[ Print Article! ]

Catalyst 5.13: ATI's Video Domination?
December 16, 2005

Summary: ATI owners (and prospective owners) must have been good this year because Santa has brought the thirteenth Catalyst driver release of 2005. With this new version, ATI claims that they've raised the bar in video performance and set the stage for worldwide "video dominATIon." Our tests show that Catalyst 5.13 is the greatest leap in video quality technology for ATI since the original Mach64-VT.


IntroductionPage:: ( 1 / 12 )


Nonetheless, FiringSquad's sources brought information of a new driver that would change things for ATI. Not simply a driver that would bring ATI ahead of XGI and NVIDIA and back to the top of the PC video processing world, but a driver that would bring ATI ahead of consumer electronics companies such as Sony, Samsung, Pixelworks, and Faroudja and put ATI hot on the tail of Gennum and Silicon Optix. Seeing that twelve drivers had already been released by ATI in 2005, most readers assumed that the new driver would have to come in January 2006 as Catalyst 6.1. ATI's Secret Santa decided to release this gift earlier to PC enthusiasts with a 13th driver release in 2005 and Oh, how things have changed!





Yes that’s really an ATI-developed logo for Cat 5.13.

Modern Day GPU Design

There's no secret that today's GPUs are becoming increasingly complex devices to manufacture and build. When it comes to the video processing portion of the GPU, it's probably one of the most challenging elements of design. Not only does video processing require a substantial amount of computational horsepower, but the algorithms are unclear. Knowing what to do and how to do it is much more difficult than building silicon that is fast enough for the task.

ATI considers Catalyst 5.13 as the first driver that unlocks the AVIVO technology present in the Radeon X1800, but it's more than that. Catalyst 5.13 represents the countless hours needed to understand the problems associated with deinterlacing and image enhancement as well as the time spent generating novel solutions.



First, our testing methodsPage:: ( 2 / 12 )

Test system

- 2x AMD Opteron 252
- 4GB Corsair Registered ECC PC3200
- Tyan Thunder K8SE
- Hitachi 7K500
- Silverstone Zeus ST65ZF 650W power supply
- Lian-Li PC-V1200
- Pioneer DVR-109
- Windows XP Professional SP2

ATI All-in-Wonder X1800XL
Catalyst 5.13
ATI Multimedia Center DVD 9.10

NVIDIA GeForce 6600GT
Forceware 81.95
NVIDIA PureVideo Decoder 1.02-185
Windows Media Player 10

XGI Volari 8300
Reactor 3.03.03
CyberLink PowerDVD 6

NEC TheaterSync with HQV Processing
Firmware 1.12.2.618
Silicon Optix Realta
MyHD MDP-130 DVI (source)


Disclosure: I am a consultant to Silicon Optix and have been involved with documentation for the current HQV Benchmark DVD as well as the development of the next-generation HQV benchmark suite for high-definition video processors.

Although H.264 and 1080i deinterlacing are critical elements of high-end video performance, we will focus on standard-definition deinterlacing due to time constraints. The HQV Benchmark DVD is used as our test DVD. This benchmark was chosen given our familiarity with the tests (the second revision of the testing guide was written by me)

All of the tests were conducted using a 1280x1024 DVI connection.

Screenshots were taken with the Epiphan VGA2USB. Since we're doing a true analog capture of image quality, please be aware that differences in brightness, color, contrast, or even aspect ratio are a fault of the capture device and that the images are only designed to provide examples to go along with the text. Make sure you view the HIGH-RES version of each of the images. Feel free to post any questions in our comments section if it's unclear whether there's a difference between image quality or if it's simply an artifact from the way we captured the video.



SIDEBAR: Could the use of a Tyan K8SE be a hint at a future article?



Color Bar / Vertical DetailPage:: ( 3 / 12 )

ATI had a perfect score in the last test, and with Catalyst 5.13 there's no difference. The ATI card flickers for a second or so, but then realizes that the image is not-moving. Both NVIDIA and XGI drop resolution but do not flicker meaning that NVIDIA and XGI have a harder time detecting when things are moving or not.

[image]

<% print_image("01"); %>
<% print_image("02"); %>
<% print_image("03"); %>

Score:
1st place: ATI (10 points)
Tied for 3rd place: NVIDIA and XGI (5 points)

Interpretation:
The ATI is the best solution for picture slideshows where there's no animation.



