IPB

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Reply to this topicStart new topic
EAC T&C without C2. Overkill?
liekloo
post Nov 20 2002, 16:51
Post #1





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



If you rely on C2 info when using EAC, using 'test&copy' is better than just 'copy'.

What if you don't use C2? EAC reads each sector twice and, as a result, has the oportunity to calculate CRC twice, and compare them.


Is that correct? and is the opportunity actually used? (that would mean test&copy is overkill if you disable C2)


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
cmyden
post Nov 20 2002, 17:05
Post #2





Group: Members
Posts: 47
Joined: 7-June 02
Member No.: 2242



What if you don't use C2? EAC reads each sector twice and, as a result, has the oportunity to calculate CRC twice, and compare them.


Is that correct? and is the opportunity actually used? (that would mean test&copy is overkill if you disable C2)


Yes, this is correct. Providing that....

1) You are in secure mode, and your drive supports accurate stream. (most do)

2) You have the correct cache setting. To be safe, leave the checkmark on beside 'this drive caches audio data'. If you know for a fact your drive does *not* cache audio data you can take the checkmark off.

chris

This post has been edited by cmyden: Nov 20 2002, 17:05
Go to the top of the page
+Quote Post
Annuka
post Nov 20 2002, 17:25
Post #3





Group: Members
Posts: 333
Joined: 2-February 02
Member No.: 1233



EAC is not exact at all, if your CD-ROM drive is old or slightly broken.

My old Plextor 32TSi works fine for CD-ROMs, but makes a random error in 5% of the albums ripped at full speed. Reducing the speed to 4X reduced the errors to 1%. My new Plextor 40TSi has ripped hundreds of CDs at 17X CAV without any problems.

Using Test&Copy for the first 50-100 cds might not be a bad idea -- until you are convinced your drive works as it should.
Go to the top of the page
+Quote Post
liekloo
post Nov 20 2002, 18:16
Post #4





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



Thanks for your replies smile.gif


To cMyden:

(Yes, I should have been more carefully, of course "secure mode" is needed.)

Cache setting should be set correctly anyway, and I do understand why it is needed (BTW is determining your drive's caching properties such a tricky thing?). But why the hell is accurate stream a must? (BTW my drive has accurate stream, but I am just curious) wink.gif

This post has been edited by liekloo: Nov 20 2002, 18:17


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
Pio2001
post Nov 20 2002, 21:48
Post #5


Moderator


Group: Super Moderator
Posts: 3936
Joined: 29-September 01
Member No.: 73



QUOTE (liekloo @ Nov 20 2002 - 06:51 PM)
If you rely on C2 info when using EAC, using 'test&copy' is better than just 'copy'.

What if you don't use C2? EAC reads each sector twice and, as a result, has the oportunity to calculate CRC twice, and compare them.

With C2 info, test and copy is better than copy.

Without C2, test and copy gives you the "cache" secureness.

But beware that CRC "OK" with test and copy is the equivalent of "no errors occured - 100% quality". It is not equivalent to "no errors occured - 99.9% quality", it is superior.

Last, test and copy in burst mode, as secure as secure mode due to the CRC checking, can also read error free some CDs that show a few errors in secure mode, because the groove is read in one pass, and no seeking has to be performed by the drive, that may introduce read errors (that should be detected) on bad CDs.
Go to the top of the page
+Quote Post
spoon
post Nov 20 2002, 23:46
Post #6


dBpowerAMP developer


Group: Developer (Donating)
Posts: 2745
Joined: 24-March 02
Member No.: 1615



If I can hijack this thread -

I am currently tackling the age old problem of getting 100% accurate rips from Audio CDs, I have called my implementation AccurateRip, I will keep my cards close to my chest at the moment as to the technical details, but I will go as far as to say when the system is fully operational it will be the first program in the world to guarantee 100% accurate rips, or tell you the rip was not accurate...I hope to beta within 2 weeks.

