Stereoscopic support

[dongle]

ZDoom is my favorite DooM engine and I'd love to see it support stereoscopy, I guess with support for Nvidia 3DVision (/AMD's HD3D, though I am less clear on how that works and have no test hardware).

I will contribute to the development of this feature if someone familiar with the ZDoom codebase helps out.

[Xtyfe]

Seems like it would a lot of work for nothing but a silly gimmick

[phi108]

I would like to see it as a feature too, and besides 3dvision, to support side-by-side crosseye and parallel views too, maybe anaglyph. I can understand the obstacles to adding it though, and Hud/weapons would have to be either not drawn or offset to make them appear closer than the vanishing point. If possible, user options for eye separation distance, angle difference of eyes, and hud/weapon offset distance would be handy.

(EDIT: Also, 3D viewpoints can benefit from backing them away from the usual player viewpoint, maybe by as far from Doomguy's eyes as the player sits from their computer screen, so there is less distortion of objects right in your face. It may only be needed for side-by-side though)

And even if it doesn't become a feature, Isle did a ACS mod for Crosseye/Parallel Dooming, without Hud and weapons though:
http://forum.zdoom.org/viewtopic.php?f=4&t=12938

And I modified it some and actually learned ACS in the process, altering dimensions and decreasing separation, and also giving a peripheral view for each eye, so the crosshair is at the red mark, for less eye strain to combine images.

[Nash]

i'd like to see it done too someday (support for nvidia's 3d vision)

[Xtyfe]

Meh, I still see this as one of types of features comparable to split screens

[Blzut3]

I guess I should throw in this link to my, SDL only, huge hack for anaglyph 3D: http://forum.zdoom.org/viewtopic.php?t= ... 50#p458350

I'm not entirely certain, but ZDoom probably isn't capable of utilizing the hardware 3D support since it's a software rendered game. Unless of course it comes down to just rendering to a frame packed canvas.

[dongle]

@phi108 - I'd like to use the hooks for 3DVision since that also supports anaglyph and it 'just works' with a number of different devices (including generic 3D DLP projectors and 3DTVs).

@Bizut3 - thx for the link. yeah, it'd have to be a framepacked hack (zdoom only uses d3d for blitting, iirc).

Seems like it would a lot of work for nothing but a silly gimmick

'silly gimmick' seems a bit harshly worded. Plenty of people use 3DVision or the 3D output on their PS3/360 consoles. It's a minority of users, yes, but I don't think 3D is going away any time soon.

also, dude, even if it's a lot of work (it's not /that/ big of a deal; i do 3d programming at my day job), I'm offering to do the work with assistance from zdoom maintainers (mostly, I want to ensure I'm doing it in the smartest way possible and in a style that they will be happy to merge).

[Enjay]

'silly gimmick' seems a bit harshly worded. Plenty of people use 3DVision or the 3D output on their PS3/360 consoles. It's a minority of users, yes, but I don't think 3D is going away any time soon.

Silly gimmick isn't harsh enough IMO. I really hope its a fad - just like it has been in the past on numerous occasions. I guess the difference this time is that there is the technology in homes for it too. However:

• The picture isn't as clear - even on the best systems.
• It forces your eyes to diverge to a position they do not naturally go to during normal viewing
• It can cause headaches (mainly as a result of the above I believe)
• People with even relatively slight eye defects are often unable to see it properly (eg my son, my father)
• It's not true 3D but looks more like various flat images stacked at different distances (ie like the effect with a "viewmaster" toy or a victorian stereoscope).
• If you spot it happening, then it defeats it's own purpose because instead of marvelling at the scene, you are marvelling at the 3D effect instead
• If you don't spot it, what's the point?

Also, it is a feature that suits the people producing the stuff rather than those viewing it. For the reasons above, and more, it's not a great experience but it is one that gets more people going to cinemas - experience has shown that this effect only tends to last for a relatively short period of time - and, nowadays, spending money on hardware.

Personally, if I can, if there is a 3D movie in the cinema, I try and see the 2D presentation instead because I prefer it and, personally, I think the experience is much better.

Would I use it in ZDoom? I'd maybe to set it to use it once or twice just to see how it looked and then almost certainly never touch it again. Gimmick.

[phi108]

[*]It's not true 3D but looks more like various flat images stacked at different distances (ie like the effect with a "viewmaster" toy or a victorian stereoscope).

I've seen this problem in specific scenes converted from 2D to 3D with "simple" cutout editing; real 3D image captures are from 2 real/virtual cameras, so objects in the scene have depth to them instead of being flat cutouts. Models in modern games give this advantage, while doom sprites give the effect you describe, though if eye separation is low, it isn't horrible. (And Doom maps have plenty of architecture to benefit, if the sprites are lackluster)

[Enjay]

I've seen it in the most up to date made-for-3D movies. It's not just an old-game-with-flat-sprites thing.

Another thing I hate about 3D is when something appears quickly in the foreground when previously you have been looking at something further away. There is a moment where some of the things you are looking at are doubled and then you can actually feel your eyes moving to their new forced position.

[CaptainToenail]

I've only watched one 3D film at the cinema (Resident Evil). It gave me a serious splitting headache, and I had to wear those silly 3D glasses on top of my normal pair which was stupid and uncomfortable. And the pain in my head, it was awful.

[printz]

[*]It forces your eyes to diverge to a position they do not naturally go to during normal viewing

Indeed. If I'm not supposed to wear stereroscopic glasses, I'm not going to strain my eyes to unite two distanced apart images. I can't even do that and I find it unhealthy. And the result is nil anyway. If I want to see 3d, I look at the real world.

[randi]

I would like to know more: Is it actually possible to do a 3D image without using the full 3D pipeline?

[phi108]

This doesn't answer the question, but I found a guy on youtube who claimed to have Skulltag working with 3Dvision, though it seems broken. He said he uses a DirectX wrapper for Skulltag's GL mode, and then uses Nvidia's 3Dvision to render in 3D. The eye separation seems either too low or unseparated, so this may be a faulty method, maybe due to the wrapper. I'd prefer a 3D mode modifiable by the player, allowing user-defined separation and angle. Here's the wrapper he used: http://sourceforge.net/projects/gldirect/ and the video: http://www.youtube.com/watch?v=7dYi-tAN1_E

EDIT: There seems to be miniscule eye separation, but hardly noticeable, while the angle difference of the 2 cameras is the major difference.