Stuttering image while Panning

[Kip Kubin]

Twixtor and I'm sure several other programs will do that work as well as camera operators that no how to operate.. pan speed etc.

 

[Raul Quintanilla]

Hmm this makes me wish that cinema would be recorded at 60 fps. Yeah, I know about the cinema look, but I believe the change is gonna come anyway. It's simply more real, just like color and sound. Why don't we do it now?

 

[Bill Anderson]

Buhot; mostly because the strobing or judder, or what ever you want to call it, will only happen under certain parameters, change one of the variables, like focal length or panning speed, subject distance/speed, and it might no longer be an issue. This is not always possible of course, but I think the common mistake ( apart from exaggerating the issue) is to assume that one particular approach or setup is always the culprit. Look at Gibby's setup using faster shutter speeds; and yet there is more lag- blank space- between each frame with the faster shutter speed - something that could quite readily promote judder under many circumstances, but it works for Gibby, and heck, he shoots more than most of us combined. But slo mo is a wee bit different from real time playback- and faster shutter speeds on real time playback can create their own unique issues. Imagine a football player, or better yet, a motorcycle passing through your frame- camera locked off- the blank time between every frame (when no image is being recorded) allows the motorcycle to move a greater or shorter distance depending on your exposure. I'll avoid math- however, an example: at 24 frames per second (or any other frame rate to be played back at real time, not slo-mo) with a shutter speed of 1/1000 sec ( sharp, "frozen" image), that leaves a relatively long time till the next frame rolls around- allowing the motorcycle to move a greater distance across the frame than if you chose 1/48 sec (at 1/48 sec, the bike moves just about as much during the exposure as it does during the non image making lag ( blurry image, but a shorter blank time), and it's the blank time, the non image making lag, that generally promotes strobing. At 1/1000sec the bike can travel that bit farther- popping up no where near its last place on the screen.
Kinda like the period in this screen: [ . . . . ]

1/48 sec would be more like the period being dragged to create blur, but with the sense of a smoother continuous motion: [___ ___ ___ ___] those gaps between the periods can be a killer at times.

 

[Eki Halkka]

I was talking about "time remapping" in Adobe After Effects, for example. They compute new frames based upon the surrounding frames.

That wouldn't help unless the playback speed was set higher.

The only way to avoid judder in 24 fps delivery is to shoot properly: take into account the scene's contrast ratio etc. subject matter, DOF characteristics, shutter speed etc. and find a panning speed that looks good under those conditions.

 

[Eki Halkka]

Hmm this makes me wish that cinema would be recorded at 60 fps. Yeah, I know about the cinema look, but I believe the change is gonna come anyway. It's simply more real, just like color and sound. Why don't we do it now?

Because we're Pavlov's dogs, that's why. We're trained to associate 24 fps cadence with "movie quality" and 60 fps cadence with "soap opera quality" - even when the latter actually is technically better.

 

[Bill Anderson]

Keeping in mind that the original post wasn't
about strobing in general, it was more alluding
to strobing being more pronounced on RED than
with film- some even going as far to say that it's
more pronounced on RED than other digital cameras.
Do codecs play a role in the potential for "strobing"?
The way they handle motion, etc.. How does that work?
I don't mean to fuel more bogus claims or guessing, just curious.

 

[Eki Halkka]

Keeping in mind that the original post wasn't
about strobing in general, it was more alluding
to strobing being more pronounced on RED

Well, only reason i can imagine that could cause RED to have more strobing than other 24p sources is simply that the image is sharp and noise free: image softness and grain may hide strobing slightly on other sources. The difference should be neglible though.

 

[Denis Buhot]

Thanks Bill

Thanks Bill

and yet there is more lag- blank space- between each frame with the faster shutter speed - something that could quite readily promote judder under many circumstances, .

I have tried Gibby's advice, but I couldn't get what's working with him to work with me :-(

BTW, using faster shutter speeds is a bit of a paradox, since one can expect motion blur to decrease, increasing judder...

 

[Bill Anderson]

Buhot; I'm sure Gibby's advice is sound; although you might have a unique set of circumstances. Just remember that Gibby is often shooting at higher frames per second rates, so the extra frames inserted will easily take up the slack from the shorter exposures. It's the 24 FPS "filmmakers" that are often taxed when looking for strobing workarounds without compromising their image expectations....

 

[Steve Gibby]

I have tried Gibby's advice, but I couldn't get what's working with him to work with me :-(

BTW, using faster shutter speeds is a bit of a paradox, since one can expect motion blur to decrease, increasing judder...

If you've followed my previous recommendations for smooth motion, and you're still getting strobey or juddery playback, then I'd suggest the problem may be in the timeline playback rate you have set up in your editing system, the type of playback monitor you're using, or potentially the hertz of the monitor you're viewing your footage on.

