Do UK bluray players apply telecine ?

[hello_hello]

Nah..... it's not the TV. At least I think what's happening is normal anyway, but I can't say I've done lots of comparing. The LCD TV in this house, which is fairly new, does have 24p mode and I've been meaning to to try some test files on it (it's not my TV) but haven't got there yet.
Now I know what I'm looking for I can see the same thing using the CRT monitor. I guess it never bothered me before because the screen is so much smaller.

Didée described the effect I've called "jitter" in this post. I'm pretty sure he's referring to the same thing. A "doubling" or "tripling" effect in some motion.
http://forum.doom9.org/showthread.ph...le#post1555695
That's why I've wondered if there's something in the way a projector works which doesn't produce the same effect.

[madhatter300871]

I read Didees post, I think I know what he is talking about. So if I am right in my understanding is what you are experiencing and call 'jitter' the effect that during motion scenes, particularly panning, you are seeing this frame duplication effect ? Almost as if the trailing edge of the moving image is leaving a trail (kind off). Believe it or not I think I have seen similar, the penny might be dropping for me now.

When I play a film via the DVD player the motion is noticeably less smooth than when I play it directly through the TV. (My LG flat screen can play movies directly from a USB pen drive). I wonder if what I am seeing is what you call 'jitter', I think it might be. For me, I just don't use the DVD player anymore, unless I absolutely have to.

[hello_hello]

The "jitter" or "edge duplication effect" only seems to be noticeable at certain speeds. At least to me. It also happens to be about the same range of motion (speed) where I can see judder if the video is 24fps and the refresh rate 60hz. If it's fairly slow it looks fluid. If it's fairly fast it also looks fluid. It's just that middle ground..... probably the speed where film produces the so called "strobe" effect. That's why ages ago I said to me it seems to be one or the other, although the "jitter" is definitely preferable to "judder".

I don't quite understand why the "edge duplication effect" (assuming that's the best way to describe it) actually happens. I can see the same thing on my CRT monitor now I look for it, so I thought maybe it was just the way PC video cards work, but a while back I used the Bluray player to look at some "problem" motion scenes and they look the same as when using the PC. Maybe it's got something to do with response time, although I'm using a Plasma so there should be no problem there. I understand why film can look "jerky" when motion is happening at a certain speed... well I assume it relates directly to the frame rate.... but why TVs produce that "doubling of edges" effect, I don't know. Hopefully someone will come along and explain it.

I watch a lot of documentaries on the universe which often have animation of objects moving through space with stars (basically white dots) also moving in the background. At slow speed the motion of the stars looks smooth, then as the speed picks up each star is duplicated or triplicated (is that a word?) until finally the speed picks up until the motion of stars becomes a blurred line. At first I thought that's just the way the computer graphics were done, until I started pausing the video during motion scenes and discovered the stars aren't duplicated in the actual video..... I can see the same thing using the CRT monitor too.

I'll have to "commandeer" the LCD TV in this house one afternoon to see if it produces the same effect and if it's 24p mode gets rid of it.

I'd assume when using the DVD player you're refreshing at 60hz so I'd assume the added "judder" robs the video of smoothness at certain speeds. I still wonder if projectors might produce the image differently enough to Plasmas or LCDs to make them look smoother? I don't know, I've not had any real experience using one. Film is projected by flashing each frame two or three times to remove "flicker" etc so chances are our brains perceive a projected film image differently to one displayed by a TV.... I don't know....

[madhatter300871]

If triplicated isn't a word then it should be !

I don't think the LCD projector works differently to an LCD TV, there is still an LCD panel with an image on it, although it is feasible that LCD, CRT and Plasma work in different ways (well clearly they do because they are different technologies).

On a side note, can you explain to me your understanding of the 'film effect'.

I now have a clear understanding of jitter and judder and wonder if jitter is related to the film effect.

Just to re-iterate I don't see any jitter what so ever on my projector, but do suffer with judder when playing back 23.976fps content.

[hello_hello]

To me, jitter and the "film effect" are pretty much the same thing.... motion at certain speeds doesn't look "fluid". Well at least I assume they're the same thing.....

