Ready to HD: Performance of Contemporary Graphics Accelerators during Video Playback

Besides gaming, video content playback is another major application field for contemporary graphics accelerators. As the HD-formats are getting more and more popular, we decided to perform an investigation and find out what graphics cards suit this purpose best of all.

by Alexey Stepin , Yaroslav Lyssenko, Anton Shilov
02/26/2007 | 05:24 PM

Playing video on the PC used to be a non-trivial task in the past. A hardware decoder was required even to play such a simple (by today’s measures) format as MPEG-1 at a resolution of 352x288 pixels. It is only after some considerable growth in the computing capacity of CPUs and after the release of graphics cards capable of hardware scaling and YUV/RGB color space conversion that the playback task could be transferred to the CPU and graphics card.


The problem emerged anew with the expansion of the DVD standard that utilized the MPEG-2 video format. A MPEG-2 stream consists of three kinds of frames: I(ntra-coded)-frames, P(redictive-coded)-frames, and B(idirectionally predictive-coded)-frames. An I-frame is created by compressing a frame from the original video material and inserted into the encoded stream with a certain frequency, usually one I-frame per fifteen P- and B-frames. P- and B-frames are created using reconstructed frames and a so-called motion vector which is assigned to each 16x16 pixel block of the reconstructed frame.

The use of information from previous (for P-frames) and from previous and next frames (for B-frames) helps increase the compression degree, but requires more computing power from the decoder. Besides, the decoder has to perform an inverse discrete cosine transform (iDTC) to uncompress I-frames which are compressed using this algorithm. Thus, the graphics card must support hardware motion compensation and iDCT, besides hardware scaling and color space transformation, for the computer to be able to reproduce MPEG-2 at an acceptable CPU load.

Eventually, all graphics cards acquired features necessary to play MPEG-2/DVD easily. But now there have appeared new compression methods that ensure better quality but also require more hardware resources. Coupled with the advancement of HD video content, the ability of the graphics card to decode and play high-definition formats becomes an issue again since software playback of such video may get even the strongest CPU to its knees. So we decided to carry out a comprehensive test of graphics cards from AMD/ATI and Nvidia to check out how much they offload the CPU while playing various video formats, including high-definition ones. We’ll start out with a brief review of technologies implemented in those cards.

ATI Avivo vs. Nvidia PureVideo: Evolution

Until Radeon X1000, ATI’s graphics processors offered a rather modest video engine, but were capable of performing some decoding and processing of video on the GPU through DXVA. Such GPUs are already obsolete, so we will only be talking about the Radeon X1000 series here.

This series was released on October 5, 2005. Among the numerous innovations described in our review called ATI RADEON X1000: Brand-New Graphics Architecture from ATI Explored, it featured the new video processing engine called Avivo.

Besides hardware decoding of HD video formats, Avivo incorporates two independent processors each of which supports 10-bit color processing, overlays, color and gamma correction, image scaling and high-quality deinterlacing. Right now the entire line-up of AMD/ATI’s Radeon X1000-based graphics products is equipped with Avivo and thus supports hardware acceleration of HD video playback. AMD’s next-generation GPUs may bring about some improvements in this field, too.

Nvidia’s alternative, PureVideo technology, was first introduced in the NV40 chip and in the GeForce 6800 graphics card family on April 14, 2004. There were some initial problems, though. The video acceleration block was disabled in the first batches of the NV40 chip, obviously due to some hardware defect. When the defect was mended, it transpired that the 6800 series was equipped with a “first-generation” PureVideo processor that didn’t support WMV HD decoding. The more functional second-generation processor was introduced later, in the GeForce 6600 series. The third-generation PureVideo acquired additional features, particularly support for inverse 2:2 pulldown, but it was only in the fourth generation, implemented in the Nvidia G80 GPU, that PureVideo technology really matured.

