@RB and others with experience

[DDogg]

RB, first of all, my hat is off to you for eclCCE. Very nice indeed!

As I have been away for quite sometime I would like to get your opinion on the various flavors of CCE that are now usable with D2S. I tend to be a "if it ain't broke, don't fix it" type and have continued to use the tried and true 2.50. Because of your work I have started to experiment a little again. Before beginning the laborious and time consuming undertaking of comparison I just thought I would get your take on the subject because it kinda sounds like you are on top of it :)

Clear visual quality increases? Better use of bitrate or better bitrate distribution? Speed? CCE 2.5, at a quick first glance, still seems fastest to me, but as said that is just a first impression. If you have time it would be good to learn from your comments > Else, I totally understand and of course no problem if you do not have the time or inclination.

Best Regards,

DD

[RB]

DDogg, thanks.

To tell you the truth, I can't see any obvious differences in image quality between 2.5 and 2.66+, but then I haven't looked too close. Some people say that 2.66+ produces more vivid colors and a sharper image. I also cannot tell whether bitrate distribution is better because this would require a quite time consuming comparison of the graphs in Advanced VBR Settings in both versions using exactly the same settings. I'm too lazy :)

However, I simply like to stay up with the latest developments. Altough I don't need them, the enhanced/added filters in 2.66+ may come in handy for some. It is also a fact that 2.66+ is P4 optimized and should in theory be faster on a P4, but because for most of us (including me) AVISynth is still the limiting factor, I don't notice a speed advantage on my P4 system compared to 2.5. But they definitely increased stability in the later versions. For example, CCE 2.50 would always freeze on me at a certain point when encoding Shrek, no problem with 2.66. The possibility to use different matrices is also an advantage over 2.5.

You still decide for yourself :)

[DDogg]

Thanks for the reply. This was actually just what I was looking for. It sounds like we are still a bit away from a clear consensus so I will change my ways gradually as I run into a specific need for some new feature or such. I have two encodes on my HD from the same source and settings, one CCE 2.50.01.00 and the other 2.66.01.07, that were done while doing some testing for upcoming versions of D2V. Been so busy with that I have not actually had much time to actually look at them. If they show any clear difference I will report back to the group. Thanks again.

[Matthew]

Unlike RB I find 2.66(7) to be noticeably faster than 2.50. I have dual Xeons though :D

From memory the max I got on 2.50 was a speed of 1.8 whereas I always get at least 2.0 with 2.66(7) :) Save version of avisynth of course :)

Incidentally I just disable the filters because I don't understand them, so from that point of view 2.66(7) holds no benefit over 2.50 as far as I'm concerned.

[DDogg]

I did a quick side by side with the versions mentioned above in two instances of MPC. It does look like 266 yields a slightly brighter picture. I would guess the equivalent of 2.5 to 3.5 points of luma correction in dvd2avi. This made some colors look a little better but I did not notice anything telling. Of course, these were both highbit (2450 average 3 pass) so hell they ought to both look good. Maybe when I have time I'll tweak the brightness slightly in 2.5 and see if I can compare apples to apples. I did prefer the slightly lighter pic though. Speed was higher on 2.50 with my lightly tweaked 1.4 tbird. Around 1.60 versus 1.40 on 2.66. Nothing scientific on that, just kept an eye on the RT on both while encoding. I guess I would actually need to run a stopwatch on them to be able to really tell. Maybe some timings are buried in the log file. I seem to remember asking d2s to do that for us. Long time away from all this stuff. It is amazing how fast you forget it.

/edit log show 1.6 for 2.50 and 1.37 for 2.66