Welcome to our community

Be a part of something great, join today!

Register 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 :)

Return of Purple Fringing.

R

Rick Cook

Active member
Joined
Nov 15, 2012
Messages
32
Reaction score
1
Points
8
Location
Philadelphia, PA USA
Hey ya’ll

Anyone else experiencing more severe then usual Purple fringe with Komodo?

I am expecting lots of call outs of this just being chromatic aberration, which may be a part of this but I have spent the past few days testing every combination of lenses I own (Not a ton) but cannot find a lens that won't produce neon purple fringing. Lenses that never produced this on other bodies.

I was an original R1 owner, loved that camera dearly. During my time with that camera I experienced the documented Purple fringe issues with its sensor often. I shot with many lenses, and all in all it was inevitable that if you are shooting against a white sky, you will have severe purple artifacts.

In the years since owning that camera, I have shot mostly with the Alexa and my own FS7 and occasionally rented Helium or a buddies old Scarlet W and I have never had the issue on my FD set, Canon L glass or CP2's (CP2's did it on the R1 but never Alexa).

Could this be a lack of filtration on the sensor? What sort of filtration does the komodo have if any?
 
Rick Cook said:
Hey ya’ll

Anyone else experiencing more severe then usual Purple fringe with Komodo?

I am expecting lots of call outs of this just being chromatic aberration, which may be a part of this but I have spent the past few days testing every combination of lenses I own (Not a ton) but cannot find a lens that won't produce neon purple fringing. Lenses that never produced this on other bodies.

I was an original R1 owner, loved that camera dearly. During my time with that camera I experienced the documented Purple fringe issues with its sensor often. I shot with many lenses, and all in all it was inevitable that if you are shooting against a white sky, you will have severe purple artifacts.

In the years since owning that camera, I have shot mostly with the Alexa and my own FS7 and occasionally rented Helium or a buddies old Scarlet W and I have never had the issue on my FD set, Canon L glass or CP2's (CP2's did it on the R1 but never Alexa).

Could this be a lack of filtration on the sensor? What sort of filtration does the komodo have if any?
Click to expand...

Can you share some screenshots (or even better some R3D files) so that we can check this to help you ?
 
Have you shot something at T5.6 - T8.0? Most lenses create CA, even stopped down a little. The color science or sensor might exaggerate that spectrum and therefore you will see it more on Komodo, but in essence... it's a prime reason why good glass costs more.
Maybe you could rent something like a Master Prime and see how it behaves.
 
Rick Cook said:
Hey ya’ll

Anyone else experiencing more severe then usual Purple fringe with Komodo?

I am expecting lots of call outs of this just being chromatic aberration, which may be a part of this but I have spent the past few days testing every combination of lenses I own (Not a ton) but cannot find a lens that won't produce neon purple fringing. Lenses that never produced this on other bodies.

I was an original R1 owner, loved that camera dearly. During my time with that camera I experienced the documented Purple fringe issues with its sensor often. I shot with many lenses, and all in all it was inevitable that if you are shooting against a white sky, you will have severe purple artifacts.

In the years since owning that camera, I have shot mostly with the Alexa and my own FS7 and occasionally rented Helium or a buddies old Scarlet W and I have never had the issue on my FD set, Canon L glass or CP2's (CP2's did it on the R1 but never Alexa).

Could this be a lack of filtration on the sensor? What sort of filtration does the komodo have if any?
Click to expand...

Each set of lenses that You've mentioned - has CA. Be it .CP2, FD of Canon L - all of them are known to have CA (some more, some less). They are not MP, or even UP - the more resolution you have the more of CA you'll see. While, starting from 5d mk II Canon had built-in CA corrections on the fly, I suppose something similar could be built in on some camera bodies. Thus, never seen nothing like this on Red bodies - Helium 8K, Dragon 6k, Gemini 5k, Scarlet MX - all show the same CA on my RPZ 18-85 T2.9 cine zoom.
 
I was expecting the responses I got since it’s the typical response when you search for issues related to purple fringing on red sensors here.
Yes, the lenses I own have some level of CA, no they are not masterprimes. So is everything shot on a red body shot on masterprimes or at 5.6? As I’ve mentioned I shoot on a lot of different camera bodies and this is only a problem I’ve encountered on the the red bodies I’ve owned.
It could be color science, it could be pixel density, it could be specific to OLPF, or lack of IR cut? Hell I posted because maybe it’s my Komodo in particular.

I’m was more curious what others have done to avoid this on their sensors or if they experience it at all. I can certainly post some r3ds later.

Thanks!
 
