Zero motion blur LCD!!! (nVidia LightBoost2 hack, looks like CRT, looks like 480Hz)

[omeds]

I've concluded I'm better off just using regular 120hz 2D, its already smooth as silk and the IQ sacrifice isnt worth it imo.

 

[SirPauly]

Then, disable it when you're just using regular Windows stuff, and only enable it during FPS gaming at night when the room is dark, and you want the zero motion blur advantage during fast gaming. It's a feature that can be turned on/off.

Somebody needs to make a System Tray Button to make it easy to switch depending on your mood!

Appreciate that you enjoy it! Thanks for creating some awareness!

 

[imaheadcase]

If you have problems seeing motion blur, you will always have problems with motion blur. For most of us who don't, cheers.

 

[imaheadcase]

This.
Appreciate your defense. Guilty as charged about pushing this on five or six forums (maybe excessively so)
It's a travesty that manufacturers have ignored this, but the recent emergence of 3D also simultaneously made possible the zero motion blur LCD, and more public awareness of this is needed.

I'll try to step back (here on AnandTech), and wait for one of you to gloat about how amazing it is for FPS gaming. I give it till Dec 26th - somebody here is bound to have some LightBoosted gift under their tree.

Manufacturer ignore it because its not a huge issue for most people. Monitor manufacturer cater to the %99 of people who buy them. Investing millions in R&D for small % of people is not smart.

 

[chaojohnson]

Is Lightboost a preferred method compared to overdriving the display to reduce motion blur?

AFAIK, overdriving also reduces motion blur, but can increase input lag (bad for gaming) and artefacts. But... Lightboost apparently turns off a bunch of monitor image settings.

http://www.tftcentral.co.uk/reviews/benq_xl2420t.htm

This review for the XL2420T (go to 120Hz and 3D Support) shows that at 120Hz, without overdrive, the display has very little motion blur. Would the downside of *no* motion blur from Lightboost (such as less settings and maybe worse colour reproduction) overrule the upsides?

It just doesn't seem like the tiny reduction in motion blur is really worth the reduction in image quality.

 

[Mark Rejhon]

The tradeoffs vary from person to person. However, one BIG correction:

It just doesn't seem like the tiny reduction in motion blur is really worth the reduction in image quality.

Wrong -- the reduction in motion blur is a shocking 85% as measured by my upcoming motion test software -- an order of magnitude. Even in older motion test software, you can see PixPerAn speed-reading test at the maximum speed of 30, just like CRT:

Yes, I know not everyone likes it, but it's a BIG reduction in motion blur, for those who don't care about color quality. Some gamers don't care as much about color quality, as for the CRT zero motion blur effect. So let's not jump to conclusions.

From Swolern on HardForum (this post)

Pixel Persistence Analyzer (PixPerAn) Readability Test Results 120hz: Wow just wow! Speechless.........

-Highest readable tempo without Lightboost (LB) was 7 and even that was blurry and my eyes where straining to read it.

-Highest readable tempo with Lightboost was 30 (Only could read the first 5 letters)

So when i first turned on LB and tried the last failed tempo of 8 again (which was blurry non-sense w/o LB)and was shocked how crystal clear it was. Kept going up until i reached tempo 24 and completed the test successfully. But at this crazy fast tempo my head was hurting at all the right to left motion that my head was doing to keep up with this speed. Next I turned it up to the insane tempo of 30. The speed is so fast my eyes have trouble keeping up with the letters, but when i finally train my eyes to lock on the letters THEY WERE CLEAR! I could only get the first 5 letter out and correct, due to the insane speed of tempo 30. Maybe a trained eye like Vega can get out all the letters, but for me to be able to read 5 letters states that yes this LB hack makes my VG278H monitors have Zero perceivable motion blur.

Anyone that has done the hack must download to see the results. http://www.softpedia.com/get/Desktop-Enhancements/Other-Desktop-Enhancements/PixPerAn.shtml

On one side note there was a minimal ghosting effect with the letters. Similar to the 3d ghosting effect. The test letters are black print, the test background is white, and the ghost behind it was a very light grey/white. The ghosting was not a blur or trailing effect it was a perfect letter, just extremely light and very hard to see. The only ghost letter that was visible was the last in the grouping as im guessing the others ghost into the other black letters. But in no way did the ghosting effect the readability of the extremely fast moving letters.

See. That is NOT a "tiny" reduction in motion blur.

This is a CRT-like reduction in motion blur. (not as good _color_ as CRT, no disagreement there-- but definitely zero motion blur like CRT). With such a reduction in motion blur, you react 100-200 milliseconds faster in fast-twitch FPS videogames because you can identify small enemies while moving/turning, without stopping moving (Becoming a sitting duck for snipers in the game). This gives you a big advantage, and more than outweighs a few milliseconds of input lag (it's still less than 1 frame of lag).