Beta Testing will take place in the Beta section of the forum on dBpowerAMP.com


--------------------
Spoon http://www.dbpoweramp.com
Go to the top of the page
+Quote Post
liekloo
post Nov 21 2002, 15:56
Post #7





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



Could anyone please explain why accurate stream is important too (apart from no caching).

The relation CRC - accurate stream isn't completely clear to me. unsure.gif


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
Case
post Nov 21 2002, 17:15
Post #8





Group: Developer (Donating)
Posts: 2214
Joined: 19-October 01
From: Finland
Member No.: 322



QUOTE (liekloo @ Nov 21 2002 - 04:56 PM)
Could anyone please explain why accurate stream is important too (apart from no caching).

Accurate stream means that drive will not produce jitter errors when extracting (unwanted clicks/pops), if drive doesn't support it EAC needs to sync sector boundaries and this is somewhat slower. Btw, I don't know of any no-name regular CD-ROM drive that supports it.
Go to the top of the page
+Quote Post
liekloo
post Nov 21 2002, 17:20
Post #9





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



I am sorry, I apparently didn't put my question clearly enough

When I asked "Why is accurate stream important too, apart from caching", I meant relating to doing CRCs.

Thanks for answering though biggrin.gif


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
Case
post Nov 21 2002, 17:24
Post #10





Group: Developer (Donating)
Posts: 2214
Joined: 19-October 01
From: Finland
Member No.: 322



QUOTE (liekloo @ Nov 21 2002 - 06:20 PM)
When I asked "Why is accurate stream important too, apart from caching", I meant relating to doing CRCs.

It has no effect in CRC calculations.
Go to the top of the page
+Quote Post
liekloo
post Nov 25 2002, 19:38
Post #11





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



So, to conclude: in secure mode, with a properly configured cache setting, test&copy should be overkill, even with scratched CDs...


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
liekloo
post Nov 25 2002, 19:53
Post #12





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



But what in this situation:

A scratched CD...
if our drive did 1000 reads on a sector,
lets say 20% of the reads give value '01111111' and 80% give value '11111111'.

Then it would be possible that
the first EAC read gives '01111111'
the second '01111111'
the third '11111111'
and the fourth '11111111'
but eac will stop after the second read and think to have found the correct value, whereas this is not sure (there is a chance of 80% that EAC has mistaken)

Is this one of the situations where even EAC fails? or does EAC have a way around?

In such a situation using Test&Copy would at least enforce an extra reread...

This post has been edited by liekloo: Nov 25 2002, 20:18


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
liekloo
post Nov 25 2002, 20:20
Post #13





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



Of course, if sectors are big enough this is not a real issue...

(EAC's documentation doesn't mention how big a sector is, but it would definitely interest me rolleyes.gif )


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
Pio2001
post Nov 25 2002, 23:53
Post #14


Moderator


Group: Super Moderator
Posts: 3936
Joined: 29-September 01
Member No.: 73



The "sectors" re-read by EAC are in fact 27 sectors, that is 15,876 samples, or a little more than 1/3 of a second.
When an error is found, only the erroneous part is reread in order to try to correct the error (no need to risk damaging the data that was fine).

Some discussions about secure ripping :

[EAC beaten??] CD-Freaks - Review - Plextor's PlexTools [comment requested tongue.gif]

Error correction and interpolation in CD ROM drives : Discussion reliability DAEquality test

Is accurate C2 error reporting an Utopia?

Explanation of C2 support : Drive options => 'capable C2 info' vs 'use C2 info'

Matching CRC for errored rips
Go to the top of the page
+Quote Post
liekloo
post Nov 26 2002, 15:12
Post #15





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



Thank you very much, this is not too bad B)

Hydrogenaudio has been beaten biggrin.gif


Anyway, I have the impression that Test&Copy isn't useless after all...

Getting twice the same erroneous info is very rare, but not impossible as SatCP has experienced several times.

