anyone got any 3D glasses working with the cube?

[happy_bunny]

I was just wondering (in a future project kinda way) has anyway got any active 3D glasses working with the gamecube ? Has there been any dev work into this before? Now we have the digital port working I was thinking we can setup the gc video hardware to output a stereoscopic image (a static one at first) in EDTV . Then we just need to attach (and control) some active 3D glasses via the gamecubes joypad port to get the 3D effect working.

has anyone tried this before ? Descent video game in full 3D would be really cool

[novenary]

The active glasses need to be controlled by the TV. You'd need to output a real HDMI signal to enable 3D I think.

[tueidj]

If you actually got the hardware organized, I could easily patch Descent to do it (all the original left/right view code still exists for VR headset support).
There's also the possibility that the video port itself already has a left/right signal, since the video HW registers have a bunch of stuff related to supporting separate left+right images.

[happy_bunny]

If you actually got the hardware organized

I think there are two ways of doing this one would require hardware mods and hopefully the second way the hardware is in place.

First up

Get some cheap active 3D glasses and hack them to work with some hardware connected to the gamecube joypad port. It looks like the joypad port (libgc sub-system SI) gets passed the vsync and odd / even screen info I think it was planned to be used by a light gun. We / I would need to read these signals and then control the 3D glasses.

Second idea

Pump out the stereoscopic image over the digital port and get the 3D TV to control the glasses, not sure if a 3D TV can do this with DVI even if it can I would need to buy one lol

[tueidj]

I don't think the controller receives anything about the video signal, from memory it's just a basic bidirectional serial port. I think the best way would be a circuit that pretends to be a controller (returns the right ID, responds correctly to commands, returns fake stick/button values) and use the rumble command from the cube for controlling the glasses. It might seem like overkill when we know it's a serial port but this would be compatible with the wiiu usb GC adapter (it's limited to working with controllers only) and could also allow for more complicated VR headsets in the future, for example eye-tracking and head movement could be sent back as the controller data.

I'm not keen on the idea of a stereoscopic image, the GC already has low resolution and halving it horizontally would be horrible. Dropping to 30Hz (Descent easily manages 60Hz on GC, except for some micro-stuttering when playing MP3 music from SD Gecko which can be fixed) for separate full-res left/right renders would be preferable.

[novenary]

Isn't timing too tight to let the cube itself control the glasses ? You need to account for latency or the whole effect it ruined.

[tueidj]

No, the controller ports are fast enough to poll 6 times per frame or something so it's easily quick enough.

[novenary]

Oh I was talking about lag not polling speed.

[happy_bunny]

Big N have already done whats tueidj is talking about, under the development section of luigi mansion

http://en.wikipedia.org/wiki/Luigi%27s_Mansion

we have this little gem

All Nintendo GameCube systems support the display of stereoscopic 3D, and Luigi's Mansion was developed to utilise this.

and if you look into reference 24 on that wiki page you have

"Iwata: To go back a little further, the Nintendo GameCube12 system actually had 3D-compatible circuitry built in [...] Itoi: Nintendo GameCube did? And all the Nintendo GameCube systems around the world? Iwata: Yeah. If you fit it with a certain accessory, it could display 3D images. [...] Iwata: We already had a game for it, though - Luigi’s Mansion, simultaneously released with Nintendo GameCube. Itoi: The one in which Luigi shoulders a vacuum cleaner? Iwata: Yeah, that one. We had a functional version of that in 3D."

certain accessory... thats interesting, lets have a look at patent US6609977 for the key words 3D LCD and see what we get..... oh look under SIPOLL SI Poll Register we can control 3D LCD shutter glasses

Vblank copy output channel 0: Normally main
processor writes to the SIC0OUTBUF register are
copied immediately to the channel 0 output buffer if a
transfer is not currently in progress. When this bit is
asserted, main processor writes to channel 0's
SIC0OUTBUF will only be copied to the outbuffer
on vblank. This is used to control the timing of
commands to 3D LCD shutter glasses connected to the VI.

ok now lets look for LCD shutter in patent US6609977 we have

An LCD shutter can be supported through automatic polling and a serial control command

so putting two and two together I think that "certain accessory" was a controller for active 3D glasses connected to the gamecube joypad port. Sorry for the long post

[happy_bunny]

I'm not keen on the idea of a stereoscopic image, the GC already has low resolution and halving it horizontally would be horrible.

Dont understand the halving it horizontal ? What I mean by displaying a static steroscopic image is we put the GC in 480p and at 50 frames a minute the left eye would see 25 even frames and the right eye would see the other 25 odd frames. Thats how I thought it would work, would that actually work ?? (need to lookup how 3D glasses work )

[tueidj]

I didn't know there was a way to control the exact timing of the SI output, that would definitely be useful rather than manually trying to hit the retrace period.

When you said stereoscopic I thought you meant a side-by-side image (left+right in one frame) rather than interleaved left/right images.

[happy_bunny]

ok looks like we are going down the control some active 3D glasses ourselves route which I am fine with as I didn't really want to by a 3D TV .

The next question is which set of glasses to hack up ? I was think at first to have them wired, ie signals go from the joypad port through a uC which converts the joypad signals to something the glasses can understand. Then later once we have the stuff working with a small demo, make the glasses IR (remove the wires).

So anyone get any ideas on some cheap 3D glasses that we / I can hack up ?

[tueidj]

I dunno, I had a quick look at the bluetooth profile for controlling 3D glasses (since BT 3D glasses go for around $15/each on ebay) and it was ridiculously over-complicated.

[happy_bunny]

what about sony PS3 active 3D glasses

http://www.ebay.com/itm/NEW-Original-OE ... 1c532da08b

they are IR but for initial testing I will hardwire them and they are cheap as well

[happy_bunny]

found this

https://hackaday.io/project/580-peril-s ... er-glasses

will get two pairs of those sony glasses see if I can get something working

[Borg Number One]

http://retro-system.com/ActiveShutter3D.htm
Maybe (not) outdated. Who knows

[Papy.G]

ATI had wired 3D glasses back in the days, they were called ELSA, if i remember well. Maybe this will help.

Do we know if analog sync signal is still active when progressive video modes are activated, if it is, we could get the even/odd signal extracted off it by a LM1881 or something in the flavor.

Unfortunately, any search ends up flooded up by dolphin 3D/VR rendering of GameCube games.

What we know now:

There were two specific 3D rendering dedicated panels, the integrated mask, which didn't need any glasses (some displays, 3DS upper screen, Sony 3D cam display), and the polarized flavour, that needed appaired glasses, while not active. Those are marginal techs, and we don't even know if the GC's alternating pixels protocol will even be compatible.

There is one 3D rendering that relys on standard panel, and synchronized LCD shutter glasses, but there too, we don't know if it will support it.

So the way to go is standard TV with external sync.

IWATA san said Luidgi's Mansion had a mechanic that outputted the sync signal via the controller port, we know that beside beeing not very effective, it is very likely to only work with the Luidgi's Mansion 3D developpment version, and that it may interfere much with SWISS inner working. We know that we don't need to do it this way, as the sync signal can be got off the analog signal, or from GCVideo off the digital signal.

All we now need is:
A patch in SWISS to force the GC to render alternating left and right frame. If possible, and this is the key, just to begin with.

A sync stripper board to output, and delay, the sync signal.
Or
A new version of GCVideo that delays and outputs the sync signal on one of the CPLD's free gpio.

A pair of LCD shutter glasses compatible with the way the hardware outputs the sync signal.