• Advertisement
Sign in to follow this  

OpenGL Quadbuffered stereo in D3D10 ?

This topic is 3487 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hey there, My first post ever in this forum :) Well, as some might know, I'm coming from the evil side (OpenGL), but I'm currently converting a GL based engine to D3D10, mainly due to ATI/AMDs incompetence of delivering functional OpenGL drivers... The conversion was surprisingly painless, and works quite flawlessly. For one exception: quad buffered stereo support. I'll give a quick explanation for people unfamiliar with the term. The idea is to have two separate framebuffers, both double-buffered, where each one will be displayed on a separate DVI (or double-DVI) output of your graphics card. While rendering your frame, you may specify to which framebuffer the output should go to (typically called left and right). The two or four DVI outputs are then usually connected to a polarization based stereoscopic display or projection system. Both outputs display an excact 1:1 copy of your desktop, except for the quadbuffer-enabled 3D windows, where the left or right framebuffers will be displayed respectively. So essentially, does anyone know of a way to get such quad buffered stereo in D3D10 ? Has native support been introduced since version 10 ? Of course, there is a way to fake it (ab)using dual screens. However, this will only work in fullscreen mode, and with a custom (software) mousepointer. Unfortunately, being an industrial visualization application rather than a game, this is not an option for our software. Any ideas from the D3D pros ?

Share this post


Link to post
Share on other sites
Advertisement
I'm wondering, is this functionality widely supported in OpenGL? From your description it sounds like there's a very limited context in which it is useful.

Direct3D doesn't support stereo natively, although you can create two different view chains for two different windows and display into them. It still won't duplicate all the other windows automatically (as I said, this sounds like functionality that's not useful in consumer space).

Share this post


Link to post
Share on other sites
This seems like something that would most likely be implemented as two adapters, kind of like a dual-head video card. You'd effectively have two devices that you use, one for each of the adapters.

This is, of course, just a guess, since I have absolutely no experience with these kinds of things.

Surely makers of these types of hardware devices document how the system works?

Share this post


Link to post
Share on other sites
Quote:
Original post by ET3D
I'm wondering, is this functionality widely supported in OpenGL? From your description it sounds like there's a very limited context in which it is useful.

It's part of core OpenGL. Hardware support is available on most highend cards from NVidia (Quadro FX line). It's a very commonly used feature in highend CAD systems. These all use OpenGL, but they don't use DX10 style features (which is our main motivation for the D3D port). We'd like to avoid maintaining two different rendering backends.

Quote:

Surely makers of these types of hardware devices document how the system works?

Well, as I mentioned, this feature is supported by all highend NVidia cards (in fact, every dual-DVI Geforce could technically do it, but the feature is not exposed through the driver due to marketing reasons). It's easy to access through OpenGL, but I was wondering if there was a way to access it (or fake it reasonably well) through D3D10.

Share this post


Link to post
Share on other sites
Honestly, I'd suggest contacting NVIDIA developer support and seeing what they have to say about it. There may be some largely unknown hack to enable it in D3D 10, or it might just not be exposed at all there.

Share this post


Link to post
Share on other sites
Quote:
Original post by Promit
Honestly, I'd suggest contacting NVIDIA developer support and seeing what they have to say about it. There may be some largely unknown hack to enable it in D3D 10, or it might just not be exposed at all there.

That's probably what I'm going to do, but getting a technical response from NVidia within a reasonable delay is, uhm, a challenge... [wink]

I'm not yet familiar with all details of D3D10, as you might imagine, it's only been a couple of weeks I jumped into it. So I was wondering if there was a standard way of doing it that I somehow missed.

Share this post


Link to post
Share on other sites
There is no support for this out of the box. I expected that not enough (if anyone at all) have asked Microsoft to include this feature.

From my own experience you are the second person who asked for this in the last 5 or more years.

Share this post


Link to post
Share on other sites
There was some sort of stereoscopic viewing option in the DX 7.1a SDK.
But I wasn't interested in it. Besides, I don't even know if what they claimed was actually true stereoscopic viewing. (you'll see... I've never seen it!)
If it's worth mentioning, they used little DX specific content

Try getting a copy of DX 7.1a (samples) and see how they achieved it, and if it can be done in DX 10..... it's a long shot, but better than nothing.

Good luck
Dark Sylinc

Edit: Dont waste your time (more official DX 9.0 though)
Seems you can't do it with a special driver

[Edited by - Matias Goldberg on July 6, 2008 3:16:27 PM]

Share this post


Link to post
Share on other sites
Quote:

Direct3D 9 does not support stereo view, so Direct3D does not use the D3DBACKBUFFER_TYPE_LEFT and D3DBACKBUFFER_TYPE_RIGHT values of this enumerated type.

