Well, after 3.5 hours of testing various TVs, taking photos and noting details, plus another six hours of parsing data and playing with Photoshop, I'm ready to post. With respect to thepoohcontinuum, I'll be staying in this thread; half of my quest involved input lag and I'll be sharing my results. But I also figure anyone interested in input lag is probably a gamer - PC or console - and for such folks, the concern for accurate color reproduction can be just as significant as the concern for lag. I know it is for me.
Preface: 1. I had no tripod. I took these photos quickly and sloppily, trusting in PS to make up the difference. Also, my camera is substandard. It cannot focus closely enough to take a proper macro shot. Its LCD screen is too small to enable me to identify when my shots were out of focus. I tried to compensate by taking a LOT of photos, with some success. 2. All processing on each TV was disabled. 3. Many TVs failed to sync at 1080p60 and settled upon 1080i60. In most cases, this surely had a negative impact on input lag. Again, there's no explanation. 4. Input lag was gauged using the RefreshRateMultitool and averaging no fewer than five samples. Said samples exhibited satisfyingly little variability. 5. Laptop input lag is accurately accounted for.Samsung LN40C630
I returned to this model because I had managed to determine how to enable game mode. Again, this TV failed to accept 1080p60 and so defaulted to 1080i60. The original input lag result, without game mode on, was 108.2ms
. Here is what the modded test chart looks like with those settings:
As can be seen, there is no difference between the two shots. Without game mode, this TV displays 4:4:4 successfully. Also note the fact that the image has been deinterlaced excellently. I suppose that may be where the 108.2ms came from. Unfortunately, this TV exhibits a horizontal (if uniform) smearing of pixels, most notably red. Get used to this phenomenon, because if my experience is anything to go by, you will be hard pressed to find a TV that doesn't suffer some variant of it.
Enabling game mode provided a new result: 74.9ms
, a difference of 33.3ms
, or exactly two frames. A proper test with the TV successfully recognizing 1080p60 would help determine final input lag.
Here's the bad news:
Apologies for the particularly blurry photos. You can still tell that there are 4:2:2 shenanigans going on. Solitary red vertical lines alternate between bright and dim. It's not subtle, but it's not as bad as some other TVs.Sharp LC-40D68UT
The Sharp 42SB48UT surprised me by not only supporting HDMI 1080p60 @ 4:4:4 color (albeit smeared both horizontally and vertically) but also having a pretty low input lag. Really, the only thing missing (apart from, you know, a good panel) was 24Hz support. So I had high hopes for this 120Hz model. However, I had already suspected, based on the user manuals which I had pored over, that this TV wouldn't deliver.
Well, first we'll cover the input lag. I should mention that maximizing this TV's performance involved hunting down a remote, because there was a particular function which could not be achieved without it. Fortunately, one was lying around. The result: 20.3ms
. Plenty good; just what I expected. And here's what that looked like:
Yeah. When a not-so-high-quality panel is also subsampling the color, the results are not pretty.Sharp LC-46LE810UN
I decided that since there was such a significant difference between one of Sharp's cheapest models and the somewhat less cheap 40D68UT, perhaps it would be worth it to keep trying their TVs and hope for a good combination. Enter the 46LE810UN.
In spite of the decidedly unfriendly menu system, I discovered this TV's game mode. But first we'll cover the PC mode. Input lag: 111.1ms
. And it looked like:
Now, you may be noticing what I can remember being forewarned about. The pixel matrix is oriented diagonally. The results are just bad, and definitely not helped by the ubiquitous presence of the color smearing issue. Look at my checkerboard modification in the lower left. It's supposed to look different
from the vertical lines it's embedded in. For what it's worth, I judge that the TV is passing 4:4:4 color, though it defeats itself with the nature of its pixel matrix and the excessive smearing.
Enabling game mode gave: 32.6ms
. Well, that's more like it, although it's actually a tad high for a Sharp. Did the change affect the image?