It is now called PureVideo HD, reflecting the video processor’s ability to decode and post-process the main HD video formats H.264 and VC-1 utilized on HD-DVD and Blu-ray media. As opposed to the earlier implementations of PureVideo technology, PureVideo HD doesn’t require you to install Nvidia’s special decoder to play DVDs. Like ATI Avivo, it can work with third-party decoders and players.

Although S3 Graphics doesn’t have much weight in the world of discrete graphics solutions, this company offers a rather advanced video engine called Chromotion that we first described in our review of the DeltaChrome graphics card. It is roughly equal to AMD’s Avivo and Nvidia’s PureVideo HD in capabilities, but S3 had implemented its technology in hardware earlier. In late 2003, test samples of the DeltaChrome S8 card could play WMV HD video at an acceptable CPU load thanks to an integrated decoder (which, unfortunately, does not support H.264 and VC-1). That’s why we decided to include products from S3 Graphics into this review.

You should keep it in mind that modern video processors – Avivo, Chromotion and PureVideo – consist of both hardware and software components. And the software part is even more important than the other because it determines the quality of such features as adaptive deinterlacing, etc.

Descriptions of Video Formats

To get as much information as possible about the behavior of graphics cards while reproducing video, we utilized all the main video formats currently in use for our tests:

MPEG-2/DVD is perhaps the most popular video compression format. Its area of application covers virtually everything, from the ordinary DVD to digital television systems. We mentioned some basic principles of MPEG-2 at the beginning of the article and can add that this format allows both interlaced and progressive encoding. The maximum frame resolution for DVDs is 720x576 pixels at 25fps (PAL) or 720x480 pixels at 30fps (NTSC); the maximum bit-rate is about 10Mbps.

Notwithstanding its rather old age, this format is far from obsolete and its ATSC version supports resolutions of 1280x720 at 60fps progressive and 1920x1080 at 30fps interlaced (referred to as 720p and 1080i, respectively). Today’s MPEG-2 compliant systems support bit-rates up to 80Mbps, which is quite enough to ensure a high-quality picture.

MPEG-4 ASP/DivX is about as popular as MPEG-2 today because this format provides an acceptable image quality at a rather low bit-rate, which makes it suitable for transferring video via the Internet. The format is based on a compression algorithm developed within the framework of the MPEG-4 Part 2 standard and utilizes the Advanced Simple Profile. One of the innovations of MPEG-4 ASP is the global motion compensation technique that allows to effectively compress still scenes with a moving camera. Another important innovation is the quarter-sample motion compensation that improves image sharpness by increasing the motion estimation accuracy from 1/2 to 1/4 pixel. The use of variable-size image blocks (8x8 pixels is the smallest size possible) helps increase the compression degree considerably in comparison with algorithms that work with fixed-size blocks.

The proprietary DivX and the GNU GPL XviD are the most popular of MPEG-4 ASP codecs. Less widespread are such implementations as 3ivx, QuickTime and Nero Digital. DivX’ HD capabilities are formally limited to 1280x720 resolution at 30fps and a bit-rate of 20Mbps. These are the numbers of the official High Def profile. However, we managed to encode a video clip at 1920x1080 resolution and 25fps with progressive scan and a bit-rate of about 9.5Mbps using XviD and then reproduced it normally with DivX 6.2.

MPEG-4 AVC/H.264 is an implementation of the MPEG-4 Part 10 standard which aims at achieving very high compression levels while keeping the image quality at a high level, too. Highly sophisticated, this standard contains a number of innovations it would take a whole new article to describe. Particularly, it supports motion compensation with variable-size blocks (the smallest size is 4x4 pixels), which ensures a very accurate selection of moving areas and, accordingly, a high image quality in complex scenes with lots of details. Motion compensation in H.264 may involve up to 32 reference images, as opposed to earlier formats that used 1 or 2 images only. This improves image quality and reduces the bit-rate, too. The motion estimation accuracy is increased to 1/8 pixel while the minimum size of an image block is now 4x4 pixels. Compression of similar-color areas has also been improved in comparison with MPEG-2; the entropic encoding algorithms Context-Adaptive Variable Length Coding (CAVLC) and Context-Adaptive Binary Arithmetic Coding (CABAC) are employed.