Instead of move-pause(wait for blur to stop)-identify-shoot-move-pause(wait for blur to stop)-identify-shoot, in FPS games, you can "move-shoot-move-shoot" because you're identifying tiny faraway enemies while moving fast. You frag a lot more enemies much faster on a LightBoost enabled. Yes, yes, color quality is worse, but for Arena-type competiton FPS, you have something like more than 100ms-200ms less human brain lag in your human eyes because of lack of motion blur in scenery (no need to wait to stop moving before you react). That more than compensates for a few ms of input lag (still under one frame). So if you are serious about FPS gaming, you're already a CRT user, or should be considering a LightBoost display. The hundreds of milliseconds improvement in reacting to far-away tiny enemies without needing to stop moving first, in fast-twitch FPS games.

For some competition FPS gamers, less motion blur is far more important, for the faster reaction times, and that is why some of them still use Sony FW900 CRT's, and now we have the same effect on LightBoost displays (provided you have a GTX 680 or faster, you MUST be able to run 120fps or greater, for the best zero motion blur effect -- otherwise, LightBoost doesn't give you enough benefit)

It may be less worth it at Windows desktop, but thankfully, you can turn on/off LightBoost!
If you have a GTX 660/680 or better (for source engine games, use fps_max 240 at console) or a GTX 680 SLI or better (for Crysis engine games) capable of 120fps@120Hz, test turn on LightBoost2 and you'll definitely get more frags shooting faraway players more quickly in online FPS matches.

Don't try it just at desktop (not as big wow)
Try it in an ARENA-TYPE FPS COMPETITION (this is the really big WOW effect)
Make sure you have Geforce GTX 680 or better (unless playing older games)

The wow effect works better in some games than others:

braid
- looks blury in 2d with LB on (game is locked at 60 fps)
- blur is gone with 3d and glasses

shatter
- looks flawless in 2d with LB on (I can clearly see the balls moving in the menu, I even broke my old record for bonus mode the first time I tried...)

diablo 3
- ingame vsync locks the game at 60fps, objects/people are doubled when moving with LB on (120hz), creating a blur effect, no stuttering or tearing
- with disabled vsync the pictures are still separated even when running at more than 120fps
- with vsync enabled in the nv control panel the game locks to 120fps but the pictures are still separated...
- in 3d everything looks great... except it's unplayable due to all the text at 0 depth

titan quest
- looks incredible in 2d with LB on... if you are getting exactly 120fps
- when the framerate drops blur starts to be noticeable (there might be 1 ghost image), not very playable
- very easy to keep 60fps in 3d, so no tearing, stuttering or blur, unplayable due to text at 0 depth
- interestingly with vsync disabled in 3d (even when going over 60fps), there is stuttering and tearing but no blur
- seems to be the same case with 2d, very stuttery at 150fps, cannot tell if there is a ghost image

solar 2
- I can clearly see the background when moving, looks great

trine 2
- looks amazing in 3d with vsync on (60fps)
- stuttering with vsync off in 3d (67fps), no tearing, no blur
- stuttering and tearing in 2d (67fps), objects seem slightly blurred so there might be a ghost image, unplayable

torchlight
- perfect movement with 2d LB on at 120fps
- some stuttering when framerate drops under 120, did not notice a ghost image since I cannot get the game under 100fps

civilization 4
- everything is clear when scrolling in 2d with LB on at 120fps
- when the framerate drops there is stuttering and 1 very noticeable ghost image (or duplicate) while scrolling the map, unplayable
- horrible stuttering/jumpy (rubberband like) scrolling in 3d when dropping under 60fps, unplayable

 

[Mark Rejhon]

Here's the BENQ XL2411/XL2420 instructions.
Apparently, the driver seems to locks the Benq from being allowed to use LightBoost without buying 3D glasses. We need to override this, see below.

TuGuX from TechNGaming appears to have a solution that you could try working from, to trick a Benq into thinking it's an ASUS VG278H so that the LightBoost strobes works without glasses in 2D mode:

@[member='bojinglebells']

First, you install this .inf via device manager: http://www.file-upload.net/download-6944730/Asus-VG278H-3D-Monitor-EDID-override.inf.html
Then you reboot.

