Compare w/ 2 instances of player possible?

[kikker]

I'd simply like to compare 2 encodes by scrolling through the frames with one encode above the other. Is there a way to do this with 2 instances of the player running - each with 1 of the encodes?

If not, besides running all sorts of psnr tests and the like, I just want to go through the frames and compare them subjectively I guess.

Specifically, I have a Nero AVC Cinema encode running up against an x264 HQ-slowest encode.

[Dark Shikari]

If you want a free comparison tool for plain visual comparison, the MSU Visual Quality Tool is very very useful for this purpose and does exactly what you're looking for.

Before you do the test itself, from my comparisons of Nero and x264, here's what I've found:

1. Nero's raw efficiency is considerably lower than x264 overall.
2. However, its grain retention is much better than x264, even with AQ, because x264's AQ doesn't actually intend to deal with grain; in fact, x264 is basically optimized as a whole to not really retain grain at all. This can be a good or bad thing depending on what you think is "good quality."
3. Nero's psy optimizations are overall more advanced than x264's. The encoder itself is a limited 2004 version of Ateme's encoder.

I would strongly recommend using AQ on your x264 encode if you intend on doing visual quality tests.

[kikker]

Thanks for the tool tip. I briefly compared a 3 minute sequence of Flags of Our Fathers - scene with slow action and a lot human interaction. I found that x264 was a bit sharper, especially notable on faces. However, there were more odd artifacts in that encode wherein close parallel lines resulted in "moireing (sp?)". The Nero encode was overall a bit softer, and the colors were not quite as saturated, while brighness overall was just a notch higher. Skin tones were actually more pleasing on the Nero encode however. I really didn't come away thinking one was clearly better than the other.

[Dark Shikari]

Thanks for the tool tip. I briefly compared a 3 minute sequence of Flags of Our Fathers - scene with slow action and a lot human interaction. I found that x264 was a bit sharper, especially notable on faces. However, there were more odd artifacts in that encode wherein close parallel lines resulted in "moireing (sp?)". The Nero encode was overall a bit softer, and the colors were not quite as saturated, while brighness overall was just a notch higher. Skin tones were actually more pleasing on the Nero encode however. I really didn't come away thinking one was clearly better than the other.
If the colors were different between the two that sounds like one is using a different color conversion matrix than the other. Check your Avisynth script.

[Adub]

Colormatrix is your friend.

[kikker]

OK, I found the best test is simply to be able to view one video at a time and switch between the 2 with the 'tab' key so that the vids are in the same position on the screen. Pausing gives you the ability to inspect and compare frames. So, in this manner then, I could clearly see that the colors were indistinguishable. My comment regarding the nasty moireing actually applies to the Nero encode. And, upon closer inspection, it appeared that there was ever so slightly more noise/graininess on the skin tones of the Nero encode as well. (During the test, I had it in my mind that the 2 encodes were reversed from one another). So, it turns out that although the 2 are VERY close in quality, aside from those nasty moireing artifacts, x264 came out on top.

This probably wasn't a very good test, as the 2 encodes' source was an HD x264 encode itself, and were down-resed to 720 width. They are probably a bit bit-starved too at 850 kbps.