Home / Tech News / Featured Tech Reviews / HQV Benchmark 2.0 Analysis – ATi, nVidia and Intel

HQV Benchmark 2.0 Analysis – ATi, nVidia and Intel

Allan 'Zardon' Campbell June 16, 2010 Featured Tech Reviews, Graphics

KitGuru knows that companies spend millions creating the best HD processing set up possible only to see the final image quality shattered by a weak component somewhere in the system. We spend many hours watching High Definition content and we never stop playing with our ATI and nVidia driver panels to get the quality as high as we can. All this fiddling might be fun, but can we add any science to the process of deciding on the best picture? HQV Benchmark 2.0 aims to do just that. Square-Eyed KitGurus investigate.

Hollywood Quality Video. You are setting yourself up with a name like that. Especially for a benchmark. But the guys at http://www.hqv.com/ did a cracking job when they launched their original benchmark back in 2007. It quickly won Wired Magazine's Editor's Pick as a ‘must have' for setting up the best home cinema experience possible. The main advantage was that it could be used across a broad cross section of systems, PC and hi-fi to give a simple number which allowed you to compare image quality.

The point of the program is that it takes a series of ‘human' tests (you look at it and score it) and lets you give that experience a number. The numbers given by a diverse audience are often quite close in scoring.

HQV Benchmark 2.0 is an updated version of the original tool and it consists of various video clips and test patterns which are designed to evalute motion correction, de-interlacing, decoding, noise reduction, detail enhancement and film cadence detection.

There are two versions of the program, standard definition on DVD and high definition on Bluray. As our audience will be concentrating on HD content so will we.

This has a total of 39 video tests which is increased from 23 in the original and the scoring is also up from a total of 130 to 210. As hardware and software gets more complicated, the software has been tuned to make sure we can thoroughly maximise our analysis.

We will be comparing ATi, nVidia and Intel graphics solutions with the HQV 2.0 Benchmark to see which company really delivers the best viewing experience. As a portion of the testing will be subjective, it is absolutely imperative that we use the best quality screen possible. Many HDTV's have intensive processing which may artificially enhance the picture quality, but we need to omit this from the equation and concentrate totally on the graphics hardware and driver output. Our initial plans were to use a television such as the Panasonic Viera NeoPDP 600hz Plasma, but while we adore the picture there is a lot of processing going on behind the scenes, even with intelligent frame creation disabled.

For the purposes of this review we will be using one of the most expensive monitors on the planet, the LaCie 730, which retails around the £2,700 mark. This incorporates a 14-bit colour system (14-bit look up table and 14-bit per colour processing) which takes the 8-bit output of your graphics cards dual-link DVI port and modifies it into the monitor's output using an overall palette which has 192 times more colour shades than the graphics card's output. The colours displayed will more accurately reflect exactly what is being displayed. The colour gamut is also 23 per cent larger than that of Adobe RGB standard which is used in applications such as Photoshop.

This monitor has been carefully calibrated with LaCie's Blue Eye Pro Proof Edition software. During our calibration we have configured blackpoint, colour temperature as well as adaptation and profile type. This screen also offers a native 1:1 1080p mode for perfect pixel reproduction.

Before we continue with the article we would like to put out a thanks to ASUS who supplied an external BluRay unit for the testing. We were actually so impressed with this drive over the last two weeks that we felt it deserved to be highlighted.

This external slim-type optical disc drive performs at six times the standard Blu-ray reading speed (it also writes to a plethora of formats, including Bluray). It features True Theatre High Definition (TTHD) technology, a pioneering suite of advanced video enhancements that delivers unmatched HD quality for DVD playback on PC-based home theatre systems. The result is sharper video, significant decrease in background noise, and enhanced aliasing for a smoother image.

The drive, as you can see, is beautifully designed. It's simplistic, but extremely attractive. It is powered over the USB bus, so no need for a power adapter, we particularly love this concept. Asus supply a cable with a dual head USB plug so it can take power from two ports.

It is supplied with a stand, so you can use it either vertically or horizontally.

You remove a rubber panel from the rear of the drive and slot the stand into the housing.

Above, the front and rear of the unit when mounted onto the stand. We like the vertical mount appearance. This is one of the thinnest drives we have seen on the market at only 21mm! At 157 x 142 x 21mm the SBC-06D1S-U is only 37mm longer and 22mm wider than a standard 12cm disc!

When the unit is powered up, Windows 7 detected it immediately and as you can see in the image above it glows blue, across accented lines. There is a mirror like finish, as you can perfectly see our Sony laptop in the surface above [Don't forget the Mr Muscle for finger prints – Ed].

Pushing the button on the front opens the tray and you can then insert your disc of choice.

Here are the specifications (warning – you WILL need to scroll for this):

Color	Black
Read Speed	BD-R: 6 X BD-RE: 6 X BD-ROM: 6 X BD-R(DL): 6 x BD-RE(DL): 6 X BD-ROM(DL): 6 X DVD+R: 8 X DVD-R: 8 X DVD+RW: 8 X DVD-RW: 8 X DVD-ROM: 8 X DVD+R(DL): 6 X DVD-R(DL): 6 X DVD-ROM(DL): 8 X DVD-RAM: 5 X CD-R: 24 X CD-RW: 24 X CD-ROM: 24 X BDMV playback: 4 X DVD video playback: 4 X VCD playback: 10 X Audio CD Playback: 10 X
Write Speed	DVD+R： 8X DVD-R： 8X DVD+RW： 8X DVD-RW： 6X DVD+R(DL)： 4X DVD-R (DL)： 4X DVD-RAM： 5X CD-R： 24X CD-RW： 16X
Writing Mode	DVD+R & DVD+R (DL)： Sequential Write DVD-R & DVD-R(DL)： Incremental Recording DVD+RW： Random Write DVD-RW： DAO/Incremental Recording DVD-RAM: Random Write CD-R/RW： DAO/TAO/SAO/Packet Write
Access time	BD: 240 ms DVD: 160 ms CD: 160 ms
Interface	USB2.0
Data Buffer	2 MB
O/S Compatibility	Windows® 7 Windows® Vista Windows® XP
System Requirement	Processor: Pentium D 945 (34GHz) or higher Memory: 1GB or more is recommended Graphics Card: NVIDIA GeForce 7600 GT or ATI X1600 series or above Use HDCP Compatible display and VGA card to High Definition digital output.
Bundle Software	Cyberlink PowerDVD with TTHD Cyberlink Power2Go Cyberlink InstantBurn Cyberlink PowerBackup TurboEngine BluTuner
Disc Formats	Audio CD,Video CD,CD-I,CD-Extra,Photo CD,CD-Text,CD-ROM/XA,Multi-session CD,DVD Video
Disc Diameters	12cm/8cm
Mounting Orientation	Vertical and Horizontal (+5° ~ -5°)
Dimension	142x157x21 mm (W x L x H)
Weight	412 g(435g with stand)
Power Requirements	USB Power +5V ± 5%
Temperature	Operating: 5℃ to 40℃ Storage: -30℃ to 60℃
Humidity	Operating: 15% ~ 80% (Non-condensing) Storage: 10% ~ 90% (Non-condensing)

The Asus drive performs like only the best drives on the market can deliver. It is very quiet in operation and we found BluRay discs loaded in around 15 seconds via PowerDVD 10 Ultra. The Blue LED is gorgeous as well, however you can turn it off in software if you wish. Best of all you can keep the firmware updated via software, which is massively important in this specific sector.

The Software also offers drag and drop burning and securing data with passwords and encryption.

If you are in the market for a BluRay drive and want something which is self-powered, quiet, attractive and handles a plethora of formats then we can strongly recommend this Asus unit. It retails online for less than £140 and the code to look for on Google is SBC-06 D1S.

KitGuru says: We love this drive and will be heartbroken when Asus take it back.

Our test results today will be gathered from various systems and a broad cross section of graphics hardware.

For nVidia will be using their latest Forceware driver 257.15 beta and for ATI the June 2010 edition of its WHQL driver, Catalyst 10.6, which features new media functionality (and should just be released by the time you are reading this). Options such as Mosquito Noise Reduction and Deblocking have been added and help improve the picture quality. There is also improved Cadence Detection, Skin Tone Correction and Dynamic Contrast. I will speak about these more later during some of the tests.

To accurately test the hardware today everything is being routed via digital output through our reference LaCie 730 Monitor. We have attempted to arrange a variety of systems to suit every budget. 750w and 850w power supplies courtesy of Corsair.

We would like to thank (in no particular order): Intel, AMD Canada/Europe, Corsair, nVidia, Crucial, Kingston, ADATA, Zotac, Mesh, YoyoTech, Asus, Dell, Noctua, Sapphire, XFX for their support with this review and getting hold of various hardware.

System 1: Intel GMA X4500 (G41) based desktop system
This is an entry level system with integrated Intel graphics. It uses Intel's GMA X4500 and is equipped with a Quad Core Intel Q6700 processor which runs at 2.66ghz. There is 2GB of DDR3-1066mhz memory in this system and we are using Intel's display driver, V15.17.4.2119

System 2: Intel HD Graphics (i3)
This is based around an Intel IGP solution with a Core i3 540 processor running at 3.06ghz. We have 2GB of DDR3 1066mhz ram in this system and Intel driver V15.17.4.2119 is used.