As a result, the developers achieved a very high image quality while keeping the bit-rate rather low. The downside is that H.264 is a tremendously resource-consuming standard. This format was accepted for use on new-generation video discs, HD DVD and Blu-ray. The High profile employed is employed on such discs and allows bit-rates up to 30 and 40Mbps, respectively, at a resolution of 1920x1080 progressive and a frame rate of 30fps.

The VC-1 standard is based on the Windows Media Video 9 codec (WMV3) developed by Microsoft. From a technical point of view, VC-1 belongs to the same class of DCT codecs as MPEG-1, 2 and 4. The codec developer provided an opportunity to compress interlaced video material without converting it first into a progressive format. In fact, VC-1 is a subset of WMV3 because the only thing that’s different is that it has an Advanced profile, also known as WVC1. As opposed to the Simple and Main profiles, it supports compression of interlaced video. Providing a comparable image quality, WVC1 achieves a two- or threefold reduction in video bit-rate in comparison with MPEG-2.

HD-DVD and Blu-ray use the L3 level of the WVC1 profile with a bit-rate up to 40Mbps. Resolutions of 1280x720 at 60fps progressive and 1920x1080 at 24fps progressive or 30fps interlaced are supported. With the L2 level the maximum resolution is 1280x720 pixels at 30fps with progressive scan.

WMV HD is the market name of the Window Media Video 9 codec (WMV3). All video encoded with it uses the High level of the Main WMV3 profile. The maximum video bit-rate can be as high as 20Mbps. The highest resolution is 1920x1080 pixels at 30fps with progressive scan. Although a few movies in WMV HD format were released, as separate discs or as an addition to the DVD version, this format was never meant for wide use at home. It was rather a transitional step from DVD to the true HD formats like VC-1 and H.264. A number of demo clips in this format available for download on the Internet have a non-standard resolution of 1440x1080 pixels and thus cannot be regarded as true HD content, which is oriented at displays with an aspect ratio of 16:9. It is a so-called anamorphic format and the onscreen image will only have about 800 lines after an appropriate correction of the dimensions.

High Definition Video on the PC: What Will You Need

While the graphics card and microprocessors definitely play a key-role in media playback, there are several other devices and features that are needed to playback high definition movies from Blu-ray/HD DVD media as well as from the Internet.

Given that officially distributed HD video usually comes in 1280x720 (720p) or 1920x1080 (1080p) resolutions, the minimum monitor or TV-set required for HD playback is the one supporting 1280x800 resolution, which is not something extraordinary these days for computer displays, but may still be unavailable on cheap LCD, Plasma and other TV-sets. The so-called full-HD resolution, or 1920x1080, is available on pretty expensive TV-sets and higher-end monitors, e.g., 23” and upwards, so, if you want to experience the high definition in all its grace, you will need to have a proper output device.

One thing that should be kept in mind when acquiring equipment for future HD video playback is the support for high-bandwidth digital copyright protection (HDCP) technology: both output device and graphics card should better support it. Unfortunately, it is not easy to determine which devices support HDCP: most of high-end LCD displays do, but when it comes to lower-end and mainstream models, you have to check with the manufacturer. It is easier with the TV-sets: all the devices that are marked as “HD Ready” should support HDCP.

All of the graphics cards that have HDMI output support HDCP, but not all of the graphics cards with DVI do. There is a list of boards over an Internet forum that lists many of those devices, in spite of this, we would recommend you to double-check HDCP support on the option that you want to obtain.