Jaggies TestsPage:: ( 4 / 12 )

The three Jaggies test evaluate how well a video processor is at determining whether objects are moving or not. The first "jaggies pattern" is a simple spinning bar, the second "jaggies pattern" is a set of three bars with a waving motion (like waving your hand), and the third test is a real-world picture of the American flag in the wind. These all represent scenes that a deinterlacer will easily be able recognize as having motion.

[image]

<% print_image("04"); %>
<% print_image("05"); %>
<% print_image("06"); %>

Catalyst 5.13 improves ATI's performance on theses tests substantially. Whereas NVIDIA develops jaggies in the yellow zone, both ATI and XGI are perfect with their diagonal interpolation.

Under Jaggies 2, all three properly smooth the first two bars, but none are able to keep the third line smooth.

[image]
<% print_image("07"); %>
<% print_image("08"); %>
<% print_image("09"); %>

Although all three earn the same numerical score, ATI and XGI do better than NVIDIA in preventing the aliasing in the third bar.

[image]
<% print_image("10"); %>
<% print_image("11"); %>
<% print_image("12"); %>


In the real world flag test, ATI does an excellent job. It appears to detect edges a little bit better than the XGI's approach, resulting in fewer jaggies. Both XGI and ATI are substantially better than NVIDIA. Although ATI's Catalyst 5.13 still shows some very minor aliasing, it takes careful A/B scrutiny between ATI's Catalyst 5.13 versus DCDi or HQV to see the difference. ATI seems to fall behind just a hair compared to the HQV processing, but since our testing is done at 1280x1024, it's possible that the differences are due to the video scaling portion. For the flag test, ATI receives the full 10 points.

Score:
1st place: ATI (18 points)
2nd place: XGI (13 points)
3rd place: NVIDIA (11 points)

Interpretation:

When it comes to de-interlacing 30 fps content such as TV shows, the ATI Radeon X1800XT and XGI Volari 8300 are superior to the GeForce 6600.



Detail EnhancementPage:: ( 5 / 12 )

When it comes to detail enhancement, NVIDIA does not implement overlay-specific sharpening tools. XGI does. In the X800 comparison, the screenshots suggested that ATI was applying some level of sharpening to the image as a result of the way the MPEG-2 decoder was written. For the X1800XL I used the Multimedia Center that shipped with the All-in-Wonder X1800 XL card and it looks like ATI is now faithfully reproducing the original image without additional sharpening. This is definitely a good move by ATI because I have always said that forced sharpening is worse than no sharpening..

[image]

<% print_image("13"); %>
<% print_image("14"); %>
<% print_image("15"); %>

Score:
1st place – XGI (5 points)
3rd place: ATI and NVIDIA (0 points)


By the testing standards, NVIDIA and ATI get 0 and XGI gets 5. However, as I have said before, the numbers can be misinterpreted. A score of zero means that ATI and NVIDIA are faithfully reproducing the material on the DVD. While sharpening is a necessity, the problem occurs when you have sharpening applied at multiple stages. Some DVDs are already sharpened, meaning that you won't want any additional processing. Some televisions sharpen and so you won't want the video card to do any additional processing.

Interpretation

When outputting to a television, you will want to have the TV control the level of sharpening whenever possible because it takes into account your viewing distance and the screen size. All three cards should be great for use with a television.

When outputting to a monitor, you will typically need some level of sharpening and so the XGI is the best option here.



Noise reductionPage:: ( 6 / 12 )


Noise reduction is one of the most important elements of a video processor and even in the high-definition world, noise reduction is very important as can be seen from these real-world examples showing the possibilities of a sophisticated wavelet-based noise reduction algorithm.

Nevertheless, like sharpening, noise filtering can be performed prior to DVD mastering. Although ATI preserves the detail in these images from the HQV Benchmark, it's possible that the noise reduction causes loss of detail in a video clip where the noise is absent. More testing will be required to evaluate this possibility. Of course, it's equally possible that a user might want to be more aggressive with the noise reduction. The differences are more prominent in motion.

