Digital Wanderer

Read more →

   Having sorted out the geopolitical situation, let's move on to the mathematical part. What does "good" or "bad" quality mean? Unlike the recent discussions about Esperanto and Film vs Digital, which I foolishly waded into, there are pretty strict criteria for evaluating blanks.
To start off - a little example from the world of digital data storage.
   Let's say we need to store eight numbers, each of which is a one or a zero (as we know, computers like ones and zeros, they have this weird quirk, they can't stand other digits). Let's take these: 0 0 1 0 1 0 1 1. How do we check that when reading from the disc we read exactly what we were supposed to, and didn't make a mistake? Add a ninth one, which will control the correctness of the previous ones. But how do we do that? Bright minds came up with this trick: we take all the previous numbers, add them up, do an integer division by 2 and take the remainder. This remainder will be that ninth number, verifying the correctness of the previous ones.
0+0+1+0+1+0+1+1=4
4 / 2= 2 and a remainder of 0
So we get : 0 0 1 0 1 0 1 1 0
Now suppose that one of the numbers was read from the disc incorrectly, namely the first one.
1 0 1 0 1 0 1 1 0
Let's run the operation described above.
1+0+1+0+1+0+1+1=5
5 / 2 = 2 and a remainder of 1
But the control number we read is 0! That means either the ninth (control) number, or one of the eight was read incorrectly.
   What I've provided above - is a primitive scheme for data integrity checking. It is capable of finding only single errors - if two numbers out of eight change at once, it will show that everything is in order. And it is not capable of correcting errors - to do that, you'll have to try reading the data one more time.
  On CDs, and especially on DVDs, complex schemes of redundant coding are used, allowing you not only to find errors, but also to correct them, if their quantity is within reasonable limits. This is done to mitigate the inevitable defects in the material of the blanks, errors in the DVD drive mechanics and laser operation. If the described schemes didn't exist, DVD drives and computers in general would cost colossal amounts of money, would work extremely unreliably and slowly, because producing perfect transistors, blanks, and lasers is incredibly expensive, and sometimes outright impossible. And of course, without error checking and correction, the emergence of our beloved internet would have become impossible.

Read more →

  So, in reality, the following scheme is applied on DVDs
A block of data is recorded, at the end of which there is information for integrity checking and error recovery. An error that occurs inside such a block is called PI. (Parity Inner)  (The numbers in the picture mean nothing, they are just for beauty)
Data
Information for error correction inside the block
{1 0 0 1 0 0 1 0}  {0 1 1 0}
    But what happens if there are so many errors in the block that they cannot be corrected using the redundant data at the end of the block itself? For this situation, after every 208 blocks, additional blocks are recorded with redundant data used to correct entire bad blocks that couldn't be fixed with the data at the end of the block. (Apologies for the endless repetition of the word Block, but at least I love him like the poet)
Data blocks
Information for error correction inside the block  
Error block correction area
{1 0 0 1 0 0 1 0}   {0 1 1 0} 
{1 0 1 0 0 0 1 0}   {0 1 0 0}  
{1 0 1 0 1 0 1 0}   {0 1 0 1} 
{1 0 1 0 0 1 1 1}   {1 1 1 1}   . . . . .  etc. 208 times

{1 0 1 0 0 1 1 1}   {1 1 1 1}
{1 1 1 0 1 1 1 1}   {1 0 0 1}
{1 0 1 1 1 1 1 1}   {1 0 0 0}

  An uncorrectable error inside a block is called a PIF (Parity Inner Failure) or PO (Parity Outer), and such a block is corrected entirely using the data marked in blue. However, if there are too many bad blocks, it's impossible to correct them, and the drive cannot read the disc, throwing a CRC Error message, which actually isn't true (for optical discs). It should be saying Unrecoverable PO errors. For us, the most interesting part is that on Lite-On DVD drives, using the Nero CD-DVD Speed or K-Probe programs, you can check the error level. (Sony drives and some other brands are also just Lite-Ons). In Nero CD-DVD Speed, which is almost always included in the NERO suite, you need to select from the menu "Disc Quality" and set the speed to 4x, which is the standard for checking quality.
   So, up to a certain number of errors, the drive's microchip is capable of correcting them and continuing the read. I won't give exact calculations and explanations here, I'll just say that for our purposes it's enough to know that the PI level shown by the programs shouldn't for a long time  exceed 280, and PO (PIF) - 32. Isolated spikes don't count - they can just be reading or writing glitches. But more or less long stretches will make the disc unreadable, although in reality a lot depends on the reading drive. Of course, a lot also depends on the recording device. Plextor and Pioneer used to be famous, nowadays the cheap and high-quality burning NEC is popular. Unfortunately, the joke "the Chukchi is not a reader, the Chukchi is a writer"  applies here 100% - it happens that NEC drives can't even read the discs they burned themselves (I mean in cases of horribly low-quality discs). But the blanks are of decisive importance - even in the worst drive,  TYs will most likely burn perfectly and last for a very long time (not counting cases of drive and disc incompatibility, which does happen)

Read more →