Eizo FG2421 -- [Turbo240 is EIZO equivalent of lightboost]

[SS4]

When/if g-sync comes out for this monitor it will be very hard to not want at least one FG2421

[Neo]

Does the Eizo still use frame rate control dithering? Is that why the color goes down? Obviously it should.

[Chief Blur Buster]

Is it possible to hack this monitor so that it actually does 240hz?

Although Blur Busters strongly advocates the development of true-240Hz displays (non-interpolaced), I don't think the VA panel can do it yet at this time... However, the commercial EIZO FDF2405W (~$6K) panel does use a two-pass refresh cycle, so its TCON might actually be able to accept true 240Hz.

Also another consideration: If you like motion-blur-eliminating strobe backlights such as LightBoost, then you need enough time between freshes for clean strobing (e.g. high speed video of LCD with clean strobing). At higher refresh rates, where the faster refresh cycle starts eating into the pixel transition cycle (GtG), you have too little time between refreshes for a good strobe backlight (e.g. high speed video of LCD thats too slow for good strobing).

[Chief Blur Buster]

Does the Eizo still use frame rate control dithering? Is that why the color goes down? Obviously it should.

I'm not sure. If you're talking about the known VA panel color issues, that one has nothing to do with that -- it's low-greyscale nonuniformity behaviors of VA. The colors is much more contrasty on the Eizo FG2421 (5,000:1 strobing on FG2421 LightBoost instead of 500:1 strobing on VG248QE LightBoost), especially when playing colorful games such as Bioshock Infinite.

[flood]

quick question: for the fg2421, does the top-to-bottom scan take 1/240s or 1/120s?

[Chief Blur Buster]

quick question: for the fg2421, does the top-to-bottom scan take 1/240s or 1/120s?

1/240sec, but it doesn't help input lag in this specific situation, because it is forced to wait until enough of the input is buffered. It still takes 1/120sec for a frame to transmit from GPU to monitor, so it can't "jump the gun" to start scanning early. Otherwise it'd run out of incoming refresh data before the end of the scan-out.

On a related topic of scanout speeds being different from input speeds, LightBoost mode actually partially buffers the input (FIFO-style, approximately half a refresh worth) and then does an accelerated scan-out (approximately double speed, 1/200sec or faster). It only buffers the bare minimum; just enough so that it can finish refreshing the bottom edge of the screen at the same time the final pixels (of a refresh) arrive on the video cable. The purpose of an accelerated scanout during a strobe backlight LCD, is so there can be a longer time period to let the LCD pixels settle (finish GtG transitions) before the backlight is strobed. The blanking interval of a LightBoost refresh is approximately 3 to 4 miliseconds long, which fits both the 1-2ms GtG pixel transitions as well as the backlight strobe.
Citation: StrobeMaster's finding in the LightBoost teardown page.

[SS4]

So when is the full review of this monitor
Really looking forward to it!

[Chief Blur Buster]

So when is the full review of this monitor
Really looking forward to it!

A little bit behind schedule -- the Holidays have slowed things down quite a bit.
I've got G-SYNC Part 2, then the Eizo FG2421 (now that I have the replacement colorimeter), then the XL2720Z review (once new firmware is installed).

[PinkysBrain]

One of the reasons is that this is because strobe backlight monitors need to compress their gamut to produce some overdrive bounce margin, so they can do overdrive quicker and more cleanly in total darkness, before the strobe. If you eliminate overshoot space below black and above white, you will get increased ghosting.

We need panels which allow you to quickly (say a ms) drive all the pixels with a negative voltage before the screen refresh. The huge and intensity dependent swing would cause problems with a continuous backlight, but with a strobed backlight it would just put all the pixels in a state for which they only need positive overdrive to reach their desired state for the next strobe. You'd still lose some of the dynamic range on top, but you would no longer need to keep any margins on the bottom so you can hugely improve contrasts.

[Chief Blur Buster]

We need panels which allow you to quickly (say a ms) drive all the pixels with a negative voltage before the screen refresh. The huge and intensity dependent swing would cause problems with a continuous backlight, but with a strobed backlight it would just put all the pixels in a state for which they only need positive overdrive to reach their desired state for the next strobe. You'd still lose some of the dynamic range on top, but you would no longer need to keep any margins on the bottom so you can hugely improve contrasts.

This is already being done -- many panels are already doing this with overdrive. And the EIZO FDF2405W professional mapping monitor is doing a two-pass refresh (one heavily overdriven, followed by one non-overdriven to clean up the overdriven refresh) before the backlight strobe.

Also, there's an asymmetry between the LCD scanout and the all-at-once strobe, so you need non-linear overdrive which is a feature of LightBoost and ULMB (a Y-axis compensated response time acceleration algorithm). This is because of the fresher pixel transitions at the bottom edge relative to top edge, before the strobe. This complicates things quite a bit too. Also, it's really challenging to make all colors transit the correct color value all simultaneously, as different GtG combinations have taken very different amounts of time, and even as panel makers worked towards making more of the GtG transitions equally fast when they released 3D compatible panels (which makes it necessary to refresh all color combinations as quickly as possible, without any dramatic outliers).

Imperfections in GtG transition completeness before the next strobe, shows up as 3D crosstalk (during steroscopic 3D) or strobe crosstalk (during motion blur reduction 2D strobing). It's getting fainter and fainter in a lot of the good strobe backlights, but still a problem in many.