After this, you enable this: http://www.file-upload.net/download-6944739/ForceLightBoostWithoutGlasses.reg.html

Make sure windows is set to 120Hz
Goto NVIDIA drivers, check the box to 'Enable Stereoscopic 3D'.
Assuming it has worked, it should now list the 'ASUS VG278H' monitor as the Stereoscopic 3D display type, and not '3D Vision Discover' which is there by default.
In this same section, set the drop down box for stereoscopic 3D to 'Always Run'.

Now start a game, I tested it with CS:GO .. I had 60 FPS and real 3D on (which is not good) AND check your monitor settings (when you go in menu -> picture, there should be something like "Nvidia LightBoost", then it's working)

Next step, just uncheck 'Enable Stereoscopic 3D'. Now you should have more than 120 FPS with LightBoost (depending on your hardware), so everything is nice and working right now.

To turn it off, I guess you have to change the Always Run stuff, or completely deinstall monitor and install normal BenQ Drivers

Alternatively, a different set of instructions:

The "always" box will not appear until you run the 3d wizard setup.

I did not use any registry tweak for my setup. What i did:
checked enable stereoscopic 3d box-> ran Wizard-> ran test stereoscopic 3d in N control panel-> set refresh rate to 120hz & select Apply refresh rate to all games box-> exit test app-> uncheck Enable stereoscopic 3d box-> apply

 

[Swolern]

Hello Anandtech. Im mostly on OCN but i just wanted to share that i did in fact prove that motion blur is non-existent with my current setup with Lightboost active in 2d mode. There is a loss in some brightness, but after contrast is adjusted the PQ is great. There is a need to keep framerate close to 120fps as possible. Some games behave differently, an example is MOH Warfighter, which after its last update only needs 80-100fps for that ultra smooth CRT look and feel.

BF3 has been one of the worst game to get this effect done well, and the worst to get smooth in regular 2d mode as of lately. The slightest framerate variance from 120fps looks choppy and stuttery with my setup.

Any non-believers that have a VG278H please try my above method to get Lightboost in 2d mode and run the readability test with Pixperan. I cannot speak for other monitors other than my VG278H.

I have 3930k/Asrock Extreme 11/ GTX 670 4-way SLI/ [x3] Asus VG278H. I do have my setup in just 2-way SLI right now due to BF3 3&4-way SLI issues.

Any questions please send me a PM on OCN.

 

[Mark Rejhon]

Welcome aboard! Glad you're here too.

I hear from Vega on HardForum that the Benq LightBoost is much brighter than the ASUS LightBoost, so if people want the zero motion blur ability without losing significant brightness, the Benq XL2411T is apparently superior and plenty bright. BENQ claims 1.0 millisecond, so if that is the strobe length (when LightBoost is enabled), that would apparently match CRT phosphor decay -- possibly even more motion blur reduction than Asus. 1.0 millisecond response time is now possible even if the LCD panel is 2ms, because a refresh is 8ms long at 120Hz -- you simply wait for the LCD to refresh first, then you strobe (even for shorter time than the pixel persistence itself). Voila. Bypassing the pixel persistence as the motion blur barrier...

I'd love to see someone who owns both an Asus VG278H and a Benq XL2411T to make a comparision, since shorter strobe lengths == less motion blur. You need more brightness in the strobes, to compensate for the short strobe length...

Apparently, not all LightBoost monitors are made equal.
I'd love a blog or a magazine review to test them.
They may not be of interest to 90% of people, but they are quite impressive for FPS gaming, and actually apparently is more important than input lag.
More users here need to test out and compare!

 

[Mark Rejhon]

Manufacturer ignore it because its not a huge issue for most people. Monitor manufacturer cater to the %99 of people who buy them. Investing millions in R&D for small % of people is not smart.

Then how have I begun to make a home-made strobe backlight for less than $1,000? (Project is ongoing as a modification to a cheap 23" 3D LCD; though a bit distracted by the LightBoost stuff now) See my website (scroll down to see my construction photos) at www.blurbusters.com (my domain) about the Arduino-driven synchronized backlight modification to an existing LCD monitor - I purchased several hundreds of watts of LED's (my unboxing photos of LED). I also made a FAQ page too.

With services like Alibaba and other B2B sites where you can order small runs of monitors nowadays, even a small run of 1,000 to 10,000 premium monitors by a manufacturer can be very profitable, if the R&D costs are as cheap as I've explained. The panel technology is already here (120Hz 3D active shutter glasses LCD panel = automatically means zero motion blur capable for 2D, for the reasons already explained in my FAQ).

Panel is no problem. It's simply mostly backlight technology now. The math calculation is 150 watts strobe per square foot is needed to match the amount of photons that CRT phosphor put out. As an electronics student, if you've made a simple flashing LED project, this is just a super turbocharged version with a direct synchronization to the VSYNC (with a strobe delay adjustment and fine tuning). You simply just need high wattge in your LED's, and you need synchronization to the refresh -- a hook from a contact point in a monitor circuit, to an interrupt pin on an Arduino, etc. Or even a dongle that connects to the plug, with a wire leading to the Arduino connected to the strobe backlight.

