Page 1 of 1


Posted: Mon Nov 21, 2011 12:28 pm
by Danny Lee Fye
As a reference for my thoughts, there is an interesting article at: ... ereCS5.htm

Please note the benchmarks a little past half way down.

Since there are no benchmark charts for speed increases with Neat Video (just scattered forum results) I am considering the ones in the article as a guide for now. Are the differences shown in that article what one would generally expect with Neat Video?

As has already been shown elsewhere on this site and the article the entire system can make a difference in the results one gets. How much vs. costs is the BIG question with not so clear answers!

One interesting observance in the article is the variations in results that one can get with different video formats. When using the optimizer with Sony Vegas Pro 11 I know it uses 8 bit but what video codec is it using for the test render to test the speed? If different from what I am rendering to it seems that would make a big difference in what combination I really need to use.

Should the frame size in the Optimizer be the same as my project 1920x1080 or what I am rendering to such as 720x480 widescreen interlaced or 560x320 progressive? This is not clear in the manual. If the frame size should be the same as my target video then the optimizer needs to be revised to allow one to type in the frame size because the current drop-down box will not have the size(s) I plan on rendering to. From my test results it would be the target frame size. Is this correct? But then what codec?

One important note in the article is memory type. They highly recommend DDR5 over DDR3. I had the 430 card with DDR3 and it would not give any positive results over CPU only when it comes to rendering with Neat Video. When it comes to rendering to mp4 without Neat Video it was a lot better than CPU only. It was a little bit better with Neat video. Actual results vs. the Optimizer are somewhat different!

I also had the 550 Ti and it did give positive results over CPU only but only with certain settings. There were results that were less than CPU only with small frame sizes and interlaced video. Progressive video would be a tiny bit better with a combination of GPU and CPU. It was nothing to get excited enough to have a party about. LOL! The absolute test results do not give a clear idea on what one can actually expect because of the variations of different codec’s, frame sizes to be rendered to, if the video is to be progressive or not and other system facts. The results can be misleading! While helpful, the Optimizer has a number of flaws. Also, the dialog box keeps getting vertically larger with each use eventually causing the ‘start’ to no longer be visible and I have to resize it to get it back.

I have returned both the 430 and 550 Ti cards and I am now thinking of getting either a high end 560 or a low end 560 Ti. The high end 560 is upped by the manufacturers over-clock and the low end 560 Ti simply has more CUDA cores. The prices on some cards are the same. So I think I have decided on the low end 560 Ti but I keep changing my mind as the wind changes directions.

And yet from the above article I can expect very little difference based on the Adobe Mercury Playback Engine rendering results. What are the REAL differences beyond the Optimizer I can expect with Neat Video and Vegas 11?

There was no question in returning the 430 card but I am now second guessing the returning of the 550 Ti card. From reading the article I am left wondering if the 560 or 560 Ti will really make enough difference considering the watts used and money spent!

Considering system specs, I recently upgraded my memory to a faster type. The results are 15% faster. This is with CPU only for now because at the moment I do not have a CUDA type video card.

The fastest GPU with an otherwise poor system will certainly NOT give the desired results! The fastest horse has to hurry up and wait on the slowest horse! The same applies to the fastest CPU! And then the fastest system with the fastest GPU will give what? Another fast horse will…

My main concern is results vs. costs! I certainly could save up a few pennies (LOL!) and get the 570, 580 or even the 590 and heat up my computer room from all of the watts they use but will the ACTUAL speed increase and costs be worth it? Based on the article, NO! Based on Neat Video, the jury is still out.