The processing time for Shrink/Recode is much longer in this test because you use DEEP ANALYSIS. In NORMAL DEFAULT mode, Recode is much quicker than Shrink and CloneDVD, at least with my 600 MHz PIII. A faster PC will only yield better result. I believe you should rerun the test, with and without DEEP ANALYSIS for Shrink and Recode.
I use AnyDVD with Recode in NORMAL processing mode. After a wait of about 3 minutes to analyse the video, one simply select the VIDEO_TS folder, crop out unwanted items, and burn to the DVD (if you have Nero). Basically the same procedures as CloneDVD. The only difference, step-wise, between Recode and CloneDVD is that you have to wait a few minutes for Recode to analyze the movie. Even with the wait, Recode should put in a MUCH shorter transcode time if you don't select DEEP ANALYSIS/HIGH QUALITY. I'm not sure what you meant when you said that "DVD Copy 3 and Nero Recode are a 2 step process".
For an average 2 hr movie with 20% compression and only the main title with AC 3/6, Recode takes 29 minutes (including the initial disc analysis). Add another 10 minutes for CloneDVD. Shrink is the slowest with a time of 49 minutes. These resulted were obtained with my 600 MHz PIII running W2K SP1.
By your definition of QUANTIZATION (a lower number = will be easier and more fluid to compress further), the original should have a low number. But shouldn't a very efficient transcoder remove most of the non-essential stuffs, thus making it harder to compress the file any further (higher number)? As I see it, quantization should not be a variable that affect the picture quality of the remake DVD. It predicts the ease to which the remake DVD can be further compressed. Now if you plan to re-shrink an already compressed copy of the ORIGINAL DVD, then a low quantization number would be beneficial.
I also agree that the AVERAGE BITRATE is the most important number in this "quality" test. The variation between the top four contenders is plus/minus 2.8 percent...well within the noise level of most non-scientific tests. Yup, definitely not visible under normal viewing condition. DVDCopy 3, however, was consistently lower in both tests.
Which software did you use to obtain these numbers? Can you provide a download link so I can retest using my setup?