shmups.system11.org

The ERA where LED's are less than a half millisecond from CRT input lag is here.

1second = 1,000milliseconds
1/60th of a second = 16.67ms
CRT takes 16.67ms to refresh each FULL FRAME.
CRT's take 8.333ms to refresh halfway thru the screen's pixels.
This is how input lag is measured.

At a 4k resolution, at 60hz refresh rate,
Samsung's 75" LED dubbed "The Terrace" has LESS THAN ONE HALF OF A SINGLE MILLISECOND OF LAG BEHIND CRT's

This next calendar year, we will likely see flat panels operating at 120hz refresh rates, operating with LESS INPUT LAG than CRT's.
Many models will have an optional 8.333ms LOCK option in game mode soon.
And the ubiquity of this reality will only increase in subsequent years.

https://www.rtings.com/tv/tools/table/29642

The future may not be bright in some areas of life...
...but at least these bright colorful new flat panels have finally come up to proper speed.
Peace & best wishes everyone!

*sees icon* Poster of culture spotted. Must read whole post.

Thanks for the technology update. I paid good money for a PVM a year and a half ago and wondered how many years until LCDs could match CRT input lag. Living room in the open air next to kid with squirt gun not making me want to buy one though. CRT lack of motion blur and handling off-spec resolutions, is that all it has left? Not that I think PVM prices will crash any time soon but I don't see people growing up now wanting to play on a 20" screen they can barely lift. Similar concern with video game collecting but getting off topic.

Once LCDs beat CRT input lag, I also expect 8.333ms lock option. 480i/525i not going to work when television draws faster than console is outputting. That CRT spreads 1 frame of lag over entire vertical resolution, I assume LCDs draw much faster than that and most of the lag is in video processing and time to change colors.

There is also the notion of perceptible lag. I doubt is under 1 frame for most people.

It's still just an LCD television, though, not OLED.

I 1000% guarantee idiots are going to tell you they will notice the one-half millisecond difference because they're 1337 gamers who were Formula One race car drivers in their previous life lol

Lol! No like button available, but this is true.

Very few people are worrying about under a millisecond. Although, if you look at LCD panels with different G2G response times, you can see the difference between 1ms and 4ms response. So, you can perceive things that happen very quickly.

You can also boot up Kirby on the SNES. Turn on Kirby Super Star and play Samurai Kirby with a friend using a PC emu. Use a pair of DS4 controllers. Set one controller to 1ms polling and set the other controller 8ms polling. If your friend doesn't suck, you'll see an advantage for the person with the lower polling time.

This idea that "CRTs have lag" due to the 8.3ms measuremeant kinda rubs me the wrong way.

Yes, it's an important number to know, because it means that if an LCD display measures lower than ~16ms in 60hz mode, that means it is approaching CRT territory, but it's not the same as saying that the CRT technology itself adds any lag to what you are seeing. Any color transmitted from the source is displayed immediately on the screen, which more than qualifies as "lag-less".

Measuring a 8.3ms delay from starting a frame until you see pixels from that frame on the middle of the screen only tells us something about the way we perceive individual frames in a 60hz context, and how video games would usually handle rendering them (ie. update data for the video chip during vblank, and calculate the next frame while that frame is rendering), not counting raster effects.
So yeah, the number makes sense when it gives us something we can compare to other measurements, but it's not the same as CRTs actually lagging.

With that in mind you can get lower lag measurements than CRTs using a 120hz display rate, but that also implies a 120hz source.
If you're trying to display video from a 60hz source, like an upscaled video game console, it's not going to do you any good, you can't just show one frame twice, as that frame still takes the same ~16ms to render from the source.

It has been quite a while since lag has no longer been one of the reasons to avoid flat panels for gaming, for me at least. It's about other reasons.



Yep, it's mainly about these. Though OLEDs do not have motion blur, they still have persistence blur, and commercial panels do not have fully satisfactory solutions for it. Relative to LCDs, you'd also have to add true blacks/dark scene handling to the list. I'd also add true representation of SD and HD interlaced video.

