Possible new 3D display method for single-chip DLP....

[THEDRUMMER]

Thanks Ozone for all the help!!!

[OzOnE2k10]

@THEDRUMMER: No problem. I've just realized something though - you need to make sure that the connector on your shutter glasses don't connect directly to the shutter panels themselves, as the circuit won't work for this. Most glasses have a small (or integrated?) controller to drive the LCD panels, the panels don't usually connect directly to the plug unless there's an extra controller for it.

eg. on the Elsa glasses, there is a small box (about an inch tall) which actually generates the correct voltages to drive the LCD panels. The signal on the 3-pin VESA connector only tells the small box which eye / shutter should be blanked.

Also, if your glasses do indeed have an integrated controller to drive the panels, you need to work out the pinout for the stereo jack. I'm assuming this could be the same pinout as shown elsewhere, but you need to be 100% certain to be sure not to damage something. I found this diagram on Google, but I don't know if it applies to your glasses or not?...

http://img265.imageshack.us/i/vesaconnections.jpg/" onclick="window.open(this.href);return false;

How were your older glasses (with the stereo jack) connected before? Did they have their own controller box or dongle?

btw, I've tried simulating the circuit on the PC, but the 74HC123 doesn't simulate properly (I'm sure it's only a software problem). I think the controller could be made to work as a delay for the VESA connector on 3D DLP TV's if the second half of the 74HC123 is used. The parts arrived today to build this, so I'll have a go later on.

OzOnE.

[THEDRUMMER]

OzOnE2k10: I have Asus shutterglasses. They used to plug directly into an 1/8" stereo jack on my Nvidia Card ... Asus V3800 TNT2 Ultra 32mb. There was no in-line box, just a wire straight to the shutterglasses. I sold that old computer this summer for $10 and regret it now. On your link: http://www.int03.co.uk/crema/hardware/stereo/" onclick="window.open(this.href);return false; they explained how their circuit wasn't made for 'dumb' glasses ... like mine I assume.
My current setup has 2 Nvidia 9800 GT cards in SLI. There is a S-Video 7pin Video out connector on the video card and that's all.
This just doesn't seem to be as easy as I was hoping!
I had seen that DIN3 to stereo picture you linked to and had assumed that that was all I needed to do to connect my Asus glasses to these controllers. On Stereo3d they used to have a guide for a controller for Asus type glasses like mine but every single link of theirs is dead! I may have to buy a new pair of shutterglasses if I ever want to get this working. I just didn't want to spend much on it because I didn't really buy my projector FOR 3D. I just thought it would be time well spent if I could get it to work somewhat decently, for a few bucks.
What I really am waiting for, and why I didn't just buy the Acer H3650 projector, is a decent Head Mounted Display .. Like TDVision 720p TDVisor (if they'd ever properly market it to the general public)! Hmmm, there's an emagin z800 on ebay right now for $995 US too! But that's another matter altogether.
Well, I'll keep trying, but this isn't easy. Should I buy some EDimensional Glasses and then buy that $9 vga pass through dongle I linked to show you? It just wouldn't have any adjustable delay ... but I should be able to see if it's going to work at that point, right? .. then maybe I'll buy a cheaper 120hz 3D projector in the future???? and I could use those glasses with that one.
Well, Good luck with your project progress.
TheDrummer

[OzOnE2k10]

OK, things have got slightly complicated again. btw, I have insomnia lately, so this is a big post!...

I was going to post the stuff below when I realized that the latest ATI drivers might have been released. There have been many reports in the last weeks that ATI will add S3D support to their next Catalyst driver. I'm sure that a few sites even mentioned that it will specifically be the 10.1 version. So, I checked on ATI/AMD and sure enough, the drivers have been released. I just installed them and...... NOTHING!! I can't find ANY new controls at all for S3D (not even anaglyph).