In case one of the mentioned does not support HDCP, some Blu-ray discs or HD DVDs will not be performed in full resolution, but in to 540p (960x540), which would make the HD look is a bit better than 720x480 resolution of typical DVDs. Nevertheless, currently there is a program called AnyDVD HD that overrides both HDCP and advanced access content system (AACS) and allows to watch HD content on equipment that does not support HDCP. Unfortunately, some modern graphics cores cannot support 1080p output with HDCP, hence, the software may be required even when everything is HDCP-compliant. Moreover, Sony Vaio notebooks with Blu-ray drive cannot output Blu-ray movies via built-in HDMI port. It remains to be seen whether SlySoft’s AnyDVD HD cures this problem too.

When an appropriate output device along with graphics card is acquired, a new-generation optical drive that supports either Blu-ray or HD DVD format (possibly, some companies, such as LG, may be working on a unified device as well, but it’s not here yet).

Currently there are several Blu-ray drive models available on the market, though, at pretty pricing from $500 to $1000, but Pioneer promised to release a more reasonably priced BD in Q2 2007, which will definitely make it easier to assemble an HTPC with the Blu-ray. Unfortunately for HD DVD camp, there are no HD DVD drives in the retail, however, the external HD DVD player for Microsoft Xbox 360 produced by Toshiba and available for $199 works perfectly with Windows XP-based PC, users would still need a software player that supports HD DVD and a special driver if they want to use HD DVD for something else apart movie watching.

To sum up, for the best high definition video experience on the PC you will need:

Testbed and Methods

We tested video playback performance of modern graphics cards on a platform configured like follows:

All the modern video engines make use of the pixel shader processors to process video. Performance of pixel processors depends directly on the graphics core frequency, and so we tried to expand the range of tested devices as far as we could. Besides that, video playback performance can be affected by other parameters like the frequency of graphics memory and the width of the graphics memory bus. Therefore, we’ve got the following graphics card models on our list:



S3 Graphics:

For our tests we used the following video content:

We were unable to find a free clip in VC-1 format with a resolution of 1920x1080 pixels and we could not use commercial Blu-ray or HD DVD discs just because we didn’t have an appropriate optical drive. We will fill in this blank in our test as soon as we can.

We took CyberLink PowerDVD 7 Ultra as our software player. It supports all the formats we are interested in, including HD-DVD and Blu-ray. It incorporates a decoder that supports hardware acceleration options offered by today’s graphics cards and is one of the most popular and widespread players (together with WinDVD 8 Platinum and Windows Media Player).

We turned on hardware acceleration in PowerDVD settings, enabled the Pulldown Detection and Windows Media Video Acceleration options in the AMD Catalyst Control Center and the Use inverse telecine option in the Nvidia ForceWare control panel. The Edge Enhancement and Noise Reduction sliders were left at their defaults.

We measured the CPU load level by means of Microsoft Management Console 2.0. We didn’t disable any of Windows processes for this test, so the CPU load was fluctuating a little. It means that a difference of 1-2% is insignificant and doesn’t indicate a superiority of one graphics card over another.

Test Results

Premium Graphics Cards

Playing DVDs is long not a problem for modern, and even some not-very-modern, graphics cards. Every graphics solution can decode this format at a CPU load of less than 5.5%.

MPEG-2 being a rather simple format, the decoding of a 20Mbps video stream with a frame resolution of 1920x1080 pixels is not a problem for today’s CPUs and graphics subsystems. In the worst case the CPU is loaded by only 16%, most of its resources being left free for other tasks. The GeForce 8800 series copes with this task better than its opponents, obviously due to the extremely high frequency of the pixel processors, over 1GHz. Note also how low the 8800 series cards keep the average CPU load.

The DivX format demands more computing power even when the frame resolution is as low as 640x480. Here, Avivo technology represented by the Radeon X1950 XTX proves to be superior. Strangely enough, the new generation of GeForce cards is the slowest of all at decoding this video content. Perhaps this is due to some flaws in their drivers.

When processing a clip in a HD resolution and encoded with XviD, the Radeon X1950 XTX cannot keep the CPU load as low as at 640x480 resolution.

