1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Blur image on Leadtek A400TDH GeForce 6800

Discussion in 'Videocards - NVIDIA GeForce' started by Jagdtiger, Dec 29, 2005.

  1. Jagdtiger

    Jagdtiger Member

    Messages:
    24
    Likes Received:
    0
    GPU:
    Leadtek A400TDH GeForce 6800 128MB
    Hi
    I have been trying to solve the problem for over a month.
    What happens is that in every game I have blur image I don't know if that should exist. I will take two examples:

    First in Rome Total War I had very low performance around 15 only when zooming in on units. I solved that problem and now I have around 30 fps (sometimes 50).
    The problem is when units move and I look at them by side I get blur image the closest units have more blur than those away from the screen ( image looks like you take your palm and move it very fast in front of you up and down). Blur image is higher if more units are on battlefield! (Very similar happenes with Earth 2160)

    Second is America's Army. I take a training and come close to sergeant and move mouse left and right quite slowly and when doing it if I look at his face for example I see it blur( example is that I see around 10 eyes around the real eyes just like flickering very fast around the actual position of eyes, this is not only eyes BUT the whole head or body and every object!!!!)

    I tried many drivers for Motherboard, graphic card, tried hundreds of detail levels in games or preferences turning off AA, AF, when I set everything at low details I get less blur but it still exists.

    And all other shows same results (some less some more). I need help please
    I DID NOT NOTICE THAT ON MY ATI RADEON 9700 (I sold it)
     
  2. Jagdtiger

    Jagdtiger Member

    Messages:
    24
    Likes Received:
    0
    GPU:
    Leadtek A400TDH GeForce 6800 128MB
    I have discovered this:
    Look at the picture on this link this happenes on all games when bought GF 6800

    http://www.beyond3d.com/articles/tbuffer/index1_4.php

    T-Buffer technology tries to solve the problem I described above. Instead of having a single isolated perfect snapshot in time, we are trying to recreate the summation over time created automatically by a Video Camera. If we can simulate this summation over time we should see an increased quality of the animations generated by the 3D card. Now, how can we generate such a summation over time in single frame?

    The easiest way to do this is by simply combining several rendered images into one single frame. Now, at this point, we have to decide what we want: more true frames per second (being true perfect snapshots shown) or a combination of several images (sub frames) in one single frame. To help us decide we have to known that the human eye and brain combination is capable of identifying at least 30 individual frames per second. This means that if you are shown 30 frames/sec you can identify them as being separate images (there are special tests done to come to this conclusion. For example, you are shown dot patterns at a rate of 30fps and at the end you have to say whether a certain pattern was present in the series or not. if you move above 30fps you see several of those patterns as one making it impossible to identify whether a pattern was present or not, while at frame rates of 30fps and below you can identify the individual image contents and thus give the correct answer). What happens is that if you increase the frame rate the brain starts to see multiple frames as one single frame. So, what happens is that our eyes and brain add something similar to motion blur (real world sample: move your hand quickly left to right in front of your face notice how the hand looks "smeared" and "blurry"). The problem is that different people get this effect at different frame rates. While, for some people, this brain-induced "motion blur"-effect kicks in at 30fps it might kick in at 70fps for others. So, instead of relying on the brain to combine images and create motion blur, we could use the computer to combine images and thus create "fake" motion blur. What T-Buffer technology actually does is it takes several images rendered and combines them into one single frame. By doing this, the brain gets confused, and it suddenly sees a motion blurred image. The brain links this effect with speed and the feeling of realism is increased dramatically. So, it's very possible that an animation at 30fps using motion blur looks much more realistic than a perfect, still frame animation at 60fps simply due to the effect that the brain isn't fooled so easily to add motion blur at 60fps.

    IS IT POSSIBLE TO AVOID THIS MOTION BLUR?
     

Share This Page