Grid like flare from OLPF

[Ed Watkins]

I am getting some bad flaring from (what I assume is) the OLPF, a question about this was raised a while ago but no solutions were posted.

Every time I try and shoot sunrise/sunsets I get the effect seen in the attached image.

It occurs on PL lenses, and all of my Nikon lenses.
It happens with and without filters on the front of the lenses.
It is worst on my 17-35mm and 80-200mm zooms.

What is causing it? (I'm guessing it is light bouncing around the RED coloured OLPF)
How can I stop it?

I am currently working on a project that requires a number of sunrise/sunset shots and currently they are all unusable because of this grid like flare.

Any help would be greatly appreciated.

 

[Dylan Macleod CSC]

I am getting some bad flaring from (what I assume is) the OLPF, a question about this was raised a while ago but no solutions were posted.

Every time I try and shoot sunrise/sunsets I get the effect seen in the attached image.

It occurs on PL lenses, and all of my Nikon lenses.
It happens with and without filters on the front of the lenses.
It is worst on my 17-35mm and 80-200mm zooms.

What is causing it? (I'm guessing it is light bouncing around the RED coloured OLPF)
How can I stop it?

I am currently working on a project that requires a number of sunrise/sunset shots and currently they are all unusable because of this grid like flare.

Any help would be greatly appreciated.

The only thing I can suggest is trying different lenses. I think that if the back of the lens element is fairly large and flat, this is more likely to happen.

How are you handling the black spot sun?

Dylan Macleod
Cinematographer
Toronto, Canada
www.dylanmacleod.com

 

[Ed Watkins]

The only thing I can suggest is trying different lenses. I think that if the back of the lens element is fairly large and flat, this is more likely to happen.

How are you handling the black spot sun?

Dylan Macleod
Cinematographer
Toronto, Canada
www.dylanmacleod.com

Thanks Dylan,

I have tried a whole bunch of lenses, and it occurs to some degree on all of them. (Nikon: 17-35, 35, 50, 105, 80-200, Zeiss standards: 16, 32)

The black spots I am simply tracking in post (AE) and masking with white circles. Not ideal, but easy to do.

Anyone from RED got any suggestions to reduce the grid pattern? Could it be a back focus adjustment issue?

 

[I Bloom]

I think the back reflection is actually from the OLPF not the back of the lense. I believe this because when RED moved the OLPF the character of the flares also seemed to change. Can anyone verify this?

IBloom

 

[Colin Hubick]

weird grid

weird grid

We too have been shooting some sunrise shots for a tourism campaign, and experienced these grids. Any info would be nice.

Colin

 

[mikeburton]

I get the same thing. I have been testing some of my Lomo Round Front anamorphics that I just recently got for their flaring properties and those OLPF dots keep showing up for me too. It happens on any kind of direct light through the lens especially sunlight.

I took this last night quickly in my garage, handheld and wide open on a 75mm Lomo Round front anamorphic. You can plainly see the red dot about eye level in the center of the screen. There is more that is going on in this shot to be concerned with ie. soft focus, a lens repair needed etc. but this image and the shot demonstrates these nasty dots that appear when a direct light source hits the OLPF.
What's the use of using anamorphics if you can't flare them cause of this dot problem?

 

[A. Bastaki]

howcome none of this shit happens to me?!

i keep waiting for stuff like this to happen.. like the skewing thing... i love it when im on set and go like.. ohhhhh did you just seeee thaaat!!!! it skewwwedd... oorrr did you just see thaaat.. thats LPABCDEFG!!! Yeaaaaaaah!!! and everyone on set goes like... woah!! we get to gain experience using this thing.. now we can brag how we undo the issue on our next/other gig and tell em how kick ass we are.

 

[mikeburton]

howcome none of this shit happens to me?!

You shouldn't be upset about that

 

[hunterrichards]

It hasn't happened to me either. I can see where this would be an unwanted effect, but on the contrary I actually like the way it looks and wouldn't mind if the lens caught an angled flair and this popped up for a bit. It is odd that some people see it and some people dont.

 

[mikeburton]

It hasn't happened to me either. I can see where this would be an unwanted effect, but on the contrary I actually like the way it looks and wouldn't mind if the lens caught an angled flair and this popped up for a bit. It is odd that some people see it and some people dont.