I know that they only said the drivers would be "more compatible" with IZ3D and the Bit Cauldron shutter glasses, and also they apparently added support for decoding the coming 3D Blu-Ray titles (using an updated PowerDVD?), but many people got the impression that they would include other S3D options in the drivers too (just like nVidia do)?? So unless there is a registry hack to enable some new S3D options, it's VERY disappointing as usual. I was hoping they might still have added some other ways of generating a stable L/R signal with the new drivers which could be taken advantage of with homebrew controllers (I'm pretty sure it's only nVidia who control the SDA pin to make connecting glasses easier.) Oh well, hopefully someone will hack the Zigbee stuff soon.

Do ATI/AMD not realize that there has always been massive interest in stereo 3D and I'm sure they lost a LARGE amount of market share in the past for not including S3D in the modern drivers! I only changed to ATI about two years ago because they apparently had better support for HD decoding at the time, and the drivers may be a bit more stable too. It was a real shame to loose the direct shutter control with nVidia. Then again, my glasses only worked well with "bench-breaking" CRT monitors, and I'm NOT going there again!

With IZ3D though, all we're really asking for is a simple way of getting a stable L/R sync signal out of the PC. Why is this so difficult?? I know it's not IZ3D's fault, but is it not possible to make a basic USB dongle for this and control it from the IZ3D API??

Anywho, rant over. I'll have to take a deeper look into these Catalyst drivers now. @THEDRUMMER: About your glasses...

Ahh, I see - it looks like the Asus glasses are wired directly to the jack plug, and they don't have an integrated panel driver. I just found a site which shows a basic circuit for connecting the Asus glasses to an SGI Workstation... The SGI uses the same Left / Right, +5V/0V, 50/50 signal that you find on a VESA connector, BUT - I didn't realize you have nVidia cards!!...

So all you will need is the opamp circuit from the site below, and connect the input of the circuit to the SDA pin on one of your nVidia cards. The nVidia drivers already support stereo 3D on almost all modern cards, and the drivers should output the L/R signal on the SDA pin (of one, or both of your VGA sockets, I'm not sure?). Here is the SGI site with the opamp circuit...

http://www.roosmcd.dds.nl/oldsite/" onclick="window.open(this.href);return false;

The Asus TNT2 card must have had a similar LCD panel driver on board, and this is most likely why they used the jack plug instead of a 3-pin plug? (The main difference is that the panels need a higher voltage to blank sufficiently, hence the extra circuit.)

I don't think you'll need the E-Dimensional dongle... I don't know a huge amount about them, but if it is just a simple pass-through for the SDA signal, then you will still need to find a 3-pin plug to connect to the opamp circuit. However, I think the E-Dim dongle needs to be activated with on-screen colour codes to set it to specific modes (sync doubling etc, which you probably don't need anyway). What I would do is buy a short VGA extension cable, carefully remove a small section of the insulation, find the wire which connects the SDA pin, then splice a wire into the SDA wire, and hook up to the opamp circuit.

OR, super-simple method - buy a pair of glasses which already have a splitter cable + shutter drive, and work with nVidia cards! The Elsa's are very cheap on eBay UK, no doubt they appear often in the US too? They are a cheap way of getting started if you just want to try S3D out.

You might have to do a bit of Web searching to see if the nVidia stereo mode works in SLI (or even if it outputs a sync signal from your cards) because I'm not 100% sure. You might only be able to use one card at a time in stereo mode - does anybody else know??

The simple opamp circuit won't have the variable delay of course, but it might work well anyway and is a much easier solution than the blue-line stuff. Things like blue-line decoding was only necessary for ATI cards, because ATI were short-sighted enough to not include S3D support as standard.

Same here about the projector situation - I've owned about six different projectors over the years and have upgraded each time I can afford to. I couldn't possibly afford one of the 120Hz projectors yet until the prices come down a lot. I've always loved the idea of S3D, so it's frustrating that there are so many "60Hz only" projectors / DLP TV's out there which are way too flickery for shutter glasses imho. (I've still yet to try ColourFlip mode on a 2x speed projector because it turns out my old projector's PSU has failed again.) If I can remember, I'll give it a try on my sister's X1 soon.

