Strobing / flicker effect when panning in 24p

[Alister Chapman]

OK, Take a look at this clip:

http://www.ingenioustv.com/clips/ex1-motion-test.mov

I shot this with a Sony F350 XDCAM HD camera and an EX1, mounted side by side, following the same car down the road on the long end of the lens, both cameras set up and framed to give matching shots.

Both cameras were set to 25P with 1/50th shutter. I chose 25P over 24P to eliminate any pull up issues as what we are looking at is the way the EX1 capture motion, not how monitors or edit suites add or subtract frames. The F350 has been around for a couple of years and has been used for many programmes shooting at 24P and 25P and no one has complained that it doesn't handle motion correctly.

The clip is not trying to show a perfect shot. The pan is too fast for 25P which should exaggerate any differences. What I believe it shows is that both cameras handle the motion exactly the same. If there was a problem with the EX1's progressive motion handling I don't believe it would have passed Discovery's extensive tests which do involve a lot of motion tests as the one thing they look for is how the codec handles rapidly. changing images.

[Steve Phillipps]

Well done Alister, that's what's needed.
No argument here, they look identical to me.
Hmmmm!
Steve

[Alister Chapman]

I thought I would see more skew on the lamp posts but even that seems so slight as to be practically invisible.

[Eric Pascarelli]

Good job, Alister.

I have been waiting for someone to do that. I hope this issue has been put to rest.

[Piotr Wozniacki]

Much appreciated, Alister, that you have proven the issue is not existing, and the whole buzz about the EX1's low frame rate, progressive modes being more stuttery than on other cameras is simply unjustified.

Which of course only makes the right shooting technique even more important - without it, the motion may indeed appear unpleasant, to say the least.

Also worth mentioning is the importance of the viewing device; watching 25p stuff on a 60 Hz PC monitor is not the best way to judge motion. I can see a huge difference between what I can see on my 24" LCD when compared to my 100Hz plasma, which displays each frame 4 times!

[Dennis Joseph]

Originally Posted by Alister Chapman

OK, Take a look at this clip:

http://www.ingenioustv.com/clips/ex1-motion-test.mov

I shot this with a Sony F350 XDCAM HD camera and an EX1, mounted side by side, following the same car down the road on the long end of the lens, both cameras set up and framed to give matching shots.

Both cameras were set to 25P with 1/50th shutter. I chose 25P over 24P to eliminate any pull up issues as what we are looking at is the way the EX1 capture motion, not how monitors or edit suites add or subtract frames. The F350 has been around for a couple of years and has been used for many programmes shooting at 24P and 25P and no one has complained that it doesn't handle motion correctly.

The clip is not trying to show a perfect shot. The pan is too fast for 25P which should exaggerate any differences. What I believe it shows is that both cameras handle the motion exactly the same. If there was a problem with the EX1's progressive motion handling I don't believe it would have passed Discovery's extensive tests which do involve a lot of motion tests as the one thing they look for is how the codec handles rapidly. changing images.

Allister, thank you for that footage.

A couple fo things I would like to note.

A. It was in 25 and not 24p (not much difference but 24p is where my gripe is)

B. The pan was pretty slow and the majority of the image was over 100 feet+.

If you pan at that speed or just slightly quicker , you will see an annoying flicker effect if the image is closer.

The only thing that this test proved (and yes it did prove something valuable) is that pretty much all HD video cameras have the flicker effect.

This is why I refused to believe that 35mm film 24p is the same as video.

[Eric Pascarelli]

Dennis,

For reasons I tried to explain above, all 24p 180° shutter footage with reasonably linear light sensitivity (film, video, whatever) will exhibit the same tendencies, if played back using the same methods. If there is an anomaly, it's highly unlikely it's in the acquisition and much more likely to be in the playback.

So if you shoot film and progressive video at 24p 180° and transfer/dub them to the same method of playback, the will be more or less identical in their depiction of motion.

What do you envision that film does that video does not to decrease the stutter?

[Bob Grant]

There is in fact a significant difference in how different recording systems may cause more motion judder than others, even all else being equal. The edge enhancement used in video cameras can make 24fps acquisition look much more juddery than say 35mm at the same everything including projection. The issue seems to be so significant that the EE can cause perceived negative steps in movement. Please don't ask me to explain this but there's an excellent paper from the BBC which explains this in great detail with lots of maths.
The way to minimise if not avoid this is to turn down edge enhancement in the camera.

[Dominik Seibold]

Originally Posted by Steve Connor

I can make a HDW750 exhibit the same effect if I max out the detail level.