System 3: HD Radeon 4200 (RS880 IGP)
An AMD Integrated graphics system, which utilises a HD4200 processor with an Athlon X4 945 Black Edition which runs at 3 ghz. We have 4GB of Crucial Ballistix 1600mhz DDR3 in this system. We are using Catalyst 10.6 for testing (V 8.741).

System 4: HD Radeon 5450/5550/5770/5850/5870 (Discrete)
In this AMD system we are using an Intel Core i7 875k Cpu running at 2.93ghz. We are using 4GB of ADATA DDR3 2000mhz ram running at 1600mhz. We are using Catalyst 10.6 for testing (V8.741).

System 5: nVidia G210/G220 (Discrete)
This is a system based around an Intel Core i5 750 processor running at 2.66ghz. The machine was equipped with 4GB of Crucial DDR3 memory running at 1600mhz. Display driver 275.15 beta was used for testing.

System 6: nVidia GTX465/GTX480 (Discrete)
This system is based around a Core i7 980x system running at 3.33ghz. The machine was equipped with 6GB of Crucial DDR3 memory running at 1600mhz. Display driver 275.15 beta was used for testing.

HQV contains 39 tests, organised into four major Classifications. Each of these comprises a number of chapters and each chapter offers specific tests. You run the test, look at the results, compare what you see to the guidance book and give your score. Easy! We had laid out the chart below for ease of use and we will look at each later in the article.

Test Class	Chapter	Tests
Video Conversion	Video Resolution	Dial Dial with Static Pattern Gray Bars Violin
	Film Resolution	Stadium 2:2 Stadium 3:2
	Overlay On Film	Horizontal Text Scroll Vertical Text Scroll
	Cadence Response Time	Transition to 3:2 Lock Transition to 2:2 Lock
	Multi-Cadence	2:2:2:4 24 FPS DVCam Video 2:3:3:2 24 FPS DVCAM Video 3:2:3:2:2 24 FPS Vari-Speed 5:5 12 FPS Animation 6:4 12 FPS Animation 8:7 8 FPS Animation
	Color Upsampling Errors	Interlace Chroma Problem (ICP) Chroma Upsampling Error (CUE)
Noise and Artifact Reduction	Random Noise	SailBoat Flower Sunrise Harbour Night
	Compression Artifacts	Scrolling Text Roller Coaster Ferris Wheel Bridge Traffic
	Upscaled Compression Artifacts	Text Pattern Roller Coaster Ferris Wheel Bridge Traffic
Image Scaling and Enhancements	Scaling and Filtering	Luminance Frequency Bands Chrominance Frequency Bands Vanishing Text
Image Scaling and Enhancements	Resolution Enhancement	Brook, Mountain, Flower, Hair, Wood
Video Conversion	Contrast Enhancement	Theme Park Driftwood Beach at Dusk White and Black Cats
Video Conversion	Skin Tone Correction	Skin Tones

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is high quality then the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

Scoring:
In the first test, a long white, rectangular dial rotates inside the white outline of a graduated circle. To determine the quality of field interpolation we analyse the rotating dial and try to ascertain at which angle the long edges of the dial start to distort. The rule of thumb is that the lower the angle at which the distortion begins, the better the quality of field merging or field interpolation.

The ‘fail' image above details the rotating dial with jaggies and the ‘pass' image is almost perfect, with relatively smooth edges.

When we score this it is important to judge the midpoint of the dial and when exactly it begins to distort. This takes some time to analyse as it moves rather quickly and is difficult to judge correctly. The scoring below shows that there is only 5 degrees separating the top two scores. We ran this test for a long time on all the cards today to ensure we reported, as accurately as possible the correct results.

Dial (Score either 5, 4, 2 or 0)
Dial arm is smooth without artifacts above 5 degrees – Scores 5
Dial arm is smooth without artifacts above 10 degrees – Scores 4
Dial arm is smooth without artifacts above 20 degrees – Scores 2
Artifacts above 30 degrees – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Dial	0	4	0	4	4	4	4	4	0	2	4	4

Most of the cards do well in this test, apart from the Intel GMA X4500, ATI HD4200 and nVidia G210. None of them reached what i would class as a perfect 5 degree result.

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is successful and high quality the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

Scoring:
The second test is identical to the previous one however the addition of a circular white mesh appears just within the circle outline. By including this white mesh the local motion adapticity capability of the video processor is being stressed. In practice this means the video processor has to draw the rotating dial with smooth edges but also has to redraw the mesh without any flickering distortion, jaggies or missing lines and without any additional distortion of the rotating dial.

Dial with Static Pattern (Score either 5, 2 or 0)
Mesh is stable and dial quality is as before. There may be a small halo around dial on the mesh. – Scores 5
Mesh is stable but some degradation of the dial or noticeable artifacts around the dial on the mesh. – Scores 2
Mesh is flickering or dial quality has degraded. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Dial with static pattern	5	5	5	5	5	5	5	5	5	5	5	5

We noticed that the NVG210 and Intel GMA X4500 suffered from a slight ‘halo' effect preceding the dial as it rotated through the white mesh. This halo while distracting is not classed as a failure in this test. As long as no lines are broken or missing, this counts as a ‘halo' and not an ‘artifact'.

A strong motion detection algorithm will use more fields to minimize the number of weaving artifacts on moving content. Various de-interlacers have different kinds of wake variances behind the objects and some even in front or in both positions. The size of his Halo effect (or motion trail) might be considered better as it often is connected with the number of fields that are used by the motion detection algorithm.

For a perfect score of 5 out of 5 the Mesh must be stable with the dial quality as before.

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is successful and high quality the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

Scoring:
The third test uses four sets of rotating bars, which range from bright white to dark gray. To score well in this test it is imperative to clearly see all four sets of bars to make a good final call. The monitor must be correctly calibrated to give a good judgement in this specific test.

Gray Bars (Score either 5, 3, 2, 1 or 0)
No artifacts on all gradients and all bars. – Scores 5
Artifacts only on darkest bars. – Scores 3
Artifacts on second darkest bars. – Scores 2
Artifacts on third gradient level bars. – Scores 1
Artifacts on brightness gradient levels. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Gray Bars	0	3	0	3	5	5	5	5	0	2	5	5

This test exposes weaknesses with some deinterlacers in that they will not perform consistently at all brightness levels. It shows issues between foreground and background seperation brightness variables. This means the end result can show weaving artifacts, jaggies sometimes appear on darker bars, as they did with many of the lower end solutions on test today.

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is successful and high quality the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

Scoring:
This test shows a girl playing a violin and in this sequence she moves back and forward which forces the violin strings to move randomly through various low angles. If the video processor can reproduce the strings without artifacts then it will score high. This targets the processor's spatial interpolation as well as adaptive directional filtering capabilities.

Violin (Score either 5, 3 or 0)
No artifacts on all strings. – Scores 5
Small artifacts during motion transitions. – Scores 3
Noticeable artifacts during most of the sequence. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Violin	0	3	3	5	5	5	5	5	0	3	5	5

Getting a score of 5 in this test requires all the strings to be free of artifacts. A score of 3 should be given if small artifacts during motion transitions are observed. A score of 0 means that artifacting is seen throughout most of the sequence. Both nVidia and ATI hardware scores high in these tests, although we were very disappointed with the performance of the G210 and expected a little more.

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is successful and high quality the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

Cadence tests all deal with the ability of the video processing unit to detect and lock onto the cadence of the video clip and then display it with minimal distortion while preserving maximum detail in problem specific areas.

The video industry use different cadences to increase or decrease the amount of video content. A ‘cadence' is a ratio which describes the relationship between the original film and the converted end result.

Scoring:
These tests assess the video processor's ability to detect and lock onto two basic cadences, 2:2 and 3:2. For both tests, the video slowly pans across seating inside a stadium. If the cadence detector is faulty, a moire pattern will be seen in various seating sections of the stadium. The curves of this pattern may also distort and the pattern can ‘shift' as the scene pans across the various rows of seats. This pattern may present itself throughout the whole sequence and may even flicker in and out of view.

Stadium 2:2 (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second. – Scores 5
No moire pattern observed in stands within less than 1 second. – Scores 3
Moire pattern observed in stands intermittently or constantly through the clip – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Stadium 2:2	0	0	3	5	5	5	5	5	0	0	5	5

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is successful and high quality the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

Stadium 3:2 (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second. – Scores 5
No moire pattern observed in stands within less than 1 second. – Scores 3
Moire pattern observed in stands intermittently or constantly through the clip – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Stadium 3:2	0	3	3	5	5	5	5	5	0	3	5	5

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is successful and high quality the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

Text that is overlaid onto film can have a cadence that is very different to that of the underlying film. In these cases the video processor must correctly identify the two cadences and adjust the local video processing to account for the different cadences.

If the video processor locks onto the cadence of the background movie but not the overlaid text then it can appear to shred. On the other hand if the video processor locks onto the cadence of the foreground text and not the movie then the background can artifact by appearing to shred also.

Scoring:
We have to pay attention to the point when the text enters the screen which will be lower edge of the screen for vertical testing and lower right corner for horizontal testing. Shredding which does not appear at all, gets a score of 5. A brief appearance caused by delay in processing gets 3 and remains throughout the clip, scores 0.

