Thanks for the great scans!
As usual, CDR-Info is pushing the limits and finding out new stuff, instead of just trusting blindly the kProbe results. I really applaud you for doing this. I couldn't afford it myself (my cd-r testing cost me enough as it was).
Now, I've been thinking about this issue as well and your test seem to correlate with my hypothesis.
- Causal reasings (PI/PO/PIF/POF) error rates are caused by mismatch between low level values of the disc (jitter, reflectivity, etc.) and the tolerance margins for those levels by the reader.
- LiteOn dvd-drives have relatively high jitter tolerance margin (sometimes twice that of lesser drives)
- Many dvds burned with Lite-On dvd burners measure/read quite well on LiteOn drives, but are utterly and totally unusable in almost any other drive (i.e. disc not recognized, video image becomes erroneous, etc.)
- CATS scanners measure the disc in accordance to published (and sometimes in-house determined) specifications and limits. These limits or guideline values are imposed on the low level disc performance values. Error rates are just a result of these. Most CATS operators look at the low level results and based on those, adjust their production line. PI/PIF values are an overall error level check, not the be-all-end-all of testing.
1. At lower rotational speeds LiteOn drives can easily compensate for bad low level values on the burned disc. For instance, jitter, which usually rises towards the end of the recording (outer rim of the disc) is more easily to keep under check at 1x
2. CATS scanners mark DC jitter values over 9,5% (? at least this was done on the C't tests). On a general level DC jitter should remain below 15% average. Lite-On drives can tolerate jitter up to 20% average.
3. At higher speeds both servo control accuracy and jitter tolerance margins come into play much more, especially at the end of the disc (were jitter values often rise). This results in rising causal error rates when higher rotational speeds are used and the average jitter tolerance of the drive is diminished. The end result is a significant rise in error rates towards the end of the disc at higher scanning speeds.
Now, jitter is also not be-all-end-all low level measure. I'm postulating that other low level measures in consumer dvd-rw drives have somewhat similar correlation: the drives themselves are able to tolerate much bigger low level deviations as readers than what the specifications really need.
I'd suspect it's easier (at the time of this writing) to produce a dvd-rw writer that reads the disc burned on itself properly (i.e. high reader tolerances) than actually make a good burner that produces so small low level deviations on the burned disc that the resulting disc can be read in any drive.
This is one of the reasons I don't recommend LiteOn for a beginner (on dvd-burners): it can burn crap discs, but it can still read them back faultlessly and even produce misleadingly good measurement results. Or to be more honest, the readings are not necessarily misleading, it's just that the people take them for what they are not. They are not an universal measure of disc readability or burn quality, like many people in many forums tend to think.
I'd rather recommend a drive that is an average reader, but an excellent burner: it will catch its' own crap burns much faster and provide much better compatibility over a larger number of drives. Two such drives IMHO are LG GSA-4081/4082 and BenQ DW800A/DW822A (based on low level measurement data on various discs, as published by Pc Professional and C't Magazine).
I think it is great that CDR-Info has again pushed the mark and shown us how to do things right.
Please provide more scans, if you can.
What I learned with my CD-R test (kProbe/PlextoolsPro/CATS correlation) was that:
1) Plextor Premium, LiteOn 48327S/52327s can still read error free disc that CATS even refuses to recognize
2) Sometimes Plex Premium vs LiteOn 52327s scans were totally different. More often CATS scan would reflect the values that the Plextor got.
3) Not all badly burned discs were always most readable on LiteOn. Some discs were more readable on Plextor Premium and horribly useless on a LiteOn. CATS scan would then correlate more with LiteOn. This seems to indicate that the drives in question had different low level tolerance margins for various low level measures. Plextor may excell on some values, while LiteOn on others.
If you can, please also test on Optorite, because apparently it has much lower jitter tolerance margin than what LiteOn burners have. It should (hopefully) provide PI/PIF scan more in line with CATS scans.
Also, I'm sure you know this, but CDSpeed 3.0 beta can provide some sort of jitter measures from Nexperia chipset drives (NuTech, Philips, BenQ). However, nobody knows if these jitter values are anywhere useful/accurate or just "pretty graphs" that tell us pretty much nothing :)
So, if you can do it in the future, consider doing low level scans and comparing those as well.
Thank you again for a great discussion opener on the issue!
PS My status went to "new member, 1 post"? Was this because I had to reset my password?
< Message edited by Halc -- 5/14/2004 6:12:34 AM >