[image]

<% print_image("16"); %>
<% print_image("17"); %>

Score:
1st place: ATI (10 points)
Tied for 3rd place: NVIDIA and XGI (0 points)

Interpretation

ATI is the first of the big GPU manufacturers to implement noise reduction into their DVD playback. This is an important step for PC video enthusiasts. Nonetheless, ATI should provide users with the ability to disable any noise reduction should a user want to reproduce a "faithful" image from the DVD.



3:2 DetectionPage:: ( 7 / 12 )


The Radeon X800 XL with Catalyst 5.12 failed miserably on both of these tests. The Radeon X1800 XL with Catalyst 5.13 flies through these tests with flying colors. Both XGI and ATI require a little bit of extra time to detect the 3:2 cadence in comparison to NVIDIA, but it's still fast enough to score a perfect score.

[image]

<% print_image("18"); %>
<% print_image("19"); %>
<% print_image("20"); %>

Score
Tied 3rd place: ATI, NVIDIA, XGI (10 points)

Interpretation

Catalyst 5.13 makes the ATI Radeon X1800 XL a great choice for watching movies. NVIDIA and XGI are equally good for playback of DVDs films as well. NVIDIA does seem to lock on a fraction of a second earlier (we were unable to capture the video output from the ATI to make a true frame by frame comparison).



Unusual cadencesPage:: ( 8 / 12 )


Of all the cadences, 3:2 is the most critical. That's your Hollywood blockbuster film. That's the bulk of what you'll end up watching. If your system doesn’t reliably detect the 3:2 cadence, your video processor may be discarding half of the resolution available in Hollywood movies.

2:2 cadence can refer to putting 24fps film onto 50 fps PAL. However, it also refers to 30 frames/sec progressive content on DVD or 1080i. This is the second most important cadence because you'll find it being used in the documentaries and many TV shows that are put on DVD. If you watch TV shows on DVD or have a Media Center PC, you'll want good 2:2 detection.

The 5:5, 6:4 and 8:7 cadence refer to cadences used in both western and eastern animation (8 fps, 12 fps). Vari-Speed Broadcast (3:2:3:2:2) is important if you enjoy watching movies on broadcast TV. Lastly, the 2:2:2:4 and 2:3:3:2 cadences are the least popular formats used. If you're involved with indie-films, make your own videos with software like Final Cut Pro, or need to review digital dailies, it’s important. Otherwise, it's a nice to have on a checklist but is the least essential for typical viewers

ATI detects all of these cadences perfectly. This is a substantial feat from ATI because if you encounter a video with an unusual cadence, ATI is able to preserve all of the detail in the original images. While NVIDIA detects Vari-Speed Broadcasts, it cannot detect any of the other cadences. Thus, if you are watching 8 fps anime, you may be throwing away half of the resolution available in the image.

With Catalyst 5.13, ATI seems to detect all of the known cadences tested by the HQV Benchmark. Amazing!

1st place:ATI (35 points)
2nd place: NVIDIA (5 points)
3rd place: XGI (0 points)

Interpretation

If you watch a lot of documentaries that were shot on 30 fps film, or a lot of animation, the ATI Radeon X1800 XL with Catalyst 5.13 is the unanimous choice. When it comes to unusual cadences, ATI AVIVO completely outclasses PureVideo and TrueVideo.




Mixed film/videoPage:: ( 9 / 12 )

Mixed 3:2 content is the least important feature to test for unless you're watching movies shown on MSNBC or when you have 24 fps film content mixed with 30 fps special effects. Here, ATI and NVIDIA do similarly well. They appropriately de-interlace but fail to apply 3:2 pulldown to the film portions of the screen. XGI doesn't handle mixed content very well at all.


[image]

<% print_image("21"); %>
<% print_image("22"); %>


Score:
Tied 2nd: ATI and NVIDIA (15 points)
3rd place: XGI (0 points)

Interpretation

You may see artifacts when watching movies on TV on CNN or MSNBC or during the credits sequence with the XGI. You don't see artifacts with the ATI or NVIDIA cards under such conditions, but you still lose half your resolution.


Total HQV Benchmark Scores

ATI Radeon X1800 XL with Catalyst 5.13: 103
"With Catalyst 5.13, ATI becomes the first PC GPU to break the 100 point barrier on the HQV Benchmark DVD. This makes ATI AVIVO an excellent choice for DVD playback."

