Which Drive Do I Trust?

For the longest time ive been using a LiteON 166s DVD-ROM with K-probe to do all my scanning. I bought a BenQ DW1655 recently and ran a scan with nero CD speed that was contradictory to the scan that I got from the LiteON. Frustrated I went back to the shop and exchanged it for the most expensive burner I could find, the Plextor PX-716a. Same problem though.


Here is a TYG03 disc scanned by the liteon


And the same disc scanned with the plextor.


The problem is that whenever the errors are low on the LiteON they are low on the other scanner as well.

Which scans do I trust? For backing up movies that I already own reliability really isnt that important, but for data that cant be replaced, having a false sense of security that the plextor scan gives me if its false would freak me out badly.

Can somebody make sense of this? I dont like scanning and I need the peace of mind that this plextor (or that benQ) isnt lying to me.

its the dvd-rom thats lying, only burners make reliable scanners.

you should have kept the benq, at least then you could have scanned at 8x.

BlastClas, welcome to the forum!

I agree with ako's comments.

Also I suggest that using Sum1 scans on your Plextor burner is an even more important scanning method than Sum8, as it will show you Parity Inner Failures (PIF) which are more important than Parity Inner Errors (PIE).

Have a look in the following thread for more information on PIE/PIF quality scanning: Interpreting PI/PO error scans

I went out and bought a liteon DVD burner for scans today (since the plextor is so slow) and boy does it make a difference in the error rates, from tripple digits down to single digits. I think my ROM just hated burned discs or something.

Now my next problem, and I know you cant solve this one: I burned over 150 discs of RITEKG05. I wish I started reading this forum earlier, good thing that none of them have started falling appart yet.

No drive is 'lying', they just report the error levels THEY see. Any time I discuss PI/PIF testing, I use the term 'accurate' loosely, as they are all accurate in that they are reporting only what they see. In terms of useful comparisons with scans you see here, a burner such as a Liteon, Plextor, or Benq would make a much more 'accurate' tool than any DVD-ROM drive. Regarding DVD-ROM drives and burned media, some do a worse job or have a harder time reading the discs than others. For example, I have a Liteon DVD-ROM drive as well and it would struggle with some discs that scanned and read excellent in all of my DVD burners. DVD-ROM drives are fine for ripping discs and playing discs in general, but burners are a better choice for testing your media, especially if you want to make meaningful comparisons to scans posted on this site by other users.

ALSO, your first scan was tested at max, you should be testing at 4x for the most 'accurate' comparisons. 4x is the generally accepted 'standard' for Liteon DVD-ROM drives and the majority of their burners (although a few of their latest burners scan similar at 8x as they do at 4x), Benq is 8x, BTC 4x, NEC 5x, Plextor I can't recall.

I would personally buy the 1655 over the Plextor, it's an excellent drive from testing to burning to ripping.

This is the Liteon I ended up getting for the 8x scans. It was cheaper than the BenQ and im only going to use it for scans.



The only media I will stick to is TY, because of the whole G05 thing.

Can you provide any evidence to support the statements? To my knowledge, existing data indicate only that most DVD writers are less sensitive than Lite-On DVD-ROM drives to errors, which means that those DVD writers may be better readers, but not necessarily more reliable scanners.

Quote:

Originally Posted by scoobiedoobie No drive is 'lying', they just report the error levels THEY see. Any time I discuss PI/PIF testing, I use the term 'accurate' loosely, as they are all accurate in that they are reporting only what they see. In terms of useful comparisons with scans you see here, a burner such as a Liteon, Plextor, or Benq would make a much more 'accurate' tool than any DVD-ROM drive.