But that's probably about it. I'm finding that a GBSControl + OLED can trim that list down to a single item, and the OSSC Pro likely will do even better. That said, if you're used to the color accuracy and quality of a PVM or BVM CRT, you may only be 100% satisfied with one of those devices combined with an OLED equivalent...

There's still the issue of LED off->on response times. A CRT can get excellent black levels by simply not drawing to those parts of the screens. With something like OLED, the LEDs take time to go from off to on. It might not be super noticeable on a TV, but I notice this all the time with my Pixel 3a's OLED--I'll have a chunk of black on the screen, then scroll, and it'll look like the black section takes longer to move up the screen compared to everything else.

Not sure what may be going on with the OLED in Pixels, but at least the OLED panels by LG as well as the pro panels that Sony used to make are advertised as having near-instant response time, which is what eliminates motion blur (again, not to be confused with persistence or sample-and-hold blur). This fast response time does have a disadvantage in causing visible stuttering during slow panning shots in low frame rate/film (24fps) video content, though I think it can even be a bit of an issue on higher frame rates. Black frame insertion, strobing, or rolling scan may help to alleviate this to some extent.

I don't like the look of rolling scan for film. I prefer Sony's motion flow interpolation. Sony got a head start dealing with fast response times from their high end projector line--and those projector guys are almost singlehandedly obsessed with how a projector handles film content.

Content framerate is the ultimate limit here. Yes - we are seeing a lot more gaming at 120Hz framerates (refresh rate is not the preferred term here) which will bring faster gaming with less lag than 60Hz CRTs can support. But at the same time there will still be lots of 60Hz content where flat panels have no latency advantage over CRTs. It is interesting to note that many obsolete CRT monitors were made with support for 120Hz and beyond. Indeed, it's probably hard to find a 90s or newer PC monitor which doesn't support at least 75Hz. Personally, I think that flat panels are more suitable for ultra high framerates and resolutions than CRTs. But if you feed both display types the same image, they will both be using the same raster-scan technology and both will still take 16ms to draw that entire single frame at 60Hz.

You could theoretically burst or chunk an image to transmit it more quickly and decrease lag, or start to draw lower parts of the image sooner. But just getting 4K out of a single flexible cable and economical hardware that doesn't burn the house down has required a lot of technology investment, and that cable is transmitting continuously (and this was true for every previous resolution and framerate adopted, too). [Edit: Wrong info due to QFT & VRR adoption, thanks orange808: We probably will see device manufacturers use additional bandwidth to instead use additional bandwidth to increase resolution, framerate, or color depth. But we likely won't see the adoption of a "less laggy" 60Hz option, in no small part because it would break compatibility with existing 60Hz content (and existing 60Hz devices couldn't make use of it at all).]

Other notes:
I don't like to use "120Hz refresh rates" when talking about input lag, because TV manufacturers like to label their counter-blur technology as 120Hz, 240Hz, or beyond. Applied to 60Hz content, the image will be clearer, but it won't be faster. It's essentially the same situation as classic movie projectors running 24Hz content - but using the shutter mechanism to flash the image two or more times for a 48Hz or 72Hz refresh rate or more.

As far as seeing 1ms differences in the gray-to-gray specifications for a monitor, this effect is more about wide differences in color purity than temporal perception. The brain and eye do some tricky things to adapt to fast moving content, but g2g blur can be spotted by devices too. I think it's probably about right that our brain operates at ~20Hz, but you can notice (well after the fact) if a color region has a fringe on its trailing edge from sloppiness in the gray-to-gray transition.

People are still calling raster time, the time it takes to draw a frame to the bottom typically @60hz, and display lag the same thing huh.

But I'll give credit to where its due, we're finally seeing huge panels with display lag comparable to "gaming" monitors.

With QFT in the HDMI 2.2 standards, I think it makes some sense. It's always complicated.

