Jump to content


Photo

Concave or Convexed Distortion Problem.


  • Please log in to reply
5 replies to this topic

#1 Tom Sykes

Tom Sykes
  • Basic Members
  • PipPip
  • 56 posts
  • Other
  • Manchester, UK.

Posted 01 May 2012 - 01:39 PM

How do Folks.

I've run into a problem which I haven't really come across before and wondered if some of the more experienced of yourselves could hopefully shed some light on the situation.

We are closing on the final edit of a film i've just shot, and doing all the post work.

Now, there is one particular shot that is treating us with some difficulty, which is a handheld "steady" rig shot, walking backwards, down a hill, with lens distortion.

Now what we're attempting to do; is figure out, whilst smoothing out the rolling shutter issues, how we could go about getting rid of the lens distortion, therefore making the image "flat".

For the shot we used a Canon 16-35mm USM L II f2.8 lens, and in the image the subject appears to come closer to us in the center as opposed to the sides.

What we want to know is whether the lens distortion is convexed, or concaved, and whether it the number of elements and their positioning which affect this, and also how to go about correcting it.

We are trying to correct in AE and I would appreciate a lot if anyone has any information at all about how to go about this, I've been loomed with the task of researching into the mechanics of the lens and supplying that information. I just thought if anyone on here knows, then best to ask as I need to find out ASAP.

Filters we can use in AE? Concave or Convex, any other information at all would be appreciated.

Let me know if you require more information, I suppose this will open up the discussion for now.

Thanks,

Tom.
  • 0

#2 Chris Keth

Chris Keth
  • Sustaining Members
  • 4427 posts
  • 1st Assistant Camera
  • Los Angeles

Posted 01 May 2012 - 01:50 PM

I think generally the way to do this would have been, in prep, to shoot a grid pattern with that lens at the focal length you wanted to use. That would of course distort that grid in a way that it could be used to create a custom filter to help flatten the footage. Since you did not do that, you will want some kind of concave applied filter. The parameters of that filter? Who knows. You will just have to experiment.
  • 0

#3 Dom Jaeger

Dom Jaeger
  • Basic Members
  • PipPipPipPip
  • 1599 posts
  • Other
  • Melbourne, Australia

Posted 01 May 2012 - 09:58 PM

Sounds like barrel distortion, which is convex. Typically with a zoom the amount of distortion varies depending on the focal length, and can sometimes go from barrel to pincushion. It can also vary depending on the focus distance, close focus in particular. It's an aberration often found in large-range zooms or ones that go to a wide angle. With wide angles (especially at closer focus distances) the center of the subject actually is much closer than the edges, so when un-corrected and flattened onto an image plane the middle will seem magnified. Fisheye lenses are like this. Rectilinear lenses are corrected within the optical design to maintain straight lines, even though with very wide fields of view this causes another type of distortion where the objects at the edges can seem stretched out. A zoom needs to try to correct the natural distortion at the wide end without introducing distortion at the long end, as well as maintaining this through the focus distance range. Some are more successful than others.

There are plenty of plug-ins/programs to correct this sort of radial distortion, but it's not something I'm familiar with. What Chris said makes sense. Shooting a grid would also identify whether the distortion is more complex, such as a flat field at the edges with barreling only in the centre.
  • 0

#4 Dom Jaeger

Dom Jaeger
  • Basic Members
  • PipPipPipPip
  • 1599 posts
  • Other
  • Melbourne, Australia

Posted 01 May 2012 - 11:10 PM

After a very quick google I found Ken Rockwell's review of this lens, where he mentions the distortion and even gives some parameters for correction using Photoshop CS2's lens distortion filter.

http://www.kenrockwe....htm#distortion
  • 0

#5 Tom Sykes

Tom Sykes
  • Basic Members
  • PipPip
  • 56 posts
  • Other
  • Manchester, UK.

Posted 02 May 2012 - 03:58 AM

Chris.

Thank you for the recommendation of shooting a Grid, naturally I hadn't planned on shooting that wide at any point, but unforeseen circumstances made the shot more plausible by doing so, which is why we didn't do a test in prep.

I have now shot a grid test at different focal lengths and sent the files over to our man at the other end and he's on it now and has said he should be good with that. So hopefully...

Dom.

Yes, I did actually find this page as well, it was indeed helpful. The wonders of google eh?

Cheers for the help guys.

Tom.
  • 0

#6 Tom Sykes

Tom Sykes
  • Basic Members
  • PipPip
  • 56 posts
  • Other
  • Manchester, UK.

Posted 02 May 2012 - 04:05 AM

Sounds like barrel distortion, which is convex. Typically with a zoom the amount of distortion varies depending on the focal length, and can sometimes go from barrel to pincushion. It can also vary depending on the focus distance, close focus in particular. It's an aberration often found in large-range zooms or ones that go to a wide angle. With wide angles (especially at closer focus distances) the center of the subject actually is much closer than the edges, so when un-corrected and flattened onto an image plane the middle will seem magnified. Fisheye lenses are like this. Rectilinear lenses are corrected within the optical design to maintain straight lines, even though with very wide fields of view this causes another type of distortion where the objects at the edges can seem stretched out. A zoom needs to try to correct the natural distortion at the wide end without introducing distortion at the long end, as well as maintaining this through the focus distance range. Some are more successful than others.

There are plenty of plug-ins/programs to correct this sort of radial distortion, but it's not something I'm familiar with. What Chris said makes sense. Shooting a grid would also identify whether the distortion is more complex, such as a flat field at the edges with barreling only in the centre.


Apologies Dom, totally missed this post.

Very useful information, Thanks. I think we shot it at about 20mm so the distortion isn't the greatest, but is a necessity to get rid of really. Our subjects weren't too close to the lens for the shot which helped as the focus distance has helped us out in that sense.

Cheers again.
  • 0


Aerial Filmworks

Tai Audio

The Slider

Glidecam

Abel Cine

rebotnix Technologies

Metropolis Post

Media Blackout - Custom Cables and AKS

Willys Widgets

FJS International, LLC

Rig Wheels Passport

Wooden Camera

CineLab

Paralinx LLC

Ritter Battery

Gamma Ray Digital Inc

Opal

Technodolly

Visual Products

Broadcast Solutions Inc

CineTape

Tai Audio

FJS International, LLC

CineTape

Gamma Ray Digital Inc

The Slider

Abel Cine

Ritter Battery

rebotnix Technologies

Visual Products

Technodolly

Wooden Camera

Paralinx LLC

Broadcast Solutions Inc

Media Blackout - Custom Cables and AKS

Glidecam

Opal

Willys Widgets

Aerial Filmworks

CineLab

Metropolis Post

Rig Wheels Passport