There is an opportunity to use higher-quality higher-CRI LED's (or even RGB LED's) and improve the color quality of a strobe backlight significantly. There is also an opportunity to do a strobe backlight with an IPS LCD, given sufficient engineering to enlarge a blanking interval long enough to wait-out the pixel persistence before strobing before the next reresh. (this one will require more R&D, since no 3D IPS LCD panels yet exist at 24" monitor sizes, but IPS 3D active LCD exists for living room sized HDTV's.

Thanks to the solved panel problem, engineering this is getting easier and easier for the correctly knowledgeable -- not much engineering is needed if you source pre-existing 3D 120Hz LCD's. (The biggest challenge is backlight optics/diffusers, they are hard to source, but you can find them if you look hard enough.) To be honest, I'm glad manufacturers are beating me to the punch. It's slowing down my project by maybe a few weeks, while I temporarily focus on testing and analyzing LightBoost monitors.

 

[Mark Rejhon]

Vega just posted on HardForum that the BENQ has better motion than the FW900 widescreen CRT he owns:

Oh my, I just got Skyrim AFK camera spinning (which I used to test LCD's versus the FW900) to run without stutters and VSYNC locked to 120. This Benq with Lightboost is just as crystal clear if not clearer than the FW900 motion. I am in awe. More testing tomorrow. Any of my doubts about this Lightboost technology have been vaporized! I've been playing around with this fluid motion on this monitor for like 6-hours straight, that is how impressive it is.

Apparently, he also says the BENQ XL2411T is brighter than the ASUS:

It is also very bright in Lightboost mode. The only issue I have is the color shift with a maroon tint while in LB mode.

Coincidentially, the BENQ XL2411T (supposed to be better than XL2420T) has a big sale on Overclockers UK now until Boxing Day (no VAT when shipping to USA). I just ordered one for myself just under an hour ago, to compare to my Asus VG278H. I need to do comparision tests, especially as I would appreciate a brighter LightBoost picture, too!

Although I hear ASUS has better colors, this might not hold true for LightBoost: The BENQ might have a better picture/color than Asus during LightBoost mode. I'm going to find out, personally!

 

[imaheadcase]

Then how have I begun to make a home-made strobe backlight for less than $1,000? [/url]).
.

You answered own question. No one drops $1000 on most things.

 

[Swolern]

You answered own question. No one drops $1000 on most things.

Why is there so much negativity on this forum? The OP is clearly an display enthusiast that has a great finding and just trying to help other people out to get the same great results.

 

[philipma1957]

Why is there so much negativity on this forum? The OP is clearly an display enthusiast that has a great finding and just trying to help other people out to get the same great results.

I hear what you are saying.

The best pc monitor i ever saw was a 21 inch

4 by 3 ratio with .21 dot pitch great colors no blur . It came out in 1999 oh that is right it was a crt and weighed 75 pounds. It was 20 inches deep and hot. still it had a better picture then any lcd/led I have seen.

 

[omeds]

Any tips for the VG278H? Image is not good with the strobing (I assume) making it dimm and difficult to view in 2D, as expected, and I can't seem to find any decent way to compensate.

Also, are the controls on the BenQ locked when in 3D mode? Or are you free to alter all display options? Most options are locked on the VG278H which is annoying.

 

[tweakboy]

I find this very interesting. I recently bought a 2ms LCD screen to replace my 5ms screen. Because I don't like the motion blur. The result was pretty disappointing. There was a little improvement, but not much.

Got a few questions.

1) For the forseeable future, we can not expect this technology to work on IPS screens, right ? Because all 120Hz screens seem to be TN, and there are no 120Hz IPS screens.

2) What is it about 120Hz that makes this trick work, and not on 60Hz ? As far as I can understand, the trick is about dimming the pixels while they are changing color, right ? And only displaying them (bright light behind the pixel) when the pixel has its final color. Can't you do that on a 60Hz screen too ? (Edit: I guess it's because we need LightBoost2, which so far has only been implemented on 120Hz monitors, right ?)

3) Your post here mentions an InfraRed emitter that you connect to your PC. Why ? What does that have to do with anything ?

4) I think I read somewhere that this trick will make your LCD use more power. I can't find that reference anymore. So how big is the power increase ? Your monitor will be in this state full time, right ? No easy way yet to enable it only during (specific) games ?