At some point, a QFT sample and hold display could be displaying a complete (top to bottom) "fresh" frame from a 60Hz source (that supports QFT!) 8ms after the source sent the frame information up the wire. I want rolling scan, motion clarity, and ~16 2/3ms of lag from 60Hz, but some would prefer to see the frame sooner.

OK, QFT is new to me . Judging from what the HDMI Forum says, and what Blur Busters says, QFT is actually transporting the entire frame much more quickly and just increasing the blanking interval to compensate. It's the "burst transmit" option. All that is needed is for the content source and display to agree, i.e., by using VRR tech. Using this you definitely could see much faster display of content that isn't at the limit of the cable bandwidth, such as 60Hz games. Per Blurbusters, It also improves 60Hz consoles which may appear even slower on 240Hz monitors because the signal has to be buffered for multiple scanouts.

https://forums.blurbusters.com/viewtopic.php?t=4064
also https://www.hdmi.org/spec21sub/quickframetransport

I'd argue that's not a question of perceiving things that happen quickly but a measure of how good you are at detecting contrast. Since video games are constantly changing what's on screen, a screen with a slow G2G will be blurrier and lower contrast wherever there's motion. That's what you are seeing, not how fast the pixels themselves are changing.

I like turtles.

In the end aren't we all playing games though? So splitting hairs about who is causing the lag doesn't change the fact that when playing on a CRT the experienced lag is 16ms (at the bottom of the screen). And keeping in mind the inherent lag caused by transmission helps interpret the significance of any differences in the lag between display hardware (ie a difference of 4ms is swamped by the inherent lag of transmitting the signal at 60hz).

As far as I know, "instant response time" only refers to changes while the LED is on, and does not include going from an off state to an on state, nor going from off to full brightness.

If you have any data that shows that that isn't the case--that these panels have LEDs that can instantaneously go from completely off/black to full brightness instantaneously, please share it.

CRTs have a lot of lag in the "white->black" direction. Which is to say a bright object on a dark background leaves a noticeable trail. I forget the hard numbers but it's several frames before white phosphors decay completely to black. The effect isn't notable on anything other than a very dark background though. I recently measured this with my piLagTesterPRObecause I wanted to validate the lag measurement against a known quantity, and here is what I observed on a pretty high end computer CRT (a NEC Diamondtron, the last good CRT commercially available in the early 2000s). Black->white took about 1ms, and white to 90% black took about 4 ms. That last 10% doesn't sound like much but against a fully black background it's quite visible and the decay rate steadily decreases; so 100% black takes "forever".

Yeah, back when I had a CRT for a computer monitor, I remember the trails that bright items would leave behind on dark backgrounds, so I was always a bit confused why people used to claim that CRTs had this super amazing response time. I mean, the whole concept of how CRTs work is that they have significant phosphor persistence.

Last few posts seem to be referring to display lag, but this topic refers specifically to input lag, which is only partially related to display lag issues specific to flat panels.

What can be said for sure is that no one ever feels like actions they are taking on a controller don't feel as responsive as they should be when using a CRT.

But, it's more than the trail behind images on LCD, isn't it?

I am not seeing contrast. That isn't it.

How do I know that? Ghosting. Superficially, foreground objects fade in and fade out.

We call it "ghosting" because moving objects appear to become phantoms. Ghosting is a direct result of slow response time and it affects every change that happens on the panel. Contrast causes the issue by pushing the LCD pixels beyond their abilities. The root of the problem is the slow LCD response time--and it affects more than the trail behind a moving object.

Worst case scenerio: an entire moving object in the foreground will appear to be transparent--an average of the object color and the background color.

Let's imagine a solid color foreground object and a contrasting background. As the foreground object moves faster across the screen, more of the "leading" object surface area will appear transparent. At a certain speed (on slow LCD) the entire object will be transparent. We could take a picture and it would be right there exactly as we "see" the artifact. If we toggle the entire object randomly across the screen without repeating any background surface that was recently occupied by our foreground object, our foreground object will never appear solid at any time! It can't be done.