NVIDIA GeForce 6600: 51
"Still going strong after one year, PureVideo represents a good video solution that provides great image quality with Hollywood films. Although NVIDIA's video deinterlacing was once considered the best among PC solutions with XGI and ATI now offering better diagonal interpolation, NVIDIA PureVideo is due for an update."

XGI Volari 8300: 38
"The XGI's low score on the benchmark is due to the fact that the Volari doesn't handle uncommon scenarios well. But common things are common, and for Hollywood movies and pure interlaced content, it's actually just as competitive as ATI's AVIVO with the exception of noise reduction and static images."



Real world testingPage:: ( 10 / 12 )


Real-World Cadence Challenge #1: MegaZone 23 Trailer

First up was the trailer for Megazone 23. For the non-anime fan out there, Megazone 23 is one of the classics of 1980's anime and was the basis for Robotech: The Movie. Although this is actually a 3:2 sequence, and the X1800 was unable to detect the appropriate cadence in our real-world test. We saw plenty of jaggies and feathering artifacts and our findings were confirmed by ATI. What's interesting is that NVIDIA's PureVideo, which only scored 51 points on the HQV Benchmark actually does well here (although it occasionally drops the resolution). NVIDIA GPUs are able to detect this cadence appropriately and for this clip, NVIDIA actually locks onto the cadence faster than even HQV solutions. The XGI Volari 8300 was not tested.

According to ATI, in turns out that one of the challenges with this clip is that there's a lot of noise in the clip which causes the GPU to assume that the fields are different when they are not. Their detection thresholds are apparently conservative in order to detect bad-edits appropriately. Fortunately, the feathering artifacts are due to a bug in the driver which is correctable and should make its way into a future driver.

Real-World Cadence Challenge #2 and 3: More Anime

With the experience of the Megazone 23 trailer, I decided to challenge the video processors with newer anime. In retrospect, using anime as an "unusual cadence" test makes a lot of sense because there are a lot of fast cuts and edits which makes it very challenging for video processors to lock onto the right cadence. On that same DVD were trailers for Magical Shopping Arcade Abenobashi (trailer) and RahXephon (trailer). In these tests, I found that the HQV processor handled the cadence detection well whereas the ATI X1800 had substantially more difficulty resulting in feathering artifacts. NVIDIA had no feathering artifacts, but frequently dropped to half-resolution when it was unable to appropriately detect the cadence. The bug fix ATI mentioned for a future driver release would take care of the feathering artifacts although it's unclear if tweaks to their cadence detection mechanism can be done.

Interpretation? Although ATI Catalyst 5.13 has raised the bar when it comes to PC video, it doesn't do nearly as well with the unusual cadences in some of the real-world challenges that we threw at it. I know what some cynical readers might be thinking, but this is really a reflection of how challenging video processing is. There's no brute-force solution.


SIDEBAR: We were unable to grab good screenshots with our VGA2USB device. What we really need is a DVI to HD-SDI converter and a PCI-X HD-SDI capture card. Donations are accepted.


ATI shortcomingsPage:: ( 11 / 12 )


HD Challenge: Deinterlacing

For the deinterlacing challenge, I used a test clip where a spinning bar is superimposed over a SMPTE resolution pattern. A good deinterlacer should be able to detect that the bar is spinning (and apply appropriate processing to those regions only). NVIDIA has always done well with this test, generating artifacts in the squares near motion only. ATI's Catalyst 5.13 brings their HD deinterlacing capabilities to a similar level. (Both ATI and NVIDIA still produce artifacts due to the size of the processing region).

HD Challenge: 3:2 Cadence

As the 3:2 cadence used for 24 fps represents the most important cadence, we wanted to evaluate how well the graphics cards did with HD cadence detection. This is important for watching movies that are broadcast in HDTV or those on D-Theatre. Likewise, it's not entirely clear if Blu-Ray and HD-DVD will be recorded as 1080i or 1080p (remember, the supported output doesn't tell us what's actually recorded).

While both NVIDIA and ATI apply region-based motion-adaptive deinterlacing in high-definition, neither are able to appropriately detect the 3:2 cadence of our 1080i test clip.







