Question: Weak sectors again

vbimport

#1

Sorry for bringing this up again. I think I understood what “weak sectors” are and how they work: Some data patters produce, when scrambled, a regular pattern which can not be burned without errors by some burners. I saw some of your guys samples and I started to play with it. Its pretty interesting…
My question is now: The guys that use this for copy protection of their cds have to write those data once without errors, right?

As I understand it, they burn the “weak sector data” error free, the “bad guys” read it, can’t burn it propperly and the cd contains errors.

How do THEY write the data? Do they use special hardware? Can I do somthing similar with my of-the-shelf-burner?

Thanks for any “enlightment”

Maik


#2

How do THEY write the data? Do they use special hardware?
A Toshiba burner (such as am SD-R 1312 for 80 Euro) can burn even more “effective” weak sectors than what Macrovision is currently using.
here is some (badly structured!) c++ source code which creates weak sectors, scans for weak sectors and which can take a guess on how weak a certain regular pattern probably is


#3

Originally posted by MaikVan
Do they use special hardware?

You mean the cd pressing machines games manufacturers use for commercial cds? :confused: With a cd press you can write anything on a cd…


#4

But don’t they need to burn it on cd-r before pressing?
(for testing for example)


#5

testing like that? proly not :wink:


#6

Originally posted by Supi Suomalaine
But don’t they need to burn it on cd-r before pressing?
(for testing for example)

Maybe, but a CDR burner can’t emulate the glass master cutter, that creates the master used to manufacture the CDs.

Originally posted by MaikVan
Some data patters produce, when scrambled, a regular pattern which can not be burned without errors by some burners.

Once the data is scrambled, we get the data that is going to pass through the CIRC encoder, then the EFM encoder, then the NRZ encoder. Has someone already looked at the pit/land pattern these three encoders produce ? It must be very, very far from what the scrambled data look like.


#7

The CDs are first cut using Laser Beam Recorders onto a glass master, these are hugely expensive pieces of kit and are extremely accurate. Therefore they can burn weak sectors reliably.

The glass master is then used to stamp the blank discs. Obviously stamping gurantees a 1:1 transfer of data, so no problems with weak sectors there.

The applications have to be burned to a master CDR, but the protection is not burned to that CDR, the protection is applied during creation of the glass master. The machine which makes the glass master has special SafeDisc software installed on it.


#8

The CDs are first cut using Laser Beam Recorders onto a glass master, these are hugely expensive pieces of kit and are extremely accurate. Therefore they can burn weak sectors reliably.
???

It doesn’t matter at all how accurate these lasers are. If they are fed with a pit-land-sequence of a high DSV (whatever “high” means), the pressed discs would be unreadable…


#9

In the old discussion about weak sectors (but I have not read this forum for a long time), we did not reach a conclusion. Someone said that weak sectors had a high DSV.

So I suggested that glass masters cutters used illegal merge bits in order to keep a null DSV on weak sectors (that would be impossible with a proper NRZ encoder), and someone suggested that it had something to do with the periodicity of the pattern.


#10

So I suggested that glass masters cutters used illegal merge bits in order to keep a null DSV on weak sectors (that would be impossible with a proper NRZ encoder), and someone suggested that it had something to do with the periodicity of the pattern.
Any weak sectors which Macrovision is using at the moment can be written without any “illegal” merging bits.


#11

Isn’t the DSV a bit high ? (though I didn’t see anywhere how we came to this conclusion).


#12

It is possible that the SafeDisc mastering software works something like Betablocker / AWS. Unlikely, but possible.


#13

@Pio, Atomicx: Take a sheet of paper, a pencil, ECMA-130 and encode 1212, 14B9, 3F65 and 7D65 on your own.
Then, do the same with 659A and 04B9.


#14

With or without scrambling ?


#15

Originally posted by Pio2001
With or without scrambling ?

Without.


#16

I assume they are one sequence of 8 bytes repeating, and another of 4 bytes repeating… Or are they 6 different sequences of 2 bytes alterning ?


#17

They shall be 6 different sequences of 2 altering bytes


#18

I can’t calculate the P and Q bytes.
Should I assume that the calculus are done modulo 256 ? Should I set alpha=00000010 (2), and ignore the GF(2^8) stuff ?

Edit : if alpha=2, alpha^8 and up modulo 256 would all be 0.


#19

Nah…forget about it. You really don’t need these (if they were of any importance, weak sectors could not work anyway…)


#20

Well, something else: Would this whole weak sector thing also work with audio? Audio doesn’t get scrambled, right? Could one just make a audio file with the according unscrambled pattern in it?