Is that me seeing contrast? Well... no. It isn't.

So, why do moving objects appear transparent on a slow LCD? Because the LCD pixels don't have time to reach the color in the source signal. They need time to change. Everything appears to fade in and fade out. I can see it, it happens very quickly, and it isn't a contrast thing. The only reason contrast matters is because the slow LCD panel shows it's weakness more clearly when there is a great deal of difference between the color of the previous frame and the current frame--and that stresses the response time beyond the limitations of the panel; when that happens, the result is: ghosting and motion blur trails.


----

Brighter parts of the lines on a CRT will glow a little longer. So, the blank time before the update may not always appear to be uniform. However, the line will be blank before the next update. Foreground objects don't appear to fade in. I don't see them fade out, either. To my eye, they just have a soft glow. (Take that phosphor glow away and you can see scanlines playing Zanac on a Sony OLED.)

(Sorry for all the edits: it's a high effort post.)

I thought input lag is about how fast the controller processes button presses and how fast the console and game react to that. Is that not the case?

I would say you are correct; where we as gamers are concerned, between a button press and its corresponding action appearing on screen, "input lag" is how you describe it--the time from the button press to the frame being output from the console--and display lag is the time between the frame being received by the display to when the frame is displayed on screen, or perhaps the time from the frame being output at the console to when it's displayed on screen (total display lag of one's video processing chain).

Which is part of how interlaced video worked as well as it did. The relatively long phosphor decay rates, in combination with how human vision works, enabled interlaced video to work.

I've also noticed that PC monitors, HD consumer CRTs, and newer PVM/BVM seem to have faster phosphor decay and require higher refresh rates to not be headache-inducing for those of us sensitive to flicker.

A good example I've seen for phosphor decay in CRTs is the attract screen on Ms Pac-Man in a dark arcade, especially if it still has an original 80s CRT with slower phosphor decay, aka higher persistence.

Actually the phosphor persistence is too short to help with interlacing - with a 4ms persistence, that's 12ms of each frame where each pixel is dark. That was just the one CRT I tested, but if you take a picture of a consumer TV you can see that it's pretty much showing only one field at a time. The human visual system has much higher persistence, as long as you are over the flicker threshold (generally around 60hz) the apparent brightness is just the integral, so long and dark can be the same as short and bright. Actually, there's a slight amplification for short and bright because the visual system likes change (eg, robinson and de sa, 2008).

So, in fact I think it's the other way around: the short persistence is why interlacing worked. If you play an interlaced video on a LCD, which has massive persistence, it looks like shit, with weave artifacts all over the place. Hence the proliferation of lag-inducing deinterlacing algorithms in the LCD age. It may be that the amount of persistence in the human visual system is "just right" for making interlacing look ok, although I can see interlaced artifacts on my consumer CRT if I watch for them.

I have a lot of CRTs – bias alert.

That said, I picked up an LG CX48 this December and played around with the OSSC awhile. Black frame insertion is great, but it killed the brightness so much it wasn't much of an enhancement over even a consumer grade CRT. Also, without the black frame mode, I had a good bit of eyestrain.

It just wasn't as enjoyable. I still like it, don't get me wrong, but for older games – nah, not really.

Also, for me, input lag isn't much of a problem as long as it is consistent. So it definitely wasn't that for me.

Just to bring things back to the top, we've had this for a while. The LG C9 and CX, for example, uses a frame buffer, so input lag varies depending on the refresh rate. Having said that, it is a 120 Hz panel, so when feeding it 120 Hz content, you get 8.333ms (a single frame) input lag. This is all assuming you are using game mode, of course, but the point is just that we already have the technology.

Of course, computer CRTs supported 120hz for a full decade before their demise. The final CRT we had in our lab went to 240hz although that was a bit of a lie: the phosphor decay was too long for it to completely disappear in 1/240th a second, so, for instance, 3d shutter goggles had some ghosting from the previous frame.