You are inaccurate in using the word “accurate”. Moreover, it is self contradictory to say that “a burner such as a Liteon, Plextor, or Benq would make a much more 'accurate' tool than any DVD-ROM drive” and that “they are all accurate in that they are reporting only what they see”. Further information may be found in the following thread:

Precision, accuracy, and reliability of disc quality (PI/PO/jitter) tests

Quote:

Regarding DVD-ROM drives and burned media, some do a worse job or have a harder time reading the discs than others. For example, I have a Liteon DVD-ROM drive as well and it would struggle with some discs that scanned and read excellent in all of my DVD burners. DVD-ROM drives are fine for ripping discs and playing discs in general, but burners are a better choice for testing your media

Those observations indicate that the Lite-On DVD-ROM drive is more sensitive than your DVD burners to errors, and may even suggest that the Lite-On DVD-ROM drive is a better choice for testing media rather than the opposite.

Quote:

ALSO, your first scan was tested at max, you should be testing at 4x for the most 'accurate' comparisons.

How do you know that testing at 4X is the most accurate?

i think this thread proves dvd-roms are not reliable scanners, plus if you searched you would find many more threads like this 1.

This thread does not prove it. Moreover, I have not seen any evidence after very extensive reading for more than a year, so can you provide any link?

Muchin, of course DVD-ROMs are not less "accurate". As both scoobiedoobie and yourself put it, they just report what they "see" (though there is room for discussion in this area mainly with NEC burners). But they are considered as unreliable scanners for a different reason: as most report errors very differently than burners do, the usual "standards" (max PIF, max PIE etc.) can't be applied at all to the scans they produce.

To compare different burns, DVD-ROM should be just as usable as burners. But considering the discrepancy between Benq, Nec, Plextor and LiteOn burners in the way they report, which is already extremely confusing, adding DVD-ROMs to the picture would render the whole thing totally unmanagable (I mean even more than it already is *LOL* ). Get my point?

Cheers from Belgium

You know exactly what I mean, pretending that you don't is not amusing. As I said, I use the term 'accurate' loosely when speaking of PI/PIF testing, as the drives are just reporting what they see so each drive is reporting accurately what it saw in terms of error levels. But throwing in a test from a DVD-ROM drive in with tests from burners at their 'standard' speeds is a misleading representation of that disc.

When I say 'accurate' and burners are more 'accurate' testers and more 'accurate' at certain speeds, what I'm suggesting is that using burners at their 'standard' speeds makes for the most meaningful comparisons between burners. For example, comparing a 16x PI/PIF scan on my Benq burner with a 1x scan on a Liteon burner would be pointless. Compare an 8x scan on a Benq with a 4x on a Liteon and you'll get much more agreeable results. Every single drive made is going to report errors differently than the next, but sticking to 'standard' test speeds and using burners will at least result in decent comparisons, although NEC and Pioneer drives are not necessarily much better than DVD-ROM drives for testing.

Quote:

Originally Posted by Francksoy Muchin, of course DVD-ROMs are not less "accurate". As both scoobiedoobie and yourself put it, they just report what they "see" (though there is room for discussion in this area mainly with NEC burners). But they are considered as unreliable scanners for a different reason: as most report errors very differently than burners do, the usual "standards" (max PIF, max PIE etc.) can't be applied at all to the scans they produce.

Sorry, Franck, you have misread my words, I did not say that "they just report what they see". I consider it established that certain drives are incorrect in their PI/PO/jitter reports. Contrary to common belief, I am afraid that the usual "standards" (max PIF, max PIE etc.) can't be applied to the scans obtained with many DVD writers instead, in view of the comparisons others and I made between scans with AudioDev CATS and some Lite-On DVD-ROM drives, in addition to other information.

Quote:

To compare different burns, DVD-ROM should be just as usable as burners. But considering the discrepancy between Benq, Nec, Plextor and LiteOn burners in the way they report, which is already extremely confusing, adding DVD-ROMs to the picture would render the whole thing totally unmanagable (I mean even more than it already is *LOL*).

To my knowledge, Lite-On DVD-ROM drives often give PIE/PIF counts not very different from Plextor 712 and earlier Lite-On writers at appropriate speeds. So Benq and the more recent Lite-On writers are the ones to blame for the confusion, IMO.

