GtG vs BtB

[TheInfernal]

Which is slower, GtG or BtB?

According to this article: http://www.xbitlabs.com/articles/monitors/display/lcd-monitor-buyers-guide-spring2010.html

Home-oriented TN monitors fall into two large categories in terms of response time: 2 milliseconds (GtG) and 5 milliseconds. They differ sharply because the former category is equipped with Response Time Compensation whereas the latter is not. If measured using the same method, the response time of the 5-millisecond models is going to be 12-15 milliseconds (GtG). The real difference is very large as you can see.

However this contradicts with the specifications of these monitors, which say "5ms (2ms GtG)", implying that BtB is slower than GtG:
http://id.benq.com/products/specificationsPrinter.cfm?product=1457
http://docs.google.com/viewer?a=v&q...pNOI5J&sig=AHIEtbTNqUKBT0HiuwE6WeZ67hOroelqtw

 

[GuyClinch]

^^^took me a while to figure out what you were thinking..

Those are 2ms monitors you linked (2ms GtG). The xbit labs are talking about how in the real world monitors with the 5ms GtG are much slower then 2ms GtG. The BtB number would be slower too.. It would be like 13ms BtB and 5ms GtG. Some monitors only list the GtG numbers Xbit labs is just letting you know - don't be fooled by the 3ms difference. It's hype.

If you read the Xbit labs reviews this will all become clear. The thing is the so called 5ms monitors have all these slow transitions everywhere - whereas the 2ms ones with RTC have alot of faster transitions.

The 2ms to 5ms makes it seem like there is only a 3 ms gap - but its much larger for other transitions so in reality the 5ms monitor is much slower.. The worst actually are the old PVA montiors which frequently had 8ms GtG transistions listed but had these HORRIFIC slow transitions for other color switches.

Anyway a so called 5ms monitor might have a graph with many transitions up near 20ms. Whereas many of the 2ms ones have none over 8ms. None of this applies to your monitors - as they are both fast...with 2ms GtG.

 

[TheInfernal]

You are saying BtB is slower than GtG. If I understand correctly, xbitlabs is saying the exact opposite.

You said:

It would be like 13ms BtB and 5ms GtG.

The xbitlabs review says:

the response time of the 5-millisecond models is going to be 12-15 milliseconds (GtG)

In another part of the article:

For 2-millisecond models, the average response time for all halftones is measured (the so-called GtG method). For 5-millisecond models, the white-black-white transition is measured only. If a 5-millisecond monitor is tested using the GtG method, its response time will be 13 to 15 milliseconds, which indicates its real-life speed.

 

[GuyClinch]

Hmm I hope this post clears it up:

What you see on most monitors at places like newegg is monitors with just 5ms stated (its not clear what kind of transition this is) and 2ms monitors marked (GtG).

What Xbit labs clearly says is if you measure these so called "5ms" monitors in the same ways you measure the 2ms (GtG) ones the speed of the 5ms (unknown transistion) ones jumps to 12-15ms (GtG).

http://www.newegg.com/Product/Produ...5&cm_sp=Cat_Monitors-_-Hot_Deals-_-24-009-255

For example this monitor is just listed at 5ms. We don't know what kind of transition that is. These so called 5ms panels often have much slower transitions when you look at the GtG numbers.

Whereas a regular 2ms monitor is usually listed like this:

http://www.newegg.com/Product/Produ...sp=Cat_Monitors-_-Month_End_Sale-_-24-236-102

The 5ms people aren't using ISO standard Black to White to Black (this would take a long time). They are using WHITE TO BLACK. This is how they pull up that 'better then GtG" number. They are just using half the official transition.

So yes for actual properly labeled monitors the black to white to black transition (ISO standard) would be slower then GtG. But most manufacturers use neither and kind of cherry pick a really nice transition. Xbit labs mentions this in their article somewhere....

Thus neither party is wrong here - its just some manufacturers are more misleading then BenQ..

Like I said this becomes clear if you read Xbit monitor reviews. You wonder how the hell are these 5ms montiors really 5ms!? I think the answer is they cherry pick a non-standard transition..

 

[elvn]

I always figure the real color transitions in actual usuage (including blacks and whites) are considerably higher than even the g2g numbers. Lately I've been posting in a few 120hz threads and I'd be really interested in knowing what the real response times are on the popular 120hz monitor models when gaming.