Damn, that's exactly what I would've needed. I guess this hasn't changed with D3D10 ? That's a shame, because it would've been pretty easy to expose.

Matias: I know about these special stereo drivers, but even if they worked on DX10, they're far too primitive to be useful for us. They're mostly designed to make typical 3D games stereo-like on shutter glasses. We're doing things like IR head tracking and adaptive stereo frustum convergence, so we need full and direct access to the two separate backbuffers.

Oh well, thanks for your help guys. I guess I'll see with NV devrel if they have some kind of undocumented feature somewhere. If not, we'll have to maintain two different rendering backends. Ugh.

Share this post


Link to post
Share on other sites
I've been working on something similar for the past month or so, we have our opengl wrapper for offaxis projection and stereoscopic view but couldn't do the same in directx9 since there is no quadbuffer.

I found a way to do it, it's not optimal yet but at least it's partially working for now. Basically you have to simulate the quadbuffer.

Our application is a directx wrapper since we're using UT3 for our demo. So we only have access to the surface once it's done rendering. The basics is to have 2 devices in 2 different threads, one for each adapter. You intercept the Present call, copy the backbuffer into memory and then get it with your thread. The application act as a feeder and the thread is the renderer. We had to be at 96Hz so when the application isn't capable of feeding the frame fast enough we're reusing the last frame and present it.

So the basics:

- Copy Backbuffer into a surface in the memory
- memcpy that surface into another owned by the second device
- Copy it to the Backbuffer of the second device
- Present

The thread is on the first adapter and the application is on the second, application doesn't present at all, only the thread does.

There is some little tricks to it too. Locking and unlocking surface is taking too long for an application like ours, so I Lock and unlock the mem surface only at the initialization so I can get the rect pointer then use it in the memcpy.

It's not optimal yet, I got to stabilize it because it's losing sync depending on the app framerate but I can get to a constant 96hz with that method. Using directx10 or directx9ex could be better too since you have access to shared surfaces.

Share this post


Link to post
Share on other sites
might work on a single screen nontracked setup, but for head tracking and offaxis view you have to customize the driver. And since nvidia doesn't release the source for the driver you have to code your own.

Share this post


Link to post
Share on other sites
Eti307, sounds interesting, but I assume this will only work in fullscreen mode ? Or do you somehow manage to duplicate the desktop and mousepointer on both devices ?

Namethatnobodyelsetook, this is the stereo driver I talked about earlier. This is a very different type of stereo ('home user stereo' as they call it) than the one used in professional systems. We cannot use this driver.

Share this post


Link to post
Share on other sites
works on both fullscreen and windowed mode, most of my tests were in windowed mode actually. I have no problem duplicating the mouse since it's managed by the game engine. The only thing needed is the backbuffer just before presenting it so the mouse position will be the one were it was drawn at that frame (if the game window has focus).

I only duplicate the game window though, not the desktop. So since I'm doing my own stereo only the game window will be flipping the right and left eye. So you'll basically see a 3D stereoscopic window and a standard 2d desktop. Have to make sure you use vsync though or the 3d window and the desktop won't have the same refresh rate and you'll lose the 3D effect

But that method is used for an active stereoscopic setup (1 projector with shutter glasses), is yours passive? (2 projectors for displaying the right/left eye)

Share this post


Link to post
Share on other sites
Depending on how adventures you fell you can “hack” the Vista DWN to do what you need. But this would require some work and experience in Direct3D hooking.

Share this post


Link to post
Share on other sites
Quote:
Original post by Eti307
works on both fullscreen and windowed mode, most of my tests were in windowed mode actually. I have no problem duplicating the mouse since it's managed by the game engine. The only thing needed is the backbuffer just before presenting it so the mouse position will be the one were it was drawn at that frame (if the game window has focus).

Ah, I see. We use the standard hardware cursor, but there might be some way to simulate that.

Quote:
Original post by Eti307
But that method is used for an active stereoscopic setup (1 projector with shutter glasses), is yours passive? (2 projectors for displaying the right/left eye)

Yep, mine is passive. Which makes it even more problematic, since I have to direct the left and right outputs onto separate DVI connectors. Well, actually I have to direct each eye onto a pair of DVI outputs, to be precise (but that's technically the same). Graphics cards are two Quadro FX 5600 SLI.