It really isn't attractive when in motion! I hope this is a problem easily solved quickly!

 

[Colin Hubick]

grid flare

grid flare

No it isn't a desirable look. Ugly actually.

 

[Ed Watkins]

It's not a pretty effect, neither is an angry director.
A beautiful sunrise I got up at 4am to film is pretty.

howcome none of this shit happens to me?!

I am quite possibly the opposite of Akube, I seem to have a knack for finding problems with my setup.
I feel terribly sorry for Travis who has to keep dealing with me at RED, he must be quite sick of me by now.
He's a very nice chap though.

 

[mikeburton]

It's not a pretty effect, neither is an angry director.
A beautiful sunrise I got up at 4am to film is pretty.


I am quite possibly the opposite of Akube, I seem to have a knack for finding problems with my setup.
I feel terribly sorry for Travis who has to keep dealing with me at RED, he must be quite sick of me by now.
He's a very nice chap though.

Hey Ed,

If you get an answer to this issue one way or anther from RED, please let us know. I would like to find out if this is something they are working to improve or solve completely.

 

[Ed Watkins]

Hey Ed,

If you get an answer to this issue one way or anther from RED, please let us know. I would like to find out if this is something they are working to improve or solve completely.

I haven't contacted RED about this yet, I was going to see if the forum could solve the issue before bothering Travis again.

If I have no luck here I'll contact them.

 

[Dan Hudgins]

Grid pattern...

Grid pattern...

I am currently working on a project that requires a number of sunrise/sunset shots and currently they are all unusable because of this grid like flare.
Any help would be greatly appreciated.

You might put a heavy ND/IR filters on the camera and shoot at f/1.2 that might defocus the grid into a general flair, then underexpose to help knock the flair into the black.

The best thing would be to take the OLPF out of the camera, shoot 4K with a random "black dot" filter in front of the lens, blur the file one or two pixels, then downrez to 2K and then rez back up to 4K using a multi pixel resize to knock out some of the color moire you will pick up. Also shooting at f/1.2 should reduce the color moire from the lack of OLPF.

You get this type of grid flair if you shoot through a fiberglass "silk", or use a holographic star filter, which is sort of what the OLPF is. Some use crystals to break up the light so your image is made into four superimposed images that can interfere, so that may be where part of the rainbow flair comes from.

The OLPF is used to split the light into four beams so that the points of light go to all for colors of pixel, as you know the Bayer filter is a box with one red corner, one blue corner, and two green corners, the OLPF divides the light so that all four pixels in each group get some of the light to reduce color moire and to reduce the resolution by sort of half to reduce alaising artifacts.

The sensor will work without the OLPF if the lens is wide open so the edges are a little out of focus and you trun off all sharpen features. If you sharpen without the OLPF you will get bad color moire and aliasing artifacts.

You might try a "black dot" filter over the lens, it has a random pattern rather than a box pattern like a silk or holographic filter. With the "black dot filter" over a wide open lens, you might get away better without the OLPF.

Some of the other reflections look like they are internal reflections from the lens elements, some lenses are bad with these. Before lenses were coated the desiginer tried to have the reflections very out of focus. Computer programs that design lenses now do not optimize the out of focus reflections, and so you can get upsidown reflections on bright objects the multi coating is not enough to make up for the bad computer programs used to design lenses now that mostly look at the infocus image.

You would not shoot a sunset through a window screen...

You can also get some box grid artifacts from resizing images if you use software that does not do a good job, so check the source of the problem and see what you can do. The RED ONE (tm) shoots well, but sometimes you might want to just shoot film and have it scanned, pointing the camera into the sun does not seem like a good idea anyway since if you burn the sensor or fade/burn marks into the bayer filter you need to have the camera fixed?

 

[Colin Hubick]

Go Ed Go

Go Ed Go

I haven't contacted RED about this yet, I was going to see if the forum could solve the issue before bothering Travis again.

If I have no luck here I'll contact them.

let me know what you find.

 

[Ed Watkins]

Thanks for the very detailed post. It is very helpful.

You might put a heavy ND/IR filters on the camera and shoot at f/1.2 that might defocus the grid into a general flair, then underexpose to help knock the flair into the black.