The flagship products from AMD and Nvidia take up about the same amount of work when DivX HD content is being decoded. A load of 40% is not low for a rather advanced dual-core CPU, but the playback conditions are comfortable anyway. At least there is no jerkiness or lost frames. The GeForce 7950 GX2 is somewhat worse than the other premium-class solutions because SLI technology doesn’t provide any advantages for decoding video whereas the frequency of this graphics card’s cores is 500MHz which is far lower than that of the Radeon X1950 XTX and GeForce 8800.

H.264 is a complex format, but the hardware support for its decoding implemented in modern graphics cards works perfectly. Note the exceptional performance of the GeForce 8800 in both resolutions. The Radeon X1950 XTX and GeForce 7950 GX2 keep the CPU load at about the same level and not higher than 50% in the 1080p resolution.

On the other hand, we use a clip with a rather low bit-rate, thus making the decoder’s job simpler. The load is going to be higher if a HD-DVD disc with H.264 video content is played because the bit-rate may be as high as 30Mbps then.

The only clip in the VC-1 format that we have at our disposal makes up for its resolution of 1280x720 pixels with a high frame rate and a bit-rate of 15Mbps. Thus, it is indeed a very hard test. The Avivo engine processes this file less efficiently than Nvidia’s PureVideo HD does, yet the results are acceptable since the CPU’s cores are each loaded by half only.

WMV HD is a rather simple format to decode and every premium-class product handles it with ease, keeping the CPU load below 25%. The Radeon X1950 XTX and GeForce 7950 GX2 are equals in the 1080p resolution, and the GeForce 8800 GTX has the best results.

So when it comes to highest-performance graphics cards available on the market, each of them has all the capabilities necessary to decode video of various formats, including HD content, and can take up most of the job even in the hardest case, allowing the CPU to be loaded by no more than 45-50%.

The average CPU load varies from 6% to 36% depending on the format, except for DVD and DivX SD.

High-End and Performance-Mainstream Graphics Cards

Playing DVDs is no test for such graphics cards. The CPU load is as low as with the premium-class products.

The Radeon X1900 XT and Radeon X1950 Pro keep the CPU load at the same level, although the latter card has worse technical characteristics. The slightly higher load must be due to the fluctuations of the system during our measurements because the Radeon X1900 XT’s core and memory frequencies are higher while the number of pixel processors doesn’t seem to play any big role at video decoding. The GeForce 8800 has a very low minimum of CPU load, but the average and maximum levels are comparable to those of the AMD solutions.

Decoding DivX SD takes about the same amount of CPU resources as the previous task. The difference is no bigger than 1%, which is negligible considering our measurement method.

The same goes for DivX HD, but the difference is somewhat bigger here, from 2% to 5%. Strangely enough, the GeForce 7950 GT loses to the dual-processor GeForce 7950 GX2 although works at higher frequencies.

We don’t see the above-described anomalies when playing H.264 video: the GeForce 7950 GT quite naturally performs better than the GeForce 7950 GX2 in both 1280x720 and 1920x1080 resolutions. The Radeon X1900 XT and Radeon X1950 Pro perform in a similar manner. An interesting fact, it is when processing the 1080p format that the Radeon X1900 series cards have poor results.

Nvidia’s PureVideo HD engine turns in good results again, contrary to the Radeon X1950 XTX, Radeon X1900 XT and Radeon X1950 Pro that line up in order of descending computing power. However, the Radeon X1950 Pro is quite capable of decoding VC-1 video without putting a big load on the CPU.

These results correlate well with the results of the premium-class products. The CPU load level is generally low, but lower with Nvidia’s than with AMD’s graphics cards.

Graphics cards from the performance-mainstream category are just as good at decoding various video formats as the more advanced products are. Like with the premium-class solutions, the peak CPU load is never higher than 50%, and the average CPU load varies from 6% to 35-37%.