I actually have an extra chip on my controller for driving glasses like yours directly. It's more complex to drive it though, as I wanted it to work like the Elsa controller where the voltage on each panel also swaps polarity every other time a shutter is blanked. (to stop the LCD panels apparently degrading over time if the drive voltage is always the same polarity.)

You will probably get away with using the opamp circuit, but I would just make sure that niether of the shutters stays blank for too long (eg. if the S3D game / movie has been closed, or the opamp is still powered when the L/R input signal is removed.)

This is why I added the extra chip to my controller, so it could drive almost ANY generic glasses directly. If I could get the damn blue-line decoding working it would be great, but it's not playing ball. I could just order some PCB's for the current design, then add blue-line decoding at a later date (the controller works fine when running from VSYNC with an L/R swap switch, but it's a tad annoying getting the swap correct).

Would anyone still buy a controller which does the following?...

• Basic shutter control via the VGA VSYNC signal (will need to toggle the L/R swap switch occasionally when starting S3D games / movies).
• Probable future support for blue-line decoding (once working properly, and if ATI/AMD don't come up with a simple solution for sync output).
• No direct HDMI support, as too complex and expensive to implement atm. (although, workaround for HDMI support is possible if HDMI / VGA ports stay in sync).
• Support for the ColourFlip mode. (Only really needed for fixed 60Hz DLP projectors / TV's, more testing to do,
  looks like the shutters will be TOO fast with 4x DLP projectors, and still requires method for configuring for different DLP projectors / TV's.)
• Manual variable phase delay (to minimize ghosting with different DLP displays and possibly plasma TV's).
• Will still work for refresh rates above 60Hz (maximum refresh dependant on type of shutter glasses).
• Integrated shutter driver for controlling "dumb" shutter panels directly.
• VESA 3-pin output (to control any standard glasses with their own integrated shutter driver, or to control an IR emitter).
• VESA 3-pin INPUT (for adding the variable delay to existing VESA outputs on 3D DLP TV's. Could still use ColourFlip on older 60Hz DLP TV's).
• USB / mains powered.
• Upgradable firmware for new modes and features.
• Cost roughly GBP £30 / US $45 ?

Any takers, or suggestions (apart from: "Man, you type WAY too much!" )

OzOnE.

[slipstream]

This thread is interesting for dlp 60hz users.

http://forums.nvidia.com/index.php?show ... ntry992265" onclick="window.open(this.href);return false;

[iondrive]

hi OzOnE2k10,

Forgive me for not reading all 5 pages, but I was wondering if you saw this thread:

http://www.mtbs3d.com/phpBB/viewtopic.php?f=26&t=4372" onclick="window.open(this.href);return false;

It starts out with the same idea but I've given up on it. My oscilloscope shows that it takes about a millisecond to transition from dark to clear so there is a max freq that you don't want to get to and the higher you go, the more percentage of time is spent in transition and during those times you're vulnerable to more ghosting as a result so don't go too high. Anyway, good luck with all that.

--- iondrive ---

[OzOnE2k10]

Ohhh - it seems I wasn't the first to think of this. I won't have to worry about the patent after all!

Well done for being the first to spot the potential benefits of this method. I still think it could hold promise, so you were definitely on the right track. Don't give up with it yet, I should might be able to help you with this....

If you get something like a PIC chip to measure the colour wheel timing with a simple sensor, it's actually not too difficult to figure out the colour wheel config. The good thing is that ColourFlipping is only really usefull at 60Hz, so the colour timing shouldn't change at all once the controller has worked out the delay values...

To configure the controller for different projectors / colour wheels, you just need a simple photosensor or small solar cell which is placed in front of the projector lens. You then start the test pattern program on the PC (very simple to write now, thanks to Peter!), then push the "learn" button on the controller (the test pattern program will display a black screen for a few seconds to give you a chance to push the learn button on the controller). The test pattern is simply a full screen of Red, Green, Blue, then possibly White (probably best with Black frames in between). The controller only needs to know the timing for the "starting point" of each colour. It can then do a simple "sort" of these starting points (for different colour segment orders). It will then know when to "flip" the shutter to the opposite eye for a whichever colour you want.