In these few situations in which EAC gets twice exactly the same erroneous info, Test&Copy enforces a third read, and since a third same erroneous sector read will 'never' happen, Test&Copy 'eliminates' this kind of statistical hazards (getting twice the same erroneous info I mean)...

It seems that the argument C2 defenders wink.gif use, is that C2 info can track this kind of hazards (getting twice the same erroneous info I mean). But if I am not mistaken, Test&Copy will 'always' find it too. Or does someone have an extra argument supporting the claim that using C2 can sometimes give more accurate results?

This post has been edited by liekloo: Nov 26 2002, 15:16


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
Volcano
post Nov 26 2002, 22:42
Post #16





Group: Members (Donating)
Posts: 916
Joined: 30-September 01
From: Berlin, Germany
Member No.: 112



Just my 0.02:

Test & Copy without C2 is certainly not overkill in my experience. EAC most of the time reports a clean rip with a track quality of 99.9% on my NEC DV-5700B. I thought, if EAC reports no errors, the rip should be very accurate - but when I started using Test & Copy a week or so ago, I had to discover that the test/read CRCs often didn't match. On scratched CDs, I found it impossible to get matching CRCs with this drive - although there were no errors reported for each rip.

Without Test & Copy, I never would have noticed this, so IMHO it is definitely worth using if you don't mind the speed decrease.
Go to the top of the page
+Quote Post
liekloo
post Nov 27 2002, 09:38
Post #17





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



If Volcano's EAC is correctly configured (caching), which is probably the case, and he didn't correct offset or so, this is indeed a problem:

Although CRC mismatches in secure mode happen only very rarely (so that you could say T&C is overkill, and btw CRC mismatches should not have audible effects),

this changes when inserting scratched CDs, as Volcano found. Why? A single scratch will cause an enormous amount of errors on the surface, drastically increasing the chance of what I called earlier "hazards" (getting twice the same erroneous info).
The question is: how drastical can the increase be?? My own CDs are all in good condition, and I never had a CRC mismatch, but I heard from several people that with scratched CDs T&C is more than just 'luxury' or 'overkill'.


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
liekloo
post Nov 27 2002, 09:43
Post #18





Group: Members
Posts: 456
Joined: 22-March 02
From: Belgium
Member No.: 1596



QUOTE
I found it impossible to get matching CRCs with this drive - although there were no errors reported for each rip.


I am quite surprised to see someone (Volcano) having so many CRC mismatches (?)
Is it possible you ripped those scatched earlier in a non-secure mode (often done with badly-scratched CDs)?
BTW CRC mismatches should have no audible effects, that's why the EAC LOG doesn't consider them as 'errors'. Even if there are some CRC fails, the LOG will still say 'no errors occured'.

Thx for reply smile.gif


--------------------
"E S S E N T I A L" Guide for E A C :

http://users.fulladsl.be/spb2267/
Go to the top of the page
+Quote Post
user
post Nov 27 2002, 10:16
Post #19





Group: Members
Posts: 873
Joined: 12-October 01
From: the great wide open
Member No.: 277



" On scratched CDs, I found it impossible to get matching CRCs with this drive - although there were no errors reported for each rip.

Without Test & Copy, I never would have noticed this, so IMHO it is definitely worth using .. "





This is typical on scratched CDs, independent of used drive, or brand.
And T&C is not worth using it, if you are in secure mode, cashing "feature" correctly set, NO C2.

Volcano gives the point:
"I never would have noticed this"

Has somebody ever heard a bad tune in a song, when eac did not use t&c and reported no errors occurred ?!

It goes further on:

Even when eac reports: "There were errors", then you are often not able to detect a wrong tune at "Suspicious position".


