Strobing / flicker effect when panning in 24p

[Steven Thomas]

Originally Posted by Dominik Seibold

I did:
http://www.dominik.ws/zoom.mov
This is a 100% frame-crop. On the upper right you can see the whole picture.
24p, 180°-shutter, f2.4, full MF, linear 10s-shot-transition from full wide to full tele.
Each frame is perfectly in focus.
But this was shot on a tripod. Why should I hold the camera while doing this??

Dominik,
I'm not sure why this test was presented. It mentions focus while zooming not the imaginary extra judder.

This test is implying focus is changing during zoom which would only happen if back focus was out and not in-and-out from frame to frame.

[Steve Connor]

Just been looking back at airshow footage filmed by Alister at 25p and 50p with NO problems with the motion as others have been describing.

Surely if this was a real issue then more than a few people would be complaining about it considering the thousands of cameras that have been sold!

[Patrick Williams]

Has anyone compared 24p in SP mode to 24p in HQ mode? The SP mode should have the 3:2 pulldown encoded to make it HDV compatible, and I'm wondering if SP looks smoother on a HD monitor than the HQ mode which I understand is adding 3:2 right before the output to the monitor. I'm getting an EX1 next week, and I will probably want to use HQ for almost all of my shooting and editing. Dominik's chair test eases my mind, now that I know the camera records 24 equally spaced frames per second. Perhaps the HQ component out isn't getting the true 3:2 pulldown it needs to for smoother motion on 60i monitors.

[Steven Thomas]

Originally Posted by Patrick Williams

Perhaps the HQ component out isn't getting the true 3:2 pulldown it needs to for smoother motion on 60i monitors.

Possibly,
This is why I never trust previewing. The JVC HD100 component out looked a bit strange when hooked up to my Pioneer plasma.

The HD100 and EX1 rendered and played back look great.

[Eric Pascarelli]

Originally Posted by Nick Baker

First I would like to say I totally disagree with Eric Pascarelli. I have worked with 35mm film for 8 years, both still and motion. The EX1 has flicker ever on slow moving shots and people with a good eye can tell. If you have an EX1 do the following (24p 180 shuttle) shoot while zooming, not too fast and make sure you are at a steady rate of zoom speed on a person or sharp defining object such as a book cover (do this while holding camera). when you play back, freeze frame, then play one frame at a time you will notice that on a wide to tight shot some will be in focus and some wont. the fact of the matter is there are many different shots that show flicker people just look at your subjects if you do not notice flicker and out of focus get your eyes checked out. Some people are uneasy with coming to terms that their new EX1 as well as all 24p HD cameras on the market suffer from rolling shutter as well as 23.98 not true 24 frames . I am one who has a love hate relationship with my EX1.

Nick,

I don't think that amount of experience has anything to do with it. It's just understanding how video/film acquisition works. There is not much room for subtlety in acquiring moving images. There's the time interval (the fps), the exposure time (shutter angle) and the response curves (gamma etc.). Yes, there's rolling shutter, but even film cameras have that to a certain extent (at least more so than CCD cameras).

As long as the camera accumulates photons and records them reasonably linearly, and each exposure is the same duration, and the spacing of exposures is consistent, there is not much you could really do to make one camera stutter more than another, all of the above being equal between cameras. Even if you wanted to, you really couldn't.

The voodoo is in the compression and even more so in the viewing.

The EX1 does a great job of compressing and introduces almost no visible artifacting. Examining stills frames proves that out.

So if there is a perceived difference in stutter, it's probably in the viewing. The interface between the digital image and the brain is where there is the most room for voodoo. There might be something in your viewing setup that's creating the appearance of stutter.

As you probably know, theatrical film shows each frame twice (48 Hz) to make the brain think the motion is smoother. Broadcast shows each frame an average of 2.5 times (3-2 pulldown). Even Sony CRT HD broadcast monitors flicker at 48Hz with a 24p signal. But your computer LCD does not. All you see is a succession of frames, one after another at 24Hz. For some this is too slow to be perceived as smooth motion.

If you have ever done traditional process photography with rear projection you've seen the immense amount of judder in the projected image when viewed live. This is because the projectors used are synchronized to the shooting cameras and have single shutters rather that two bladed shutters and show their film at a true 24fps. It looks very similar to true 24p on an LCD monitor.

BTW, the difference between 23.976 and 24 is imperceptible and does not cause judder. Unless you've watched the above mentioned rear projection, you've probably never actually seen anything at a true 24fps. This is because all video in the NTSC world is 23.976/29.97/59.94 and most every "24p" pro and consumer camera actually shoots at 23.976. And projected film is 48Hz.

[Dominik Seibold]

Originally Posted by Steven Thomas

I'm not sure why this test was presented. It mentions focus while zooming not the imaginary extra judder.

This test was the answer to Nick Bakers assertion that the ex1 has out-of-focus frames while doing a zoom. (I hope I understood him right)
I guess I already showed that the ex1 (at least mine) does sampling in the time-domain at 24p accurately, so this question is answered for me.
The question about judder-perception of some individuals is too ambiguous for me.

