfitz
KCK wrote:I find it hard to believe that KProbe's running at about half of the desired speed has gone unnoticed so far.

That's because it hasn't. :wink: Well, at least it hasn't gone unnoticed for very long. This is a new bug that was introduced in 1.1.13 (which wasn't available for very long before it was replaced with 1.1.14).

I was a little surprised to read the comments about KProbe testing at half speed (i.e. 4x instead of 8x), because although I normally only test with the MAX box checked, I have tested in the past at 8x and it seemed to take the right amount of time. However, I never sat down with a stopwatch and timed it, so I thought I would give it a go. Sure enough, as you reported, KProbe 1.1.14 is testing at speeds slower than the speed set in the configuration box. On my LTR-48246S, KProbe 1.1.14 tests using the correct speed when the MAX box is checked, but takes 19:51 to test a 79:27 disc (an average rate of 4x) at the 8x setting.

This didn't square with what I remembered in previous tests at 8x, and until today I hadn't done any tests at anything other than MAX using 1.1.14, so I reverted to 1.1.13 and 1.1.12 to see how they do at other than MAX. Version 1.1.13 also has the bug, but with 1.1.12 everything tests out fine:

Code: Select all
KProbe 1.1.12 on LTR-48246S  SS0E

set speed    elapsed time   actual speed    C1 avg
----------------------------------------------------
   4x          19:51            4x CLV       0.157
   8x           9:56            8x CLV       0.165
  12x           6:37           12x CLV       0.195
  16x           4:57           16x CLV       0.144
  24x           4:38           24x CAV       0.403
  32x           3:20           32x CAV       0.684
  40x           2:37           40x CAV       0.767
  48x           2:14           48x CAV       0.262
  MAX           2:14           48x CAV       0.264


In this rare case my memory was correct. :D Older versions did test at the proper speed as set, and the problem is a new one introduced in 1.1.13.

Alexnoe over at CDFreaks offers an explanation of how it may be possible to set the speeds mistakenly because the speeds are set in units of bytes/sec rather than units of "x":

http://club.cdfreaks.com/showthread.php ... post437828

Until Karr corrects this, you can either go back to an older version, or set your speed to the next level up.

cfitz

<edit>
I got a copy of 1.1.13 and was able to test it also and confirm that it too has the bug. I have updated this post accordingly.
</edit>