[Scanning Backlights] About Full-Array Backlights and Strobe Crosstalk

[ClarinD]

So I had a thought the other day...

Would it be possible for a monitor with full-array backlight to switch on and off each row of LEDs sequentially in sync with pixel transition in order to minimize strobe crosstalk? Sure it's nowhere near CRT levels of clarity but would still be a lot better than any strobed LCD on the market right now. Seems to me like any monitor with both backlight strobing and local dimming features should, in theory, already be capable of this.
Any reason manufacturers haven't made this happen? Would love if Chief someone similarly knowledgeable could enlighten me.

[Chief Blur Buster]

So I had a thought the other day...

Would it be possible for a monitor with full-array backlight to switch on and off each row of LEDs sequentially in sync with pixel transition in order to minimize strobe crosstalk? Sure it's nowhere near CRT levels of clarity but would still be a lot better than any strobed LCD on the market right now. Seems to me like any monitor with both backlight strobing and local dimming features should, in theory, already be capable of this.
Any reason manufacturers haven't made this happen? Would love if Chief someone similarly knowledgeable could enlighten me.

Excellent question!
(which is worth asking again and again once in a while -- it's our specialty)

I'm moving this to Area 51 Display Science, Research and Engineering as my answer will be excellent refreshers for newer audiences unfamiliar with Blur Busters history, we were born because of this (Arduino Scanning Backlight Experiment).

Blur Busters used to be called ScanningBacklight.com in 2012

We still have www.scanningbacklight.com which redirects to Blur Busters. I created a Scanning Backlight FAQ (Year 2012) that explains a lot of this, as well as Electronics Hacking: Creating a Strobe Backlight (Year 2014) that explains a lot of the problems with strobe crosstalk.

The Short Answer: Internal Backlight Bloom Was a Major Problem

Scanning backlights are MUCH more bloom-sensitive than local-dimming bloom, you need much, much, much better bloom control (light bleed from ON section to OFF section).

In a FALD picture, the backlight OFF areas also has LCD OFF (panel black). But in a strobed picture, you can have an OFF backlight behind non-black pixels (including fully white). That means easier to see fainter bloom, because pixels are not also simultaneously OFF. Because of the contrast ratio between LCD black and LCD black, you literally need between 10x-1000x better bloom control than many of today's FALD, to reduce strobe crosstalk.

Even back in the CCFL tube days (multiple fluorescent tubes flashed in sequence) -- this has been attempted many times but the problem is that old scanning backlights in the past was prone to internal backlight scatter which worsens strobe crosstalk.

Global Strobe Is Historically Superior In Less Crosstalk Than Scanning Backlights

Global flash backlights completely sidesteps the "bloom creating strobe crosstalk" problem, at the penalty of requiring LCD GtG to be hidden in the blanking interval's time interval (on modern panels, this is now actually relatively easy to do at ~120Hz while running on good 240Hz fast-GtG LCDs, thanks to the Refresh Rate Headroom Trick). Some GtGs are a bit slow, so that's why there's still crosstalk at the top/bottom screen edges.

However, Denser FALD At Thousands Of LEDs Will Help

Putting LEDs much closer to the LCD glass, and blocking diffusion between the LED rows, helps bloom significantly.


(Credit: billburrbuster from this thread)

Future MicroLED FALDs will keep bloom in much better control, allowing less strobe crosstalk during scanning operation. Also, it raises costs dramatically since you need a full array backlight. Most full array backlights are local dimming optimized rather than strobe-timing optimized, so they are sometimes unable to be correctly programmed to scan.

However, I expect a re-emergence of scanning backlights when we reach thousand-LED-count MicroLEDs, since a huge number of LEDs much closer to the LCD display surface, means much better control of backlight diffusion. The backlight controller will have to be upped in precision, to allow real-time control of LEDs at extremely precise times (sub-refresh timing at less than 1 microsecond precision backlight timing per row required)

These will still require LCD GtG to be faster than a refresh cycle (even if they no longer need to fit in the VBI), so strobe crosstalk might become bad again at 360 Hz, even if scanning can help 120-240Hz. In all situations, the least-crosstalk-area of a global strobe (e.g. middle of the screen) is almost always superior in strobe crosstalk versus any part of the screen of all scanning backlights. On the other hand, the strobe crosstalk of a scanning backlight is generally exactly identical for top/center/bottom.

Also, scanning backlights also eliminates the problematic latency behaviors of strobe backlights, so that's another bonus for esports -- reduce strobe latency. However, blur reduction by brute frame rate is actually superior, since ~500fps+ at ~500Hz+ with 0ms GtG looks like LightBoost/ULMB at default pulse widths without needing a strobe backlight. 2ms MPRT is achieved either by 2ms flashes (strobe) or via 2ms unique-frames refresh cycles (500fps 500Hz) with exactly the same amount of motion blur (assuming GtG wasn't bottlenecking). Don't forget GtG and MPRT are two different pixel response benchmarks.

So as long as ultra high Hz and the GPUs are cheap enough to generate that, then strobing/scanning backlights is obsolete for that material. But, the problem is legacy frame rates, like 60fps and 120fps material. Those would still hugely benefit from a scanning backlight.

Related Reading In Other Historical Area 51 Articles & Threads

• Article: Blur Busters Scanning Backlight FAQ (Historical 2012 Article)
• Article: Electronics Hacking: Creating A Strobe Backlight (Historical 2014 Article)
• Thread: Blurbusting Technology Suggestion: Scanning Backlight
• Thread: Global BFI or Rolling BFI, Which Is Better? (Explains Strobe Latency Behaviors)
• Thread: Refresh Rate Headroom Trick For Reducing Strobe Crosstalk (Cram LCD GtG hidden in VBI)
• Thread: Global Refresh Methods (Background Information Of Why Top-To-Bottomk Scanouts Occur)
• Other Blur Busters Area 51 Display Research

[ClarinD]

Wow... I asked and Chief delivered! Thanks Chief! Time to go on another reading rabbit hole.
Now, the Neo G9 apparently uses 2048 LEDs, the days of scanning backlights must be close