Horizontal Text Scroll (Score either 5, 3 or 0)
No shredding artifacts can be observed in the text or the background. – Scores 5
Observable Artifacts in the text or background when subtitle first appears but quickly corrected. – Scores 3
Shredding observed on the text or background throughout the majority of the sequence – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Horizontal Text Scroll	3	3	5	5	5	5	5	5	3	3	3	3

We found that the ATI solutions gave the best result and were quite surprised to see even the HD4200 effectively dealing with all forms of shredding on screen. The nVidia solutions seemed slower to react which is sure to be fixed in a driver update.

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests stress the video processors ability to process data accurately, and is based around both stationary and moving elements within specific scenes. If the processing is successful and high quality the images will be rendered crisply and in a very clear manner. On the other hand, if the processing is poor then jaggies and artifacting will appear on the edges of moving objects.

The text's motion can be unsteady and this will be perceived as juddering motion to a greater or lesser degree that will depend on the refresh rate of the display device.

Vertical Text Scroll (Score either 5, 3 or 0)
No flicking or serration artifacts in the top and bottom edge of the characters nor in the background. – Scores 5
Flicking or serration artifacts in the text and background when the subtitles first appear but quickly corrected. – Scores 3
Flicking or serration effects in the top and bottom of the edge of the characters or in the background – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Vertical Text Scroll	3	5	3	5	5	5	5	5	5	5	3	3

The ATi discrete cards score well again, as does the Intel HD graphics solution. The nVidia cards also maintain good clarity throughout with only minor serration artifacting appearing on the GTX465/480.

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests focus on the performance of the cadence detector at varying speed of motion. They are more severe stress tests than the Stadium 2:2 and 3:2 on the last couple of pages.

Scoring:
When the globe is moving through the right side of the screen, it is displayed with a 3:2 cadence, while on the left, the cadence is 2:2. A scoring guide is provided on the screen, with a green zone that begins right at the cadence boundary, a yellow zone which starts a half-second later, and a red zone that begins one second into the cadence zone.

Response time – Transition 3:2 Lock (Score either 5, 2 or 0)
Scores less than 1/2 second to transition from 2:2 to 3:2 – Scores 5
Takes more than 1/2 second to transition from 2:2 to 3:2. – Scores 2
Takes more than one second to transition from 2:2 to 3:2– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Transition to 3:2 Lock	0	5	5	5	5	5	5	5	0	0	5	5

The G210 and G220 have a problem with this test which will hope to see fixed in a later driver. The Intel GMA X4500 also puts in a poor showing with this specific test. ATI solutions all score 5/5.

Overview:
The first section is based around Video Conversion tests, followed up by a long series of Cadence related tests and lastly we get to examine colour upscaling errors.

Notes:
These tests focus on the performance of the cadence detector at varying speed of motion. They are more severe stress tests than the Stadium 2:2 and 3:2 on the last couple of pages.

Response time – Transition 2:2 Lock (Score either 5, 2 or 0)
Scores less than 1/2 second to transition from 3:2 to 2:2 – Scores 5
Takes more than 1/2 second to transition from 3:2 to 2:2. – Scores 2
Takes more than one second to transition from 3:2 to 2:2– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Transition to 2:2 Lock	0	5	0	0	0	0	0	0	0	0	5	5

This test cause issues for all the solutions on test today with ATI scoring 0 with all their products. Only the Intel HD graphics and GTX465 and GTX480 can handle this specific transition.

Overview:
This chapter focuses on various cadences which can occur in video content from a plethora of sources. Severe moire is used as an indication that cadence lock has not been achieved. It is the amount of time it t akes to achieve a cadence lock, indicated by a reduction in the moire distortion, which is the scoring mechanism for this specific test.

Scoring:
We have to pay attention to moire pattern throughout and document the time it appears (if any). No moire is a perfect result, slight moire appearance in specific sections means the result is intermediate, and if moire is noted throughout, the graphics solution is awarded no points.

Multi Cadence: 2:2:2:4 24 FPS DVCam Video (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second – Scores 5
No moire pattern observed in stands within less than 1 second– Scores 3
Moire pattern observed in stands intermittently or constantly through the clip– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
2:2:2:4 24 FPS DVCAM Video	0	0	5	5	5	5	5	5	0	0	5	5

ATI all score 5/5 in this test with both Intel solutions and the lower end nVidia boards showing a lot of moire noise throughout the test. The higher end GTX boards score as well as the ATI discrete solutions.

Multi Cadence: 2:3:3:2 24 FPS DVCam Video (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second – Scores 5
No moire pattern observed in stands within less than 1 second– Scores 3
Moire pattern observed in stands intermittently or constantly through the clip– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
2:3:3:2 24 FPS DVCam Video	0	0	5	5	5	5	5	5	0	0	5	5

Again ATI all score 5/5 in this test with both Intel solutions and the lower end nVidia boards showing a lot of moire noise throughout the test. The higher end GTX boards score as well as the ATI discrete solutions.

Multi Cadence: 3:2:3:2:2 24 FPS DVCam Video (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second – Scores 5
No moire pattern observed in stands within less than 1 second– Scores 3
Moire pattern observed in stands intermittently or constantly through the clip– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
3:2:3:2:2 24 FOS Vari-Speed	0	3	5	5	5	5	5	5	0	3	5	5

ATI keep up their perfect record in this series of tests and the Intel HD Graphics and nVidia G220 manage to get some points in this test as we noticed reduction of moire in this specific test.

Multi Cadence: 5:5 FPS Animation (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second – Scores 5
No moire pattern observed in stands within less than 1 second– Scores 3
Moire pattern observed in stands intermittently or constantly through the clip– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
5:5 FPS Animation	0	0	5	5	5	5	5	5	0	0	5	5

A similar result to the first 2 in this series of testing. ATI score full marks, with the two high end nVidia discrete solutions keeping up.

Multi Cadence: 6:4 12 FPS Animation (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second – Scores 5
No moire pattern observed in stands within less than 1 second– Scores 3
Moire pattern observed in stands intermittently or constantly through the clip– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
6:4 12 FPS Animation	0	0	5	5	5	5	5	5	0	0	5	5

Identical results to the 5:5 FPS Animation test on the previous page.

Overview:
This chapter focuses on various cadences which can occur in video content from a plethora of sources. Severe moire is used as an indication that cadence lock has not been achieved. It is the amount of time it takes to achieve a cadence lock, indicated by a reduction in the moire distortion, which is the scoring mechanism for this specific test.

Multi Cadence: 8:7 8 FPS Animation (Score either 5, 3 or 0)
No moire pattern observed in stands within less than 1/2 second – Scores 5
No moire pattern observed in stands within less than 1 second– Scores 3
Moire pattern observed in stands intermittently or constantly through the clip– Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
8:7 8 FPS Animation	0	0	5	5	5	5	5	5	0	0	5	5

Identical results to the last two pages. A perfect showing for ATI with the nVidia GTX models equaling the results.

Notes:
Both of these tests check for colour artifacts which would be introduced by upsampling of previously downsampled video. When video is compressed the colour data is condensed and this truncated colour information can result in artifacts if the image is not correctly upscaled from the remaining information.

Scoring:
To score this correctly we have to analyse the patches which should have no colour lines in the dark areas around the colour bars. The colour bars should be smooth.

The image above shows heavy artifacting in the black areas between the bars which has been highlighted to make it easier for you to notice. This would result in a lower score. We need to carefully analyse chroma bleed.

Upsampling Errors: Interlace Chroma Problem (ICP) (Score either 5, 2 or 0)
No Visible colour lines can be observed in the dark region around the colour bars of patch. The colour bars should be smooth. – Scores 5
Very faint colour lines can be observed in the dark region around the colour bars of patch. The colour bars have some mild ‘stair' stepping. – Scores 2
Highly observable colour lines in the dark region around the colour bars of patch and within the colour bars. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Interlace Chroma Problem (ICP)	2	2	5	5	5	5	5	5	0	2	5	5

The ATi cards all score 5 in this test and again the higher end GTX boards from nVidia match the results. The Intel solutions have some noticeable stair stepping in the results as does the GT220. The G210 gives the worst result in this test as colour lines are really noticeable.

Scoring:
To score this correctly we have to analyse the patches which should have no colour lines in the dark areas around the colour bars. The colour bars should be smooth.

Chroma Upsampling Error (CUE) (Score either 5, 2 or 0)
No Visible colour lines can be observed in the dark region around the colour bars of patch. The colour bars should be smooth. – Scores 5
Very faint colour lines can be observed in the dark region around the colour bars of patch. The colour bars have some mild ‘stair' stepping. – Scores 2
Highly observable colour lines in the dark region around the colour bars of patch and within the colour bars. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Chroma Upsampling Error (CUE)	2	5	5	5	5	5	5	5	0	5	5	5

The ATi cards all score 5 in this test and again the higher end GTX boards from nVidia match the results, we were pleased the report that the G220 gave a good showing in this test also.. The Intel solutions give improved results in this test with the HD solution holding its own at the top of the pile.. The G210 again fails this test miserably.

Overview:
This second chapter focuses on the video processors ability to reduce noise throughout the image in a variety of ways. This is critical to giving a good, clean image as the human eye is very susceptible to picking up noise within a video file.

