LCD driver: What changes in 3-D mode?

Advanced display talk, display hackers, advanced game programmers, scientists, display researchers, display manufacturers, vision researchers & Advanced Display Articles on Blur Busters. The masters on Blur Busters.
Post Reply
istemihan
Posts: 2
Joined: 11 Dec 2015, 13:01

LCD driver: What changes in 3-D mode?

Post by istemihan » 11 Dec 2015, 13:07

Facts:
  • We know that a system needs high data transfer speed if it works with +120 Hz refresh rate and +1080p resolution(This mostly requires DVI dual link).
  • We know that in a 3-D system with shutter, the method is to flash for Right and Left eye arbitrary.
Image

For right eye, shutter shuts the left eye and display shows the image for right eye and then for the left eye, shutter shuts the right eye and opens up the left eye and display shows for the left eye and on and on goes like this. This is how 3-D visual is done with shutter.

Now, the most popular 3-D system: lightboost setup:

Image

As the lightboost released on the market many monitor vendors also released +120 Hz monitors with lightboost support(such as VG248QE the most popular one).

Before I reveal my question, I feel like I have to explain technical details of DVI dual link and LCD driver a little bit. There are some missing parts of them which I could not understand.

This is DVI dual link pinouts:

Image

I found this on the adress: http://www.alciro.org/alciro/conectores ... 269_en.htm

explained as follows:
The DVI connector is divided mainly into three areas:

TMDS. (Transition Minimized Differential Signaling) or Transition Minimized Differential Signal similar to communication systems balanced. Marked on the figure by the brown color has four twisted pair cables one for each primary color video red, green and blue. Each two pairs have a mass of specific protection marked in dark brown. The video signal is synchronized with a pair of twisted wires in light brown line has its own mass. With a video link can transmit high-definition HDTV signal (1920 × 1080) at 60 Hz (139 MHz) for better definitions and better refresh the DVI connector includes a second link in combination with the former provides resolutions higher than HDTV HDTV (1920 × 1080) at 85 Hz (2 × 126 MHz) or WQXGA (2560 × 1600) at 60 Hz (2 × 135 MHz), among others.

Pug & Play. Marked in blue has a number of lines that highlight the DDC (Display Data Channel), serial connection with pin I2C Bus SDA (data) and SCL (clock). Are used to detect and display settings when connecting to a computer.

Analog. Analog system, marked in green features a pin (pin) for each RGB primary color red, green and blue analog sync with the digital connector makes it compatible with the analog system, reaching a bandwidth of 400 MHz

Image

So, 3-D visualization method seems not that hard. Send the pixel datas for Left eye than send the datas for Right eye and on... Communication protocol also seems not that hard! You simply send the datas not encapsulated as you can see below:

Image

As you see, with DVI, we send datas bit by bit and the LCD driver of the monitor decodes received data. So, my question is this: Why do I need lightboost(Or any other 3-D vision product)? What lightboost(3-D stereo feature of it) actually does? Why don't we just use the same communication with monitor and simply send preconstructed Left and Right pixel datas to the monitor and then it decodes the data and do the same job as in lightboost?

To open up my question a bit more, I have to indicate where my question actually relies. It is between the graphic board and the monitor. You see, lightboost requires both graphic board and monitor lightboost compitable. What does it mean? In lightboost 3-D mode, does communication between graphic board and the monitor change? If so, why? Why doesn't the graphic card first construct 3-D content and send them in the same way before? For instance, when try to use lightboost on some other machine with old graphic cards, I get the warning that the graphic card does not have 3-D stereo feature. If 3-D content is already prepeared(like 3-D movie or anmation), why can't I use any graphic card(which can drive 120 Hz 1080p) to send them to the monitor? After all, the method looks simple:

(1)drive pixels with left eye data, then strobe backlight

(2)drive pixels with right eye data, then strobe backlight

then it goes like 1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2....

Why do I need 3-D stereo? What 3-D stereo feature does? How it effects the communication between the monitor and the graph card? Are there primarily differences in the hardware structures of graphic cards which have 3-D stereo feature and which don't have? If there are, what are they? I firstly, wonder about these. I could not reach any detailed documentation about the topic.

Secondly, say it has to be done in that way, communication between the monitor and the G.board(via DVI dual link) has to change in 3-D mode. Then why is that?What changes? In what way they manage to switch to 3-D mode? I firstly, want to know the reason for this and secondly, what is the method for this.

Thank you already.

RLBURNSIDE
Posts: 104
Joined: 06 Apr 2015, 16:09

Re: LCD driver: What changes in 3-D mode?

Post by RLBURNSIDE » 11 Dec 2015, 13:41

"Why do I need 3-D stereo?"

You don't, it's just fun to play games and watch movies in 3D. Millions of years of evolution have conferred upon us the benefits of binocular vision and it's absurd how many people hate 3D, it's like a bird hating its wings and preferring to stay on the ground to spite its own natural capabilities.

