3D Vision Blog

A normal user's look into the world of 3D Stereo Technologies

3D Vision Blog header image 2

ViewSonic VX2268wm has Just a Bit more Input Lag than a Good CRT

November 8th, 2009 · 17 Comments · Other S3D Tech

analog-splitter-input-lag


My second video splitter, this time compatible with analogue displays arrived so I took the opportunity to connect the ViewSonic VX2268wm along with the good old Samsung CRT to a single video output of the card in order to test for input lag. As I already said in the article Samsung 2233RZ and ViewSonic VX2268wm are With the Same Input Lag and based on my previous experiences, testing with two displays connected to the two outputs of a video card and using a clone mode does not give accurate results – they simply vary too much! So in order to compare displays and have an accurate judgment on the presence or lack of input lag you must use a passive video splitter that is capable of splitting a single video signal to two displays. If you try to use an active splitter (with its own electronics and needing external power) you can again get inadequate results as this additional hardware may or may not introduce input lag too…


crt-lcd-input-lag-1


So, now lets get to the point where I compare the 120Hz LCD and the CRT display, using the analogue splitter cable. The following pictures were taken with 1/1000th of a second shutter speed, so that the camera can register even a 1 millisecond delay between the two monitors. At first look the numbers on the counter look the same on both displays, which is good, but the second and more detailed look reveals the fact that the millisecond counter on the LCD does not show absolutely clear numbers as on the CRT.


crt-lcd-input-lag-2


crt-lcd-input-lag-3


crt-lcd-input-lag-4


crt-lcd-input-lag-5


As a result of this test I can conclude that in fact there is a very little input lag on the 120Hz LCDs like ViewSonic VX2268wm compared on what you get on a good CRT display, but still there is some, even if it is very little. Now I’m talking about a very minor input lag and by that I mean something like 1 or 2 milliseconds at most, but then again we can also consider that the “ghosting” of the numbers on the LCD can also be caused by the slower response time of the pixels – taking pictures at 1/1000th of a second while the response time of the pixels is at 3-5/1000thof a second. But anyway to answer the question if a 120Hz LCD display is a good alternative to a good CRT display I can already say they perform as close as possible in the critical areas like input lag or the lack of such, but the LCD also has some clear advantages over the CRT technologies. So if you consider replacing your gaming CRT monitor you should definitely go for a 120Hz LCD, especially if you also consider using this display together with 3D Vision for viewing stereoscopic 3D content.

Tags: ······


17 responses so far ↓

  • 1 Fredric // Nov 9, 2009 at 23:59

    A million thank yous for performing this important test. Being an old CRT junkie (I now run a SONY GDM-FW900, and I own three of them :-)) this is as far as I know a breakthrough in the LCD market. The “ghosting” you see in the pictures above must come from the response time just as you say, if it would be input lag you would not see the exact same numbers on both screens, which you do above. This means <1 ms input lag in the Viewsonic, impressive. Of all shots you took, did you at any time see any difference greater than 1ms?

    Thanks again!

  • 2 Bloody // Nov 10, 2009 at 02:06

    No, no time difference in a bit over 100 pictures I took. Jut the response time creating the ghosting effect of the numbers on the LCD on all of them…

  • 3 satan // Nov 10, 2009 at 22:28

    really thx for the test. i have the monitor till 20/10/2009 and i can confirm that playng quakelive i see no difference froma a old crt…

  • 4 Just Skilled // Apr 17, 2010 at 23:22

    Hi!

    I’m searching already for 10 hours for 24″ LCD with no input lag, and I found this review very interesting even we are talking about 22″. I like it, but now I’m really confused. You are using 1/1000 shutter for testing input lag, and in all pictures input lag is 0.

    But, then I also saw this reviews:
    http://www.digitalversus.com/article-358-6823-41.html
    http://prad.de/en/monitore/review/2009/review-viewsonic-vx2268wm-part13.html

    Where guys claiming an input lag 14 and 20ms.

    Where is the catch? What to buy now?

    Thank you

  • 5 Bloody // Apr 18, 2010 at 02:02

    Well, I’ve had similar results when using a cloned image with both of the displays connected to each of the VGA output ports, but that is just not the best way to measure input lag. After using a splitter to drive both displays from a single output the results were different as you can see.

  • 6 Just Skilled // Apr 19, 2010 at 18:32

    If I can maybe ask you for your opinion…
    VX2268wm or Samsung PX2370 ?

    I only care about input lag and ghosting, so I search for best options to get new LCD, to put it near to mine HP LP2475w

    Thank you

  • 7 Bloody // Apr 20, 2010 at 10:24

    Haven’t seen this Samsung and have not tested it, so it is hard for me to compare…

  • 8 onizu // Apr 25, 2010 at 00:32

    The Samsung PX2370 is LED ! I wonder how LED compares to a CRT when not in it’s native resolution. Would it become blury like any LCD or stay cristal clear like any CRT ?

  • 9 onizu // Apr 25, 2010 at 00:37

    Bloody, first off, thank you for this review! Very usefull to me.
    Did you try to see what the input lag does when the VX2268wm is not in its native resolution compared to that CRT ?

    Thanks

  • 10 Bloody // Apr 26, 2010 at 11:21

    onizu, with the so-called LED monitors it is just that the backlight is different from the traditional CCFL one, the panel is still LCD. So in the end the LED backlight can improve on some things like colors, but can’t help you get better clarity of the image when you are not in native resolution (LCD technology related issue). I haven’t compared non-native resolutions with a CRT, but the idea is good, so I will try it… thanks for the suggestion ;)

  • 11 Jon Austenaa // Jul 20, 2010 at 23:13

    Hi what cpu, os, graphics card and drivers did you use? If you repeat the test by using two video outputs instead of one maybe you get similar results as the other reviews?