Here are some quick examples from my stock footage library of high speed pans of sports action and wildlife I shot with RED. I've listed the RED camera setup data for each shot, which you can also read on the page the clip icon is on. Simply click on the clip icon on each of the pages and the QuickTime clip will play.

Surfing shot at 2k, 75 fps, shutter 1/196th second - Playback of the clip is at 23.98 (read data on page)
http://www.mhdp2.com/XRED1/REDL18/MHD_rSG4_r0073.html

Surfing shot at 3k, 59.97 fps, shutter 1/125th second. Playback of clip is at 23.98 (Read data on page)
http://www.mhdp2.com/XRED1/REDL18/MHD_rSG4_r0099.html

Surfing shot at 3k, 59.97 fps, shutter 1/125th second. Playback of clip is at 23.98 (Read data on page)
http://www.mhdp2.com/XRED1/REDL18/MHD_rSG4_r0102.html

Surfing shot at 3k, 59.97 fps, shutter 1/125th second. Playback of clip is at 23.98 (Read data on page)
http://www.mhdp2.com/XRED1/REDL18/MHD_rSG4_r0107.html


Pelicans in flight, 2k, 75 fps, shutter 1/196th second. Playback of clip is at 23.98 (Read data on page)
http://www.mhdp2.com/XRED1/REDL03/MHD_rSG4_r0015.html

Pelicans in flight, 2k, 75 fps, shutter 1/196th second. Playback of clip is at 23.98 (Read data on page)
http://www.mhdp2.com/XRED1/REDL03/MHD_rSG4_r0016.html

Pelicans in flight, 2k, 75 fps, shutter 1/196th second. Playback of clip is at 23.98 (Read data on page)
http://www.mhdp2.com/XRED1/REDL03/MHD_rSG4_r0017.html


----------------------------

The above are just simple, quick, and available online examples (surfing and Pelicans) of a tiny fraction of my high speed action RED library. I think you'll see from them that RED can be extremely smooth (no judder or strobing) in high speed pans if you have the right combination of frame rate and shutter speed. I think you can also tell from the above clips that I am very experienced at smoothly tracking fast action subjects with a camera. BTW - the tripod head used in all of the demo clips above was a Miller Arrow 55 - an excellent head for smooth action following.

These clips above are just very simple examples of my work. For a much more complex example of my work, this past year I was one of the Executive Producers for a National Geographic Channel wildlife documentary in Alaska and almost all of the program involved fast moving wildlife - and yes the footage and program came out extremely smooth. As you may know, I was DP a few months ago for a 10-RED camera production of the classic rock back Journey in concert in the Philippines, and all then cameras were setup to shoot Quad HD, 29.97, and there was tons of fast motion pans of the singer and guitarists moving around on the stage and thrust ramp - and it all came out extremely smooth. The DVD of that will be out next month worldwide if you'd like to check it out. As a DP for the production, I didn't want to suspend reality in the viewers with a juddery mess - but rather to achieve an organic, fairly filmic, almost 3D feel to the footage - and we succeeded in that. In reviewing the footage, me, the producer, director, editor, and the band's management were all way satisfied with the smooth "look and feel" of the footage.

I've owned several RED cameras for two years duration of time - and though I shoot a lot of other genres, I've shot huge amounts of high action footage (sports, wildlife, etc.). My recommendations are very accurate. If you're using RED camera setups I have recommended, and you aren't getting the results I am, and that the above linked clips demonstrate, then you're probably setting up the timeline on your editing system wrong, your playback monitor is set up wrong, your monitor has a wrong hertz rating to give you a true view of what your footage looks like, or there is some other tech issue related to your playback or editing system - that doesn't relate to RED One or the properties of your R3D files themselves.

Hope this helps...

 

[Denis Buhot]

Bill ; never meant

Bill ; never meant

Gibby's advice wasn't sound ! Just that I didn't find it working for me. With me, judder occurs in very ordinary situations... nothing special really. And similar shots or much more shaky ones are silk smooth on the big screen. However, the issue is substantial, making nice and difficult shots hardly usable. Ok with focal length, panning speed, angle, etc. accounting for some of the judder. But I don't think it can account for all the judder I see. I'm worried that some people here, including red people, might be simply denying the fact instead of adressing it/explaining it, especially if it is of no practical consequence at the end. Could have to do with monitoring, sreen frequency or whatever will disappear when finally transcoded/watched on TV or on a very high-end screen. I simply don't know.

 

[Steve Gibby]

Buhot,

See the post I just made 20 minutes ago. From what you just posted, I get the felling you didn't see my post...