Notes:
We need to check noise reduction features for all the graphics solutions on test today, and measure the results while ensuring that fine detail is not compromised. Image quality can be reduced by blurring and smearing techniques which means a reduction in the overall scoring. Many HDTV's have noise reduction features built in which can improve image quailty, but we have omitted this today to ensure we are correctly analysing the hardware and driver solution.

In Catalyst 10.6 ATI have several sliders for reducing unwanted noise. The De-Noise slider which we find worked best on our panel at around 75%. They also have new options called Mosquito Noise Reduction and Deblocking which we set to around 3/4 full. Mosquito Noise artifacts helps reduce artifacts that appear on the edges of structures and surfaces while Deblocking helps to remove those that can appear within a surface.

Intel's G41 panel has an on/off switch for noise reduction with no fine detail over settings. On the i3 IGP, these controls are set in the Media panel/image enhancement section. This is set to AUTO, automatically.

nVidia have a noise reduction control is found in the “Adjust video image settings” area of their control panel. We find setting it around one third to one half in gave best results.

Scoring:
To score these tests correctly we needed to adjust each panels options ‘on the fly' to find the ideal area of noise reduction without losing detail on the image, it is a fine art which can take some time to get perfect. Removing all noise for instance can tend to make an image look slightly out of focus which we don't want.

Random Noise: Sailboat (Score either 5, 2 or 0)
Noise well reduced and details not softened or smeared. – Scores 5
Noise lightly reduced and details not softened or smeared. – Scores 2
Noise not reduced or details softened or smeared. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Random Noise: Sailboat	0	0	5	5	5	5	5	5	5	5	5	5

Both nVidia and ATI solutions handle this test with relative ease, putting in an almost reference result.

nVidia have a noise reduction control is found in the “Adjust video image settings” area of their control panel. We find setting it around one third to one half in gave best results.

Random Noise: Flower (Score either 5, 2 or 0)
Noise well reduced and details not softened or smeared. – Scores 5
Noise lightly reduced and details not softened or smeared. – Scores 2
Noise not reduced or details softened or smeared. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Random Noise: Flower	0	2	5	5	5	5	5	5	5	5	5	5

Both nVidia and ATI solutions handle this test with relative ease, putting in an almost reference result.

nVidia have a noise reduction control is found in the ‘Adjust video image settings' area of their control panel. We find setting it around one third to one half in gave best results.

Random Noise: Sunrise (Score either 5, 2 or 0)
Noise well reduced and details not softened or smeared. – Scores 5
Noise lightly reduced and details not softened or smeared. – Scores 2
Noise not reduced or details softened or smeared. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Random Noise: Sunrise	0	2	5	5	5	5	5	5	5	5	5	5

Both nVidia and ATI solutions handle this test with relative ease, putting in an almost reference result.

nVidia have a noise reduction control is found in the “Adjust video image settings” area of their control panel. We find setting it around one third to one half in gave best results.

Random Noise: Harbor Night (Score either 5, 2 or 0)
Noise well reduced and details not softened or smeared. – Scores 5
Noise lightly reduced and details not softened or smeared. – Scores 2
Noise not reduced or details softened or smeared. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Random Noise: Harbour Night	0	2	5	5	5	5	5	5	5	5	5	5

Both nVidia and ATI solutions handle this test with relative ease, putting in an almost reference result.

Highly compressed digital video displays artifacts on playback and these can appear as noise, blockiness or ringing on the edges of objects. The tests on the coming pages give an indication of how the graphics processor can remove compression artifacts while preserving image detail.

Scoring:
Sometimes it can be hard to see compression artifacts and the tests have moving items which can bring in judder and pulsating artifacts in the images.

Compression Artifacts: Scrolling Text (Score either 5, 3 or 0)
Noise around the edge of the characters is significantly reduced and the characters are not softened. As well, the blocks in the background are softened and not distinct. – Scores 5
Noise between the characters is substantially (not fully) reduced around the edge of the characters. As well, the blocks in the background are softened and not distinct. – Scores 3
No noise is reduced around the characters or in the background. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Scrolling Text	0	0	3	3	3	3	5	5	0	0	0	3

This is a particularly demanding test and we noticed the best results with the high end ATI graphics cards such as the HD5850 and HD5870. nVidia cards struggled with this test.

Highly compressed digital video displays artifacts on playback and these can appear as noise, blockiness or ringing on the edges of objects.

Scoring:
Sometimes it can be hard to see compression artifacts and the tests have moving items which can bring in judder and pulsating artifacts in the images.

Compression Artifacts: Roller Coaster (Score either 5, 3 or 0)
Noise around the edge of the structures and coaster is significantly reduced and blocks in the sky are softened and not distinct without loss of detail. – Scores 5
Noise around the edge of the structures and coaster and blocks in the sky is reduced but not fully. – Scores 3
No noise is reduced around objects or details are smeared or softened. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Roller Coaster	0	3	3	3	3	3	5	5	0	3	3	3

This is a particularly demanding test and we noticed the best results with the high end ATI graphics cards such as the HD5850 and HD5870. nVidia cards did better with this test however are still outperformed by the ATI boards.

Highly compressed digital video displays artifacts on playback and these can appear as noise, blockiness or ringing on the edges of objects.

Scoring:
Sometimes it can be hard to see compression artifacts and the tests have moving items which can bring in judder and pulsating artifacts in the images.

Compression Artifacts: Ferris Wheel (Score either 5, 3 or 0)
Noise around the edge of the structures and coaster is significantly reduced and blocks in the sky are softened and not distinct without loss of detail. – Scores 5
Noise around the edge of the structures and coaster and blocks in the sky is reduced but not fully. – Scores 3
No noise is reduced around objects or details are smeared or softened. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Ferris Wheel	0	3	3	3	3	3	5	5	0	3	3	3

Highly compressed digital video displays artifacts on playback and these can appear as noise, blockiness or ringing on the edges of objects.

Scoring:
Sometimes it can be hard to see compression artifacts and the tests have moving items which can bring in judder and pulsating artifacts in the images.

Compression Artifacts: Bridge Traffic (Score either 5, 3 or 0)
Noise around the edge of the bridge and cars is significantly reduced and without loss of detail in either the road, structure, or vegetation. – Scores 5
Noise around the edge of the bridge and cars is somewhat reduced but not fully or some loss of detail in either the road, structure or vegetation. – Scores 3
No noise is reduced around objects or details are smeared or softened. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Bridge Traffic	0	3	3	3	3	3	5	5	0	3	3	3

Upscaled Compression Artifacts are most common when watching Internet Video on big panels, such as plasma or LCD televisions 40 inches or above. As the image is scaled the artifacts become bigger and more noticeable.

The tests here assess the video processors ability to remove artifacts while keeping the detail as high as possible. As the images are upscaled, the noise reducer has more work to do.

Scoring:
We need to analyse these tests with noise reduction turned off, then on and fine tune the result to get the best image possible.

Upscaled Compression Artifacts: Text Pattern (Score either 5, 3 or 0)
Noise around the edge of the characters is significantly reduced and the characters are not softened. As well, the blocks in the background are softened and not distinct. – Scores 5
Noise between the characters is substantially (not fully) reduced around the edge of the characters. As well, the blocks in the background are softened and not distinct. – Scores 3
No noise is reduced around the characters or in the background. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Text Pattern/ Scrolling Text	0	3	3	3	3	3	5	5	3	3	3	3

The only cards to get the highest score are the two 58xx series cards from ATI.

The tests here assess the video processors ability to remove artifacts while keeping the detail as high as possible. As the images are upscaled, the noise reducer has more work to do.

Scoring:
We need to analyse these tests with noise reduction turned off, then on and fine tune the result to get the best image possible.

Upscaled Compression Artifacts: Roller Coaster (Score either 5, 3 or 0)
Noise around the edge of the structures and coaster is significantly reduced and blocks in the sky are softened and not distinct without loss of detail. – Scores 5
Noise between the characters is substantially (not fully) reduced around the edge of the characters. As well, the blocks in the background are softened and not distinct. – Scores 3
No noise is reduced around objects or details are smeared or softened. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Roller Coaster	0	3	3	3	3	3	5	5	3	3	3	3

Again, the only cards to get the highest score are the two 58xx series cards from ATI.

The tests here assess the video processors ability to remove artifacts while keeping the detail as high as possible. As the images are upscaled, the noise reducer has more work to do.

Scoring:
We need to analyse these tests with noise reduction turned off, then on and fine tune the result to get the best image possible.

Upscaled Compression Artifacts: Ferris Wheel (Score either 5, 3 or 0)
Noise around the edge of the structures and coaster is significantly reduced and blocks in the sky are softened and not distinct. No detail is lost. – Scores 5
Noise around the edge of the structures and coaster and blocks in the sky is reduced but not fully – Scores 3
No noise is reduced around objects or details are smeared or softened. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Ferris Wheel	0	3	3	3	3	3	5	5	3	3	3	3

Identical results to the Roller Coaster test.

The tests here assess the video processors ability to remove artifacts while keeping the detail as high as possible. As the images are upscaled, the noise reducer has more work to do.

Scoring:
We need to analyse these tests with noise reduction turned off, then on and fine tune the result to get the best image possible.