  • 12 Jon Austenaa // Jul 20, 2010 at 23:53

    Ah i guess you already have. I will inform others about this detail, to test the input lag at same output on the graphics card.

    Also looking for a 24″ 1920×1200 120hz display. 19:10 os best for computer use, especially with Windows 7 and its doublesized task bar. Old games work better in 1200 vertical then 1080, not to mention RTS gaming.

  • 13 Allen Pestaluky // Sep 8, 2010 at 17:02

    Hey dude, thanks for the test! This looks like the best method used yet, but I would like some clarification on the splitter that you used:

    First, could we have a model number for this splitter?

    Second: I assume that this splitter works by essentially creating a “DVI-I Dual Link” and a “VGA” connection based off of the “DVI-I Dual Link” port on your graphics card. This means that it’s essentially acting as a DVI-VGA adapter AND a DVI passthrough /at the same time/.

    This idea of splitting seems interesting to me… To be truthful, the biggest surprise is that your graphics card was OK with you using both the analogue (VGA) AND digital (DVI) signals at the same time! I guess it’s obvious from the tests that it /did/ work, but I assume that the hardware path taken to create the signals for the DVI and the VGA are different (specifically, the VGA signal needs to have passed through a DAC)…

    To go even further with this, if the analogue has a more complex path (needs to go through the DAC), then shouldn’t the VGA actually have MORE lag than the DVI? Is there a way of measuring the DAC lag?

    Random thought about assumptions you made: I expect that we cannot actually assume that all graphics cards will pull from the exact same frame in the frame buffer for both the digital and analogue signals of a DVI-I as you did during this test — But, that said, because you saw the same exact numbers (frames) in all pictures and the refresh rate and resolution were the same, you can probably assume that the graphics card is reading from the frame buffer /and then/ splitting that frame to the digital and analogue paths that are used for the DVI-I. Again, reinforcing that the VGA /should/ be slower than the DVI because of a more complex path after the split that happens after reading the frame.

    Aaaaactually, here’s a cool thing I noticed about the first picture:

    Note how the bottom of the numbers HAVE NOT YET DRAWN on the CRT, but they are completely there on the LCD!!! This is probably caused by the DAC lag that I was talking about. MAN! That LCD just beat the CRT because of that DAC lag! That’s NUTS!

    Can you try to capture some more shots where the frame can be seen on the LCD before the CRT like the first picture to verify these results?

    Really cool test, it’s probably the best one that I’ve read yet. :D

  • 14 Allen Pestaluky // Sep 8, 2010 at 17:55

    Actually, sorry, one more comment: it would be very good to see what graphics card you are using with that splitter, as it seems that there may be many graphics cards that do not support both analogue and digital output from the same DVI-I port at the same time. (http://sewelldirect.com/gefen-dvi-dvi-and-vga.asp)

    Also, since it seems that we have found a situation where the graphics card’s DAC could be the limiting factor, it would be good to know which graphics card you are using so we can compare with other tests that use different graphics cards and see if there’s a drastic difference between lag introduced by different DACs (well, most of todays cards probably use the same DAC component I would expect, but whatever). Anyway, I guess at this point it would be great to talk with a hardware guy at ATI or NVIDIA to see how the DVI-I port works.

  • 15 nksq // Sep 10, 2010 at 15:51

    Not even close to scientifically rigorous, as you hopefully know.

    It would have been more interesting if you included a “standard” LCD as well, so as to debunk everyone who thinks there is a measurable performance different between “120hz” LCDs and “60hz” LCDs. I put the hz in quotation marks because if you read about it, you know that LCDs don’t quite have refresh rates in the typical sense.

    Anyway, if you’re bored, plug in two LCDs and let’s see the difference there. Until then, don’t recommend a specific LCD without comparing the others.

  • 16 Allen Pestaluky // Sep 11, 2010 at 22:17

    Hey again,

    Just took a second look at the screenshots: I was very wrong about my earlier DAC-lag assumption. I noticed that the previous frame on the LCD that can be seen in the ghosting is 31ms earlier: This would prove that your software is actually drawing the same frame three or four full times instead of drawing a new frame every time. (505-474 = 31ms = a little less than 4 frames at 120Hz) A screen update should have happened 8.33ms before the current frame, not 31ms before, if the monitor was running at 120Hz. This is what gives you the result of the LCD being “on par” with the CRT in terms of lag.

    In fact, the CRT is actually drawing that same frame /a second, third, or fourth/ time, and it can be fully seen on the LCD because it’s already been drawn during the previous frames. The ghosting we see on the LCD of that screenshot is most likely not the previous frame, but is actually TWO OR MORE frames ago.

    Here’s a good article on input lag testing procedures: http://www.prad.de/en/monitore/specials/inputlag/inputlag.html

    I think your use of that passive splitter with your specific graphics card could be a good contribution if you have a tonne of different monitors to back it up — but only if you were to use software like this or by running 3D Mark or something that actually updates every frame:

    http://www.prad.de/new/monitore/testsoftware/smtt.html

    I’m very certain that the ghosting in your images totally disproves the zero latency results: Since the same timer numbers are being drawn more than once, then all the data is actually very misleading: you should make an update to let people know that this test is not valid. Sorry to bear the bad news!

  • 17 RocKien // Nov 16, 2010 at 17:56

    XL2410T vs VX2268WM , 2233RZ

    Input Lag test

Leave a Comment

Current ye@r *