Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Eyefinity triple Monitor issue

GrubbsGradyGrubbsGrady Member UncommonPosts: 371

I just bought three monitors and set them up, initially they were all fine and once I turned on eyefinity the one monitor is raised. What I mean by that is the taskbar runs even along the bottom of the monitors, but the third one is about an inch above the bottom.

It is basically thinking my monitor starts from a spot higher than it does, and the mouse will actually go off screen to the top as if there is still more space up there. What should I do to make the screen actually fit the third monitor properly?


  • RidelynnRidelynn Member EpicPosts: 7,060

    I'm assuming all 3 monitors are identical?

    What happens if you plug them in a different order? Does the same monitor keep the high spot, or does it stay with a specific cable or port on the video card? What happens if you tell Windows to use a different frequency and/or resolution (just for testing, not that you want to keep it that way)?

    Most monitors have an "auto configure" option that tries to set the picture - it may be that is getting messed up and you need to play with menu options.

  • QuizzicalQuizzical Member LegendaryPosts: 22,096
    Are all of the monitors exactly the same resolution? If not, then setting a taskbar some number of pixels from the top could easily cause that. Also, in Catalyst Control Center, you can go to Desktop Management -> Arrange Desktops and see how the video drivers think your monitors are arranged.
  • steelwindsteelwind Member UncommonPosts: 352
    The Eyefinity wizard within the Catalyst control panel should resolve this issue. Also adjusting the bezel will also help, also found in the Eyefinity settings. It can even adjust for different size monitors.

    Eyefinity is not like simple dual monitor and must be setup after monitors are plugged in.
  • GrubbsGradyGrubbsGrady Member UncommonPosts: 371
    I got it, for some reason I had to manually change the settings on the third monitor. I searched through some menus and was able to shift it downward (the option was disabled on the other two monitors). They are all three the same model so I don't really know what the deal is.

    Either way, its working now. Trying to get used to the side monitors looking a bit stretched in games...not sure if I am gonna wanna keep the 3 monitor setup after all. Maybe two will be fine for productivity boost and I can return the other for a little bit of savings.
  • RidelynnRidelynn Member EpicPosts: 7,060

    My guess would be a different connection type?

    If 2 are hooked up DVI, and the third HDMI, or if they are all hooked up HDMI but the third one is using an different port type on the card (that would vary card to card), then it would be getting a slightly different signal, or rather, a signal that is in a different format than the other two, and that would cause it to behave differently.

  • QuizzicalQuizzical Member LegendaryPosts: 22,096
    Originally posted by GrubbsGrady
    Trying to get used to the side monitors looking a bit stretched in games.

    There's good news and bad news on that front.

    The good news is that the stretched stuff on the side monitors is completely fixable on a game by game basis.  That is, a particular game can take steps to fix it.  Better yet, a complete fix probably runs tens of lines of source code, or maybe hundreds if you count copy/pasting a block of 10 lines or so separately in every single geometry shader that you paste it into.  Speaking of which, you need geometry shaders to fix it, which means DirectX 10 or OpenGL 3.2 or later.

    The bad news is that fixing it requires doing explicit computations in real-projective space RP^3.  The reason that's bad news is that you likely know more about real-projective space than the median person who does computer game graphics, simply on the basis that, after reading this post, you've heard of it.

    The computations aren't hard if you have the appropriate math background for it.  The problem is that a degree from a game design college or even a BS (or probably, for that matter, a PhD) in Computer Science from a prestigious university is far insufficient.  Trying to parse the meaning of a quotient topological space isn't likely to end well if you've never seen a topological space at all, nor a quotient of anything under an equivalence relation.  For the latter, you probably want to start with quotient groups, which means you need groups, and that means abstract algebra, aka, the point at which higher mathematics becomes really bizarre and people don't think it should qualify as "math" anymore.

    So the fix is for a game studio to hand the problem to their in-house mathematician, or probably physicist.  I didn't see the relevant mathematics until grad school, so an undergraduate degree likely isn't enough, though a fairly advanced undergrad with a geometry bent may have the appropriate background.  And if you don't happen to have one of those people on staff, or even if you do have one and don't know that this is a problem that you need to hand to them, then the game isn't going to fix the problem.

