A lot of time have passed since our last publication and we got a great deal of letters with your comments and questions. Many our readers ask us to make the results more comprehensible for an average user, others complained about a little number of recorded discs and only two CD recorders. So, we are glad to introduce a new cycle of CD-R testing which will be more objective and informative. Our testing technique has undergone some changes. First of all, now we have 4 CD recorders - Yamaha CRW2200E, TDK CyClone 161040, TEAC CD-W516E and Plextor PX-W124Tsi. Secondly, a level of CD-R recording was lifted to 4x speed, and now each drive writes discs at 4X, 8X and 12X. Moreover, after we consulted with experts from the Urals Electronic Plant, TDK Recording Media and Rostok Media we decided to remake an algorithm of final data calculation. It doesn't mean that the earlier results are wrong. On the contrary, all manufacturers who asked to retest the discs confirmed the correctness. Nevertheless, the in-depth investigation of the test parameters showed that it is impossible to define with an absolute accuracy whether a parameter belongs to one or another disc characteristic in question - mechanical, optical or a characteristic of the number of errors on a disc. All measured parameters characterize a level of errors on a disc. It is not correct to estimate a weight of each parameter or of their groups as any parameter can bring in an error which would cause culling out a disc. For example, BLER and E11–E32 determine the number of found errors on a disc. The less the level of errors, the higher the probability of data integrity. Besides, you should account for algorithms of data reading of different types of CD players: for example, CD drives installed in computers try to read data from a disc at the highest possible speed, and if it is impossible to read without errors they decrease the speed, thus, causing a decrease in a data transfer rate into a computer. If a speed is too low a user would hardly enjoy, for example, watching a movie. At the same time, a disc can contain information for which a reading speed is not very important: programs, drivers etc. A reading speed doesn't matter much for digital audio data as well (of course, within the certain frames). Audio players in audio systems read data in the digital audio format, and a low error level is not important for them as far as data integrity is concerned. So that a form of data representation will be more comprehensible, and according to the advice of CD-R makers we have adopted the following scheme of representation of the results: 100 scores mean an ideal disc with BLER=0, an ideal symmetry, lacking crosstalk noise, E11–E32 errors etc. If a parameter is bounded below and above, an average value between the maximum and the minimum specified in the Orange Book is to be considered a reference one. E32 errors almost halve the result, though such a disc can be considered "dead" as it definitely has unreadable areas. ParticipantsAt first sight, discs are not the most up-to-date. But you should take into account that such a huge test requires a lot of time starting from a recording process (each recorder needs a rest after each disc) and to digital data input into the program of calculation of the results. Besides, a lot of users have plenty of yesterday's discs or recorded data are kept on CD-Rs which were taken for examination. Test resultsHere you can look at the combined diagrams plotted according to the average results of the discs obtained on each recorder. Test results of the discs recorded at 4x More is better
Test results of the discs recorded at 8x
Test results of the discs recorded at 12x
The majority of experts dealing with CD-R production consider that most of brand-name discs keep on closing in in quality. The test results prove it. We do not mean noname CD-R media whose manufacturers are doing their best to gain users' wallets. But to maintain the lowest possible prices manufacturers should save on something: polycarbonate, dye, protective varnish. Another way is to glut the market with disc supplies with the minimal extra charge. But such a way has some drawbacks: a production line for CD-R media has a limited production speed, and "overclocking", for example, from 400,000 to 500,000 discs a month will worsen their quality. By the way, some discs in question have unstable results, i.e. they are defective. Our specialists have found a casting defect in the SKC discs, that is why their BLER factor was 36–40, and some CD-R discs had E32 errors. For thorough test results take a gander at all printouts of the CDCATS SA3 test station.
|