Cheers from Taipei

Quote:

Originally Posted by scoobiedoobie You know exactly what I mean, pretending that you don't is not amusing. As I said, I use the term 'accurate' loosely when speaking of PI/PIF testing, as the drives are just reporting what they see so each drive is reporting accurately what it saw in terms of error levels.

Pretending? No, I am not sure about what you wanted to convey exactly by the word “accurate”, though I guess that you used it to express “indiscriminative”, “insensitive” or the like, which is radically different from accepted definitions of “accurate”.

I used to believe that “the drives are just reporting what they see and each drive is reporting accurately what it saw in terms of error levels”, too. However, I have not seen any evidence for it, not to mention proof, after very extensive reading at several major websites. On the other hand, evidence against those viewpoints has been identified very recently. It is true that every drive reports what it sees in transfer rate test, in which data are indeed read. In contrast, signals are treated in some way different in disc quality tests, and even the testing software plays some roles in the errors computed. In the thread I cited before, it has been mentioned that DVDInfoPro prior to the newest version gives incorrect PIE/PIF numbers at least when using certain drives. There are other cases of incorrect reporting of error levels by some drives, but the causes remain unknown.

Quote:

But throwing in a test from a DVD-ROM drive in with tests from burners at their 'standard' speeds is a misleading representation of that disc.

Please see my reply to Francksoy.

Quote:

When I say 'accurate' and burners are more 'accurate' testers and more 'accurate' at certain speeds, what I'm suggesting is that using burners at their 'standard' speeds makes for the most meaningful comparisons between burners. For example, comparing a 16x PI/PIF scan on my Benq burner with a 1x scan on a Liteon burner would be pointless. Compare an 8x scan on a Benq with a 4x on a Liteon and you'll get much more agreeable results. Every single drive made is going to report errors differently than the next, but sticking to 'standard' test speeds and using burners will at least result in decent comparisons, although NEC and Pioneer drives are not necessarily much better than DVD-ROM drives for testing.

It seems that you are unaware of the frequent disparities between scanning at 8x with a Benq and at 4x with a newer Lite-On writer. For example, see scans in a recent review.

There is not enough information to say whether all NEC and Pioneer drives are inaccurate or not in writing quality tests, IMO.

He's well aware of that, he said it, and he deliberatly put the word betwen quotes. Aren't you being kinda nitpicking, here? (no offense intended).Once again, there is one accurate between quotes and the other without quotes. - please pay attention to such things, they are used to limit misunderstanding/confusion, so ignoring them easily leads to endless arguments. Why is that? Assuming one wants to apply ECMA standards to scans, the best scanner would not be the scanner the most (or the less) sensitive to errors, far from it. It would be the scanner which error reporting is the closest to Audiodev's CATS, if such a scanner exists (CATS vs. Homemade scans).

And if one doesn't want to apply ECMA standards, any consistent scanner (i.e. reporting consistently with several passes of the same disc) will do, after all, as without standards, scanning can only be used to compare burns, and not in the least to judge of the quality of the burn.Indeed, I see it now. Sorry about that.Good points, but it doesn't lead me to different conclusions than Scoobiedoobie's: no drive is more accurate than the next, as no drive actually reports errors that are on the disc whatsoever, but reports "something" about the way it retrieves information from the disc. I think it's the core of the misunderstanding here: as long as the idea that drives report physical errors on a disc prevails, we'll all have a hard time to get each other's point. I also think that the three of us (and also DrageMester and rdgrimes) basically look at the same facts, understand it correctly, but draw different conclusions because of different perspectives and different habits.

The way I see it, considering the widely spread misconceptions about scanning on this board, what we need as scanners are not accurate scanners, but scanners that are the closest possible to the way CATS report errors, so we can apply ECMA standards to these scans. OR scanners that are consistent, and get used to the way they report errors. But this is an art, relies on empirical testing and experience, and can't be recommend "as is" to users who are only trying to know if their burns are OK.