Mainstream Solutions

Despite modest technical characteristics of mainstream graphics cards, it is only the GeForce 7900 GS and GeForce 7600 GT that produce different results when playing MPEG-2 1080i. The former graphics card has a low core frequency while the latter is probably limited by its 128-bit memory bus.

When playing DivX SD on the GeForce 7900 GS, the peak CPU load is abnormally high, but it is rather the consequence of our not-very-accurate measuring method because we don’t see anything like that while playing DivX HD. The CPU load is 34-36% on average then.

The Radeon X1600 Pro has the worst result, most likely due to an imperfect implementation of H.264 decoding in the driver because this card is only 2% behind the Radeon X1900 XT at the 1080p resolution.

The memory bus is not an important factor here: the GeForce 7600 GT offloads the CPU better than the GeForce 7900 GS that has a 256-bit memory bus and 20 pixel processors. A direct relation between performance of the PureVideo HD engine and the GPU frequency is obvious.

The decoding of VC-1 in 720p@60fps format produces about the same results as with the graphics cards from the higher class, except that the peak CPU load has become 1-2% higher with Nvidia’s solutions.

The GeForce 7600 GT is in the lead again when processing WMV HD as well as H.264 content. We don’t see the mainstream graphics cards from AMD and Nvidia being much weaker than the more advanced products.

Thus, modern graphics cards from the mainstream category are capable of playing video of any existing format, including full-HD formats, no worse than more advanced and expensive solutions from higher product categories. And now let’s see how the cheapest products behave.

Entry-Level Graphics Cards

There’s no difference when it comes to DVDs: the GeForce 7300 GS and Radeon X1300 Pro do their job just as well as any other modern graphics card from AMD and Nvidia does. It’s somewhat less good with MPEG-2 1080i, but the peak CPU load isn’t higher than 30%, the average being 8-10%. It means these graphics cards are sufficiently fast to perform most of the work associated with decoding and playing back this format.

The Radeon X1300 Pro seems to suffer a big performance hit when it starts playing DivX SD, yet this results in a peak CPU load of only 13%. The average CPU load remains on the same level as with the more advanced solutions from AMD. The GeForce 7300 GS is about as successful as the Radeon X1300 Pro in this test.

When the DivX HD 1080p clip is being played on the GeForce 7300 GS, the peak CPU load reaches 50%. It would probably reach 100% and there would be dropped frames if we had a single-core CPU. The Radeon X1300 Pro has a more acceptable result – a peak CPU load of 43%.

It is when the 1920x1080 clip is being played that we see the CPU load exceed 50% for the first time. There is some driver-related problem with the Radeon X1000 series: the Radeon X1300 Pro has a higher CPU load than the S3 Chrome S27 which does not incorporate a hardware H.264 decoder. The GeForce 7300 GS, on the contrary, performs well and is no worse than any other model from the GeForce 7 series.

The Radeon X1300 Pro shows the same behavior when processing the VC-1 video clip, although the peak load is lower here than with the H.264 1080p content. The Nvidia GeForce 7300 GS is less confident here than in the previous test.

WMV HD uses a less resource-consuming profile than VC-1 and can be played successfully on the Radeon X1300 Pro, although the CPU load is much higher in the 1080p resolution than with more advanced Radeons.

Video Processors and Image Quality: HQV Tests

Besides checking out performance of AMD Avivo and Nvidia PureVideo HD video engines, we want to see how they affect the image quality. To do this, we will utilize the popular HQV test ( developed by Silicon Optix. It is intended for a comparative quality analysis of DVD playback. This test is an ordinary DVD with a handy menu system and onscreen tips, so it can be used not only on the PC, but in any device that can play DVDs.

The downside of this universality is in the certain bias of the end result because the user has to evaluate the quality of each feature visually, basing on reference images and assigning points. Fortunately, the test comes with detailed documentation that contains all the reference images, and if the tester is attentive enough, the influence of the subjective factor is rather low, and the results can indeed be used for making comparisons.