The controller just stores this timing info into EEPROM, and uses it when ColourFlipping mode is activated. It would be easy to download and analyse the timing data from the controller to the PC in cases where the colour wheel timing is complex, and doesn't auto-detect properly.

It still appears that it's best to swap the GREEN channel, so the controller just needs to swap the shutters when the Green segments start, then keep them swapped until the start of the next colour (Red in case of my projector). This sounds complex, but could be made to work quite easily (compared to the pain of writing the rest of the code in the first place.)

I should really give an update on the project as well. I didn't post for a while because I've come across a bit of a problem with the controller hardware... The good news is that I FINALLY got blue-line decoding to work , but there are a few issues with that as well.

The problem is that the PIC chip I'm using still isn't fast enough to get accurate timings for different resolutions. I can tweak the code and get blue-line decoding to work for say 1280x720 at 60Hz, but if I change the resolution, the point where it "samples" the blue line is not in the middle of the line any more.

I started off trying to measure the length (duration) of the entire blue-line code. I was blanking the right shutter if the blue-line was short, and blanking the left if the blue-line was long (this might need to be swapped depending on your display). This method didn't work properly at all, so now I'm just testing the middle of the blue-line to see if it's blue or not (black). This works great apart from the resolution problem mentioned above!

So, for my projector's native res of 720p, I can finally keep perfect shutter sync using the blue-line mode in IZ3D or StereoScopic Player WITHOUT the need to constantly keep toggling the LEFT / RIGHT swap when a game or video is first started! (also don't need to worry about sync glitches ever again.)

I'm running the PIC chip at the fastest it's designed for (20MHz), but this is divided by 4 internally, giving a 5MHz instruction clock. No matter how few instructions I use to detect the middle of the blue-line, the VGA signal "line" is already one-third of the way across the screen before the code even has chance to sample the blue-line. This can be made to work, but when you change resolution the decoding can go apesh*t again.

What I really wanted to use from the start was an FPGA chip, but I've looked EVERYWHERE for a small and cheap FPGA chip (or board) which would do the job. There doesn't seem to be such thing as a small and relatively cheap FPGA. Using an FPGA would have been ten times easier than trying to code the PIC chip because FPGA's are perfect for this type of work (which requires accurate timings).

I currently only have one FPGA board, which is the Pluto-II from http://www.knjn.com" onclick="window.open(this.href);return false; . Even this board is overkill for what we need, but the main problem is that it costs $60, which is probably more than most people would be prepared to pay for the entire controller box.

So, what I will try next is to buy a much faster PIC chip. The current PIC chip only costs around £4 (USD $6.27), but the whole controller is already working out at around £40 (USD $62). A faster PIC would probably work, and would only cost around £8 (USD $12.53). The main advantage of the PIC is obviously cost, but also that most of the code is already working and the PIC contains a lot of useful hardware which helped reduce the number extra components needed.

I've spent a long time figuring out how the controller box would connect up, and designing the PCB layout for it. I just need to find a chip which will work for all resolutions now. I've attached an image of the current layout to give you all an idea of how it would connect. It is powered from USB, and can also be configured via USB (for ColourFlipping or variable delay settings). This gives a nice "solid" +5V for the controller to work from.

The VGA passthrough dongle will be connected to the controller via a 6-pin mini-DIN connector (similar to the 3-pin VESA connector, but the plugs can't be swapped by accident). The controller has a 3-pin VESA OUTPUT for controlling glasses with integrated shutter drivers, but the controller also has it's own on-board shutter driver for controlling cheap glasses directly (via a 3-pin stereo plug.)

When designing the board, I suddenly realized that using a 3-pin VESA (DDC) input would be fairly useless for most people who have a 3D ready DLP RPTV. The reason being that any 3D ready TV is already designed for flicker-free 3D, so ColourFlipping would be fairly pointless. I can't think of many good reasons to include a 3-pin VESA input on the controller when it's only really 3D RPTV's and some graphics cards which have this output (you might as well use the VGA passthrough on the graphics card anyway?) The only thing which might prove handy is if people want to try adding a variable phase-shift to the shuttering to reduce ghosting, although the other features of the controller (blue-line decoding and ColourFlipping) wouldn't even be required when doing this.