That's pretty much what I ended up doing (on a 50mm, f/ 1.4), it was still vaguely present but not as ugly. The shot I attached was just one of the tests I shot to see what was causing the grid.

The OLPF is used to split the light into four beams so that the points of light go to all for colors of pixel...

So, if I'm reading correctly, what you're saying is that the pattern is due to the way the OLPF refracts the light for the bayer sensor. And the only way to solve this is to remove it? Unfortunately, taking the OLPF is somewhere I'm not willing to go.

Some of the other reflections look like they are internal reflections from the lens elements, some lenses are bad with these....

The image attached was a test to see what made the grid worst, there are definitely a few other flares caused by the lens present.

But, why do only some people see the grid shooting single point light sources and others don't? I have the effect on a number of different lenses (old and new) I doubt all of them could be compared to shooting through a window screen.

You can also get some box grid artifacts from resizing images if you use software that does not do a good job...

The image was a screen grab from RED CINE the image hadn't been re-sized, also I saw this in the LCD when shooting.

 

[Dan Hudgins]

Point light source...

Point light source...

So, if I'm reading correctly, what you're saying is that the pattern is due to the way the OLPF refracts the light for the bayer sensor. And the only way to solve this is to remove it?

The whole point of the OLPF is to make four images, one shifted about a pixel to 1:30 on the clock, one shifted to 4:30, another shifted to 7:30, and the last to 10:30. The star filter effect is an artifact that may not show up as bad with some OLPF that fit closer to the sensor maybe, I have not tried all the brands of OLPF myself to say for sure. I think on the SI-2K site they said you could take the OLPF out of their camera and shoot without it when you want a sharper result (or no OLPF artifacts),

http://www.siliconimaging.com/DigitalCinema/SI_2K_full_specifications.html

Quote:

[Removable optical low pass filter with IR cut filter]

Does the OLPF in your camera have a height adjustment?

But, why do only some people see the grid shooting single point light sources and others don't? I have the effect on a number of different lenses (old and new) I doubt all of them could be compared to shooting through a window screen.

This may be due to the power of the back group in the lens behind the iris, some lenses like telephoto have negative power in the back group, and others like retrofocus have positive power in the back group. This may change the "size" or "angle" of the iris "pin hole" from the distance of the OLPF, the smaller the iris seems the sharper the quad flair maybe?

That's why when you open up the lens the quad flair should get more amorphous...

Try putting your eye close to the back of the lens, like at the OLPF distance, and see how big the iris looks when the lens is stopped down to 22...

There also seem to be two types of OLPF used on RED ONE (tm) cameras of different serial numbers, maybe someone can say why the change was made?

The window screen note was that when you shoot through a screen you also get a quad rainbow diffraction pattern on a point light source, just at clock points 3, 6, 9, 12 rather than the 45 degree angle of the OLPF that gives you 1:30, 4:30, 7:30, and 10:30 image offsets. Anything with a small pattern can cause diffraction, like the "groves" in a CD disk make rainbows, that's why you do not get aliasing with film, the grain is random and changes from one frame to the next (and the resolution is low). I was focus testing on some 7360 film with its high resolution on a window screen, if you held the film up to your eye you could see the quad patterns in the frames that were in focus and not in the soft ones, its a great way to check the backfocus, moire on a sensor is maxed out when the lens is in focus also if you have a pattern with the right pitch.

 

[mikeburton]

This may not make sense or be correct at all but don't you think this may have something to do with the build possibly? I remember this being an issue back in build 11-12 I believe and then it vanished for 13-14, and 15 I think? Now it seems to be back in 16.

Like I said, I could be wrong but I haven't seen this issue in a while until build 16.

 

[Dan Hudgins]

Black level...

Black level...

This may not make sense or be correct at all but don't you think this may have something to do with the build possibly? I remember this being an issue back in build 11-12 I believe and then it vanished for 13-14, and 15 I think? Now it seems to be back in 16.

Like I said, I could be wrong but I haven't seen this issue in a while until build 16.

I think there were some changes to the black level (or curve), the flair may show up more in Build 16 since the black level is higher maybe to reduce some of the compression artifacts in the dark parts of the image?