Out of interest for your views, I've been following closely your input on this board for some time now. The problem IMO in the way you put things, despite the fact that most of the points you make are sound , is that I feel you're making the whole thing even more confusing than it already is, partly for the pleasure of contradiction (once again no offense intended, I recognise I myself tend to do that sometimes ). So actually I'm a bit frustrated, because I feel our conversations about scanning could be much more "productive" than they've been up to this point, if only we'd all make more efforts to be more "practical" and less "rethoric" (mind the quotes ) in the way we exchange ideas.

And talking about rethorics, I hope this long, very rethorical post of mine will be the last of its kind before long *LOL*

PLEASE, leave CATS equipment out of these discussions. They really are irrelevant unless each party to the discussion actually has a CATS machine at his/her disposal. The only real "benchmark" for the end user as far as scanning "accuracy" goes is a 1x scan. ONLY a 1x scan can be compared to ECMA specs.
Given scanning speeds in excess of 1x, another good benchmark for comparison might be a scan of a pressed disc. In any case, there's no purpose to comparing scans from different drives or programs unless you have a detailed understanding of how those drives/programs differ. Kprobe and CDSpeed now offer reports of the sample rate, which is probably the most critical variable to be considered apart from ECC sums.

You make excellent points Francksoy. Even so-called "experts" (mostly self-proclaimed) often don't really understand error scanning. But even those that do, often disagree about what conclusions to draw. The most common scenario you will see around here is that the poster has already decided what results he/she wants to get and uses scans to "support" that conclusion. Welcome to human nature. Often the same scans can just as easily support a different conclusion. Welcome to error scanning.

Anyone with any experience knows that ROM drives are unrelaible scanners. You don't need "proof" to draw this conclusion. Some ROM drives do manage to produce consistant results, but most cannot even approach that hurdle. This has been true since C1/C2 scanning began some years ago with WSES and the other CDR scanning applications. Even TRT can be dramatically different in a ROM drive when compared to a burner. We don't need to know why, or offer proof that this is true, it just is.

Another interpretation of the discrepancies between LiteOn and BenQ scans in that review, is to say that the BenQ drive may be slightly defective - there are almost no burns that scan well in that particular BenQ drive in the review.

I agree that different drives show different scans, but I don' agree that the difference in general is as big as in the review you linked to.

Many drives cannot read slower than 2x, and most users don't have the patience to wait 1 hour for a scan to complete, so I think your campaign for 1x scanning is a lost cause.

The next best thing IMHO is to use a CLV scanning speed if available, which is e.g. 2x in Plextor drives, 4x in LiteOn drives (or "2x" which is actually 2.55x in my LiteOn 1635S). NEC drives seem to have 1x as the only CLV speed.

Also if a scanning drive always reports the same PIE/PIF levels when scanning faster than 1x, there is no real reason to continue scanning at the theoretically more correct 1x speed.

That's your opinion. But the ECMA specs can only be compared to 1x scans. there's a number of very good reasons for this. I did not suggest that everyone scan at 1x. A 1x scan would be a benchmark for comparing scans at different speeds or different drives or whatever. It may well be inconvenient, but it's the only was to get a comparison that relates to ECMA spec. If you are not able to do a 1x scan, then a scan of a pressed disc might be a suitable benchmark. We can only assume that pressed discs at least approximate a "standard" for disc quality.

Point being that if someone wants to compare scans from different drives or whatever, a reasonable benchmark must be the starting point for each drive or scenario.

Using the transitive law of logic, if "y"x scanning on drive A always produces equivalent result to 1x scanning on drive A, and 1x scanning on drive A always produces equivalent results to 1x scanning on CATS, then "y"x scanning on drive A always produces equivalent results to 1x scanning on CATS.

I think the transitive law of logic is a little more than just my opinion!