Ok, this post is long enough now I think. Basically, I'm still working on it. If you have any questions or suggestions, I'll try keep the replies shorter!

OzOnE.

[parasupaman]

# VESA 3-pin output (to control any standard glasses with their own integrated shutter driver, or to control an IR emitter).
# VESA 3-pin INPUT (for adding the variable delay to existing VESA outputs on 3D DLP TV's. Could still use ColourFlip on older 60Hz DLP TV's).

I am REALLY REALLY keen to find a timing delay controller to use with my eDimensional shutter glasses, with the hopes of reducing the ghosting I am experiencing. I would attempt to build it myself if I could find a fairly simple straightforward circuit for it.

[gonthorn]

I am very interested in this project, I have the same model projector as you (acer ph530) and while its a great 2d projector I have not had much luck with 3d. So far I have been playing around with 60hz flicker via an edimensional dongle and wired shutter glasses.

I would be very interested in a solution like this, either a custom dongle to increase the shutter speed, or a software mode (BLC ?) that does this colorflipping as you experimented with that could be used with say a home built controller (sell the schematic ?).

I would pay 75-100 dollars for a working dongle that would do this, as that would eliminate the need to purchase a 120hz dlp, like the new acer 120hz series I see talked about.

Don't let this project die, there are others out there that would jump on it im sure. Of course I would be willing to help with any testing regarding the acer ph530, but it seems like you are past that point.

[atariguy]

I'm looking for something simpler and wondering if you can help.
I just want a circuit to drive the old style 3D LCD glasses, like the ones that came with the sega, and the cheapies that come with the 3D field sequential DVDs and trigger them from the VGA sync pulse, or from the output of the DVI sync extraction circuit that is on these boards.

The old VGA passthru circuits from the 90s would do it, but they are all broken links these days.

I'd be happy to pay for a simple wired glasses controller that was driven by the VGA or DVI sync pulses.

I just want glasses like this to work with Peter Wimmers 3D player on the PC

http://tinyurl.com/2667pcc" onclick="window.open(this.href);return false;

[tritosine5G]

...wow I know ozone from diyaudio.com !!! :

Cool, im goin to contact you. We'd want some powerled 120hz afaik?

[Hornet]

Hi,
little about VGA pass through controler (synchronised with synchr.impulses of VGA signal) is here:

http://www.stereo3d.com/3dhome.htm" onclick="window.open(this.href);return false;

in section homebrew.

I use one of them with older DVD player with VGA output (but it is years ago and i dont know the concrete type of controler) It works good with projector too, or as enabler of 3D for Z800. (it is syntethised DCC signal to one pin in output VGA conector).

[OzOnE2k10]

Hi,

First thing's first - Merry Christmas and a Happy New Year to all !!

There's some great stuff going on in Petrus' original "DIY Shutterglasses Controller" thread atm - Sounds like they're getting very close to finding the IR protocol for the Samsung glasses...

I've been waiting for around three months now for a Chinese company on eB*y to release their new pocket oscilloscope. Unfortunately, it doesn't look like it will be released until next year. (The idea was to grab the Samsung IR protocol from a 3DTV in a shop).

Anyway, I just thought I'd chime in to say that I finally got my grubby hands on not one, but TWO "3D Ready" projectors! (slightly OT, but I'll get to the point soon)...

Firstly, I won a ViewSonic PJD6211 (£75) with no lamp, so sold as "parts only". The lamp is only around £100, and the projector powers up fine, so I should get that working soon. Less than two hours later, I won the "Golden Nugget" - an Acer H5360 for £100. This was also sold as "parts only", but when it arrived, it powered up for a while then shut down again. Amazingly, it still had a lamp inside which looked almost perfect !