 

[martinnoweck]

I heard that maybe it's the 1:2 problem, that if you will shoot 16:9, it wont show this problem? is that true?
Ariel.

is this true?

best wishes,
martin

 

[Steve Gibby]

is this true?

best wishes,
martin

No, that's not true. As examples, click on the links in my post above. Some of them were shot in 2:1 (2k) and most of them were shot in 16:9 (2k, 3k, Quad HD). All of the clips play back smoothly - except the ones in the Buffalo Ridge piece I noted which were slow pans shot in 23.98, shutter 1/48th second, in which you can see some slight judder.

In our work we've naturally been producing judder with 23.98 fps 1/48th second (180 degree shutter) - a camera setup which by choice or need we rarely use in our work genres. At frame rates of 29.97 or higher and shutter angles between maybe 220 degrees and 360 degrees (we very rarely use 360 degrees), judder has not been an issue, whether we've shot 2:1, 16:9, or Quad HD.

Some genres want/need 23.98 or 24 fps production. If you shoot in them, you simply need to observe the need for slower pan and tilt speeds, avoiding high contrast areas in a pan/tilt, etc, which have been accurately suggested by many posters earlier in this thread. If you shoot 29.97 fps or higher, especially when using shutter angles above 180 degrees, then fast pans, tilts, and fast subject motion through the frame are very smooth.

Some genres of production seek judder, but many do not. You can set up RED One to shoot for any genre "look" you want or need.

 

[Tim Whitcomb]

No, that's not true. As examples, click on the links in my post above. Some of them were shot in 2:1 (2k) and most of them were shot in 16:9 (2k, 3k, Quad HD). All of the clips play back smoothly - except the ones in the Buffalo Ridge piece I noted which were slow pans shot in 23.98, shutter 1/48th second, in which you can see some slight judder.

In our work we've naturally been producing judder with 23.98 fps 1/48th second (180 degree shutter) - a camera setup which by choice or need we rarely use in our work genres. At frame rates of 29.97 or higher and shutter angles between maybe 220 degrees and naturally 360 degrees (we very rarely use 360 degrees), judder has not been an issue, whether we've shot 2:1, 16:9, or Quad HD.

Some genres want/need 23.98 or 24 fps production. If you shoot in them, you simply need to observe the need for slower pan and tilt speeds, avoiding high contrast areas in a pan/tilt, etc, which have been accurately suggested by many posters earlier in this thread. If you shoot 29.97 fps or higher, especially when using shutter angles above 180 degrees, then fast pans, tilts, and fast subject motion through the frame are very smooth.

Some genres of production seek judder, but many do not. You can set up RED One to shoot for any genre "look" you want or need.

no offense, but did u see that wave in the first clip? sweeeeet shot Steve!

 

[James Brundige]

Pardon me if I missed some part of the discussion; I have not read the entire thread.

I encountered this problem with any footage shot at 24 (or 23.98) and played at 29.97.
FCP was doing a very bad pull up, and does not tell you this. Opening the files in Cinema tools, and transforming them to the correct frame rate remaps the frames instead of doing the bad pull up. This fixes the problem completely, but is of course not useful for sync footage, since it is sped up slightly.

 

[Steve Gibby]

no offense, but did u see that wave in the first clip? sweeeeet shot Steve!

Thanks! Glad you liked it...me and my crews shoot a lot of that kind of stuff...

I shot that clip with RED #8 on the North Shore of Oahu in late 2007. Those wave clips I posted links for are just some of my minor stuff to illustrate frame rates. In the wave section of my RED footage library I have tons of big, grinding, tropical location waves, and huge temperate location waves too. I have a lot of empty, gaping, round, 25-30' barrels in perfect light, in glass smooth slow motion - big enough to drive a bus through

Beyond the footage for my TV productions over the past two years, about 90% of my personal RED footage library, shot with my REDs all over the world, has never been seen by anyone. Its my "private reserve", but someday soon I'll bust a lot of it out as stock footage - and for some documentaries I have in development.

Some here may be interested in what lenses I used in the linked clips. The lenses used in the Buffalo Ridge piece were Tokina 11-16 (with Nikon mount), Nikon 28-70, Nikon 80-200 AF, and Nikon 50-300 AIS. I shot most of the surfing clips and all of the pelican clips using my RED 300 tele. I also shot at least one of the surfing clips using my Canon 150-600 (with Century PL mount conversion).

 

[Bill Anderson]

Great stuff Gibby, and, once again, thanks for sharing your experiences with RED.

 

[CharleyWeston]

I have run into this problem as well. Thanks for the info and the links!

-Charley
www.hdsanfrancisco.com
www.hdpost.us

 

[Denis Buhot]

thanks, Steve

thanks, Steve

Buhot,

See the post I just made 20 minutes ago. From what you just posted, I get the felling you didn't see my post...

I eventually came to the conclusion that it has to do with my monitor frequency... my time line setting being 25 fps, I'll try to make the frequency a multiple of 25, 75 hertz, in case it helps !