That's also why I have to duplicate the desktop (and all other running application windows, ugh) onto both channels, because otherwise they will only be visible on one eye. Same problem with the mouse pointer. If it is not duplicated, then it will only be visible on one single eye, which is very unpleasant and can generate headaches and nausea.

I also thought about the clone mode, where the graphics card auto-duplicates the framebuffer onto two separate outputs. But I would have to display different 3D backbuffers on each of the cloned displays, one for the left and one for the right eye. I don't think this is possible without modifying the driver.

This is all extremely straightforward with native quad buffer support. I don't understand why it wasn't included in D3D10, especially since the enumerants were already defined (but unused) in D3D9.

Share this post


Link to post
Share on other sites
Quote:
Original post by Demirug
Depending on how adventures you fell you can “hack” the Vista DWN to do what you need. But this would require some work and experience in Direct3D hooking.

Ah, I'm interested. Do you happen to have more information on that you might be able to share ?

Share this post


Link to post
Share on other sites
ah I see, passive is easier for fullscreen stereo (you don't have to worry about frame rate and synchronization with the shutters with a passive setup, lucky you :))but yes it'll be complicated if you want to duplicate everything in windowed mode. One thing that you could try is to be in clone mode and hook your direct3d application. you then create 2 devices, one on each adapter and present them with their respective frames. I don't know if it will work (no idea if you have access to the second adapter in clone mode)

If you want to look at how to hook directx, there's a guide on this site:
http://www.gamedev.net/community/forums/topic.asp?topic_id=359794
I'm doing similar to that but I'm overriding the d3d9.dll so it's directly injected each time I start the application (and no need for a launcher)

Share this post


Link to post
Share on other sites
Quote:
Original post by Yann L
Quote:
Original post by Demirug
Depending on how adventures you fell you can “hack” the Vista DWN to do what you need. But this would require some work and experience in Direct3D hooking.

Ah, I'm interested. Do you happen to have more information on that you might be able to share ?


The Vista DWM (Desktop Window Manager) is a Direct3D 9Ex application. It consumes all application windows and 3D back buffers (D3D and OpenGL) as textures. That was one of the reasons why 9Ex supports shared textures.

I haven’t tried this yet by my own but it should work.
- Make sure that both displays run the same resolution.
- Create your Application on the primary screen and the 3D window inside your application Window.
- Create a second window on the second display.
- Create an additional swap chain (same device) for this window.
- Hook the DWM Direct3D present method
- Inside your hook (before calling the original function) copy the back buffer from display one to display two. But only the array that is not covered by your second 3d Window.

There would be some possible problems with this setup depending on how your costumer use your application. As example if the run other windows applications at the same time.

Share this post


Link to post
Share on other sites
Check this out for hacking the DWM:

http://siwu.info/66/hacking-into-vistas-desktop-window-manager-dwm.html

I also hooked the entire d3d9.dll for a display wall system:

www.cpsc.ucalgary.ca/~pennere/projects

Eric Penner

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement
  • Advertisement
  • Popular Tags