So, why was it shutting down again? - Someone had left the colour wheel sensor unplugged! It had obviously had a lamp explosion previously, so there was a lot of dust and glass bits inside the optics. There was also some lens grease all over the inside of the main lens. After some careful dismantling and cleaning, I now have a fully working (and very bright) H5360 with apparently only 270 hours on the lamp.

Obviously, this has got me interested in the DIY Shutterglasses controller again. For some reason, Blue-Line decoding has been removed from the latest version of Stereoscopic Player?? It still didn't work too well with my controller anyway (especially at 120Hz), so I just used VSYNC toggling instead. The S3D on the Acer is very good of course, but having to use my old wired Elsa glasses is a serious PITA and takes some of the fun out of it.

There is a tiny amount of flicker at 120Hz, but excellent in comparison to the headache-inducing 60Hz (with my old PJ). The H5360 is around 30% dimmer when running at 120Hz, but still comfortable to watch. Watching 2500 lumens @ 60Hz is actually a bit TOO bright! There is almost zero ghosting at 120Hz with the Elsa glasses!

So, I'm now very interested to find the correct protocol for my Samsung SSG-2100 glasses. This would be a killer-app for a DIY controller, since I'm sure these glasses will start to drop in price quite dramatically over the next year. I will try playing with the software today to see if I can get my SSG glasses working and see how the frame sync works.

The SSG glasses do actually turn on when I put them close to my PC's LCD monitor! (Dell SP2208WFP). I noticed right away that the shutter rate was high (so, very likely fixed at 120Hz). The glasses also stay on for a few seconds even when moved away from the monitor. This suggests that the glasses only require a sync signal every so often to keep them going. This is a simple solution for reducing the possibility of glitches (when the IR signal is blocked for a short moment etc.).

I personally think that the nVidia kit is expensive for what it is, and I don't really want to buy yet another GPU for 3D. I don't really play games much, so I'm more interested in watching 3D Blu-Ray etc. It's nice to see that ATI are finally introducing quad-buffered support for improved playback, but this also means buying a new GPU series apparently?

For the time being, software pageflipping is working quite well for me for 720p 3D MKV playback (very few glitches, and frame-sync never flips). I just need to get these damn SSG's working now.

It looks like Stereoscopic Player does keep the left-right sync very well even if frames are dropped. So, I might have a look at decoding the e-Dimensional activation signal next (so you don't need to choose "Normal" or "Invert" mode when you first start a movie).

I can't promise much success with the SSG glasses, but I'll mess around with the code. It sounds like the guys on the other thread are working hard on the SSG stuff, so fingers crossed!

OzOnE.
P.S. I hope some of that post was useful to someone... I might be a tad drunk.

[cybereality]

Peter Wimmer had to remove blue-line coding from the Stereoscopic Player due to licensing issues.

[OzOnE2k10]

Ahhh, thought as much. There wouldn't be much point in removing it otherwise.

That means the "best" version with blue-line code is now 1.6.5. Oh well, I'll see if I can get frame-sync working with eDimensional / X3D type frame marking.

This would just require monitoring the Red and Green lines of the VGA connector instead. It might actually be easier to decode than blue-line and should stay in sync as long as the PC doesn't glitch and swap frames?

OzOnE.

[tritosine5G]

2200* 0.32 * 0.5 *matte white (1) = not too bright , should be twice as bright, very washed out colors bcause white+ untreated room

700*0.32*0.5 *matte white (1) = no chance,washed out colors bcause white+ untreated room

700*0.32*0.5 * flat silverscreen (2) = this is acceptable but not quite there yet, Im using it this way now. If projectors black level perfomance would be better, it would be a good scenario. Ambient light rejection is so nice, even 2d picture becomes 3d easily ( MFT , mind routine transform ), like POP!...

700*0.32*0.5 * curved silverscreen(10) = This is what I want, since black level isnt stellar anyway, I choose brightness.

Sony shutters are interesting , cause theres no outer layer polarizer on those ones, but you can ask sony to give it to you for free. So it can be traditional shutter , or if you have a polarization preserving screen , you can omit the outer layer. I want to try this, it'd cut down ambient light very nicely. Its like room correction in audio .