Video Playback (FPS) Vs. Monitor Refresh Rate (Hz) - Any problems?

[prankstare]

I was just reading this very interesting blog entry which got me pretty much surprised and of course uncertain (as always, hehe).

My current system is using Windows XP in 1440x900 resolution @75Hz on a 19" widescreen LCD. I wonder if I am not properly watching PAL movies (25FPS) because of my current refresh rate setting (75Hz or practically 75FPS), as mentioned in the blog entry above, or maybe there's some feature on the present video decoders that sort of "fix" this kind of problem.

Thank you very much.

[guns1inger]

Re-read the blog. What is it basically saying is that so long as the refresh rate is a multiple of the framerate, you should be OK. This is basically how modern HD televisions work as well.

In your PAL example, the refresh rate is a multiple (X3) of the framerate - 25 X 3 = 75.

[prankstare]

So that means if I watch a 30FPS NTSC movie (75 / 30 = 2,5) I am good?

Damn, then would somebody give me an example when I'd be in trouble watching @ 75Hz refresh rate.

I could not quite understand that blog.

[guns1inger]

Again, from the blog, NTSC @ 23.976 is optimal at 72, NTSC @ 29.970 is best at 60. if you can divide and get a whole number, you should be OK.

Of course, before you read that blog, you didn't care, so why worry now ?

[prankstare]

Hum, must be a whole number.

Well, it's interesting because when I had my 17" CRT it was set @85Hz (good old Samsung 796MB Plus series) and I would notice nothing strange playing PAL or NTSC or NTSC-Film videos (85/29.970=2,8 ; 85/25=3.4 ; 85/23.976=3.6).

[Cornucopia]

That blog was inconsistent and inaccurate.
That 3:2:3:2:2 pulldown is certainly not the predominant method for converting 25->30fps. Nothing was mentioned about fields, which can often alleviate the judder (because the timebase is now more like 50->60Hz). And there ARE a number of types of blended/interpolated frame methods (some work better than others depending upon the material). That 2nd picture is totally false--look at the predominant color in the 2nd/25th second--it's still yellow.

Don't get quagmired in the FUD. Learn how it all works and understand the gives and takes. Science is your friend.

Scott

[prankstare]

OK, thanks Cornucopia.

[jagabo]

Fields don't eliminate judder. 23.976 with 3:2 pulldown still has judder. Duplicating fields can be smoother than duplicating frames. For example converting 24 fps to 30 fps by duplicating every 4th frame, then displaying 30 fps one field at a time, is jerkier than using 3:2 pulldown on an interlaced TV.

Judder comes from different frames (or fields) being displayed for different amounts of time. To play 24 fps video on a 60 Hz monitor half the frames are displayed twice, half are displayed 3 times, alternating between the two. This can be quite noticeable on smooth medium speed panning shots.

If you are running your display at 60 fps see the video in this post:
https://forum.videohelp.com/topic369790.html#1981421

[prankstare]

Hum, just seen that video. Really weird. I've decided to switch back and forth between 60 and 75Hz and could not notice any difference in the 24FPS object in the video (in full screen mode that is). Should I have seen any difference by changing the refresh rate or is that normal?

Damn, I decided to also see this video on my old 29" CRT TV with the help of DIVX standalone player Philips DVP5990K/55. The frames are ridiculous there - I guess I've never seen much flickering as I did watching that thing on there. My question: how do these things work on standard CRT television sets and standalone players? At which refresh rate do these two work? I'm keen curious!

[Cornucopia]

Fields don't eliminate judder. 23.976 with 3:2 pulldown still has judder. Duplicating fields can be smoother than duplicating frames.

That's what I was saying. "Alleviate" means make less objectionable, not eliminate. Judder with smaller time step sizes is less objectionable than judder with larger time step sizes.
We agree.

I think the OP needs a good test clip with which to try out his/her system pragmatically. You've got one, don't you jagabo?

Scott

[jagabo]

Fields don't eliminate judder. 23.976 with 3:2 pulldown still has judder. Duplicating fields can be smoother than duplicating frames.

That's what I was saying. "Alleviate" means make less objectionable, not eliminate. Judder with smaller time step sizes is less objectionable than judder with larger time step sizes.
We agree.

Sorry, I thought you were saying that displaying 24 fps film at 60 fields per second has less judder than displaying it at 60 frames per second.

We differ in terminology too. I call the small time step problem judder as it looks like the picture is shaking. You can't count the skakes, they're too fast. The larger time step problem I call jerkiness. When 24 frames are converted to 30 frames by duplicating every sixth frame (then displayed as 60 fields or frames per second) you can see and count the six little jerks every second.

[jagabo]

Hum, just seen that video. Really weird. I've decided to switch back and forth between 60 and 75Hz and could not notice any difference in the 24FPS object in the video (in full screen mode that is). Should I have seen any difference by changing the refresh rate or is that normal?

With 60 Hz refresh you should have seen the bottom row of circles (60 fps) moving very smoothly across the screen. The middle row (30 fps) should move fairly smoothly but you might say it flickers or judders. The top row (24 fps) has a definite judder. The circles don't seem to be moving at a constant speed across the screen but rather speed up and slow down several times a second. Try tracking the circles with your eyes as they move across the screen. They just don't appear where your eyes expect them, they seem to move back and forth.

At 75 Hz the bottom row should show four small jerks per second. The middle row too, although with the flickering it's harder to notice. The judder of the top row should be visible and the additional little jerks difficult to notice.

Note that other playback issues may interfere. If your computer isn't fast enough to play 60 fps video at 60 Hz refresh perfectly you'll get jerkiness, even on the 60 Hz row. Other Windows housekeeping tasks may occasionally interrupt and cause jerks. The clip is only three seconds long. If you have your player set up to repeat the video indefinitely there will probably be a jerk every three seconds as the player "rewinds" back to the beginning.

[prankstare]

jagabo, yeah. My PC is a Pentium 4 2,4GHz, 1GB DDR333 with an ATI Radeon 1600PRO running on WinXP Pro SP3 - I know this isn't much but I don't think it would interfere in the playback. And yes, I've seen all the things you said above, but the most interesting to me is that when I changed the refresh rate from 60 to 75Hz, surprisingly the top row became little less jerky than when in 60Hz. Well, there is also something about this video that doesn't seem to make much sense - according to the blog and to what 'guns1inger' said, as long as the refresh rate is a multiple of the framerate the video should be okay. Then why when I play this video either in 60 or 75Hz, the middle row which is 30FPS, shows itself not very smooth like it's flickering or jerky? Mathematically, 60 / 30 = 2, which is perfect round.

Another question here: in which refresh rate do the old 20", 29" CRT's (not monitors) operate?

[jagabo]

Well, there is also something about this video that doesn't seem to make much sense - according to the blog and to what 'guns1inger' said, as long as the refresh rate is a multiple of the framerate the video should be okay. Then why when I play this video either in 60 or 75Hz, the middle row which is 30FPS, shows itself not very smooth like it's flickering or jerky? Mathematically, 60 / 30 = 2, which is perfect round.

The point is that displaying the 30 fps video at 60 fps is as smooth as it's going to get. 30 fps is inherently jerky to your eyes. Note that that particular video was made especially to accentuate this problem. The bright, high contrast objects with a medium speed motion are one of the worst scenarios.

Another question here: in which refresh rate do the old 20", 29" CRT's (not monitors) operate?

Interlaced PAL CRT TVs always operated at 50 fields per second. NTSC at 59.94 fields per second.