Upscaled Compression Artifacts: Bridge Traffic (Score either 5, 3 or 0)
Noise around the edge of the bridge and cars is significantly reduced and without loss of detail in either the road, structure, or vegetation. – Scores 5
Noise around the edge of the bridge and cars is somewhat reduced but not fully or some loss of detail in either the road, structures, or vegetation. – Scores 3
No noise is reduced around objects or details are smeared or softened. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Bridge Traffic	0	3	3	3	3	3	5	5	3	3	3	3

Identical results to the previous tests with ATI leading the way again.

Overview:
This third chapter focuses on Full HD resolution test patterns to test the scaling and filtering ability of the video processor.

Notes:
It is important when doing these tests to ensure that scaling is disabled both on any panels and in the driver software. Luminance Frequency Bands appear as alternating black and white stripes. The higher the frequency the thinner the stripes will be.

For this test the top half of the screen displays a set of eight boxes of vertical luminance bands and the frequency of the bands increases from the lowest frequency box in the top left of the screen to the highest frequency box in the second row on the right side. The bottom half of the screen displays the set of complimentary horizontal luminance bands.

Scoring:
Ideally, both sets of frequency bands should appear as a series of increasingly thinner lines. The lines in each frequency box should form a uniform pattern, evenly spaced with no artifacts or distoration.

Luminance Frequency Bands (Score either 5, 2 or 0)
Uniform and full amplitude on all bands and without ringing or ghosting of the lines outside the pattern. – Scores 5
Some non-uniform or attenuation only on the highest band or mild faint ringing or ghosting. – Scores 2
Some non-uniformity or attenuation on any bands or heavy ringing or ghosting. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Luminance Frequency Bands	2	5	5	5	5	5	5	5	2	2	5	5

The majority of graphics processors score 5 in this test however we noticed some mild ghosting with the X4500 solution and ghosting with the nVidia G210 and G220.

Overview:
This third chapter focuses on Full HD resolution test patterns to test the scaling and filtering ability of the video processor.

Notes:
We It is important when doing these tests to ensure that scaling is disabled both on any panels and in the driver software. Luminance Frequency Bands appear as alternating black and white stripes. The higher the frequency the thinner the stripes will be.

The Chrominance Frequency Test also consists of sets of horizontal and vertical frequency bands. Since this is a chrominance test, the lines are orange and blue instead of black and white.

Chrominance Frequency Bands (Score either 5, 2 or 0)
Uniform and full amplitude on all bands and without ringing or ghosting of the lines outside the pattern. – Scores 5
Some non-uniform or attenuation only on the highest band or mild faint ringing or ghosting. – Scores 2
Some non-uniformity or attenuation on any bands or heavy ringing or ghosting. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Chrominance Frequency Bands	5	5	5	5	5	5	5	5	2	2	5	5

The Intel X4500 handles this Chrominance test better than the Luminance test on the last page. The G210 and G220 again exhibit some mild ghosting.

Overview:
This third chapter focuses on Full HD resolution test patterns to test the scaling and filtering ability of the video processor.

Scoring:
Per the scoring criteria, this is a relatively straightforward test. Examine the edges of the characters and note whether striping, stair stepping and/or ringing artifacts are “noticeable” (score 0), “slight” (score 2) or not present at all (score 5).

Vanishing Text (Score either 5, 2 or 0)
Outlines around the characters are smooth without any striping or stepping artifacts. – Scores 5
Slight striping or stair stepping on the edge of the characters or ringing artifacts around the characters. – Scores 2
Noticeable striping or stair stepping on the edge of the characters or ringing artifacts around the characters. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Vanishing Text	5	5	5	5	5	5	5	5	5	5	5	5

All the graphics solutions on test handle this test with relative ease.

Overview:
The third chapter focuses on Full HD resolution test patterns to test the scaling and filtering ability of the video processor.

Notes:
As we said earlier upscaled video has two issues to contend with. The presence of upscaled colour and compression artifacts and a tendency for the video to soften with a lack of detail. These tests focus on the video processors abilities to handle detail enhancement. All the graphics processors on test today have various methods of enhancing detail via sharpening.

Scoring:
All the tests are combined into one result, giving a total figure. The score is based on evaluating the sequence of images. Note that fine detail would not be observable on small screens or displays which are not fully HD compliant.

Resolution Enhancement (Score either 15, 10, 5 or 0)
Fine details are more distinct and 3 dimensional. No bright or dark outlines are introduced around characters or objects. – Scores 15
Fine details are more distinct and 3 dimensional. Some bright or dark outlines are introduced around characters or objects. – Scores 10
Fine details are more distinct and 3 dimensional but mild bright or dark outlines appear around characters and objects. – Scores 5
No enhancement of detail was observed or the enhancement looks course and artificial. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Resolution Enhancement	15	15	15	15	15	15	15	15	15	15	15	15

We found that all the manufacturers have clearly spent some time fine tuning their hardware via drivers as the results were all equally impressive.

Overview:
The fourth chapter focuses on analysing the video processor's ability to enhance the image through ‘adaptive processing'.

Notes:
The first set of results examines contrast enhancement features. These results concentrate on expansion of image contrast ratios while maintaining detail in dark and white sectors. This test is not targeting image sharpness, detail, flicker or noise, just contrast enhancement.

In this particular test we look out for expansion in image contrast without the loss of fine detail in the dark areas as well as overexposure or washing out in the highlighted zones. The parked cars bear significance.

Contrast Enhancement: Theme Park (Score either 5, 2 or 0)
The overall image contrast expanded without loss of detail in the white or dark regions. – Scores 5
The overall image contrast expanded with slight loss of detail in the white or dark regions. – Scores 2
The overall image contrast expanded with moderate to high loss of detail in the white or dark regions. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Theme Park	2	0	2	5	5	5	5	5	2	5	5	5

We had an unusual issue with the Intel HD graphics with considerable loss of detail in the highlighted regions. A similar experience was given on the G210 although it was much less noticeable.

Overview:
The fourth chapter focuses on analysing the video processor's ability to enhance the image through ‘adaptive processing'.

Contrast Enhancement: Driftwood (Score either 5, 2 or 0)
The overall image contrast expanded without loss of detail in the white or dark regions. – Scores 5
The overall image contrast expanded with slight loss of detail in the white or dark regions. – Scores 2
The overall image contrast expanded with moderate to high loss of detail in the white or dark regions. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Driftwood	5	5	2	2	2	2	2	2	5	5	5	5

The nVidia cards score particularly well with this specific test, as do the Intel solutions. There is a slight loss of detail with the ATi boards in the highlighted areas.

Overview:
The fourth chapter focuses on analysing the video processor's ability to enhance the image through ‘adaptive processing'.

In this particular test we look out for detail in the vegetation and dark beach areas which are lower and right zones on the screen. Watching carefully for any loss in fine detail in the brighter regions – the sky and the clouds. It is also important to monitor overall scene brightness which should remain rigid.

Contrast Enhancement: Beach at Dusk (Score either 5, 2 or 0)
The darker beach and vegetation area has more detail without increasing overall brightness or loss of detail in the bright regions – Scores 5
The darker beach and vegetation area has more detail but overall brightness increased or some loss of detail in the bright regions. – Scores 2
Overall contrast increased but with loss of detail in dark and bright area. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Beach at Dusk	2	2	5	5	5	5	5	5	5	5	5	5

Intel solutions failed to deliver the same quality of results as the ATI and nVidia boards with this particular test. We noticed loss of detail in bright regions.

Overview:
The fourth chapter focuses on analysing the video processor's ability to enhance the image through ‘adaptive processing'.

In this particular test we look for overall reduction in contrast to expose greater detail, especially with the black fur of the cat. Bright areas in the scene such as the white cat and background textures should also have improved detail.

Contrast Enhancement: White and Black cats (Score either 5, 2 or 0)
Overall contrast is reduced so more detail is visible in the bright and dark areas. – Scores 5
Detail is improved on either one of the dark or bright area without loss of detail in the other area. – Scores 2
Overall contrast remains the same or is increased and detail is loss in either the bright or dark areas. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
White and Black Cats	5	5	5	5	5	5	5	5	5	5	5	5

All the hardware/drivers on test today managed to deliver an almost reference set of results in this particular test.

Overview:
The fourth chapter focuses on analysing the video processor's ability to enhance the image through ‘adaptive processing'.

Notes:
One aspect of good movie watching is believing that the people in the scenes are ‘real and 3 dimensional' which is helped in part by accurate looking colour in regards to skin tones. If a face is too yellow or too red then it gives a very fake impression and can ruin the immersive experience. This test shows a group of people with reference hues and then with too much yellow and a third time with too much red. This gives a good indication of positive and negative hue shift. Obviously this is not as straighforward as it sounds, as there are literally hundreds of different skin tones. The video processor has to analyse the image and fix problems which have been deliberately introduced into the 2nd and 3rd images.

Scoring:
The scoring takes into consideration the effectiveness of the fleshtone correction feature. Scores vary from a 10, which is ‘substantially closer to the original skin tone' to a 0, which is ‘no improvement'.

Skin Tone Correction: Skin Tones (Score either 10, 7, 3 or 0)
The off hue skin tones appear substantially closer to the original skin tone and does not affect other colors. – Scores 10
The off hue skin tones are somewhat corrected but hue differences are still discernable. Others colors are not affected. – Scores 7
The off hue skin tones are somewhat corrected but hue differences are still discernable. Affects other colors. – Scores 3
No improvement can be observed. – Scores 0

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Skin Tones	0	0	3	7	7	7	7	7	3	3	3	3

ATi have the slight edge in this test, probably helped by the dedicated controls they offer in their panel to correct flesh tones. We found a setting around 70% in this panel seemed to give the best results.

