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)
DVD+R Verbatim Advanced AZO 8x Made In Taiwan
This disc is relatively decent, aside from the spike in errors at the end - a common phenomenon with cheap blanks. If the disc had been filled to the very end, without 100-200 MB of free space, this hill of errors would have become an intimidating mountain.
Pay attention to the "Disc" line (second from the top). There you can see the code of the disc's real manufacturer. In this case, it's MCC02 , which stands for Mitsubishi Chemical Corporation. The code is sometimes faked, but not that often.
This is a Taiyo Yuden disc
This is a Taiyo Yuden 8x disc , burned at 8x speed. As we can see, the very sparsely located POs (PIFs) don't exceed level 2. A PI level of 10-25 is perfectly normal. If I had an NEC burner, this level could have dropped to 5-15, but no lower.
Taiyo Yuden 8x , burned at 4x
As you can see, in this case, writing at 4x probably did some good, since the error level dropped even further (check the Total Count line to be sure). But generally speaking, this can go either way. Further on in the screenshots we will observe an increasing deterioration of quality, and at the end a total horror-show.
TDK 16x, burned at 8x
This TDK blank is actually manufactured by CMC Magnetics, and it's expected to be pretty awful. Luckily, nothing particularly terrible is visible yet - just an unhealthily elevated PI level, and the PIFs are going back-to-back, with their total count being 8 times higher than TY's.
Verbatim 16x, made in India
And here this beauty demonstrates a perfect Mount Everest of errors at the end. I've seen worse, but even here, the last chunk of the disc will be fundamentally impossible to read.
Verbatim 16x, made in India
What can you say here. The picture is the same as in the previous case, but the mountain didn't have time to grow, because there is less data on the disc, and the scan finished earlier.
Let's sum up. Why is a low error level important and what is so good about it?
1) There's a guarantee that your discs will be readable everywhere, including consumer DVD players, and not just on good computer readers like Lite-On.
2) If you scratch the disc, which is inevitable, the error level in that spot will rise sharply. If it was already high, the disc simply won't read, because the drive won't be able to cope with the correction. If it was low, like with TY, it will merely rise to menacing proportions, but the chances of reading it are high.
3)A low error level also indirectly proves that the quality of the blank's materials is high, and it will serve for many years if you protect it from the sun and scratches. TY gives a 100-year guarantee, but in reality, no one knows how long they'll last. Aging chamber tests have shown that after 25 years TY will still be readable.
I could also write about how the drive's firmware has built-in burning strategies for blanks from different manufacturers. And if a blank is unknown to it, it uses a "generic" burning strategy. The results in this case are much worse. And if you buy new blanks for an old drive, it definitely won't know them (which, apparently, is exactly what happened with my 16x Verbatim).
I could also write about how the drive self-learns, and each subsequent blank from the same batch by the same manufacturer will be burned better and better (in theory). And some models can even learn on the fly, improving the recording quality of subsequent chunks on the exact same blank. But I'm not going to load the general populace with an abundance of technical info, since the provided data is enough to make a reasonable choice.
Happy Burning!
P.S. I want to express my gratitude to my sister for the blanks, the fridge magnets and the tasty cherry sore throat spray. The mail works like clockwork - the package arrived in less than a week.