[Dominik Seibold]

Originally Posted by Eric Pascarelli

But your computer LCD does not. All you see is a succession of frames, one after another at 24Hz.

That's not right. LCDs also do have a refresh rate (although there isn't a synchronized electron-ray, the electronics behind need it). 60Hz are typical, so if you watch 24p on a LCD there is hopefully something resulting in a 3:2-pulldown.
(but the rest of your post I absolutely support :) )

[Eric Pascarelli]

Yes, there's a refresh rate but no flicker. So essentially you are seeing one frame after another at 24Hz, but technically there would be a cadence introduced by the beating of the refresh rate with the frame rate, a form of 3-2 pulldown (assuming 60 Hz). And the cadence may contribute to the appearance of judder (though not more so than in standard 3-2 pulledown).

But there is no flicker on an LCD. It's the flicker introduced by CRT's, plasmas and projection that fool the eye into seeing smoother motion where there is not. Not so on the LCD.

Anyway this is all part of the "voodoo" of getting the image from computer to brain. It's somewhere in this part of the workflow that the appearance judder is being introduced. It's not specific to the acquisition or camera.

Thanks for (otherwise) agreeing with my post.

[Dominik Seibold]

Originally Posted by Eric Pascarelli

Yes, there's a refresh rate but no flicker.

Yes, there is flicker! Of course there's some smoothing, because of not insignificant response-times, but a perfect LCD-display with a zero response-time running at 60Hz shows exactly 60 frames per second each with a duration of a 1/60 second. And 1/24 isn't an integral multiple of 1/60, so you can't do anything but 3:2-pulldown => flicker. (Flicker = variable duration of consecutive frames. Common sense?)
Again: The electronics behind do work with whole frames and the count of frames/second running through your DVI-data-cable is fixed, so you have to deal with that.

[Eric Pascarelli]

Originally Posted by Dominik Seibold

Yes, there is flicker!

There is not.

3-2 pulldown does not equal flicker. It's a cadence that can occur with or without flicker.

It means that the duration of every other frame is different (3 refreshes vs 2).

LCD's do not go to black between frames as a projector does. This "blanking" is what is commonly known as flicker.

[Dominik Seibold]

Originally Posted by Eric Pascarelli

but technically there would be a cadence introduced by the beating of the refresh rate with the frame rate, a form of 3-2 pulldown (assuming 60 Hz). And the cadence may contribute to the appearance of judder (though not more so than in standard 3-2 pulledown).

Sorry, my last answer was too hurried. About that kind of judder (3:2-judder) I tried to talk. :)
But I don't think that the 60Hz-flicker of CRTs does contribute more to a 24p-perception-issue than to a 60p-perception-issue. Do you know what I mean?

[Eric Pascarelli]

Originally Posted by Dominik Seibold

But I don't think that the 60Hz-flicker of CRTs do contribute to the 24p-perception-issue.

You may be correct.

But the point of my post, which i think you agree with, is that there is no magic to camera acquisition, and that cameras won't differ much in their "rendering" of motion as long as certain basic criteria are met.

And the EX1 most definitely meets those criteria.

There's much more magic involved in the way our brains stitch images together to form motion. It's affected by things like cadence, flicker, etc. and is the likely part of the chain where differences in judder are being introduced.

The examples I cited (48hz projection, etc.) are ways that technology has been developed to help us better see smooth motion from a series of still images.

To be more clear, if you are seeing judder in a scene where you wouldn't expect it, it's probably happening in playback, your screen, your chosen frame rate, shutter angle, etc.

It's not because you chose one brand of camera over another.

[Dominik Seibold]

Originally Posted by Eric Pascarelli

But the point of my post, which i think you agree with, is that there is no magic to camera acquisition, and that cameras won't differ much in their "rendering" of motion as long as certain basic criteria are met.

And the EX1 most definitely meets those criteria.

There's much more magic involved in the way our brains stitch images together to form motion. It's affected by things like cadence, flicker etc. and is the likely part of the chain where judder is being introduced.

I absolutely agree! Let's convince the others! ;)

[Steve Phillipps]

All this talk about LCDs and computer screens and how and why they flicker is not really that relevant I don't think, as I put the same sort of shots through my Macbook from an HPX2100 and an EX1 at the same frame rates etc. and it was only on the EX1 that I saw motion problems. Never seen the like of it from any other camera I've used either. Just my observations.
Steve

[Dennis Joseph]

Originally Posted by Steve Phillipps

All this talk about LCDs and computer screens and how and why they flicker is not really that relevant I don't think, as I put the same sort of shots through my Macbook from an HPX2100 and an EX1 at the same frame rates etc. and it was only on the EX1 that I saw motion problems. Never seen the like of it from any other camera I've used either. Just my observations.
Steve

I am interested in seeing some footage from Jeremy Hughes as he seems to have a perfect camera.