As I said, it's something I only see now and then, but here's a little sample.... which is why it took me a while to reply. I remembered the scene well as the judder stood out more than I think it ever has before (pre 50hz refresh rate, now it's jitter), but I only just got around to finding which episode it was in. This sample (already 25fps).... the camera pan doesn't look anything even close to being fluid to me even with a 50hz refresh rate. It's "jittery". Where the camera pans past the legs of the guard with the gun, the light coming from behind gives it a strobe effect and the edge duplication/triplication of the guard's dark uniform is fairly obvious. If I pause there and step through one frame at a time I can see the edge duplication isn't in the video itself. Each frame looks fine. At 50fps it looks good, but that's a little too much of a speed increase.

PS. If it looks fluid to you using your projector, what about using something else such as a PC monitor? Playing it on my CRT with it's 85hz refresh rate makes it look even less fluid.

[Didée]

Sample. And. Hold.

http://www.mediafire.com/?xrda455a7w1kkbb


(German) explanation: klick

[Didée]

A quick try in english -

When the eye is tracking a moving object, the eye is moving continuously. In contrast, the picture on a sample&hold display (most prominently LCD screens, but more generally also plasma screens, or projectors) is standing still for ~40ms, then switching the next picture, again standing still for ~40ms, etc.

First consequence is that the image on the eye's retina is getting blurred. It's the same principle as a static photo camera shooting a fast-moving object, just the other way round: a moving camera (eye) is shooting a static object (sample&hold image frame).

However, it's not the same as the "motion blur" that's catched by a film camera with long shutter. It is much worse than that. A film camera is sampling over continuous motion, hence the motion blur is "natural" and evenly distributed.

When watching a 24/25 fps film on a sampl&hold display, the eye is presented with fractional motion, and it does a dynamically "running window" sampling over that fractional motion. The resulting effect is that the motion blur building up on the retina in fact is creating "backwards motion" during the integration time of one frame, plus "double image'ing" when the transition from one frame to the next is happening.

Retina persistance is playing a big role in the process, which also explains why the effect is most prominent on high-contrast detail, and much less in dimmed scenes. (The bright parts sting much light energy onto the retina, forcing a much stronger persistance, compared to the low-energy of dimmed detail / dimmed scenery.)

[hello_hello]

I'm not sure I fully understand your explanation yet, but the "sample and hold" effect in your example is what I sometimes see, only I guess in your example it's somewhat exaggerated? I think I can see the same effect in the "real world motion" example when it's displayed on the Plasma, only it's very minimal.

[Didée]

That video is intended to be a slow-motion analysis of what's actually happening in your bio-camera. (And for that purpose, the 60fps version isn't important.)

A more brutal video for the "I-dont-want-my-TV-to-do-any-motionmagic" fraction is this lighthouse video.
(Yessss, it is lacking motion blur. But when I made the same sequence *with* motionblur, the visual effect was almost the same.)

[hello_hello]

Some questions regarding your lighthouse video.....

On my plasma (50Hz) at 25fps it's "jittery" with obvious edge duplication/triplication, but why when I play it at 50fps is there still a "ghosting" effect? The motion is smooth at 50fps but there's still edge duplication or "blur".

I tried the same thing using my CRT monitor (85hz). When playing the video at 42.5fps it looks the same as the Plasma does when the frame rate and the refresh rate are an exact match (motion is smooth but there's "edge blur"). When I increase the frame rate to 85fps it's perfect on the CRT. No blurring or edge duplication, the pictures slide past looking exactly as they would when still.

In other words when using the plasma, and when the frame rate and refresh rate match, it looks the same as when using the CRT with the frame rate being half the refresh rate. So the plasma can't do what the CRT does, which is for motion to look perfect when the frame rate and refresh rate match. I'm just trying to understand the "why" when it comes to the difference between the displays.

My plasma doesn't do frame interpolation.

[Didée]

Well, many of the "hardcore" pasma fans are telling that it's only those darn LCD screens that have problems with smooth motion, and plasma and only plasma and nothing but plasma could do. The truth is that ... this is not true. Plasma displays are somewhat more forgiving with this problem, but not by that much. In essence they face the same technical problem as LCDs do. (I assume a bigger point is that most plasmas have a darker picture than LCD panels. And/or: many LCD screens are set too bright. More light energy => bigger problem due to stronger retina persistence. Darker picture => less problem because of less retina persistence.)


