Is it better to have higher PI Error average and low Pi failure or reverse order?

vbimport

#1

Hi,
I have scanned a lot of discs (TYG02) burned recently with my burner (LG 4163B) and I have noted the following facts:

  • 4X burns results with PI Error Average between 10-20 and PI Failure very low (0.00 to 0.02 average and total PI Failure always under 50!)
  • 8x burns results with PI error Average between 2-5 and PI Failure higher (0.02 to 0.05 average and total PI Failure between 200-750)

Is it better to have an higher PI Error average with very low PI Failure OR
lower PI Error AVERAGE and more PI Failure.

Thanks


#2

1st, the total count for either type error is not a measure of quality, nor is there any specification for total counts. Given max values under 4, 50 is no different than 500 in terms of the “readability” of the disc.

Most folks prefer to see lower PIF even if PIE are a bit higher. But in your situation, both cases are perfectly fine discs. Unless you have a really picky DVD player you won’t see any difference.

At error levels this low, differences have less and less meaning. Burn them at 8x and enjoy.


#3

PIE can be as high as 300 and PIF as high as 10 for good readability.
I suggest 12x burning for best speed and quality.


#4

If you’re scanning for PIE per 8 ECC blocks (e.g. Plextor, LiteOn, BenQ scans) then a maximum of 280 is within acceptable limits according to ECMA-337 specifications.

If you’re scanning for PIF per 1 ECC blocks (e.g. Plextor, LiteOn scans) then a maximum of 4 is within acceptable limits according to ECMA-337 specifications. A disc with a maximum of 10 PIF might be readable but it might also be a coaster.

If you’re scanning for PIF per 8 ECC blocks (e.g. BenQ scans) then a maximum of 12-16 is within acceptable limits according to general consensus, but it’s not possible to translate directly to the ECMA-337 specification.

In my experience however, a PIF maximum of 10 per 8 ECC blocks is closer to the limit of 4 PIF per 1 ECC blocks, so I would use 10 as the equivalent limit but that’s just my opinion.


#5

What about Jitter %? What shoould the limit be for the maximum?


#6

Cheers for that, Drage…I only just saw this, and though I’m “OK” with BenQ scans, I’m still learning :wink:


#7

As drage has already pointed out, averages & totals mean next to nothing, except when you are getting excellent results and you’re nitpicking to differentiate burn quality between discs.

The maximums are what you need to watch out for.

280 PI & 4 PIF maximums for liteons and other 1ECC burners.
280 PI & 16 PIF maximums for Benq’s & other 8ECC burners.

The 16 PIF (8ECC scan’s) though are not a good indicator, as you theoretically could have 16 PIF on a single ECC block & the other 7 be perfectly fine (although it’s highly unlikely), so the limit has been dropped to about 10-12. It’s not an exact science though.


#8

Just a little nitpicking:

What you mean is 1 ECC scanners and 8 ECC scanners.
(The burners are irrelevant here.)


#9

Can’t scan with a reader, the results are inaccurate. Therefore the scanner has to be a burner.


#10

[QUOTE=DrageMester;1399883]If you’re scanning for PIF per 1 ECC blocks (e.g. Plextor, LiteOn scans) then a maximum of 4 is within acceptable limits according to ECMA-337 specifications. A disc with a maximum of 10 PIF might be readable but it might also be a coaster.

If you’re scanning for PIF per 8 ECC blocks (e.g. BenQ scans) then a maximum of 12-16 is within acceptable limits according to general consensus, but it’s not possible to translate directly to the ECMA-337 specification.[/QUOTE]

I read that scanning at 1ECC will collect 8 times as many samples as scanning at 8ECC. So doesn’t that mean that there could be around 8 times more PIFs @1ECC than @8ECC? Just wondering why the acceptable limit for PIF @1ECC is less than the acceptable limit @8ECC.


#11

[quote=sl_dvd;2236509]I read that scanning at 1ECC will collect 8 times as many samples as scanning at 8ECC. So doesn’t that mean that there could be around 8 times more PIFs @1ECC than @8ECC? Just wondering why the acceptable limit for PIF @1ECC is less than the acceptable limit @8ECC.[/quote] In principle
[li] all of the ECC blocks on a disc are scanned. When performing 1 ECC scanning there is only one 32kB ECC block in each sample, and when performing 8 ECC scanning there are eight 32kB ECC blocks in each sample.
[/li]
This means that there are 8 times as many samples when scanning at 1 ECC compared to the number of samples when scanning at 8 ECC. The total number of ECC blocks scanned is the same, and all else being equal and ignoring random variations, the total number of PIE and PIF will be the same.

The number of PIE and PIF per sample however, will on average be 8 times higher for 8 ECC scanning than for 1 ECC scanning, since there is 8 times as much data in each sample.

The maximum number of PIE and PIF per sample in 8 ECC scanning compared to 1 ECC scanning is not possible to predict accurately - it can only be estimated based on statistical or probability analysis.

[*] Some types of drives report all samples and other types of drives skip some samples depending on system setup and scanning speed.


#12

Thank you for the explanation, DrageMester. So if a drive supports both 1ECC and 8ECC scanning, would it be better to use 1ECC or 8ECC, or does it not matter?


#13

[quote=sl_dvd;2236519] So if a drive supports both 1ECC and 8ECC scanning, would it be better to use 1ECC or 8ECC, or does it not matter?[/quote] 1 ECC scanning is better because the reported results are more fine-grained and can be compared with ECMA standards.


#14

[QUOTE=DrageMester;2236515]In principle[COLOR=gray]The number of PIE and PIF per sample however, will on average be 8 times higher for 8 ECC scanning than for 1 ECC scanning, since there is 8 times as much data in each sample.

The maximum number of PIE and PIF per sample in 8 ECC scanning compared to 1 ECC scanning is not possible to predict accurately - it can only be estimated based on statistical or probability analysis.[/QUOTE]
So effectively, if you get 1000pi errors on an 8ECC scanning drive, you can’t determine whether there is 125 PI errors per block and the disc is usable, or whether all 1000 are in a single block and the disk is useless :wink:


#15

[quote=debro;2236627]So effectively, if you get 1000pi errors on an 8ECC scanning drive, you can’t determine whether there is 125 PI errors per block and the disc is usable, or whether all 1000 are in a single block and the disk is useless ;)[/quote] Pretty much right - especially for PIF. :iagree:

The ECMA standards specify acceptable PIE per 8 ECC blocks, while acceptable PIF are specified per 1 ECC block.

When using a 1 ECC scanning drive, the scanning software can trivially add up PIE values for 8 x 1 ECC samples at a time.

When using an 8 ECC scanning drive, the scanning software cannot know how to divide the reported sum of PIF for 8 ECC blocks into how many PIF were in each of the 8 x 1 ECC blocks, however.

This is why I call 1 ECC scanning “more fine-grained”.