Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Ugly Pink Outline - Help!

Billy McCannon

Billy McCannon

Well-known member
Joined
Dec 28, 2006
Messages
266
Reaction score
0
Points
0
Location
Dublin
The attached still is from footage shot at 4k 16:9/25fps.

Lens: Canon 35mm f1.4L shooting wide open. + uv filter

The pink outline appears around everything in focus against the window, including the window frame.

I checked it on both RA and RCine and the artifact is visible on both.

This still is at iso 1600 - the pink line is more apparent here but it is visible at 500asa also. At ISO 500 it is possible to make out detail outside. In this render the highlights are obviously gone.

I also shot on Canon 24 - 70L but the light outside was less and the stop was 2.8. There is no visible outline on this footage. The 35mm Canon is brand new - I'm not familiar with it but I've never seen this kind of aberration from a lens.

Has anyone else seen this? Know what is and how to avoid it.

Is it a sensor issue or user issue?

Welcome any suggestions.
 
That's chromatic aberration due to your lens quality. It also shows up more when you open up.

The 24-70 didn't have CA because I'm assuming it's the L version (does Canon have a 24-70 non L-version?) Anyway, that's one of the benefit of L lenses--no to very little CA.

http://en.wikipedia.org/wiki/Chromatic_aberration

There are plug-ins that correct it. I don't know how well they work. The only one I heard about, though, is pretty high-end. Maybe AE has one? Lightroom and PS can fix it.

The best way to avoid it is to make sure to buy good quality lenses.
 
the CA is more prominent since you have an extreme case of contrast in this shot. the
window is grossly over exposed, notice how everything within the window has the pink fringe. I think this has more to do with the window exposure than the lens.
 
From what I've seen, it's a lens thing. Usually when the lenses are fully open.
 
The lens was in fact a Canon 35mm 1.4L lens. I've edited the original post to correct that error. The Canon 24-70mm 2.8L - also L.

Perhaps it's not a good copy of the lens?

Putting a scrim on the outside of the window wasn't an option - this is low budget documentary. As I said in the original post - the exterior is not was not very overexposed when viewed at ISO 500 - but the pink outline is still visible.
 
I've seen it on the Canon 85 F1.2 L, other Canon glass too. But I've also seen it on Cooke's and Zeiss. It's a lens thing. Wide open. I think all lenses do it.
 
This is not CA (chromatic aberration) If it were there would be a blue/green offset as well. If it were CA it would also show up more uniformly across the photo, and tend to be more exaggerated towards the outer edge of the frame.

This is electron bleed (goes under some other names, sensor saturation, pixel interference) They look very similar but are very different.

When a chip becomes extremely saturated (over exposed) the variance in the charged state of the pixels starts to effect the pixels next to it. Think of it like speaker wires that are to close to power wires and develop hum. The reason in this case it goes away when stopped down is the voltage discrepancy on the chip becomes less, and more within the range the chip can deal with.
 
What you've seen on film is most likely CA. Like I said they look very similar, but in this specific case it is not CA
 
I believe Mike Prevette is right on this one, it is a sensor thing.

I believe you have seen CA on film that looked similar.


edit: I missed mike's next post.
 
I shot an entire day with 4K HMIs blasting in through windows at close range for very intentional major flarage, and while I saw the purple sun effect, I did not see anything close to this, nor have I seen it before in similar situations...

I agree it looks a little too perfect to be straight up CA.
 
NateWeaver said:
I shot an entire day with 4K HMIs blasting in through windows at close range for very intentional major flarage, and while I saw the purple sun effect, I did not see anything close to this, nor have I seen it before in similar situations...

I agree it looks a little too perfect to be straight up CA.
Click to expand...


The trick is to saturate a large area of the sensor. That makes the difference between the voltages leaving the sensor greater and more likely to bleed into one another.
 
I've had the issue a couple of months back as well, and both Graeme and another Red team member instisted it was CA, I still don't think it is CA, but a sensor thing, CA looks different and has a softer edge , so let's turn it into a poll.....
 
I've seen the purple on small highlights which would disprove the large area saturation theory. Has anyone tried comparing stopped down v ND down to see what happens? If it's CA, stopping down will help, but ND will not. I've always seen this is a type of CA
 
Sorry but why did you post a still that was rendered at ISO 1600 when you shot it at ISO 500? Definitely you will have over exposure and the CA caused by extreme contrast.
 
I usually find CA to shift to one side. This wraps around anything that is overexposed. Weird. What is the color temp set to? I had someone show me some overexposed streetlights that they were worried about that they shot with my camera. The overexposed areas turned opaque cyan. When I raised the color temp in RA they turned white again and everything looked normal. Probably not the issue, just trying to think of what it could be if not CA.
 
I'd like to see a full resolution frame, but this looks like purple fringing to me, which is essentially a form of CA. But purple fringe occurs around high contrast edges and is not as directional as typical CA, nor will it be offset by cyan casts. PF also differs from typical CA in that it tends to be more consistent throughout the entire FOV, whereas typical CA often worsens as you move outward from the center of your image.

A good way to test CA and fringing on a lens is to shoot WFO (and at max focal length if it's a zoom) a backlit chart.

Here's a 1:1 crop of the lower right corner of a 4K 2:1 frame shot REDCODE 36, Build 18 beta. Cooke 18-100 @ 100mm T3. Notice the CA. This is as bad as it gets on this lens. And yes, the chart was upside down. It's really just a chart image printed on paper so I could do a back-lit test. The lower corner is more out of focus because the paper wasn't mounted completely flat, just taped top and bottom and the camera wasn't entirely perpendicular... blah, blah, blah... But it shows the CA and that was the point.
535_1228530748.jpg
 
Another thing we could be seeing here, since you say this still was developed for ISO 1600... is an artifact related to the "purple sun" over-exposure anomaly on the RED One. For some reason, the note about ISO 1600 didn't register the first time I read the original post. If this is the case, then Mike Prevette is absolutely right about the overflow. Anyone who has ever over-exposed highlights, the sun, a light / glare, etc.. and then played with it in post, adjusting exposure will notice the purple/magenta ringing that can occur at a certain over-exposure point, before it completely blacks out. I've got some images of that effect somewhere too... I'll see if I can dig them up and post one.
 
You must log in or register to reply here.
Back
Top