|
|
|
286
 
Group: Forum Members
Last Login: Yesterday @ 19:08:01
Posts: 484,
Visits: 3,253
|
|
| Slipstream has been a strong advocate of ATI picture quality being > Nvidia, but equally, many remain in the corner of the higher FPS. As you can see from my sig, slippy won me over, but I would really like to see a detailed test on the issue. Difficult as picture quality is subjective I hear you say. How about this, 2 identical rigs, one with a 8800gt, one with a 3870. A sample of games, maybe the performance test like in CoH, then show each rig at 3 different levels, min, med and max to a number of gamers (the higher the number the better). You would have to muddle up the order of the rigs, otherwise the person watching would know that 1, 2 and 3 were from card A and 4,5 and 6 from card B. Or even better, have the two rigs next to each other running the same settings on two monitors next to each other, each person must then pick if one looks better than the other or if they look the same. I personally would be quite interested to see the results.
My Rig  E6600 @ 3.2, P5KC, 2Gb 6400 @ 4.4.4.9, HD3870, 2X500Gb
|
|
|
|
|
Pentium
Group: Forum Members
Last Login: Yesterday @ 23:29:02
Posts: 5,720,
Visits: 5,674
|
|
|
|
|
|
Pentium
Group: Forum Members
Last Login: Yesterday @ 23:00:16
Posts: 1,527,
Visits: 2,484
|
|
You may have to limit the frame rates too, as the NVidia will invariably be faster, though obviously at the expense of image quality!! 
PC No. 1
C2D E8200@3.6ghz - Gigabyte GA-P35-DS3 - SapphireHD3870- 2GB Crucial Ballistix PC6400 + 2Gb Patriot PC6400- SEAGATE BARRACUDA 320GB - SEAGATE BARRACUDA 500Gb - HIPER TYPE-R 580W - Antec 900 - Running Vista x64 SP1
|
|
|
|
|
286
Group: Forum Members
Last Login: Yesterday @ 19:08:01
Posts: 484,
Visits: 3,253
|
|
| I don't think you should limit the frame rates, that is part of the test. If the person notices slow down, then that card is clearly not up to it.... Or maybe MM could set up a blind online survey..... E.g. 6 pictures from a game, 3 from each card, one at each setting, we then have to rank them in the order of what we think looks best...
My Rig E6600 @ 3.2, P5KC, 2Gb 6400 @ 4.4.4.9, HD3870, 2X500Gb
|
|
|
|
|
Pentium
Group: Forum Members
Last Login: Today @ 00:18:24
Posts: 2,175,
Visits: 5,048
|
|
While it would make an interesting article I don't think there is very much difference.
I base my conclusion on the image quality being slightly subjective and there being different implementations. i.e. 2xAA may be optimised for speed on the ATI card and quality on the NVidia card so you are not comparing like for like.
gaming: E4400@2.66GHz / P5K-E / 2x1GB PC8000@533MHz / 2x80GB D'Max 9-RAID0 + 320GB / 8800GTS 512MB / ViewSonic VX2835wm
server+media: P4 2.8GHz @ 2.4GHz / 775Dual-VSTA / 2x512MB PC3200 / 1000GB+500GB+250GB / HD2400pro
also: P3 for the mrs and a linux box and probably enough bits to build another one
|
|
|
|
|
Pentium
Group: Forum Members
Last Login: Today @ 01:04:17
Posts: 8,025,
Visits: 14,935
|
|
You couldn't do it with video Andy, only stills. 
For a video to give a true representation of the original input it would have to be in an uncompressed video format and the files would be absolutely gigantic.
It's very difficult to do with stills accurately as well due to the Temporal Antialiasing used by almost all ATI cards for some time now. You'd need to show a mixed version of up to three consecutive frames to give a true impression of just how good this additional method of antialiasing really is. AFAIK, nVidia have nothing similar. Please correct me if I'm wrong on this as I don't want to spread untruths.
