Edwood,
scanning through cdr.cz site I think they used CD Speed for that test. So, ideally in that bar chart you want a drive that has 100% good and 0% unreadable. They did not include the percentage figures for "damaged, but readable" blocks and that's why the Good + Unreadble don't add up to 100%. The missing portion is the "damaged, but readable" part.
To interpret the figures more fully, one can conclude:
- Ideally you want a drive that has 0% unreadble. This way all data is always able to be read properly (if the drive has working error detection/correction circuitry)
- Ideally you'd also want 0% damaged and 100% good. However, this is not as simple as it seems. A drive with lousy error detection accuracy may report a higher number of "good" and less "damaged", but still report bad data! As such, a graph like that must always be combined by looking at the statistical c2 accuracy of the drive. Another way to test is to use the CDRLabs method of ripping twice and comparing the number of differences. Ideally you'd want 0% differences. However, there are drives that report 0% bad, but still from difference detection they show as much as 78% of data being different. This means the reading is totally unreliable. The difference between reads also varies as a function of read speed.
- So combined, you'd want an ideal drive with 100% c2 accuracy. 100% good blocks read (hence 0% bad, 0% damaged) and 0% difference between two consecutive DAE rips. However, such a drive does not exist.
The best audio cd ripping dvd burner drives from CDRLabs/CDRInfo combined testing results are:
Sony 700A (diff: 0.16%, good 49%/unreadable 0%, c2 acc: 99.7%
Plex 712 (diff: 0.66%, good 30%/unreadable 0%, c2 acc: 74.69%)
LG 4082B (diff: 0.86%, good 15%/unreadable 0%, c2 acc: 75-79.6%)
* c2 accuracy data from LG 4081B (not 4082B, as results for that does not exist)
However, Sony 700A (LiteOn OEM) managed that by scanning at c. 2x, when others scanned at 16x - 24x. It also completely failed the second A-bex test disc and failed to read properly scratches longer than 0.75mm (when the Plextor managed 1.124mm scratches). LG failed even more in the scratche length test and as such I'd not get it for ripping badly scratched discs.
Again, the results are slightly differents from cdr.cz and CDR Info results, probably due to statistical nature of cd reading, different units, firmwares and most of all, different test discs.
The biggest problem with the above is that those tests only use EAC tools to rip audio. Plextools Pro useses a propietary command set on Plexto drives that is able to extract audio off really scratched discs by recovering best bytes and concealing errors. This works remarkably well.
The problem with PX-712 is that EAC does not support it's c2 accuracy calculation, although it exists on the drive.
Of course, if I could have all, I'd have:
- error concealment circuitry of Sony DDU1621 (superior up to 1.5mm scratches)
- C2 reporting accuracy of LiteOn drives
- Ripping speed of Plextor drives
- Plextools Pro in addition to EAC compatibility regardless of who's drive I choose
But again, such a drive does not exist.
So I choose several drives myself and recommend Plextor with Plextools to those who want good audio cd ripping and don't want to get lost in the world of cd audio ripping testing methodologies.
psxguy85,
it should not matter. I have used both and both work ok. However, you may experience higher cpu utilisation rates with external USB drive (this depends on your motherboards USB controller).
regards,
halcyon