Articles: Graphics
 

Bookmark and Share

(1) 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 ]

Instantaneous Performance and Image Quality

Our tests have so far showed you that the game doesn’t like Nvidia’s hardware. To pinpoint the bottleneck we performed another test. You will now see if you can lower the graphics quality settings to increase the frame rate at the expense of the game’s visuals. We used another test scene that contained an attack on a fully constructed UN base and a subsequent use of the secret weapon the game developers called Proton Collider for some reason. The name follows the overall Red Alert style well enough, though.

In this test we used a resolution of 1680x1050 pixels and added another graphics quality mode to the in-game ones. This mode was the same as Ultra High but had 4x rather than 2x MSAA. We recorded the instantaneous frame rate for 60 seconds to build comparative diagrams for two popular mainstream graphics cards, ATI Radeon HD 4850 and Nvidia GeForce 9800 GTX+.

Besides the instantaneous performance data we also used Fraps 2.9.6 to capture a series of screenshots to check out the visual difference between the graphics quality settings. We got the following data:

 

 

 

 

 

 

Oddly enough, it makes sense to use 2x MSAA instead of the popular 4x level in this game. The difference in image quality is insignificant – 4x MSAA is a little better at reproducing micro-geometry but you can hardly spot the improvement with a naked eye, especially during actual play – but the GeForce 9800 GTX+ gets much closer to the Radeon HD 4850. Its average frame rate grows up by 56% and becomes nearly comfortable. We don’t mean that full-screen antialiasing is a bottleneck of Nvidia’s solutions because we don’t see them being so slower than ATI’s GPUs in other games. However, we can presume that the game engine just doesn’t match the raster back-end architecture of Nvidia’s GPUs well enough.

The transition to the High profile doesn’t affect the overall image quality much. Small details suffer from the lack of antialiasing, but everything else looks just as good as in the Ultra High mode. At least, we wouldn’t say that the game looks far worse then. As for the speed factor, the Radeon HD 4850 and GeForce 9800 GTX+ now match each other’s speed, confirming our supposition that the Red Alert 3 engine does not run well on Nvidia’s GPUs when you turn antialiasing on.

The Medium profile provokes a terrible degradation in image quality. A number of small but conspicuous elements such as stairs or stones in the pavements just disappear. The depth of water is largely lost and the sea doesn’t merge into the land as smoothly as in the previous modes. Many shadows are not displayed. The lighting model is simplified, too. As a result, playing in this mode is not such a pleasure for your eyes. The game doesn’t look pretty anymore. And this sacrifice is not justified in terms of speed. The graph only shows the lack of a slump at the moment the Proton Collider is used.

The Low and Very Low graphics quality modes are no good at all. The game looks ugly in these modes as you can see in the screenshots. Perhaps you’ll be able to achieve a playable speed at these settings on an entry-level card or even an integrated graphics core, but do you really want to play a strategy game with visuals like those of games from the year 2000?

Thus, we guess the High profile is the optimal choice in this game. It is but little worse visually than Ultra High but the game runs equally fast on the ATI Radeon HD 4850 and Nvidia GeForce 9800 GTX+ in it.

If you’ve got a weaker graphics card, you can try the Medium profile. You’ll lose a large share of the visual appeal of the game, though. The Low and Very Low profiles provide such a low level of detail that they are only going to be used by hardcore RTS fans who don’t care about the beauty of a game scene but need the highest frame rate possible to have more chances to win through smoother control over the units. Well, the game has a frame rate limiter, so it cannot be accelerated infinitely.

 
Pages: [ 1 | 2 | 3 | 4 | 5 | 6 | 7 ]

Discussion

Comments currently: 1
Discussion started: 12/27/08 09:33:18 PM
Latest comment: 12/27/08 09:33:18 PM

View comments

Add your Comment