Bear in mind that any screenshots that you're likely to have seen so far will not be showing the effects of Temporal Antialiasing as they've all been individual single-frame screengrabs.
All recent(ish) ATI cards including the X1950Pro support this feature at a level of x3. Each multiplication doubles the effectiveness of FSAA. x1 = off, x2 = a doubling, x3 = a quadrupling of total antialiasing effectiveness.
It's safe to assume that any image you see displayed as a single-frame screengrab would look four times worse in terms of antialiasing than the actual image appears to the eye on a "live" screen.
As I keep saying, you simply can't compare apples and oranges.
Cheers, Slipstreem. 
System specs: "Phoenix" - Intel C2D E4500 overclocked to 3GHz with ACF7Pro HSF on Volt-modded ASRock 775Dual-VSTA mobo (modded BIOS rev 3.10A and VNB=1.65, Vagp=1.8), 2x1GB Crucial Ballistix DDR2-800 RAM (3.0,3,3,8,1T @546MHz), Sapphire ATI HD3870 512MB GDDR4 PCIe graphics card overclocked to 850MHz GPU & 2.4GHz RAM. Powered by Hiper Type-M 580W PSU. Guess who likes overclocking on a budget. 
MP3 Encoding for Audiophiles
Fun MPEG-4 Encoding Race
MPEG-4 Playback Enhancement Using FFDShow
How good is the Arctic Cooling Freezer 7 Pro HSF really?
Boosting ATI Framerates with CCC (X700 on)
Optimise ATI Image Quality And Framerates With ATT (X1XXX Series under WinXP)
   
|
|
|
|
|
286
Group: Forum Members
Last Login: Yesterday @ 19:08:01
Posts: 484,
Visits: 3,253
|
|
OK.... Back to original idea... One member of the MM team set up two comp, two screens, side by side, doesn't tell the rest of the MM team which is which... Go through lots of different combinations of live video, e.g. the CoH screen test, and the rest of the MM team vote for their choice. Not a test we could be part of mind, but would be interesting if all the MM team pick the ATI at med over the gt at high.....
My Rig E6600 @ 3.2, P5KC, 2Gb 6400 @ 4.4.4.9, HD3870, 2X500Gb
|
|
|
|
|
Pentium
Group: Forum Members
Last Login: Today @ 01:04:17
Posts: 8,025,
Visits: 14,935
|
|
They'd have to use proper displays as opposed to TFT monitors to really appreciate the benefit, but I'd be surprised if they couldn't see some difference.
Even without being able to see the benefit of Temporal Antialiasing, every screenshot comparison I've seen so far shows that the antialiasing used by ATI is highly superior. Multiply this difference by a factor of four to take into account the effects of Temporal Antialiasing and you get some idea of how vast the gap in image quality is in this respect.
As has already been said though, it's largely subjective. Many people claim that they can't see the difference. I think that their poor quality TFT displays are largely to blame for this.
Cheers, Slipstreem.
System specs: "Phoenix" - Intel C2D E4500 overclocked to 3GHz with ACF7Pro HSF on Volt-modded ASRock 775Dual-VSTA mobo (modded BIOS rev 3.10A and VNB=1.65, Vagp=1.8), 2x1GB Crucial Ballistix DDR2-800 RAM (3.0,3,3,8,1T @546MHz), Sapphire ATI HD3870 512MB GDDR4 PCIe graphics card overclocked to 850MHz GPU & 2.4GHz RAM. Powered by Hiper Type-M 580W PSU. Guess who likes overclocking on a budget.
MP3 Encoding for Audiophiles
Fun MPEG-4 Encoding Race
MPEG-4 Playback Enhancement Using FFDShow
How good is the Arctic Cooling Freezer 7 Pro HSF really?
Boosting ATI Framerates with CCC (X700 on)
Optimise ATI Image Quality And Framerates With ATT (X1XXX Series under WinXP)
|
|
|
|
| | |