Pixel response times need to be below 16.67 milliseconds in order to fully represent the bandwidth of color changes necessary for 60 Hz video. However, even when this response time is achieved or surpassed, motion blur can still occur because of the least understood blur effect: eye tracking. LCDs often have a greater motion blur effect because their pixels remain lit, unlike CRT phosphors that merely flash briefly. Reducing the time an LCD pixel is lit reduces motion blur due to eye tracking by decreasing the time the backlit pixels are on. [2] However, an instant strobe is required to completely eliminate the retinal blurring.

..

.
..Yes at 120fps or more it (the panel/pixels) would be attempting to transition with new information very fast, once per hz (8.3ms).. at 60fps the screen pixels might not have to change for every 2hz 16.67ms, and anything in between would be some other mix of repeated frames I imagine. So fully utilizing 120hz to get new action information displayed in every update by maintaining 120fps or more could actually make the blurring worse potentially - depending on the real response time which could cause it (the panel/pixels) to be unable to keep up with the high rate of screen updating (when each update is new action). 60fps might theoretically update cleaner if it's only changing the pixels every two hz/ updates, which is pretty much like frame doubling which makes movies and animation appear smoother. This might help aim slightly with 1/2 the blur - but is not making the action more precise for gaming in regard to more current/up-to-date action being displayed like a crt at 120hz would at 120+ fps. Maybe 120fps+ vs the 120hz screen updates on 120hz LCD's would blur worse and/or start missing displaying frames out of the 120fps of action happening. I'd be interested in finding that out, as well as the real response times on the current 120hz monitors.
....The response times and blurring in general seem to be a big problem and are one of the main reasons I have my doubts about getting the most out of 120hz.

 

[TheInfernal]

The 5ms people aren't using ISO standard Black to White to Black (this would take a long time). They are using WHITE TO BLACK. This is how they pull up that 'better then GtG" number. They are just using half the official transition.

This contradicts with the article:

For 2-millisecond models, the average response time for all halftones is measured (the so-called GtG method). For 5-millisecond models, the white-black-white transition is measured only. If a 5-millisecond monitor is tested using the GtG method, its response time will be 13 to 15 milliseconds, which indicates its real-life speed.

Also you still haven't explained why monitor specifications say "5ms (2ms GtG)". They should have been "0.5ms (2ms GtG)" if what you or xbitlabs were saying were the case.

 

[XoR]

NOTHING IS MEASURED

5ms is for TN without RTC and 2ms is for TN that have this feature.

6ms is for PVA with RTG second generation that doesn't "smear" that much (but have one to three adidtional frames of input lag) and 8ms is for PVA with first generation RTG that have terrible black smudges (but don't have input lag).

IPS have 6ms most of the time as they are somewhat slower than TN without RTG and that's the reason for it's 6ms response time.

ps. if it was measured thet we would have more numbers used than 2, 5, 6 and 8 for GtG
ps2. we see blur cos maybe gray to gray transitions are fast but say red to blue is god damm slow and GtG doesn't cover that

 

[Sycraft]

All depends on the display and how it is set up. Back in the day, usually black-white-black was the fastest transition. That is why it was used as the standard measurement, put the best foot forward and all that. Any other transition was at best the same or often slower.

However overdrive often changes that. It means that some of the smaller transitions can be faster hence "gray to gray". Manufacturers started listing that since it was faster. There's no standard for what points they are talking about, it is basically just whatever happens to be the fastest.

Pixel transition time is not constant and to try and graph it would be complex and require a 3D graph as it varies based on what value you are starting at and what value you are ending at. So you could very well have a monitor that is, say 6ms for 0-255-0, 2ms for 0-50-0, 8ms for 50-255-50 and so on.

Can be far more than that too. For example particularly older PVA/MVA monitors had really slow response times for darker transitions. So you'd have a monitor that could do 12ms for 0-255-0 and even as low as 6ms for something like 200-255-200, but could be 100ms+ for 0-10-0. They never listed those slow dark transition times, of course.

Now for those wondering why it is always black, white and gray not colours that is because the subpixels themselves, what actually changes and response, have no colour. That is done later by filters, so it makes no difference in response. Any colour responds the same since it is just a RGB filter over the subpixles.

Basically response time is a highly BS'd up stat these days so you can't really look at the number and get much. These days you can be assured that any modern panel, regardless of the tech, tends to be plenty fast enough for 60Hz. Past that, you just have to see the screen in action to see how you like the picture (or read good reviews). The speed won't tell you much, particularly since you can get artifacts from a high speed monitor that has too much overdrive or the like.