You can emulate an increased-exposure-time-look with heavy detail-settings? Well, I can't. ;)

[Paul Curtis]

Originally Posted by Dominik Seibold

You can emulate an increased exposure-time with high detail-settings? I don't think so...

I've only just caught up on this thread (and i've had too much coffee so putting my 2c in) but the whole 24p strobing thing is just that HD cameras have more DOF with parts of the frame in focus with high contrast areas and therefore the strobing is more obvious. Adding detail would also serve to create parts of the frame that have high contrast, hence the 750 comment.

As many people have pointed out this is completely normal and will vary depending on what's being shot. Having long shutter speeds will help but as Eric quite rightly points out, the cameras cannot really produce odd stuttering effects, what they're doing is pretty simple. Unless the EX1 internal timebase is not accurate and wavers around (which i don't believe it to be the case, it would be pretty impossible to technically screw that up anyway)

In the past people are just careful of how they shoot.

However there is the chance that playback is introducing stuttering (as opposed to strobing) which is dependant on each pipeline. Without examples of bad motion it's difficult to tell. In the case of comparing cameras you need to make sure it's like for like. Like Alister shows, same settings (shutter speed and motion) and im pretty sure that you'll see the same effect because that's just physics at work. Doing 60i -> 24p will introduce blurring artifacts which will help (resolution would be less, less detail to see the effect).

I've got stuff with strobing using a wildly spinning steadicam (don't ask!) and i can see that it's because there's too much detail visible and not enough blur, i can add motion blur in post (it's part of a heavy comp so i knew that going in).

So either some physical cameras are broken (unlikely i would have thought) or the playback has problems or we're expecting the impossible?

Perhaps it would be worth just doing some comparisons of shutter speeds and focus (or try to compare shallow DOF with wide). Or take the example footage and add motion blur in post (but that's only part of the solution, throwing the background out of focus is better)

cheers
paul

[Steve Phillipps]

On my HDW750 you get motion blur but not the stroby motion I saw on the EX1. Still a bit baffled by this, and because of the fair volume of people raising the issue I don't think it can just be disregarded. Maybe it is the monitoring though, I don't know. Thing is on Alister's tests above I siad they look identical, but they are identical in that I see horrible juddery motion on both - they are equally bad! Nothing to do with you Alister, I know you stated quite clearly that they were too fast and it was to show the effect. But I would expect the effect to be blurring but I see juddering! Again, maybe it's just 'cos it's compessed for the web or whatever?
I may get hold of an EX1 again and try to do some motion tests and get them into an edit suite, I really think this is the only way to know, and again urge anyone to take internet discussion onboard but in in no way treat it as gospel!
Steve

[Steve Phillipps]

Don't forget though Paul, that there is a radical difference in the EX1 to say Sony 750, which is that it has a rolling shutter, and it's a reasonable assumption to make that this may have an effect.
Steve

[Paul Curtis]

Originally Posted by Steve Phillipps

Don't forget though Paul, that there is a radical difference in the EX1 to say Sony 750, which is that it has a rolling shutter, and it's a reasonable assumption to make that this may have an effect.
Steve

Sure, but im pretty sure the rolling shutter won't affect the strobing if you're panning left to right because the sensor readout will be in the same place for a given vertical position, so the same distance even if the overall position is different at the top of the frame compared to the bottom. If you're panning up and down violently then you'll get the jello effect and jello effect trumps strobing each time as far as being objectional goes!

I think that we, as a group, need to quantify what the objectional strobing looks like then take a vote as to whether it is a problem or normal. Alisters examples are normal as far as im concerned. But we've had discussion about a JVC and EX providing different results and i would like to see that.

cheers
paul

[Dominik Seibold]

Originally Posted by Steve Phillipps

On my HDW750 you get motion blur but not the stroby motion I saw on the EX1.

If you're using a loooong exposure-time it gets blurry and if you use a shrt exposure-time it gets stroby. A general rule. But what's the difference between what you've seen from the ex1 and your hdcam-footage concerning that subject? It would be nice, if you would post a sample-clip from your hdcam doing a similar pan like that from the xdcam-hd/ex example.

[Steve Phillipps]

Dominik, unfortunately I couldn't post EX1 clips as I don't have an edit system so couldn't convert to mpeg or whatever. Can't post 750 clips either, but there are some on my website, also film and Phantom HD on there, and none of them have stuttered motion. I did shots on the EX1 following flying birds, also scenic pans and tilts up vegetation to birds in background, all with jittery motion. This was evident at 60P, 25P, shutter 180 degree.
Steve