    Now, the way modern 3D graphics is done does do computations in RP^3 already.  But every tutorial I've found basically says, here are the formulas you should use, and while it is technically legal to use something else, you'd better really, really know what you're doing (i.e., much more than whoever wrote the tutorial knows) or else it's going to be completely broken.

    So the real bad news is that, while it's completely fixable, the odds that fixes for it will ever become common are basically nil in the foreseeable future.  Any game that doesn't make the field of view user-adjustable (which is a partial fix for the same problem, by the way) has no real hope of addressing the stretched side screens, as a user-adjustable field of view is at least an order of magnitude easier.  Which is not to say that fixing stretched side screens is hard; one could argue that the hardest part of making the field of view user-adjustable is making a menu that lets the user adjust it.

    So why are the side screens stretched, anyway?  It's a straightforward result of making the game perspective-correct for you under two particular assumptions that are almost invariably false.  If you really wanted to, you could alter your monitor setup to make the game perspective-correct--but when I explain what you have to do, you'll see why basically no one ever does.

    The first assumption is that all of your monitors have a screen in the same plane.  The problem with this is that basically no one ever does this.  People who have multiple monitors nearly always angle them in, so that their line of sight toward the side monitors is a little closer to orthogonal to the screen.  That, after all, gives you a better view of the side monitors--and makes them look vastly better for anything using an isometric projection--which includes everything 2D, such as word processors, e-mail, most web pages, and so forth.

    The second assumption is that you'll be some fixed distance from the center of the monitor--and this distance tends to be very close.  For a 90 degree field of view on a 24", 1080p monitor, the assumption is that your eyes are a little over 10" from the screen.  If you want a 120 degree field of view, change that to about 6" from the screen.  The field of view to make a game perspective correct for typical distances at which people sit from monitors would make many games very awkward for basically everyone using a single monitor--which is to say, nearly everyone who plays the game, period.

    So games use a wider field of view, but not so wide that you get bad stretching on a single monitor.  Make it three monitors, though, and the stretching on the side monitors becomes very obvious.  Angling the monitors toward you makes it even worse.  A user-adjustable field of view would narrow this enough to make the stretching not as bad--and potentially not really much worse than on a single monitor.

  • GrubbsGradyGrubbsGrady Member UncommonPosts: 371
    Thanks a ton for that explanation Quiz, rarely do you find people who are willing to go to that depth of detail to explain something to a total stranger. Makes sens to me now. I think sticking with two monitors will be the best thing for me- maybe down the road when there is more support for it, or FoV changing is common in all games I will grab another monitor. For now I'll enjoy the upgrade from one to two :D
  • QuizzicalQuizzical Member LegendaryPosts: 22,096
    Originally posted by GrubbsGrady
    Thanks a ton for that explanation Quiz, rarely do you find people who are willing to go to that depth of detail to explain something to a total stranger. Makes sens to me now. I think sticking with two monitors will be the best thing for me- maybe down the road when there is more support for it, or FoV changing is common in all games I will grab another monitor. For now I'll enjoy the upgrade from one to two :D

    If making the field of view user-adjustable were ever going to become common, it probably would have happened a long time ago.  The relevant graphics APIs are DirectX 8.0 or later, OpenGL 2.0 or later, OpenGL ES 2.0 or later, WebGL, and probably others.  It was likely possible before that, even, though I'm not 100% sure what you could do with the old fixed-function pipeline that didn't offer any programmable shaders.

    And not just possible, but easy.  I've made the field of view adjustable in a project I'm working on, and all it takes is taking a constant making it into a variable that gets referenced in exactly two places (two consecutive lines of code, even!) rather than hard-coding it in as a constant.  It would also take setting up a menu to let the user change that variable, which I haven't done yet, but for now, I can adjust it by changing the value that the variable is initialized to.

    Now, to completely fixed the side monitor stretching issue, you do need geometry shaders (well, either that or a very inefficient work-around to avoid using geometry shaders) and advanced mathematics.  But changing the field of view is easy.

Sign In or Register to comment.