CRT response time

Z

zzz

Gawd
Joined
Oct 9, 2006
Messages
801
Can anyone explain what's going on here? This is a picture of my CRT playing UT3 at 90fps where I moved the mouse quickly. I saw it first with my eyes, so it's not a camera artifact. I didn't think CRTs had this sort of response time shadow behavior.

f_dhxm_4cba4ae.jpg
 
It affects all types of displays. CRT's are just typically much better at it than other displays.
 
My first impression upon seeing this photograph is that it does not look like a photo of a CRT display. In a photo of a CRT, it would be a near certainty that your exposure time would not perfectly coincide with a full scan of the screen. The retrace would overlap at some point, so one rectangular region of the picture would be brighter than the rest. This effect is commonly seen in photos and video of CRTs.

I can think of two possibilities, assuming it is a CRT:

1) You took a 1/NN sec exposure which just happened to coincide perfectly with the NN Hz scan rate the CRT was set to, being so perfect as to be seamless. The game itself is doing some kind of internal frame rate conversion, similar to PAL->NTSC conversion in that does weighted averages of adjacent frames.

2) You took a long exposure, so long as to make the seam from the CRT vertical retrace appear invisible to the eye, and the camera exposure averaged the frames together.

Given that you saw the effect with your eyes, #1 is perhaps the answer. Maybe a second photo would show the characteristic CRT retrace seam, but would still have the averaged-frames look.

If you had posted this photo with no accompanying text, I would think it's a photo of an LCD with an exposure time long enough to average multiple frames together.
 
blackhand1001 said:
I think he has it set to output interlaced. Thats what it looks like.
Click to expand...
That still wouldn't explain why there's no retrace overlap in the photo. Even if you manage to match your shutter speed to the vertical scan rate, there usually are retrace artifacts.

zzz: Please post another photo that does not have the EXIF stripped, so we can see what the shutter speed is.
 
Last edited:
I think #2 applies, where my eyes and the camera are averaging frames together. It also seems like images take time to fade away, but I don't have a high speed camera on hand to check that. This effect in the game seems to be the same as simply moving the mouse cursor around on a black background; you see a trail of cursors (even at 170Hz).

Here's a couple more pictures. 1 even has a scan line visible. Unfortunately my camera is a Canon 1100IS that doesn't give manual control over the shutter speed. The specs say it does 1/1500 s but I'm way above that.
#1 (1/60 s)
#2 (1/20 s)
#3 (1/10 s)
#4 (1/8 s)

Do CRTs have zero response time, where each frame has no influence from the previous?
 
zzz, you can increase your ISO and/or lower your EV compensation to get a faster shutter speed.

Your 1/60 second shot strongly suggests that it's the game or video card doing the frame averaging, not your eye or the camera or the CRT. CRTs have very fast response time, much faster than any LCD. The image of one frame has almost completely faded to blackness even before the next frame is redrawn. The only time you'd be likely to see trails left behind by moving images is against a black background, in a dark room. The photo is not really conclusive though, especially without knowing what refresh rate you have your CRT set to and using a faster shutter speed.

How about taking a screenshot (with FRAPS or something similar) instead of a screen photo?
 
MetaGenie said:
Your 1/60 second shot strongly suggests that it's the game or video card doing the frame averaging, not your eye or the camera or the CRT.
Click to expand...
I was thinking the opposite. That 1/60 shot shows 3 buildings, and at the top you can see it says "115 fps". At 115fps, 3 images will be shown in a 1/60 second duration. Additionally, the shadowing effect gets progressively worse at 1/20 and 1/10, so that's clearly what's causing it as opposed to game averaging. Also, in-game screenshots show no averaging.

In all shots, the refresh rate was at least 118Hz. Once I tried 170Hz with vsync sticking it at 85Hz, and the effect looked the same.

I still wonder if each CRT frame is independent.
 
I assumed you wouldn't be using such a high refresh rate. Obviously I'm wrong, you're taking full advantage of your CRT!

Please use a high ISO or negative EV compensation on your camera and take some more photos... try to get the shutter speed faster than the refresh rate.
 
Last edited:
it could be u using a resolution that your monitor or videocard cant support. or the refresh rate is to high or low? crts have a 0 ms response time, dont have to deal with that at all i think, that's a lcd only, problem.
 
That's a good idea. Do both ISO and EV directly affect shutter speed? When I have a chance I'll try 1600 ISO, -2 EV, and the CRT at 60Hz.
 
Yes, they both directly affect shutter speed. Doubling the ISO will halve the shutter speed, and -1.0 EV will also halve it (-2 EV will quarter it).

I was thinking you should keep your CRT at a high refresh rate... unless you still see the frame-averaging effect with your eyes when the CRT is at 60 Hz?
 
Why is there no crosshair "blur" on the first picture in this thread?

This looks like a software problem.
 
Yeah, my bad, I thought the crosshair was also supposed to move instead of having a fixed position on the screen.

Still, this doesn't look like anything CRT-like. It's as if the game itself is trying to add some motion blur.
 
those photos look normal :/. Its hard to say without actually seeing it.

edit: see if you can get a bright enough photo at roughly the same speed as the refresh rate (around 1/115 ...)
 
CRT's DO have a response time, but it's so small that it's virtually undetectable. What CRT do have that LCD's don't is pixel persistence. The phosphor coating on the screen is designed to continue "burning" after the electron gun has passed on to the next pixel to reduce flicker. Since CRTs operate over a large range of frequencies, the time required to keep a pixel burning for a 60hz refresh is much longer than the time for 170Hz. Conversely, the time required for 170Hz refreshes is much shorter and would introduce a painfully bad flicker when running at 60Hz. Since the phosphor coating can't be changed on the fly to accommodate different refresh rates, the pixel burn time is designed for a middle of the road refresh (like 85Hz or 100Hz). This allows a large reduction of flicker at 60Hz for high resolutions, but also reduces refresh overlap at 170Hz for smoother motion at resolutions capable of running that refresh rate.

I think what you are experience is that overlap, especially if it's an older monitor. Old pixels don't necessarily behave as they do when new. Try lowering your refresh rate to see if the effect diminishes.
 
I finally got back to experimenting with this. I'm pretty sure that the dominating factor responsible for the picture at the top of the thread is: camera exposure time. And my eyes average frames in exactly the same way. It surprises me that my eyes are the limiting factor. It's like everything we see has a response time of 5ms give or take!

I did a simple experiment. With the CRT at 170Hz and a black background, I moved the mouse around constantly. I see a trail of cursors. Using my compact digital camera, it also captures the trail. But by setting it to -2EV, 1600ISO, and aiming it toward a 100W bulb, it forced the camera to take a 1/1000 s exposure, and it captured just 1 cursor on the CRT. It was hard to believe, so I tried it 5 more times and it was the same every time.

ryan, it sounds like there's no disadvantage to maximizing burn time, assuming that doesn't impact the time to transition when the gun is applied. Ideally speaking, if the burn time were 17ms, 60Hz would be always-on like an LCD. Do you know how long a typical burn time is?
 
You must log in or register to reply here.
Back
Top