For the differences - it's a puzzle of many small pieces, you can't break it down that easily to one definite answer.

First, the (playback) framerate does make a difference. When you play the sequence at 25fps, then each picture is lasting for 40ms. At 50fps, each picture is lasting for only 20ms, and at 85fps, its only 11.7ms. Obviously you cannot (shouldnot) compare "the effect" directly to each other.

The strength of the CRT at its native refresh rate is that the CRT is *halfway* an impulse-type display. It isn't truly one because of the phosphor afterglow (and nobody would bear to watch a *true* impulse display), but still, each pixel is most bright in the moment it is activated by the cathode ray, and fades to dark(er) afterwards. When framerate is matching the refresh rate, then the "peristance" of the shown picture is dimished because of the brightness fade. The next time the ray comes along, it produces a new motion state. This is basically "impulse type" (and practically halfway impulse because of the afterglow).
However, when dropping framerate to half refresh rate, then much of this beneficial effect gets lost: a pixel is activated, fades out, the ray comes again and again activates the same pixel at the same old motion state. Tadaaa!, sample+hold -problem even on a CRT.

[hello_hello]

Originally Posted by Didée

First, the (playback) framerate does make a difference. When you play the sequence at 25fps, then each picture is lasting for 40ms. At 50fps, each picture is lasting for only 20ms, and at 85fps, its only 11.7ms. Obviously you cannot (shouldnot) compare "the effect" directly to each other.

Yeah, you're right. Bad/lazy assumption on my part. I dropped the CRT refresh rate to 60hz (I couldn't stand it that low for a monitor but when watching video it's okay) and tried again, also comparing it to the plasma again.

The plasma at 25fps/50hz does look a little more jittery than when it's running at 30fps/60hz but either way it's still horrible.
The plasma at 50fps/50hz definitely has more edge duplication than it does at 60fps/60hz but it's still nowhere near as good as the CRT at 60fps/60hz.

The CRT at 30fps/60hz doesn't look much different to the plasma at 30fps/60hz.
The CRT at 60fps/60hz looks much better than the plasma at 60fps/60hz. There's no significant "ghosting" and no edge duplication. I can see a little blurring of the blue where it turns to black but I assume that's just the phosphor afterglow and nothing I'd notice when watching "normal" video.

So yeah, what you're saying makes sense.
I'll be keeping your video for future testing. While the PC and the plasma seemed happy to play together at 50hz.... not so much at 60hz. Even Reclock didn't help with the odd dropped/duplicated frame or the tearing when resizing video which I thought I'd fixed, but as it turns out the fix (telling the renderer to wait for flushes) was fixing the tearing but causing more dropped/duplicated frames. Playing with the vsync offset in MPC-HC basically got everything running smooth and fixed the tearing, so your video is definitely handy for revealing problems. Mind you these days I use Reclock to play everything in "PAL cinema" mode anyway (25fps/50hz). At least it eliminates the 60hz judder.

[hello_hello]

One other question.... I've basically assumed what you're calling the sample and hold effect is the same thing as film 24p "strobe" or jitter. Well not so much assuming they're the same thing, but assuming where the film 24p strobe/jitter effect would be noticeable is generally the same motion/speed where the sample and hold effect becomes obvious on a Plasma/LCD display.

Would it be fair to assume the film 24p strobe and Plasma/LCD sample and hold effect are noticeable around the same range of motion/speed and therefore related in that way, or does one not have much to do with the other?

[hello_hello]

It's a pity SED or FED displays never happened. They sound like they would have been a nice compromise between a CRT and a Plasma/LCD.

[ramicio]

It's not a sample and hold problem. It's a combination of larger display sizes of the present times, low FPS, and lack of motion blur.

[Didée]

display size - no
low fps - sure, that's why we are talking at all
lack of motion blur - basically no

Display size doesn't matter, resp. you must take into account the viewing distance. My monitor is 24", my TV is 40", but on the bigger TV the problem isn't worse than on the smaller monitor. (Because of viewing distance the picture of my monitor is "bigger". Intercept Theorem, you know.)