5) So what do we need to buy to get this to work ?
*) Any LightBoost2 monitor. I understand it's not possible at all on any 120Hz monitor, right ?
*) A Dual-link DVI cable.
*) Any nVidia videocard (faster is better, obviously).
*) Do we really need that separate USB IR emitter ?
*) Do we need 3D glasses ? I guess some 120Hz monitors can be bought without the glasses ? Which saves some money.

I hope some vendors will start supporting this. With proper drivers and software support. So it can be enabled/disabled with a mouseclick. Or it will be enabled automatically when the screen goes into D3D/OGL mode. Right now I find it a bit tricky to buy a new screen just for this, without having it seen with my own eyes. And without official support. But it looks very promising.

The only way it will be like a CRT is if your monitor was 1ms ,,,,,,, mines instant mouse vsync on...... smooth as butter.

 

[exar333]

So, what about AMD users?

Out of luck...

 

[EliteRetard]

Glad I still have my CRT...1920x1440 @ 72-85Hz.
Simply superior when considering all factors at once.
No bugs, glitches, weird issues or frame requirements.

LCD is finally starting to catch up though, about as good as a crappy bug ridden CRT with 120fps requirement.
Once we get good quality colors and higher DPI without a 120fps requirement then I'll get one...unless my CRT dies before that.

 

[Swolern]

Any tips for the VG278H? Image is not good with the strobing (I assume) making it dimm and difficult to view in 2D, as expected, and I can't seem to find any decent way to compensate.

Also, are the controls on the BenQ locked when in 3D mode? Or are you free to alter all display options? Most options are locked on the VG278H which is annoying.

Yes I turn my contrast up to 92 and LB at max on my VG278Hs. Upping contrast helps a lot with brightness with LB 2d mode. It is still not as bright as non-LB mode, but I find it very comfortable on my eyes and just the right 2d brightness. But my room is dark when gaming.

Controls are locked on the BenQ also, but from what I read it is brighter with LB than the Asus.

 

[Mark Rejhon]

You answered own question.

My point is that it's now possible for a monitor manufacturer to make a profitable small 1000-monitor run, amortizing the cost of now-fairly-lowR&D cost of a strobed backlight, and sell them at roughly Alienware-style prices. There's a market for that. The emergence of LightBoost is proof of that too.

 

[Mark Rejhon]

Here's some oscillscope photodiode measurements of LightBoost, for an Asus VG278H. The strobes are 2.4 milliseconds long. (I hear the BENQ is about 1ms, I've ordered a BENQ to test out).


(Large Version of Image)

Also, here's a video I've made for forum readers, captured at 480fps and 1000fps:

High Speed Video Of LightBoost

It also proves that strobing can be used to bypass pixel persistence.
Enjoy!

 

[3DVagabond]

I am dying to say that there is no "lightboost 2". It is "3D Vision 2" with lightboost technology. Also, I am wondering if the 3d light is designed to be on all the time?

I was wondering how much additional wear and tear this puts on the monitor. If it runs hotter, it's likely going to shorten it's life span. Have the manufacturers chimed in on this yet?

 

[Mark Rejhon]

I was wondering how much additional wear and tear this puts on the monitor. If it runs hotter, it's likely going to shorten it's life span. Have the manufacturers chimed in on this yet?

LED's have virtually no wear and tear when they are flashed. Computer monitors are already using PWM for dimming.

LCD wear and tear is not affected, since the LCD panel doesn't operate faster than it normally does without LightBoost. The pixel persistence is simply completely bypassed, by keeping the pixel persistence in the dark, as proved by this high speed video on YouTube. It just have to be a 3D compatible panel, that's capable of practically finishing pixel persistence before the next refresh, to allow fully-refreshed frames to strobe through. The strobe length can be shorter than the pixel persistence length.

One minor related factor (but separate from the strobing): Response time acceleration technology can lead to image retention in certain panels, especially in early implementations from a few years ago, but this is less common now.

My VG278H actually uses less power in LightBoost mode, BTW. The picture's also somewhat dimmer. Probably because of the dark period between the strobes. It's running cold, even though it's been on all day. The back of the panel isn't even warm. BENQ is supposed to have much brighter LightBoost, looking forward to my XL2411T when it arrives, so I can test against it.

 

[EliteRetard]

Wow that's terrible...120hz doesn't look to have improved things at all over a 60hz monitor.
And your strobe doesn't help at all, I can still clearly see a ghost image after the strobe.

 

[Ben90]

It's actually overdrive trailing you are seeing. I'm curious as to how they are overdriving the display. If they strobe during the vertical blanking space, the top pixels won't need any but the bottom pixels are going to need to get hit pretty hard.