Some people, like me, would love to play more 3D games but NVidia is such a dumb company, 3D Vision is poorly supported and doesn't work at 1080p most of the time (no support for 60hz 3D unless you pay big bucks for Quadros plus the 3D plugin software as well), and doesn't support SBS or O/U so many people with 3D TVs and projectors can't even use it to game at 1080p. 720p is ugly as sin.

Stereo 3D has nothing whatsoever to do with Lightboost.

I suggest you read more about 3D modes especially FramePacking vs Side-By-Side or Over/Under. Many 3D displays will auto-detect when your signal contains such information. For example, FramePacking Blurays have both frames sent at the same time, at 24fps, in full resolution one on top of the other, with a small black gap of pixels in between. This is very easy to detect, since it's just a custom 2D resolution. SBS and O/U are also fairly easy to detect, but many displays (such as my w1070 projector) require you to manually switch into the appropriate 3D mode as the signal is just a normal 1080p 2D signal, from the point of view of the HDMI input / output chips.

Nobody cares / uses DVI so that's mostly pointless to discuss anyway, it's dead / obsolete tech (due to its limitations in terms of bandwidth) so I wouldn't even waste my time worrying about it.

If you want to do strobing (or black frame insertion) with 3D content, you need a display with multiples of 48hz in terms of refresh rate, namely 96hz at least (for 50% strobing). 192hz for 25%, etc.

spacediver
Posts: 505
Joined: 18 Dec 2013, 23:51

Re: LCD driver: What changes in 3-D mode?

Post by spacediver » 11 Dec 2015, 18:46

RLBURNSIDE wrote: Stereo 3D has nothing whatsoever to do with Lightboost.
Well, lightboost was originally a solution to minimize cross talk between adjacent frames (while maintaining enough luminance), to allow for better stereoscopic presentation, no?

User avatar
masterotaku
Posts: 436
Joined: 20 Dec 2013, 04:01

Re: LCD driver: What changes in 3-D mode?

Post by masterotaku » 12 Dec 2015, 03:56

I recommend playing games in 3D to everyone with two working eyes. One of the best advancements in gaming since polygonal graphics, I'd say (and ignored by 95% of the gaming population).
RLBURNSIDE wrote:3D Vision is poorly supported and doesn't work at 1080p most of the time
What? Only with 3DTV Play, which you mentioned later. With my 3D Vision 2 monitor, I can play all games at the resolution I want.

Very few developers care about 3D nowadays, but at least we have many community fixes (helixmod).
RLBURNSIDE wrote:If you want to do strobing (or black frame insertion) with 3D content, you need a display with multiples of 48hz in terms of refresh rate
That's the problem with 3D Vision 2. It works at 100Hz and 120Hz. I've found only two ways to make 24fps videos smooth: speeding up the video with reclock to 25fps and using 100Hz (and I sometimes use black frame insertion to have 25Hz per eye :p), or using motion interpolation at 120Hz (with Smooth Video Project, for example).

But 3DTV Play has a 1080p@24Hz mode, right? The 720p for 60fps per eye artificial limitation in 3DTV Play and my myopia (I would hate wearing glasses on top of other glasses) are the reasons I don't use a projector.
CPU: Intel Core i7 7700K @ 4.9GHz
GPU: Gainward Phoenix 1080 GLH
RAM: GSkill Ripjaws Z 3866MHz CL19
Motherboard: Gigabyte Gaming M5 Z270
Monitor: Asus PG278QR

istemihan
Posts: 2
Joined: 11 Dec 2015, 13:01

Re: LCD driver: What changes in 3-D mode?

Post by istemihan » 12 Dec 2015, 07:42

@RLBURNSIDE
thank you for such detailed explanation. So what I understand here is only difference of lightboost is baclight modulation and lightboost compitable monitors are just know how to detect backlight information(some where in the encoded data).

But, other than that lightboost and other 3d products have standart 3D modes(which I have to do a research on now).

Now that I found such informative users, may I ask another question?

Ive tried many 3D augmented reality games animations vs, all of them were based on this concept: your head is on one reference point, standing still, they only allow you to turn your head and look around, maybe arms and legs are included but you are not allowed to walk around or change your position. Right?

Is this because graphic processor has enough competing power to look around but not enough power to calculate depth perception when you also move? Because when you calculated, 120hz*1080*1080px*3(RGB)*8bits =3359232000 bits data transfer per second is required. When you are standing still, 3+Giga ram may be enough to store calculated other direction of view datas to cache but when you are moving it may not be enough right? Because your processor will has top 3 Ghz clk speed and even in most RISC processors calculating new depth perception will require minimum 2-3 assembly instruction step so the chip would not be able to support such calculation because even if it achieves to calculate new perception per pixel in 2 steps it will decrease the construction of pix speed lower to 1.5 Ghz and the monitor will require minimum 3.3 Ghz(I am not counting the dimming datas of BLU or RGBs).Am right? Or not?

Post Reply