The reason for all this is the physical condition of the CD's surface:
If the scratch reflects the light in other ways, then eac (and any other program) will even think, that these wrong data is the correct one, because the programs read again and again the SAME values (with or without t&C, with t&c you are only doubling the times of reading, that is not much, remember: eac itself reads up to 82 times (5 red lines), even in copy-only mode, without test).
Next reason:
The safety of reading twice (without test, NO c2) for error detection is high, it is not like 50% (would be very unsafe), as eac compares sectors, which contain a lot data, I think, the safety is around 1:10.000 or so ?


--------------------
www.High-Quality.ch.vu -- High Quality Audio Archiving Tutorials
Go to the top of the page
+Quote Post
Pio2001
post Nov 27 2002, 15:01
Post #20


Moderator


Group: Super Moderator
Posts: 3936
Joined: 29-September 01
Member No.: 73



QUOTE (Volcano @ Nov 27 2002 - 12:16 PM)
" On scratched CDs, I found it impossible to get matching CRCs with this drive - although there were no errors reported for each rip.


Liekloo is right. A minor scratch usually produce an enormous amount of errors. One of them is always detected. However, there are two cases in which CRC can mismatch without a bad rip reported :

1 - the scratch is light enough for all errors to be corrected in less than 80 rereads. Then EAC displays "no errors occured". But as thousands of errors were corrected, it is highly possible that one of them gave more than 8/16 times a wrong value, that was taken for right.

2 - A bad drive combined with a bad media. For example the Hitachi GD-7500, that always finds one or two errors on brand new CDs, reading a Memorex black CDR, will find tenths of isolated errors, all of them corrected in one row. With this combination, on most of the tracks, CRC mismatch while "no errors occured".

QUOTE (user @ Nov 27 2002 - 12:16 PM)
Has somebody ever heard a bad tune in a song, when eac did not use t&c and reported no errors occurred ?!


Yes, in Halcyon's test http://www.digital-inn.de/showthread.php?t...or&pagenumber=2
12th message in the link above :

Sony CRX140E
C2: OFF (not available)
Accurate Stream: ON
Disable Audio cache: ON

Copy OK status (from EAC log):
track 3: OK

Track audible quality (listening test):
track 3: Two loud pops 6 seconds from the start of the track

CRC (from EAC test & copy):
track 3: B601AA2C <> EA0EA86C


QUOTE (user @ Nov 27 2002 - 12:16 PM)
If the scratch reflects the light in other ways, then eac (and any other program) will even think, that these wrong data is the correct one, because the programs read again and again the SAME values (...) The safety of reading twice (without test, NO c2) for error detection is high, it is not like 50% (would be very unsafe), as eac compares sectors, which contain a lot data, I think, the safety is around 1:10.000 or so ?


The safety of C2 is very high in theory. From BobHere's documentation http://www.digital-inn.de/showthread.php?t...15&pagenumber=5

QUOTE
The much quoted figures for the 1st generation players (Philips CD100) with their much simpler CIRC implementations are:

sample interpolation rate:
one sample every 10 hours at 10exp-4 BER (Bite error rate)
1000 samples per minute at 10exp-3 BER.

Undetected error samples (clicks) less than 1 every 750 hours at BER 10exp-3, negligible at BER 10exp-4


Reading twice safety is impossible to evaluate. It must be between 1/8 (jitter errors - E32 chipset) and 1/1,099,511,627,776 (damaged CD, E52 chipset). (quoting myself). Explanations in the same link.

This post has been edited by Pio2001: Nov 27 2002, 15:03
Go to the top of the page
+Quote Post
Pio2001
post Nov 27 2002, 15:10
Post #21


Moderator


Group: Super Moderator
Posts: 3936
Joined: 29-September 01
Member No.: 73



By the way, this thread is interesting : (http://www.digital-inn.de/showthread.php?t...?threadid=14600 above)

Safety of reading twice, caching, and C2 were deeply discussed. Unfortunately we were very long to reach a result about C2. It was this discussion that started the DAE quality analysis project ( http://www.exactaudiocopy.de/eac13.html ).
Since we only have a theoretical value of the C2accuracy, I've been trying for two monthes to get experimental results. Work still in progress. I'm writing another analysis program.

