So, in reality on DVD the following scheme is applied
A block of data is written, at the end of which there is information for integrity checking and error recovery. An error occurring 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 within 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 written with redundant data meant for correcting entire bad blocks that couldn't be fixed by the data at the end of the block. (I apologize for endlessly repeating the word Block, but then again I love him like the poet)
Data blocks
Information for error correction within 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 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 doesn't correspond to reality (for optical discs). It should be saying Unrecoverable PO errors. For us, the most interesting thing is that on Lite-On DVD drives you can check the error level using the Nero CD-DVD Speed or K-Probe programs. Drives from Sony and some other brands - are also Lite-On). In Nero CD-DVD Speed, which is almost always included in the NERO suite, you have to select from the menu "Disc Quality Test" 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 can correct them and continue reading. 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 displayed by the programs shouldn't stay above 280 for a long time, and PO (PIF) - above 32. Isolated peaks don't count - they might be read or write glitches. But more or less long stretches will make the disc unreadable, although in practice a lot depends on the reading drive. Naturally, a lot also depends on the recording device. Plextor and Pioneer used to be famous, now cheap and high-quality-writing NECs are popular. Unfortunately, the saying "the Chukchi isn't a reader, the Chukchi is a writer" works 100% here - it happens that NEC drives can't even read the discs they burned themselves (I mean the case of horribly low-quality discs). But blanks are crucial - even in the worst drive, TY will most likely burn perfectly and be preserved for a very long time (excluding cases of incompatibility between the drive and discs, which happens)
DVD+R Verbatim Advanced AZO 8x Made In Taiwan
This disc is relatively decent, apart from the error growth at the end - a common thing with cheap media. If the disc were filled to the very end, without 100-200 MB of free space, this hill of errors would have become a menacing mountain.
Pay attention to the "Disc" line (second from the top). There you can see the real manufacturer code of the disc. In this case, it's MCC02 , which means Mitsubishi Chemical Corporation. The code is sometimes faked, but not all that often.
This is a Taiyo Yuden disc
This is a Taiyo Yuden 8x disc , burned at 8x speed. As we can see, the sparsely scattered POs (PIFs) don't exceed level 2. A PI level of 10-25 is perfectly normal. If I had an NEC burner, this level might have dropped to 5-15, but no lower.
Taiyo Yuden 8x , burned at 4x
As you can see, in this case, burning at 4x was rather beneficial, since the error level dropped even further (see the Total Count line to be sure). But generally speaking, this can vary. Further down in the screenshots, we'll observe a progressive degradation in quality, and at the end, an absolute horror show.
TDK 16x, burned at 8x
This TDK blank is actually manufactured by CMC Magnetics, and should be pretty awful. Luckily, nothing particularly terrifying is visible yet - it's just that the PI level is unhealthily elevated, and PIFs are running solid and their total count is 8 times higher than on TY.
Verbatim 16x, made in India
And this beauty shows us a prime Mount Everest of errors at the end. I've seen worse, but even here the last chunk of the disc will be basically impossible to read.
Verbatim 16x, made in India
What can you say. 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 ended earlier.
Let's sum up. Why is a low error level important and what makes it so good?
1) There's a guarantee that your discs will be readable everywhere, including on household 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 spike sharply. If it was already high, the disc simply won't be read, as the drive won't cope with the correction. If it was low, like with TY, it will merely jump to threatening proportions, but the chances of reading it are high.
3)A low error level also indirectly testifies that the quality of the blank's materials is high, and it will last many years if you keep it away from sun and scratches. TY gives a 100-year warranty, but in reality, no one knows how long they will last. Tests in an aging chamber have shown that after 25 years TY will still be readable.
I could also write about how the drive's firmware has hardcoded burning strategies for blanks from different manufacturers. And if a blank is unknown to it, it uses a "generic" burning strategy. The results then are much worse. And if you buy new blanks for an old drive, it definitely won't know them (which, apparently, is what happened with my 16x Verbatim).
I could also write about how the drive is self-learning, and each subsequent blank from the same batch by the same manufacturer will be burned better and better (in theory). And some models even know how to learn on the fly, improving the burn quality of subsequent chunks on the same blank. But I won't overload the populace with an abundance of technical info, for 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, fridge magnets, and tasty cherry sore throat spray. The postal service works like clockwork - the package arrived in less than a week.