Yes. It's an irony that the image was already so indistinct that the subsampling of the color to 4:2:2 can be said to have imposed only a subtle effect.LG 42LD520
I already put this TV through its paces. Keeping in mind that the TV defaulted to 1080i60, the input lag result was 56.5ms
. I was told to switch to ISF Expert 1, which I did. I was also told to try labeling the input as PC. This turned out to be impossible without a remote, and none was handy. However, it was said that doing so really didn't matter. The new input lag result was higher by about 1.3ms, so I'm ignoring it.
Anyway, here's what this TV looks like (keeping in mind that this is
an S-IPS panel):
What is it with LCD TVs and red, anyway? Look at that big block of red. That used to be vertical lines. Whatever. It should be obvious that this TV fails to pass 4:4:4 color. The phenomenon of alternating thickness is typical.
I need to point out that thepoohcontinuum kindly went through the trouble of performing a 4:4:4 color test on his own 32LD450 (see his earlier link) and his results clearly exhibit proper 4:4:4 color support. Conjecture: Perhaps only the 60Hz LGs enable such support, with the 120Hz models abandoning the feature as an unnoticed consequence of different hardware. This would be highly disappointing, but I intend to investigate, because another thing I noticed about that 32LD450 is a total absence of the all too familiar pixel smearing.Panasonic TC-L42U25
I tested this TV's input lag and found its performance to be a match for the TC-P42S2 (about 37ms
). The TV did not accept 1080p60. I consider things largely moot because of this:
Again we have a big red block where there should be vertical lines. The 4:2:2 subsampling is anything but subtle. Worse, look at the deinterlacing artifacts, such as on the number 2 on the right. This minimal deinterlacing is probably a big reason why the input lag I measured is about the same as what others have measured for 42" Panasonics without interlace woes. Anyway, I'm very disappointed in Panasonic, and sure won't be buying any of their 2010 models.Toshiba 40G300U
Firmly in the grasping-at-straws phase, we turn to a Toshiba. 1080p60 was not synced so we got 1080i60. Game mode off: 104.1ms
. Game mode on: 86.6ms
. I should point out that when game mode was enabled, the TV stopped deinterlacing completely, giving instead a very flickery image. The point of this observation is to suggest that 86.6ms may be what you'd get with 1080p60, assuming the TV can accept that refresh rate.
These input lag results speak for themselves. The upside is that regardless of what mode one used, or what the pixel offset was, the image looked like this:
I think of this image as being what the Panasonic 42" TVs would look like if they passed 4:4:4 color. Maybe this is what Panasonic's 2009 models look like.Toshiba 46G300U
Decided to try one more Toshiba, to see if there was variability between models like there is with Sharp. 1080i60 again. Game mode on: 90.6ms
. Looked identical to the 40".RCA 40LA45RQ
At the end of my rope, I tried out a 60Hz RCA (they had no 120Hz available). No game mode and very few actual menu options. Input lag was a respectable 28.7ms
After a whole day with displays that refused to show me the video I was sending them, I was more than a little pleased to see:
No subsampling, very little pixel smear, and a decent input lag result. Make no mistake, I was laughing.
At the end of a quite long day, I'm find I'm shaking my head over how ludicrous it is that if you want to find a TV that gives back what you give to it, you have to find something old or cheap. Look at those images. How hard can it possibly be to just serve up a faithful rendition of the input?
I will refer to this article:http://www.hardwarecanucks.com/forum...-review-9.html
It gives some game mode / non game mode shots, and particularly this one:
Now, the article makes the argument that the game naturally looks like crap unless the Samsung is allowed to perform its image processing, at the cost of the lag performance. Well, folks, I play on a humble CRT every day. It doesn't process the image. I am here to tell you that it looks fantastic. And I would have to judge that the reason for this is because
it's not modifying the image. Do you think their Samsung is just magically un-crushing those blacks and whites when the image processing is turned on?
Lastly, I have to thank thepoohcontinuum for rekindling my hope in a quest for a TV. After I left Sears, I was pretty well convinced that I'd end up having to make several compromises that would leave me unhappy in addition to $700+ in the hole. Now I know I have options remaining to me. The quest continues.
Note to De3k: They had no 32" LGs, unfortunately, and nothing below 42LD520 that I could find. Sorry.