This post has been edited by Pio2001: Nov 27 2002, 15:12
Go to the top of the page
+Quote Post
Pio2001
post Nov 27 2002, 15:48
Post #22


Moderator


Group: Super Moderator
Posts: 3936
Joined: 29-September 01
Member No.: 73



I forgot : http://www.digital-inn.de/showthread.php?t...15&pagenumber=4 , post number 6 in this page, reports the Sony crx220e1 and Asus crw5224 to have perfect 100% accurate C2 reporting.
Go to the top of the page
+Quote Post
user
post Nov 27 2002, 15:56
Post #23





Group: Members
Posts: 873
Joined: 12-October 01
From: the great wide open
Member No.: 277



For all these reasons and uncertainty, checking in eac capabilty of c2 is not recommended, if your drive should be able to.
Let's see your test results !


hmm, Halcyons test, was carried out with a special test-CD, with black stripes on it ?

Is this very realistic, or was that Sony drive so bad ?

Anyways, many thanks for reminding these test of halcyon, I remember having read it.


I have never denied the addition of safety, if you do t&c, but the question is, how much worth is this extra safety (hmm, you are just reading 2 times again by t&c, but eac has had the chance to read it 82 times in the first (or only) run..) ?
(btw, I am always talking of settings NO c2 in secure mode, c2 is not matter of interest for 'users', until there would be new results... (and they would be dependent to brand of drive etc.))

The question is:

How big is the probability, that eac reads twice (no c2, secure) and thinks: no error, all is well, but it has read twice the same wrong value for that damaged sector.

This probabilty depends on size of sector.

Is there an answer ?

(I know, the theoretical safety of c2 error detection is in theory higher than that reading twice method....)


--------------------
www.High-Quality.ch.vu -- High Quality Audio Archiving Tutorials
Go to the top of the page
+Quote Post
Volcano
post Nov 27 2002, 16:51
Post #24





Group: Members (Donating)
Posts: 916
Joined: 30-September 01
From: Berlin, Germany
Member No.: 112



user:

QUOTE
QUOTE
On scratched CDs, I found it impossible to get matching CRCs with this drive - although there were no errors reported for each rip.

Without Test & Copy, I never would have noticed this, so IMHO it is definitely worth using ..


This is typical on scratched CDs, independent of used drive, or brand.


But even with my crappy Acer CD writer I could get matching CRCs on all tracks on one of the scratched CDs. Had that not been the case, I probably wouldn't have worried and would have thought, just like you, that CRC mismatches are perfectly normal for scratched CDs.

There might be an explanation for this - the Acer reported 100% on nearly all tracks, in other words, EAC's error correction never kicked in. On the NEC however, error correction was performed far more often, which leaves more room for minor differences between the reads which aren't audible, but still exist (which a CRC check will, of course, pick up). (I guess that's what you meant with your first point, Pio2001...?)

Anyway, I'm still a little worried about the results of the NEC with Test & Copy.
Go to the top of the page
+Quote Post
user
post Nov 27 2002, 17:07
Post #25





Group: Members
Posts: 873
Joined: 12-October 01
From: the great wide open
Member No.: 277



" But even with my crappy Acer CD writer I could get matching CRCs on all tracks on one of the scratched CDs. Had that not been the case, I probably wouldn't have worried and would have thought, just like you, that CRC mismatches are perfectly normal for scratched CDs. "


Oh, that shows, how uncertain even T&C is.

My opinion, is still:

T&C in secure mode, no c2, does not increase safety by a reasonable amount ?

Unfortunately, if the surface is damaged, the original data is changed.
A Polishing may help in this case.


--------------------
www.High-Quality.ch.vu -- High Quality Audio Archiving Tutorials
Go to the top of the page
+Quote Post

2 Pages V   1 2 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 29th August 2014 - 05:14