  • Advertisement
  • Popular Now

  • Similar Content

    • By LifeArtist
      Good Evening,
      I want to make a 2D game which involves displaying some debug information. Especially for collision, enemy sights and so on ...
      First of I was thinking about all those shapes which I need will need for debugging purposes: circles, rectangles, lines, polygons.
      I am really stucked right now because of the fundamental question:
      Where do I store my vertices positions for each line (object)? Currently I am not using a model matrix because I am using orthographic projection and set the final position within the VBO. That means that if I add a new line I would have to expand the "points" array and re-upload (recall glBufferData) it every time. The other method would be to use a model matrix and a fixed vbo for a line but it would be also messy to exactly create a line from (0,0) to (100,20) calculating the rotation and scale to make it fit.
      If I proceed with option 1 "updating the array each frame" I was thinking of having 4 draw calls every frame for the lines vao, polygons vao and so on. 
      In addition to that I am planning to use some sort of ECS based architecture. So the other question would be:
      Should I treat those debug objects as entities/components?
      For me it would make sense to treat them as entities but that's creates a new issue with the previous array approach because it would have for example a transform and render component. A special render component for debug objects (no texture etc) ... For me the transform component is also just a matrix but how would I then define a line?
      Treating them as components would'nt be a good idea in my eyes because then I would always need an entity. Well entity is just an id !? So maybe its a component?
      Regards,
      LifeArtist
    • By QQemka
      Hello. I am coding a small thingy in my spare time. All i want to achieve is to load a heightmap (as the lowest possible walking terrain), some static meshes (elements of the environment) and a dynamic character (meaning i can move, collide with heightmap/static meshes and hold a varying item in a hand ). Got a bunch of questions, or rather problems i can't find solution to myself. Nearly all are deal with graphics/gpu, not the coding part. My c++ is on high enough level.
      Let's go:
      Heightmap - i obviously want it to be textured, size is hardcoded to 256x256 squares. I can't have one huge texture stretched over entire terrain cause every pixel would be enormous. Thats why i decided to use 2 specified textures. First will be a tileset consisting of 16 square tiles (u v range from 0 to 0.25 for first tile and so on) and second a 256x256 buffer with 0-15 value representing index of the tile from tileset for every heigtmap square. Problem is, how do i blend the edges nicely and make some computationally cheap changes so its not obvious there are only 16 tiles? Is it possible to generate such terrain with some existing program?
      Collisions - i want to use bounding sphere and aabb. But should i store them for a model or entity instance? Meaning i have 20 same trees spawned using the same tree model, but every entity got its own transformation (position, scale etc). Storing collision component per instance grats faster access + is precalculated and transformed (takes additional memory, but who cares?), so i stick with this, right? What should i do if object is dynamically rotated? The aabb is no longer aligned and calculating per vertex min/max everytime object rotates/scales is pretty expensive, right?
      Drawing aabb - problem similar to above (storing aabb data per instance or model). This time in my opinion per model is enough since every instance also does not have own vertex buffer but uses the shared one (so 20 trees share reference to one tree model). So rendering aabb is about taking the model's aabb, transforming with instance matrix and voila. What about aabb vertex buffer (this is more of a cosmetic question, just curious, bumped onto it in time of writing this). Is it better to make it as 8 points and index buffer (12 lines), or only 2 vertices with min/max x/y/z and having the shaders dynamically generate 6 other vertices and draw the box? Or maybe there should be just ONE 1x1x1 cube box template moved/scaled per entity?
      What if one model got a diffuse texture and a normal map, and other has only diffuse? Should i pass some bool flag to shader with that info, or just assume that my game supports only diffuse maps without fancy stuff?
      There were several more but i forgot/solved them at time of writing
      Thanks in advance
    • By RenanRR
      Hi All,
      I'm reading the tutorials from learnOpengl site (nice site) and I'm having a question on the camera (https://learnopengl.com/Getting-started/Camera).
      I always saw the camera being manipulated with the lookat, but in tutorial I saw the camera being changed through the MVP arrays, which do not seem to be camera, but rather the scene that changes:
      Vertex Shader:
      #version 330 core layout (location = 0) in vec3 aPos; layout (location = 1) in vec2 aTexCoord; out vec2 TexCoord; uniform mat4 model; uniform mat4 view; uniform mat4 projection; void main() { gl_Position = projection * view * model * vec4(aPos, 1.0f); TexCoord = vec2(aTexCoord.x, aTexCoord.y); } then, the matrix manipulated:
      ..... glm::mat4 projection = glm::perspective(glm::radians(fov), (float)SCR_WIDTH / (float)SCR_HEIGHT, 0.1f, 100.0f); ourShader.setMat4("projection", projection); .... glm::mat4 view = glm::lookAt(cameraPos, cameraPos + cameraFront, cameraUp); ourShader.setMat4("view", view); .... model = glm::rotate(model, glm::radians(angle), glm::vec3(1.0f, 0.3f, 0.5f)); ourShader.setMat4("model", model);  
      So, some doubts:
      - Why use it like that?
      - Is it okay to manipulate the camera that way?
      -in this way, are not the vertex's positions that changes instead of the camera?
      - I need to pass MVP to all shaders of object in my scenes ?
       
      What it seems, is that the camera stands still and the scenery that changes...
      it's right?
       
       
      Thank you
       
    • By dpadam450
      Sampling a floating point texture where the alpha channel holds 4-bytes of packed data into the float. I don't know how to cast the raw memory to treat it as an integer so I can perform bit-shifting operations.

      int rgbValue = int(textureSample.w);//4 bytes of data packed as color
      // algorithm might not be correct and endianness might need switching.
      vec3 extractedData = vec3(  rgbValue & 0xFF000000,  (rgbValue << 8) & 0xFF000000, (rgbValue << 16) & 0xFF000000);
      extractedData /= 255.0f;
    • By Devashish Khandelwal
      While writing a simple renderer using OpenGL, I faced an issue with the glGetUniformLocation function. For some reason, the location is coming to be -1.
      Anyone has any idea .. what should I do?
  • Advertisement