To check that scrolling scene with motion blur, I had provided the same scene WITH motion blur (change framerate by yourself to whatever you want to test, Vdub does it in a breeze, it's just an AVI), and the problem is pretty much the same. Sure, it is less obvious, because local contrast of the moving bits is reduced. But the underlying problem has not changed in any way. (Like the bad sound of a crappish musicplayer - when you reduce the volume, then the sound is less annoying. But it doesn't get better because of that.)


The problem is not so much sample&hold in itself, but rather the interaction happening on the retina. To see the difference, watch the sample out-of-focus (not exactly easy to do). While it is running, focus on a point some meters behind the screen, or hold a finger before your nose and focus the finger. While keeping that focus, put attention to the scrolling scene. Only when the screen is out-of-focus, and eyes are NOT TRACKING the motion, only then it is possible to see the raw strobing motion of film rate, but without those retina-blur/sample sideeffects.

[ramicio]

Display size, yes. The average size of a CRT was like half of the average size a person buys of a panel, and they say far away. The close you are to a screen, because you can be, because of the resolution, the more distance frames travel, thus making the tracking problem worse. Lack of motion blur, yes. The problem presents itself between slow motion and fast motion. Usually medium speed scene pans. They could also be shooting movies lately to do away with motion blur to make processing of motion interpolation on TVs more realistic, to do away with artifacts. The problem goes away when you simply move back.

There is no strobing... Strobing would be a 50% duty cycle of 24 Hz light, for actual film. Then they went to strobing at 48 Hz. You know what a strobe light is? It's flashing, not jagged video.

[Didée]

Well yeah ... is it a greenish bluething, or rather a blueish greenthing? Sure is only this: as soon as you butt in, it is, in any case, the other thing.

I congratulate you for being a native english speaker. Discuss with me in German, meet your Waterloo.


- edit -

(since it seems sports to argue for the sake of arguing)

Yes, strobing light is a change of light/darkness/light/darkness. But to go by the word, I did not speak of strobing light, but of strobing motion. Call it an artistic interpretation, but ... if there is some continuous natural motion, then a strobing light turns the continuous motion into a chain of "snapshots in time", which is pretty much the characteristics of a digital video stream.

[hello_hello]

Originally Posted by ramicio

There is no strobing... Strobing would be a 50% duty cycle of 24 Hz light, for actual film. Then they went to strobing at 48 Hz. You know what a strobe light is? It's flashing, not jagged video.

Disadvantages of 24p

"In general, 24 frames-per-second video has more trouble with fast motion than other, higher frame rates, sometimes showing a "strobe" or "choppy" motion, just like 24 frame/s film will if shot as if it's video, without careful panning, zooming, and slower camera motion. It is therefore not well-suited for programming requiring spontaneous action or "reality" camerawork.

It should also be noted that while the strobe of 24p is in many ways considered a disadvantage, it's also part of the "film look." 24 frame/s film strobes in exactly the same way."

That's basically what I was referring to when I referred to 24p strobing. You can of course call it whatever you want.

[hello_hello]

Originally Posted by Didée

Display size doesn't matter, resp. you must take into account the viewing distance. My monitor is 24", my TV is 40", but on the bigger TV the problem isn't worse than on the smaller monitor. (Because of viewing distance the picture of my monitor is "bigger". Intercept Theorem, you know.)

I'd agree with that according to the limited testing I can do. I can get far enough away from my CRT monitor for the sample and hold effect to basically disappear. I can't get far enough away from the plasma though. The room isn't big enough.

Originally Posted by Didée

To check that scrolling scene with motion blur, I had provided the same scene WITH motion blur (change framerate by yourself to whatever you want to test, Vdub does it in a breeze, it's just an AVI), and the problem is pretty much the same.

Agreed. It doesn't really look any different to the version without motion blur, except for the fact it's blurry. It still "jitters".

Originally Posted by Didée

The problem is not so much sample&hold in itself, but rather the interaction happening on the retina. To see the difference, watch the sample out-of-focus (not exactly easy to do). While it is running, focus on a point some meters behind the screen, or hold a finger before your nose and focus the finger. While keeping that focus, put attention to the scrolling scene. Only when the screen is out-of-focus, and eyes are NOT TRACKING the motion, only then it is possible to see the raw strobing motion of film rate, but without those retina-blur/sample sideeffects.

It turned out to be fairly easy to watch the sample out of focus, I just played it and thought of ramicio.
But yes, I could duplicate exactly what you're describing.