Tests from the HQV suite can be divided into several sections, each of which helps check out a certain feature of the video processor:

Each test comes with a description and a few reference images each costing a certain amount of points. The tester compares these images with those he sees on the screen and awards points. The final result is the sum of all the HQV tests. The maximum possible number of points is 130.

Here are the results we’ve got:

As you can see, there is a very small difference between Avivo and PureVideo HD. Nvidia’s video-processor has a somewhat lower deinterlacing quality, especially in the Waving Flag test, but is superior to the ATI processor when it comes to detail enhancement algorithms. Although Avivo produced a sharper-looking picture in that test than the PowerDVD software decoder it didn’t allow to adjust that parameter and thus scored less points than PureVideo HD. On the other hand, the AMD solution was better at detecting the necessity to use 3:2 pulldown. Both video processors identified it with a delay of about half a second, but Avivo remembered the necessary mode and when the content was played again, it didn’t show moir? or aliasing in the first moments. That’s why it scored more points in this test. In the other tests the video processors from AMD and Nvidia deservedly scored the maximum amount of points possible. They produced a much better picture than the software PowerDVD decoder did even with enabled CLEV-2.

Unfortunately, the Chrome S27 graphics card didn’t pass the test. It produced a jittering image, probably due to the NTSC version of the HQV test disc. Having a video processor with good capabilities, the Chrome S27 gets nil, obviously due to its faulty drivers, which are a long-time problem of S3 Graphics. Although the Chrome S27 has good technical characteristic for an entry-level product, it cannot be recommended for playing video.


Our tests show that almost each of today’s graphics cards has everything necessary to play high-definition video formats, let alone the traditional DVD and DivX. We mean both hardware and software: the tests suggest that the graphics card drivers have been polished off so that even 1920x1080 video is decoded and played without any performance or quality related problems.

The CPU is loaded quite heavily at that, especially when it comes to the current version of ATI Avivo and H.264, but that’s not a serious problem considering how much computing power modern dual-core CPUs provide. Still, we can’t explain why Avivo is very fast at decoding H.264 and WMV HD video with a resolution of 1280x720 pixels, but slows down when the resolution is increased to 1920x1080. It may be a hardware problem (some bottleneck in the Avivo architecture that shows up at high bit-rates and resolutions) or a software one (related to the noise reduction and detail enhancement algorithms that can’t be disabled).

Both technologies offload the CPU to an acceptable level, but Avivo provides a somewhat sharper and less noisy picture than PureVideo HD at the default settings. The latter, however, makes up for that by providing more flexible options to reduce noise and improve detailedness.

Video playback performance of graphics cards from different price categories varies in a rather small range, yet we can make some points here. First, this performance depends on the graphics core frequency, specifically on the frequency of the pixel shader processors. This explains the record-breaking results of the GeForce 8800 series whose shader processors are clocked at frequencies above 1GHz. It is also clear that a rather small number of shader processors are actually employed – just compare the results of the Radeon X1950 XTX against the Radeon X1650 Pro or the GeForce 7900 GS against the GeForce 7600 GT. Such parameters as the amount and frequency of graphics memory or the width of the graphics memory bus do not influence performance of the hardware video processor much.

The entry-level solutions are different. Such graphics cards either have slow memory with a narrow memory bus like the GeForce 7300 GS or have few pixel processors on board like the Radeon X1300 Pro. Their video processors are thus greatly constrained and do not suit well for playing HD formats with a high bit-rate or resolution, especially if you’ve got a single-core CPU. The rest of today’s graphics cards, from the modest Radeon X1600 Pro and GeForce 7600 GT to the ragingly fast Radeon X1950 XTX and GeForce 8800 GTX, do their job well.

So, the choice of the graphics card for a multimedia center is in fact determined by such factors as noise and your desire to play games. If you don’t want to build a gaming station, you can do quite well without a GeForce 8800-class graphics card.