Coding Errors
Michel GEISS
Has worked with Marc Lavoine, Jean Michel Jarre, Olivier Bron ...
We have to admit it: reliability of burnt CDs is not bulletproof. Who didn't notice that a CD was played without problem at home and badly or not at all elsewhere? The causes are numerous. On the one hand incompatibilities between burners and blank CD-R models, mode of burning, etc... On the other hand, of course, the error rate on a CD is an important factor. Even commercial CDs display many errors. Did you know that this error rate depends on the factories, and on the state of the matrices, and that when buying a CD you may without knowing it get one which almost goes beyond the limit... ?
The reading of a compact disk is a perilous technical operation. Imagine the difficulty, for the laser beam, to follow the tiny groove in a spiral not broader than a hair (space between whorls: 1,6 microns), and to read there the binary data materialized by microscopic " pits " of which the shortest are only 0,83 microns (thousandths of a millimeter) large, and are read by a laser beam of which the wavelength is only 0,78 microns. The mechanics of movement of the laser is a little miracle of technology, but the correct reading of a CD would not be possible without a sophisticated system of correction of errors. Those which concern us for the CDR are due to the inaccuracy of the process and the defects always present on the sensitive layer of matter. These errors, when detected by a special analyzer, are of two types: BLER and E32. BLER, abbreviation of Block Error Rate indicates the error count per second (a few tens per second correspond to good quality). Each CD burning which you will make will be bound to have an error rate greater than zero. The key factor is to try to reduce it as much as possible, in order to guarantee the integrity of your recordings up to the mastering studio or even to manufacture stage if the CD is used as master for further duplication, but also to avoid its rejection by the factory. The E32 errors, as for them, consist in a series of errors such as they exceed the authorized threshold, beyond whose a CD player cannot even use any longer its correction system, however sophisticated it might be.
Schematically, on a CD, the same information are splited, so that if certain digital blocks are deteriorated, information they contain is only part of original coding. Which makes it possible to seek the other part at another supposedly not destroyed location. A circuit of errors correction handles the rebuilding work , and thanks to it the playing of the CD is possible without digital clics, which otherwise would inevitably occur.
When this principle cannot work, when even the other pieces of blocks are illegible, a CD player can still try to fill the blanks by interpolation, thanks to an internal DSP circuit (with samples taken in the vicinity). Let us make it clear here that CD players are more or less efficient in correcting errors (whence certain differences in price...), which depends primarily on digital DSP circuits contained in those.
And then, when the interpolation cannot even occur, you feel quite bad (it is also the case for a DAT). There the sound somewhat cut short, which is particularly pittyful if it occurs in the middle of such a powerful guitar chorus of your that it makes all metalurgists on this planet cry in shame. Another considerable disadvantage: some factories, when they detect one or more E32 on a master CD reject it purely and simply! It should be noted that a burnt multisessions CD (with pause between tracks), with for example an autonomous burner, produce an error E32 between each track. Factories find CDs made this way quite hard to duplicate.
Regarding the CD-R, two conditions must be achieved to minimize the error rates. On the one hand, as the blank compact disks ain't all manufactured in the same way, one had better favour the models recommended by the manufacturer of the burner over any other.
On the other hand, it is important to burn at the authorized maximum speed (double, quadruple...). Indeed, the gyroscopic phenomenon of rotation at higher speed makes it possible to ensure a better stability of the CD, at the same time as the burning laser works only in shorter thus more constant impulses, (it can do it easily). In addition, the power of the laser is higher when burning at faster speed, therefore the prints are clearer. Let us not forget however that only compatible certified discs must be used.
Lastly, you will avoid dust, cigarette smoke... without mentioning finger prints, which in addition, in the event of police investigation, will make it possible to identify the person responsible for a lack of bass on a piece which might have become the hit of the year...
In addition, as far as DAT is concerned, the situation is not brilliant either. To such a degree that when I provide masters to MPO for manufacture of cassettes, they are grateful to me to provide them a CDR for side A and another for B, rather than a DAT. Their observation: a certain number of DATs present mechanical defects due to the encasing. They regard the CDR as Beingg more secure. And it is known that the projects often finish in last minute in the studios, whereas the installation in the stores must respect quite precise dates. A 24h delay due to a defective master can bear heavy consequences... To finish on the errors of the two supports, do not believe that the DAT is free of them. Far from there. What's responsible for this ? The tape of course, since it comprises a layer of inevitably imperfect metallic oxides. But also the recorder itself and for various reasons: worn out, or clogged head (that cannot be the case with the laser of the CDR) dust, etc...
Fortunately, the reading of numerical information also passes by a sophisticated error correction system, but this one cannot repair everything.
Few DAT players and even fewer CD players display the errors, whereas they are always existing. For example, Sony DAT 7000 series comprise a visual LED display on their front panel. Knowing the error rate is very important, for several reasons. Primarily, it reflects the quality of the transfer.
The hard disk itself is a much more reliable support for digital audio recording than CDR or DAT, since it almost does not comprise errors.
It should be noted that during the burning process, the coding of error correction is of better quality in CDROM mode that in CDR mode. This explains why a CDR can record 74 mn of audio, equivalent of approximately 740 MB on a hard disk, and only 650 MB in CDROM mode.
In any event, in audio mode, DAT and CDR not Beingg perfect, the best advice is to copy your mix on both, especially if you are in the urgency for the remaining operations to be performed next, and all the more more if your project is important. Lastly, for archiving audio, WAV or SDII/AIFF files will be in far better safety on a CDROM than on an Audio CD.
Text published with the kind authorization of
Michel Geiss
Mastering engineer