HQV Benchmark has always been a very useful application to ascertain the quality of video output from the hardware, through the driver into the panel. V 2.0 literally doubles the amount of tests on offer, giving the end user more options to fine tune and enhance their viewing experience.

The suite is designed to detail and report any weaknesses with video output and to educate and inform the user with possible driver adjustments. After many many hours of testing, our final results are shown below.

	Intel GMA X4500 (G41)	Intel HD Graphics	ATI HD4200 (IGP)	ATI HD5450 (Discrete)	ATI HD5550 (Discrete)	ATI HD5770 (Discrete)	ATI HD5850 (Discrete)	ATI HD5870 (Discrete)	NV G210 (Discrete)	NV G220 (Discrete)	NV GTX 465 (Discrete)	NV GTX 480 (Discrete)
Dial	0	4	0	4	4	4	4	4	0	2	4	4
Dial with static pattern	5	5	5	5	5	5	5	5	5	5	5	5
Gray Bars	0	3	0	3	5	5	5	5	0	2	5	5
Violin	0	3	3	5	5	5	5	5	0	3	5	5
Stadium 2:2	0	0	3	5	5	5	5	5	0	0	5	5
Stadium 3:2	0	3	3	5	5	5	5	5	0	3	5	5
Horizontal Text Scroll	3	3	5	5	5	5	5	5	3	3	3	3
Vertical Text Scroll	3	5	3	5	5	5	5	5	5	5	3	3
Transistion to 3:2 Lock	0	5	5	5	5	5	5	5	0	0	5	5
Transition to 2:2 Lock	0	5	0	0	0	0	0	0	0	0	5	5
2:2:2:4 24 FPS DVCAM Video	0	0	5	5	5	5	5	5	0	0	5	5
2:3:3:2 24 FPS DVCam Video	0	0	5	5	5	5	5	5	0	0	5	5
3:2:3:2:2 24 FOS Vari-Speed	0	3	5	5	5	5	5	5	0	3	5	5
5:5 FPS Animation	0	0	5	5	5	5	5	5	0	0	5	5
6:4 12 FPS Animation	0	0	5	5	5	5	5	5	0	0	5	5
8:7 8 FPS Animation	0	0	5	5	5	5	5	5	0	0	5	5
Interlace Chroma Problem (ICP)	2	2	5	5	5	5	5	5	0	2	5	5
Chroma Upsampling Error (CUE)	2	5	5	5	5	5	5	5	0	5	5	5
Random Noise: Sailboat	0	0	5	5	5	5	5	5	5	5	5	5
Random Noise: Flower	0	2	5	5	5	5	5	5	5	5	5	5
Random Noise: Sunrise	0	2	5	5	5	5	5	5	5	5	5	5
Random Noise: Harbour Night	0	2	5	5	5	5	5	5	5	5	5	5
Scrolling Text	0	0	3	3	3	3	5	5	0	0	0	3
Roller Coaster	0	3	3	3	3	3	5	5	0	3	3	3
Ferris Wheel	0	3	3	3	3	3	5	5	0	3	3	3
Bridge Traffic	0	3	3	3	3	3	5	5	0	3	3	3
Text Pattern/ Scrolling Text	0	3	3	3	3	3	5	5	3	3	3	3
Roller Coaster	0	3	3	3	3	3	5	5	3	3	3	3
Ferris Wheel	0	3	3	3	3	3	5	5	3	3	3	3
Bridge Traffic	0	3	3	3	3	3	5	5	3	3	3	3
Luminance Frequency Bands	2	5	5	5	5	5	5	5	2	2	5	5
Chrominance Frequency Bands	5	5	5	5	5	5	5	5	2	2	5	5
Vanishing Text	5	5	5	5	5	5	5	5	5	5	5	5
Resolution Enhancement	15	15	15	15	15	15	15	15	15	15	15	15
Theme Park	2	0	2	5	5	5	5	5	2	5	5	5
Driftwood	5	5	2	2	2	2	2	2	5	5	5	5
Ferris Wheel	0	3	3	3	3	3	5	5	3	3	3	3
Skin Tones	0	0	7	7	7	7	7	7	3	3	3	3
Total	49	111	152	173	175	175	193	193	82	114	172	175

It is worth pointing out again that HQV Benchmark 2.0 is not an exact science, the testing is subjective. While many people will be using a panel or television with heavy processing, we have attempted to negate this by using one of the most expensive panels on the market with a very wide gamut and colour depth. This ensures that the output from the hardware, via the driver to the panel is clean and pure. We also studied reference documentation from the creators of the benchmark to educate ourselves on all the tests beforehand. Knowing exactly what you are looking for is tantamount to ensuring you can accurately measure the results.

Here are the various performance figures stacked for each solution.

Once you've evaluated all the HD processing options with HQV, the various levels of quality are obvious

So what can we say about the 3 companies who compete for your hard earned bucks in this market?
.
Intel: Even though the original programme for discrete Larrabee was cancelled, it's obvious that Intel's engineers have already learned a lot about high quality video processing and each generation is definitely getting better. Comparing an integrated graphics processor against a state-of-the-art GPU from someone like ATI is never going to look good for Intel. Dedicated hardware solutions are always going to rock. But you must to take your hat off to Intel when you consider that its latest chips score more than 125% higher than its previous generation. Intel may not be a leader right now, but it is very definitely moving in the right direction – and it's doing so faster than anyone else.

nVidia: Again, progress has been made. The jump from nVidia's G210/220 scores to the present generation of Fermi processor is substantial. Also, once nVidia's engineers have had a chance to look through KitGuru's test results, it will be straightforward for them to target issues the issueswe've flagged and eliminate them. It's unlikely that the lower scores are anything to do with hardware, so we'd expect the scores to pick up within a few driver iterations. Likewise on the functionality side of things, we expect to see some of ATI's new functions mirrored in the nVidia driver panels soon enough, for example mosquito noise reduction and deblocking.

ATI: It seems that ATI's background with the production of processing chips for digital TVs still gives it some kind of advantage. Even budget cards like the passively cooled Sapphire HD 5550 generate similar video quality when compared with nVidia's flagship GTX480. Obviously a lot of this is down to driver development and advancements. For now, KitGuru can state clearly that ATI's Radeon HD 5000 series cards offer the best HD video reproduction quality available in the market.

Our findings have also shown that Catalyst 10.6 is without a doubt the best ATI driver yet, specifically in regards to the quality of video rendering. The new features incorporated into Catalyst 10.6, such as Mosquito Noise Reduction and DeBlocking, really to help to enhance picture quality output, with our reference monitor it was easy to ascertain quality differences. Just as we're sure that ATI's discrete solutions are leading the HD video quality pack right now, we're also positive that nVidia will be working hard to catch up with subsequent driver releases over the coming months.

KitGuru says: We hope that this article today has not only informed you on slight variances with hardware and driver configurations but has educated you on how to run and analyse these tests for yourself … we would like to hear your own reports in our forums.

Become a Patron!

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

Tags ATI ati v nvidia bluray optimisation Cat 10.6 Catalyst 10.6 G41 G45 HDTV quality HQV HQV 2 HQV 2 article HQV 2 preview HQV review Intel media playback news nvidia nvidia v ati radeon radeon HD

Check Also

Nvidia RTX 5080 expected to launch before RTX 5090

Nvidia is shaking things up with its GeForce RTX 50 series launch, prioritising the RTX …

117 comments

Death Dealer
June 16, 2010 at 4:12 pm

what a stunning article Zardon, im speechless.
Eric K
June 16, 2010 at 4:14 pm

I really dont know how you get time to do all these articles man, its highest quality journalism. seriously.
Tim
June 16, 2010 at 4:15 pm

Shit thats good info ! AMD for me next GFX I think 🙂
Frank
June 16, 2010 at 4:18 pm

Glad the site came back up there, I was reading this and I lost it !

