ATI and nVidia crush high-end DVD players

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

TheRyuu

Diamond Member
Dec 3, 2005
5,479
14
81
Originally posted by: rbV5
Originally posted by: xtknight
From my understanding ffdshow uses none of PureVideo (certainly not H.264 acceleration). I suspect it can accelerate MPEG-2 but it cannot use any of PureVideo's image enhancement features either.

So, basically you would be using the purevideo decoder in software mode and ffdshow for post processing...meaning the CPU. Thats what I'm getting at. Any recent vcard will support that very configuration.

It seems reasonable to include those types of configurations when doing these types of tests/comparisons since I regularly hear users tout them as being superior. Basically, you are talking about disabling the hardware feature set in favor of a software solution.

If hardware features like AVIVO and Purevideo are superior (or at least the hardware components of the feature sets); Why not compare them against software solutions? Why are users disabling the features or at least part of the feaure sets in favor of software solutions.....and yes, for H.264 and MPEG/DVD/trasport streams ect.

Yes.
You use Nvidia Decoder to decode and deinterlace (deinterlacing is VERY important).
Plus a few DVD Enhancements in the new CP
Then add ffdshow on top of that
And the DVD's look a hell of a lot better.
 

rbV5

Lifer
Dec 10, 2000
12,632
0
0
Yes.
You use Nvidia Decoder to decode and deinterlace (deinterlacing is VERY important).
Plus a few DVD Enhancements in the new CP
Then add ffdshow on top of that
And the DVD's look a hell of a lot better.

I understand how important deinterlacing is, which is specifically why I ask.

For instance, using ffdshow in the filter chain precludes Nvidia's advanced hardware Pixel-Adaptive deinterlacing for generic software Bob or Weave (in other words, you are disabling your 7900GT's Purevideo hardware for basic software deinterlacing using your CPU cycles)
Plus a few DVD Enhancements in the new CP
Again, aren't the edge enhancements and noise reduction driver features disabled when using software rendering?

How are you using ffdshow exactly to make "DVD's look a hell of a lot better"? Its a full featured audio/video decoder/post processor, so "adding it on top of that" doesn't really explain how you're getting it set up to process the video.

By better, I'm assuming you mean better than simply using Nvidia decoder in a supported DVD software like Nstant media, WMP10/11, MCE.. like used in the linked OP test?

That setup scored a damn impressive 93 with the HQV video test suite. I'm assuming that since you have the Purevideo decoder, that you've tested it with and without ffdshow in the filter chain, and to your eyes....ffdshow software processing trumps Nvidia hardware processing.

That is why I wish reviewers would test against software processing. IMHO, if anything, placebo affect would tend users to give the push to their expensive, heavily marketed hardware over software solutions, but I still hear users, even with these new hardware features, still promote tried and true ffdshow + CPU.

Anyone familiar with both (hardware/software video processing) can readily admit that both can output excellent quality video. The subjective nature of video PQ makes tests like the HQV suite very useful in that it makes it a bit less subjective with the point system (we've seen this with the vast improvement in driver support resulting in much higher scores, from the very first tests where ATI scored very poorly in comparison)

I've yet to see ANY software HQV test results for comparison, maybe there is a reason?.