Yeah, it's been some time now that I understood this, well, partly because... I found myself doing the same *LOL*. That's why I'm always interested in points of views expressed by experienced users when they disagree with each other (or with me ), because I feel that everyone (with a real experience, I mean) is, most of the time, right AND wrong at the same time.

Personally I plan to keep being a "learner" and I don't take much for granted in this area, even when it comes from the "experts".

I'm intruigued that you want to leave CATS scanning out of the discussion. If we do so, why not leave ECMA standards out as well? These standards can certainly not be applied to scans produced by drives reporting errors in a very different way than CATS does... as they were decided for CATS scanning! My point being: I think that it helps in keeping thing in perspective, to understand that the max PIE/PIF levels that everyone on this board is talking about (and applied in scanning softwares) were stated for CATS scanning, and that, then, applying these numbers to end-users scans is most of the time grossly approximative, and in some cases (like PIE levels of NEC and BENQ scans, for opposite reasons) plain irrelevant IMO.

It seems established by the long-time regulars on this board (at least that's what OC-Freak seems to imply in the "interpreting PI/PO error scans" sticky) that these standards can be applied to scans performed with many LiteOn models, but I have yet to see any empirical/real-world evidence that they can be applied to scans performed with other drives.

Personally this is something I'm also frustrated with: I can rely (almost) only on my own testing/use to find a correlation between what tests tell me about a disc and how this disc behaves in real-world. - I think I may, soon, just buy a LiteOn drive and stop hurting myself.

And now I'm tired, enough for me today

Completely wrong. Unless ALL your burned discs are identical. In other words, the fact that a drive may produce a result similar to CATS on one disc does in no way predict that it will do so on all discs, (or even any), regardless of the read speed. There's ample evidence to back that up. But this ignores the point that it really makes no difference what CATS machines have to say, unless you happen to have one at home. You would need to make comparisons of all media types and all burn speeds with each drive. And that doesn't even take into account the batch variations.

Actually, OC and myself colaborated on all that stuff when we first set up the test forums. The intent is that they are similar enough, and that the end result is the same. CATS tests are not the holy grail. Just another tool, and not one which has much meaning for the end user. It's not about the actual error values, it's about the message they send. They almost always agree that crap is crap and good is good. It's the in-between stuff where problems arise, which is why I always recommend, (and do), when in doubt - throw it out.

The ECMA-337 standard doesn't mention CATS scanners at all. The standard mentions specifications for how a disc should be scanned, and presumably CATS drives are made to adhere to those specifications. Consumer drives presumably are not made to adhere to the same specifications, or at least not as strictly.

So I think you're stating the relationship between ECMA and CATS scanners in the wrong direction; ECMA is not dependent on CATS specifications, but rather it is the other way around!

Do you have a certain mental mode reserved for cases, where you absolutely positively have to disagree with someone, even if you have to deliberately misinterpret what they say in order to disagree with it?

Your argument can also be used to dismiss comparing two CATS drives on the basis, that even if they agree on every disc tested so far they may disagree completely on the next disc tested, so it is "completely wrong" to assume that two CATS drives can be compared, even if all evidence available should point in that direction.

Maybe it's just fun for you to disagree with me on principle?

I never said that all scans in all drives at all speeds could be compared, I merely said that if scans (not just a single scan, do you really think I'm that stupid) performed at some speed "y"x in a drive are found to be equivalent to scans at 1x in the same drive, then there's no need to keep scanning at 1x since "y"x gives equivalent results.

If you have to disagree with me, please disagree with what I'm actually saying, and not with some warped version of what I'm saying!

I take it that your refer to scans performed on some precise models... or I can't agree with this! NEC drives are totally looney, for example, with the way their report PIE, and 300+ on a the most reliable NEC scanner to date (3540A) can still be totally acceptable. Not to mention the 4550/4551 drives

So I guess these statements of yours refer to LiteOn scans (something I wanted to ask you for some time, actually)? This would be a rather important precision...

Gee, I'm still posting... I can't believe it...