There's not a good single number out there to measure how a display is likely to perform, unfortunately.

 

[elvn]

Could be nice if they spec'd something like...
.
Response Times -- Most Common:5ms, Common Best:2ms, Common Worst:18ms
.
.

Basically response time is a highly BS'd up stat these days so you can't really look at the number and get much. These days you can be assured that any modern panel, regardless of the tech, tends to be plenty fast enough for 60Hz. Past that, you just have to see the screen in action to see how you like the picture (or read good reviews). The speed won't tell you much, particularly since you can get artifacts from a high speed monitor that has too much overdrive or the like.

It might be fast enough for 60hz, but I've been looking at 120hz as I showed in my last reply, and I suspect that response times probably aren't fast enough if you are utilizing all 120 screen updates of the 120hz.
.

at 120fps or more it (the panel/pixels) would be attempting to transition with new information very fast, once per hz (8.3ms).. at 60fps the screen pixels might not have to change for every 2hz 16.67ms, and anything in between would be some other mix of repeated frames I imagine. So fully utilizing 120hz to get new action information displayed in every update by maintaining 120fps or more could actually make the blurring worse potentially - depending on the real response time which could cause it (the panel/pixels) to be unable to keep up with the high rate of screen updating (when each update is new action). 60fps might theoretically update cleaner if it's only changing the pixels every two hz/ updates, which is pretty much like frame doubling which makes movies and animation appear smoother.This might help aim slightly with 1/2 the blur - but is not making the action more precise for gaming in regard to more current/up-to-date action being displayed like a crt at 120hz would at 120+ fps. Maybe 120fps+ vs the 120hz screen updates on 120hz LCD's would blur worse and/or start missing displaying frames out of the 120fps of action happening. I'd be interested in finding that out, as well as the real response times on the current 120hz monitors.

There will still be this blur issue regardless as well apparently....

Pixel response times need to be below 16.67 milliseconds in order to fully represent the bandwidth of color changes necessary for 60 Hz video. However, even when this response time is achieved or surpassed, motion blur can still occur because of the least understood blur effect: eye tracking. LCDs often have a greater motion blur effect because their pixels remain lit, unlike CRT phosphors that merely flash briefly. Reducing the time an LCD pixel is lit reduces motion blur due to eye tracking by decreasing the time the backlit pixels are on. [2] However, an instant strobe is required to completely eliminate the retinal blurring.

 

[Sycraft]

I wouldn't worry about it too much. In part because at 120Hz you are getting below the limits of human perception. That isn't to say there is no benefit, particularly since computers render discrete images, they are blurred like camera images, but it does mean that the ability to have every single frame displayed crisply may not be very important.

It would matter more for 3D, since you are rapidly switching images for eyes. That 3D seems to work well indicates that they probably are doing a reasonable job forming the images quickly.

Again, the thing to do is look at reviews, and then just check the display out for yourself. Don't worry about the stats, they aren't a big deal.

 

[elvn]

It would matter more for 3D, since you are rapidly switching images for eyes. That 3D seems to work well indicates that they probably are doing a reasonable job forming the images quickly.

... I see what you mean, and on the surface that would seem like a good indicator - but with 3D , part of the reason the frames might be isolated enough vs transitions could be the shutters on the glasses which have their own "hz" or timing. Who knows exactly how long the time window of the shutters being open is , and closed is.. vs the time pixels are in transition, and if the shutters have a transition period of their own. Perhaps the shutter is only open mostly during the "clean" part of the frame, etc. The screen updates could appear cleaner/more isolated due to the nature of the shutter vision isolation in general too.. and who knows - it could even help vs retinal retention blur. I really don't know but I'm just saying that the shuttering lens over each eye brings a whole other component into the mix in regard to isolating frames vs response times and blur, so it might not be a valid "apples to apples" comparison to 120hz forming screen updates cleanly/quickly enough when you aren't wearing the glasses.... especially when feeding the screen 120fps+ so that each screen update is a new frame of action.
..
.... Yes 3D must form a unique (at least offset) image during the 8.3 ms + 8.3ms of each 3D frame pair, but how clean that transition would without wearing glasses comes into question... vs 120fps+ ~ 120hz of high speed action gaming screen updates without glasses when gaming at 120hz at high framerates without 3D.