What an amazing read man, top job.
Stefan
June 16, 2010 at 4:21 pm

Amazing, absolutely bloody amazing. Surprised how well intel did actually considering.
Ray
June 16, 2010 at 4:22 pm

wow, ive been reading an hour and im not even half way through it, will check out the rest later, what detail ! love it guys.
Derek
June 16, 2010 at 4:24 pm

Yeah, the 480GTX did well too I think, nVidia were well known not to have great IQ in video, and this shows they are still lagging but catching up a bit.
tom
June 16, 2010 at 4:26 pm

so much detail in this id need the weekend to really take it all in. Interesting to see the Catalyst 10.6 in this. when they arent even out yet !
tri color
June 16, 2010 at 4:28 pm

staggering level of work in this chaps. most impressed
AMD4EVER
June 16, 2010 at 4:31 pm

Well I would be lying if I said I understand even half of it, but what I do get is AMD ROCKS 🙂
francois lebon
June 16, 2010 at 4:33 pm

Catalyst 10.6, I must have it. now!
breaker
June 16, 2010 at 4:35 pm

Nice writeup. good info and id like to try this myself, how much does it cost? is it out yet?
Trev mang
June 16, 2010 at 4:37 pm

Amazing editorial piece, best ive read all year by a long shot. Cant even imagine the work load in this.
Dax
June 16, 2010 at 4:38 pm

43 pages, eh? is this for real. shit. I just read the conclusion, sorry ! but ill read more later. Intel did well didnt they? i thought they would end up with 5 points.
fluffychicken
June 16, 2010 at 4:41 pm

I don’t really even know what to say after reading that. how long did this take to test? 3 weeks? insane testing and very useful to get information about driver quality. I am sure nvidia will release a set shortly to target this benchmark which is good as it will assuredly improve IQ real world for everyone.
brad
June 16, 2010 at 4:43 pm

lol. really, lol 🙂 I gave up after page 10, information overload. but I get the point from the conclusion which is all I need, ATi seem to have trhe edge.
tech head
June 16, 2010 at 4:58 pm

simply staggering detail and what a great idea for a benchmark. Lets use OUR OWN EYES and not have some coder tell us what is better. and then we get fake results when they can optimise drivers. real world testing FTW.
dauth maul
June 16, 2010 at 5:00 pm

Well I really do love this article. why? as someone said earlier, its not sythethic crap. its real world to the core. I am sure different people might get slightly different results, but ive known Zardon long enough on Driverheaven over the years to know he is pretty expert on video stuff.
Jim oldtimer
June 16, 2010 at 5:00 pm

Brilliant to read and actually understand. great the way all the tests MEAN something and a great benchmark idea from HQV. ill get myself a copy and hopefully a 5550 shortly too 🙂
Joseph F
June 16, 2010 at 5:03 pm

Amazing reading. im not a fan of futuremark style tests. what does a score of 25,000 really mean? apart from epenis on the forums. This is what I wanted to see for a long time. I think nvidia will add some new features to their drivers also. soon.
geek master
June 16, 2010 at 5:05 pm

Good to see ATI dont win ALL the tests ! seems nvidia have a bit of driver work to do. In 6 months I bet both companies are over 200 points each, battling to max out the scores.
i love porn movies
June 16, 2010 at 5:06 pm

Great testing and I liked the use of a reference monitor, rather than a processed loaded TV which would alter the raw settings anyway
dark knight
June 16, 2010 at 5:07 pm

Id like that monitor myself, any chance I could trade it for my 24 inch Dell? its a good deal, really.
Mark Ramistin
June 16, 2010 at 5:08 pm

Intel did pretty well didnt they for not really spending much time doing this kind of work? very impressed with their results in some of those tests. That said, AMD here I come 🙂
Skin Force
June 16, 2010 at 5:11 pm

that was a wicked read over lunch here. have to take it all in again later then get the new drivers and maybe order the benchmark, how much is it ?
Stefan
June 16, 2010 at 5:12 pm

I think its only $25 on their store.
Fred
June 16, 2010 at 5:13 pm

great read, but it seems a bit biased to me for ATI!
Gareth Cringle.
June 16, 2010 at 5:14 pm

lol, always one idiot isnt there 🙂 I would think you should get a copy and try it out yourself man.
horrace
June 16, 2010 at 5:15 pm

im sold, I want these drivers, cant get them on ATIs site yet ;(
Eric k
June 16, 2010 at 5:16 pm

I have been refreshing to try and get them, nothing yet 🙁
Tunick
June 16, 2010 at 5:18 pm

What a fucking brilliant editorial KG. I nkow you get a bit of stick for your news sometimes, but the reviews are top drawer.
Rob Max
June 16, 2010 at 5:19 pm

I ordered the bluray disc, need to check this out for myself 🙂
me
June 16, 2010 at 5:20 pm

10/10. Thank you for the work. bookmarked it as a reference.when I order the bluray myself.
Trey Kool
June 16, 2010 at 5:23 pm

I would like to just pass on my regards for spending all this time for something so informative and educational. This is a credit to the website and will be bookmarked by thousands I feel for a reference. I hope in 3 months time you can follow up with new drivers from all companies and measure the changes
Melachonly Freebie taker
June 16, 2010 at 5:26 pm

Posted this on the wrong article ! urgh. Thanks for the article, its the most indepth ive seen online. very helpful, even if I struggled to keep up with some of the tests.
henry
June 16, 2010 at 5:28 pm

I used to think I could be a reviewer/writer for a tech site, until I read something like this and realise I would just jump out the window or shoot myself in the head.

work overload 🙂
nerd sucker
June 16, 2010 at 5:29 pm

fabulous reading and it shows that drivers do make a massive difference. ATI have came a long way in 4 years. I remember when their drivers were a disaster.
long jumper
June 16, 2010 at 5:30 pm

stunning work guys, I have to read this more later.
Flo
June 16, 2010 at 5:30 pm

Interesting to see termingology for things ive seen with my HD movies but never understood. invaluable.
Terrance M.
June 16, 2010 at 5:31 pm

Great to see such an indepth analysis for an important subject matter.
faith
June 16, 2010 at 5:32 pm

ATI winning another video quality shoot out is not a surprise, we agree.
nVidia’s improvement was good to see, but for me the biggest story here
is just how much Intel has come on. If they ever manage to get a discrete
graphics solution to market, these boys will be dangerous.
Rick
June 16, 2010 at 5:32 pm

I am so annoyed I bought a GT220 for media after reading this. should have went for a 5550 🙁
Ron
June 16, 2010 at 5:33 pm

Glad I went for a HD5850 ! worth the cost. I think lol
Trix
June 16, 2010 at 5:34 pm

I wonder do ATI do trade in deals for HD4 series cards anywhere? id like a 5 series but just cant afford it
Stuy Lewis and the news
June 16, 2010 at 5:34 pm

I agree, Intel did very well, I wasnt expecting that at all.
Vivian
June 16, 2010 at 5:37 pm

Very very interesting. Wasnt expecting intel to outperform one of the nvidia boards. that g210 really is a piece of shit.
Flung hard
June 16, 2010 at 5:38 pm

Intel scoring so hard blew my mind, I thought their onboard solutions sucked ass.
Wez
June 16, 2010 at 5:38 pm

Zardon I love you. Really. glad to see you back and kicking ass with this level of article.
Garry
June 16, 2010 at 5:39 pm

speechless. ill come back later when I can digest a third of it 🙂
Tim Smalls
June 16, 2010 at 5:40 pm

Wow just what I wanted, im sick seeing game benchmarks, which mean nothing to me.
Stereophonic
June 16, 2010 at 5:47 pm

Amazing detail of testing there fella. cant imagine how long that took. thank you.
Raymond Ticklestick
June 16, 2010 at 6:24 pm

well that was quite the ride and very informative to someone who knows very little about IQ.
Hairy Beast
June 16, 2010 at 6:25 pm

That really got me worked up KitGuru. very good article, I shall b e back.
Paul Sheers
June 16, 2010 at 6:26 pm

While it was important you used a reference panel, I think people should know this is subjective testing and as such errors can creep in depending on the user.
Paul Sheers
June 16, 2010 at 6:26 pm

Why not use a TV, no one will have a panel that expensive.
Death Dealer
June 16, 2010 at 6:28 pm

@ Paul sheers. read the pages, explains why. you can’t test something if TV processing and enhancements are between the driver/hardware and viewer. We need to see quailty of the raw image to the panel, not work out what is being processed by the panel to improve on things.
Harry
June 16, 2010 at 10:08 pm

Brilliant. Top job – shall use this as a guide in college
francois lebon
June 16, 2010 at 10:09 pm

Wow ive been trying to read this thoroughly all night and the site has been up and down. problems? Great article, still cant get seeing it all yet as I cant load some pages, but ill try later.
Freak show
June 16, 2010 at 10:10 pm

SO useful this in my thinking about a future card. can nvidia catch up though with their drivers or could it be more hardware related?
Dereke
June 16, 2010 at 10:11 pm

Have had a real hard time getting here tonight. traffic issues? I am surprised intel did so well really compared to some of the discrete solutions, but its impressive for them. I think.
Fortitude (RF)
June 16, 2010 at 10:11 pm

This is exactly why ive always bought AMD, even when I know nvidia often make better gamign cards. I watch A LOT OF HD content and it verifies my own findings.
faith
June 16, 2010 at 10:12 pm

BIG APOLOGIES

Zardon’s HQV 2 article has generated a wave of traffic the like of which we’ve never seen before

Our dedicated server just could not cope as traffic spiked >8x normal!

Fear not, we have a solution coming in the next day or so, which will change KitGuru’s capability massively

In the meantime, please bare with us
Ricky Holding
June 16, 2010 at 10:12 pm

I dont normally post comments on sites, but I wanted to lay down a thanks for this work. its really benefited me in the decision for a card.
Harding R.
June 16, 2010 at 10:14 pm

I really rated HQV 1 and bought it, and will buy HQV2. This is a killer article however and ive noticed it linked a lot tonight. Will really help you get attention from the media audience.
Trial by fire
June 16, 2010 at 10:15 pm

Finally got to read this after 3 hours of trying ! you trying to create false demand by taking the site down ?

ITs an overrated review, not much work at all in it.

