HQV Film Cadence Test

This section contains a series of 8 tests with different cadences to see how well the processor handles each cadence. While 3:2 is the most popular cadence (used for movies), other cadences like 2:2 or 5:5 might be necessary for viewing home videos or animation. Jagged edges and moire in the newspaper and coffee cups indicate poor decoding.

ATI managed to pass all of the cadence tests and score a perfect 40 points. NVIDIA however failed all of the tests except for two: 3:2, and 3:2:3:2:2. Granted, 3:2 is probably the most important cadence (24fps film requires this cadence), so it's good that NVIDIA manages to pass this test, but this is one area that ATI does decidedly better. There are quite a few tests for this section, and all the images look pretty much the same. As such, we decided to show only one case that represented most what we see (2224). The key difference is in the ring around one of the coffee cups. Take a look at how smooth and solid it looks on ATI hardware. There is s slight smudge in the ring on NVIDIA's screenshot. The difference is actually pretty subtle and hard to see in motion if you don't know what you are looking for, but there it is.
  • ATI: 40
  • NVIDIA: 10
  • (highest score: 40)




HQV 3:2 Detection Test HQV Mixed 3:2 Film With Added Video Titles Test
Comments Locked

43 Comments

View All Comments

  • Sunrise089 - Friday, February 10, 2006 - link

    You have identical numbers for w/ AA and w/o. Also, the text's comment on the X1900AIW being playable at all reolutions with AA uses the incorrect numbers.
  • plonk420 - Friday, February 10, 2006 - link

    does the ATI decoder give you the option to ADD sharpening? or not at all? my whole reason for wanting a (hardware accellerated) software decoder is so i can have a pic quality rivaling a ("popular") $200 hardware player for whatever extra it would be for the software. supposedly free for ATI or $30 for nVidia (for my application: 2.0 sound)
  • hwhacker - Friday, February 10, 2006 - link

    here we go:

    http://www.bytesector.com/data/bs-article.asp?ID=6...">http://www.bytesector.com/data/bs-article.asp?ID=6...

    590/684...I was close.

    18+% improvement in 3dmark06, you know that has to translate to something good in gaming.


  • hwhacker - Friday, February 10, 2006 - link

    It uses 2.0ns chips from what I recall, as does newer BBA x1800xl's (instead of 1.4ns).

    There was one site that did an overclocking section on it, I forget which. The results were similar to x1800xl's, the end clock speed ending up 600+/almost 700 iirc. You know how XL's clock, i'm sure.

    So in essence, yes, it overclocks well, and I do remember the site being amazed by the performance improvement through overclocking. I still don't get how 2.0ns chips can hit 1.4ns speeds if there is a speed bin in-between for cards like nvidia's 7800gt/gtx that you would think use that supply...but i've seen quite a few cards with the newer, slower, chips hitting the same approx speeds as the old ones with 1.4ns, and i'm not complaining. ;)

    I'm sure with the overclock or ATiTool soft-vmods this thing would be killer, especially with better cooling than the known-for-sucking XL stock cooler.
  • Shadowmage - Friday, February 10, 2006 - link

    What I'm curious to know is whether the AIW can overclock to roughly XT/XTX speeds.

    What type of RAM does the AIW use?
  • Zebo - Saturday, February 11, 2006 - link

    Agreed. How can AT not include this? Lame.
  • DigitalFreak - Friday, February 10, 2006 - link

    Although I appreciate the DVD decoder tests, how is this review related to the AIW features of the card?
  • highlandsun - Friday, February 10, 2006 - link

    I would have been more interested in seeing how well it handled H.264 decoding at 1920x1080p.
  • oxid - Friday, February 10, 2006 - link

    does the 7800 gt use another video processor then the GTX? because in the last review with the HQV tests the 7800 gtx scored better then the gt in this review...
  • mpeavid - Friday, February 10, 2006 - link

    You guys need to use the exact same frame as an example for all cadence tests. Not doing so can invalidate your test.

Log in

Don't have an account? Sign up now