H.264/ConclusionPage:: ( 12 / 12 )

H.264 Acceleration

The Catalyst 5.13 drivers also bring H.264 acceleration as well as the AVIVO Video Converter. This means that you'll be able to watch those high-definition 1080p trailers from Quicktime with an X1800XL with ease. At the moment, we've only had time to do some limited tests with the X1800XL, but it seems like H.264 acceleration works very well. We'll have to spend some more time in January looking at H.264 decode performance with a wider range of CPUs – the dual Opterons I've been using obviously is an atypically powerful configuration.


Conclusion

Catalyst 5.13 represents the greatest leap in video quality technology for ATI since the original Mach64-VT. It brings sophisticated diagonal filtering, basic noise reduction, and the ability to detect unusual cadences. ATI marketing told us about "video dominATIon" and we have to agree that Catalyst 5.13 brings ATI to the top of the PC video world.

One of the best features of the Catalyst 5.13 is that the software is a free upgrade. As long as you have DVD software that supports DXVA, the AVIVO video processing will work. Owners of Built by ATI cards obviously have access to ATI's Multimedia Center as well. In comparison, PureVideo has never been free. It starts off at $20 for the "bronze" edition with basic SPDIF out support or 2 channel audio, $30 for the gold version with 5.1 analog out, and $50 for the platinum version with DTS support. When price shopping for a new GPU, if you're going with NVIDIA you need to factor in the extra $20 that you wouldn't have to with a built-by-ATI card. Although (or because?) NVIDIA PureVideo cannot detect the unusual cadences, its 3:2 algorithm is faster and is able to lock-onto cadences such as some of the anime test clips. ATI believes that NVIDIA PureVideo is faster because the PureVideo MPEG-2 decoder is able to provide more information to the GPU. Still, we'd certainly prefer that PureVideo be free for NVIDIA owners, at least for the GeForce 7 series or that NVIDIA could offer a free DXVA-only video processing solution like ATI.

Ten days ago, ATI had the worst video quality on the PC. With this new driver, ATI has jumped to the top of the class and then built a nice lead. For Hollywood films, ATI edges out NVIDIA's PureVideo. Although the cadence detection is faster with NVIDIA, ATI has noise reduction will improve the quality of films once the cadence is locked on. In other words, for most Hollywood films, ATI will look better for the 2 hours whereas NVIDIA might look better for the few seconds after a bad-edit, if such a bad edit exists. When it comes to processing interlaced video, ATI now surpasses anything that we've seen on the PC with the best implementation of diagonal filtering we’ve seen yet. When it comes to unusual cadence detection, ATI AVIVO has no peer in its price range.

There's still room for improvement though. Although ATI and NVIDIA apply a region-based motion-adaptive deinterlacing algorithm to high definition sources, they do not seem to reliably detect the 3:2 cadence in high-definition. Neither ATI nor NVIDIA have the ability to handle mixed film/video content with the same skill that some dedicated scalers can. Neither ATI nor NVIDIA have advanced detail enhancement or advanced noise reduction algorithms that can take picture quality another step forward. Even so, with Catalyst 5.13, ATI is making a serious challenge toward the dedicated video processors that cost $3000+ and they're doing it as a free upgrade. With a fully programmable architecture, there is also room for ATI to grow. Companies like Faroudja, Silicon Optix, and Gennum should take notice at this new competitor to the market.

The video engineers at ATI deserve a big holiday bonus. They've made a big leap in performance and will make Catalyst 5.13 free for download on December 22nd. Perhaps the greatest benefit of AVIVO technology is that for most of FiringSquad's readers, it's an added bonus. Most people aren't buying the X1800XL for video performance alone – they're buying for it for the flagship gaming performance.

So what does the future bring? Well, most of what I said 10 days ago still applies. XGI still offers a nice product for $50. NVIDIA's PureVideo is just as good today as it was 10 days ago, but ATI has tripled their deinterlacing performance in those 10 days. NVIDIA must now play catch up with ATI. What about ATI? Well, it turns out that this level of deinterlacing performance was planned from the get go with the X1K family of GPUs. Now that these drivers are about to be released, their engineers are working on enabling the next set of technologies for AVIVO to bring even more functionality and quality. We can’t wait until 2006!

Editors Note: The driver we tested for this article is the same driver ATI will be releasing next week, only the final driver will add WHQL-certification. ATI is working with Microsoft on WHQL-certification as we speak, with an ETA of 12/21.

© Copyright 2003 FS Media, Inc.
[ Print Article! | Close Window ]