No, im kidding, awesome, just messing around with you faith 🙂
Garabic
June 16, 2010 at 10:17 pm

Hey man, if you ever need work, let me know, ive a magazine will employ you in an instant 🙂
Trebor Mints, love em
June 16, 2010 at 10:18 pm

Had to post a thanks for this Kitguru, a considerable read by any definition of the words.
Rocksteady Crew
June 16, 2010 at 10:19 pm

Linked to this from our site,and thought id pop on and say congrats on the work, very definitive.
Anand
June 16, 2010 at 10:20 pm

polished, indepth and educational. Good read.
Fadilickier
June 16, 2010 at 10:21 pm

Hello there, very informative and useful. I like the idea of a real world ‘benchmark’ rather than a prescripted figured based thing.
Haditbutlostit
June 16, 2010 at 10:22 pm

yeah, nice work, HQV rocks and this is a showcase of how damn good it is. we need more tests like these on the market. Actually relating directly to how we enjoy movies and theater.
Robert Truhihol
June 16, 2010 at 10:23 pm

Hello from Spain! read this earlier then the site went down and my friends couldn’t see it.

Many thanks from us, this is very useful. HD5850 ordered 🙂
Lativia Horriscopic
June 16, 2010 at 10:24 pm

Hello from Spain! read this earlier then the site went down and my friends couldn’t see it.

Many thanks from us, this is very useful. HD5850 ordered 🙂
Palas Yankovic
June 16, 2010 at 10:25 pm

We like this for news and for breath of informative nature. useful in circumstances to get hardware needed.
Klaus Distorm
June 16, 2010 at 10:26 pm

HD5850 is the king of all cards, for the money its impossible to beat. nothing has changed. HQV V 2.0 is impressive.
Steve Read
June 16, 2010 at 10:27 pm

I like the style of testing they use, as it covers all forms of artificating and issues we experience with less than perfect configurations. Once they make a 52 inch version of the LaCie 730 screen so I can watch my movies I will be happy. and quite possibly broke.
Lucy Terbonne
June 16, 2010 at 10:28 pm

Thank you, I almost bought a G210 last month, I just ordered a 5450.
Darkness involves love
June 16, 2010 at 10:29 pm

Good test procedures, as the reference monitor has no filtering or interlacing. Clean image out means clearer to see the importance of the driver coding and quality throughput of hardware. 10/10
Grass is greener
June 16, 2010 at 10:30 pm

I love this writing and testing is very useful. I was surprised to see the GTX480 being outclassed by lesser powerful ATI solutions, however ATI I tink have many new driver features in 10.6 maybe nvidia will copy them and get the standard higher.
Gary Walker
June 16, 2010 at 10:31 pm

This site is now my homepage. I never knew it existed, and its almost criminal I didn’t I just read 10 reviews. KEEP IT UP PLEASE ! (ex tomshardware fan).
Hans Zimmer
June 16, 2010 at 10:33 pm

This article has helped me work out the wheat from the chaff. Decided to order a 5750 for my media center. It can also play games if I need on the TV> anything lower is fine, but not good for gaming. Thanks kitguru for an education read, I feel so much more capable in picking my hardware this time around for what I need (HD and bluray first, gaming next)
Gets out
June 16, 2010 at 10:34 pm

very overjoyed to read this. excuse english, informative good. Figures interesting. result helped buy card. HD5850. thank you.
Paul
June 16, 2010 at 10:35 pm

Fuck it, just ordered two 5850s for my system. I hate this site now 🙂
Pepe
June 16, 2010 at 10:36 pm

J’aimerais vous remercier les gens pour l’écriture ceci. J’ai voulu une bonne base de données pour mesurer l’exécution dans les secteurs clés, comme la cadence. Ceci a aidé me choisit la solution vidéo juste pour mes besoins très heureux.
Michael
June 16, 2010 at 10:37 pm

Vad ett stort jobb och kvaliteten av skrift är översta skåra. Vi tackas som för tiden och ansträngning att hjälpa oss att köpa kort.
Jurgen
June 16, 2010 at 10:38 pm

Vielen Dank für das Helfen von mir mit einem Kauf des HD5850 für meine Medien zentriert. und für ein Gaming auch(sorry bad english)
Brian
June 16, 2010 at 10:39 pm

Well above and beyond the call Z man. well done.
Fruitiger Condensed
June 16, 2010 at 10:40 pm

Quite surprised when I saw text links to 43. then saw it was actually 43 pages.

I would hate this job, but I appreciate the work for everyone. Not often you get a man nuts enough to write 43 to help for a video card purchase. bookmarked for later.
AMD really do rock my world
June 16, 2010 at 10:41 pm

nvidia doubters take note. if you want IQ and gaming performance, you dont buy green.
Stefan
June 16, 2010 at 10:41 pm

wow quite a few new faces here.
Stefan R
June 16, 2010 at 10:42 pm

Ik maakte een geïnformeerde beslissing om een 5 seriekaart te krijgen bedankt naar dit artikel. Ik kijk hoofdzakelijk MKV bestanden en Bluray. Slechts wat ik wilde lezen. Dank u kitguru
Tony
June 16, 2010 at 10:43 pm

This is way too much to read. but thank you, you do need therapy for writing all that however
Max
June 16, 2010 at 10:44 pm

I wonder why ATI did so bad in a few of the tests. they seem quite important ones too. anyone else notice it ?
Death dealer
June 16, 2010 at 10:45 pm

@ MAX – yes 2:2 lock – wondered why they failed that so badly when nvidia did so well.
Eric K
June 16, 2010 at 10:46 pm

2:2 lock can be important in some situations, but I think ATI might have it fixed by the time driver 11 is out.
Death Dealer
June 16, 2010 at 10:47 pm

I dont understand the specific test though, what would it affect in real movie terms, fast moving multicolored objects moving across axis?
eric k
June 16, 2010 at 10:49 pm

“When the globe is moving through the right side of the screen, it is displayed with a 3:2 cadence, while on the left, the cadence is 2:2. A scoring guide is provided on the screen, with a green zone that begins right at the cadence boundary, a yellow zone which starts a half-second later, and a red zone that begins one second into the cadence zone.”

It seems to be more a transitional state between areas and brightly coloured boundaries which could mean the cards have a slight delay in locking onto them for processing reasons. Maybe Terry Makedon would have info?
death dealer
June 16, 2010 at 10:50 pm

Zardon, dont know if you read these, but can you ask Terry about that test? and if they can get a score on it? transitition locks seem a rather important processing element.
Stefan
June 16, 2010 at 10:51 pm

Takes more than one second to transition from 3:2 to 2:2- Scores 0

Sounds like a processing glitch when moving from primary coloured areas to another ?
Bob
June 16, 2010 at 10:55 pm

wow. what an article. would take me a week to read it all.
Kimil
June 16, 2010 at 10:56 pm

thank you for all the testing, most helpful!
Terrance
June 16, 2010 at 10:59 pm

Definitive. if the site goes down at some stage, this will be the article to remember kitguru by!
Doorknocker
June 16, 2010 at 11:00 pm

Not sure how to really say it with feeling, but ill try.

Maximum headshot! nailed IQ. nailed the testing. we want MORE!
Charles
June 16, 2010 at 11:02 pm

I have been following Zardon since he was a moderator on Rage3d and was actually probably one of their best too. When he started DriverHeaven I loved the reviews. Now KitGuru. One of the most unashamedly talented technology writers on the planet and doesnt get half the credit he should. This is an epic piece of work. Mucho gracias.
Brent
June 16, 2010 at 11:04 pm

He was a moderator on Rage3d? didnt know that. but yeah … what charles said. where ever he goes, ill find him !
Tom Beatty
June 16, 2010 at 11:23 pm

I just ordered the ASUS bluray online, that is such a sexy little drive. had to have it.
Iluvatar
June 17, 2010 at 2:48 pm

Zardon , you may rule Middle Earth, you are wothy!!!
BigN
June 19, 2010 at 8:11 pm

The people that make HQV must be impressed. Every search engine ranks Zardon’s article as the best on benchmarking video. Maybe they should hire him 🙂
gabs
June 20, 2010 at 5:33 pm

STunning, but maybe more pictures of the tests?
DavidC1
June 21, 2010 at 2:12 am

Why do you guys use the antiquitated 8.15.10.1952 drivers for Intel while the competitors use the nearly/latest drivers? One might say that’s called a “bias”.

There has been at least 6 driver versions since 1952.
Zardon
June 21, 2010 at 8:36 am

Actually there is nothing biased about it at all. Due to an administration error the wrong intel driver was listed. it was V15.17.4.2119
Dragos
June 22, 2010 at 1:53 pm

Zardon can you please tell us what where the settings used in Catalyst Control Center, I mean besides the ones that you already wrote about (like Brither Whites, EE ?)

Thanks
EuGENIUS
June 24, 2010 at 8:13 am

No shots for denoising\sharpening?! That’s REALLY bad! 😉

I always prefer to check such things myself, just because too often I see something like “great resolution enhancement algo in nvidia Purevideo” etc., but when I try to compare… well, it’s just “cheap and fast” version of unsharp mask, and it kill too much real details and adds too much galos\noise to make sharp edges.

This conclusion is based on my old (2007) tests and may be absolutely wrong now. But I have no need for deeply re-evaluate DXVA decoding… I choose software decoding with ffdshow, because it’s sharpening\upsampling\denoising capabilities are just outstanding. 🙂