Rick Cook said:
Hey ya’ll

Anyone else experiencing more severe then usual Purple fringe with Komodo?

I am expecting lots of call outs of this just being chromatic aberration, which may be a part of this but I have spent the past few days testing every combination of lenses I own (Not a ton) but cannot find a lens that won't produce neon purple fringing. Lenses that never produced this on other bodies.

I was an original R1 owner, loved that camera dearly. During my time with that camera I experienced the documented Purple fringe issues with its sensor often. I shot with many lenses, and all in all it was inevitable that if you are shooting against a white sky, you will have severe purple artifacts.

In the years since owning that camera, I have shot mostly with the Alexa and my own FS7 and occasionally rented Helium or a buddies old Scarlet W and I have never had the issue on my FD set, Canon L glass or CP2's (CP2's did it on the R1 but never Alexa).

Could this be a lack of filtration on the sensor? What sort of filtration does the komodo have if any?
Click to expand...

Each set of lenses that You've mentioned - has CA. Be it .CP2, FD of Canon L - all of them are known to have CA (some more, some less). They are not MP, or even UP - the more resolution you have the more of CA you'll see. While, starting from 5d mk II Canon had built-in CA corrections on the fly, I suppose something similar could be built in on some camera bodies. Thus, never seen nothing like this on Red bodies - Helium 8K, Dragon 6k, Gemini 5k, Scarlet MX - all show the same CA on my RPZ 18-85 T2.9 cine zoom.
 
I am not at home atm so can’t upload any raw clips. Here is some quick dirty footage shot with a FD 50 1.4 stopped down to F2. Again, this lens does have CA but it never glows purple like this on any other camera I have shot with ranging from DSLR’s up to MiniLF. Every lens I own is glowing like this. Maybe CA is creating out of gamut color causing it to light up like this?

https://youtu.be/kbt9RunUtow
 
Last edited:
That’s ugly. I seem to remember this on some other red cam but can’t remember which cam/firmware. I know Komodo leans magenta, so that could be part of it. If you don’t see it on other cams in the same scenario then that def tells you it’s the camera.
 
Could be wrong, but that seems almost too consistent to be CA; it seems equal across the entire frame. I can even see it while viewing on my phone (the smaller screen usually hides a lot of lens performance issues), and there's no way you wouldn't have noticed it on other cameras if it was happening there (especially Helium and Scarlet-W)...Conversely, maybe it is just exaggerated/visible because there's no off-sensor-stack OLPF in Komodo? Have you shot with any other 4k+ cameras that lack typical OLPF (like the BMDs)?

Graeme had a purple de-fringing LUT specifically for this purpose, but I believe it was for Dragon and eventually was rolled into the colour science. Re-update the firmware, black shade (do you black shade with Komodo?) and try your least CA lenses to see if that fixes/helps it. Also might be worth sending off a log and .r3d still to RED and see if they have any thoughts.
 
Alex Lubensky said:
Each set of lenses that You've mentioned - has CA. Be it .CP2, FD of Canon L - all of them are known to have CA (some more, some less). They are not MP, or even UP - the more resolution you have the more of CA you'll see. While, starting from 5d mk II Canon had built-in CA corrections on the fly, I suppose something similar could be built in on some camera bodies. Thus, never seen nothing like this on Red bodies - Helium 8K, Dragon 6k, Gemini 5k, Scarlet MX - all show the same CA on my RPZ 18-85 T2.9 cine zoom.
Click to expand...

We have had this as well in the last several generations of ENG camera bodies from Panasonic and Sony for certain higher end Fuji and Canon ENG lenses. The way it worked for my last Varicam(s), I downloaded the CAC lens files from Panasonic for the applicable lenses and loaded them into the camera and when an applicable lens was attached, it was automatically recognized and the corrections applied in real-time.
 
I think a better white balance helps. I don't own any FD lenses so I can't comment on how they handle C.A. throughout the Aperture range. However, as others have mentioned usually in areas where there is a high amount of contrast C.A will be exhibited. I'm not saying that the Komodo may be to some part contributing to this, I just think the Komodo may be less forgiving with slightly off white balances than usual.


Here is a different white balanced version of your clip with some C.A. control.

446MB 4096x2160 DCI
https://sendgb.com/s4JIhFXC8jv
 
It's not that it can't be fixed in post, the problem is if this isn't CA and actually how his sensor handles high contrast edges, it would suck a lot of ass having to fix every shot (or not being able to use ColourTemp creatively)... It's not "overheat after 12mins" bad like the R5, but it's certainly not something worth paying for.

That said, no one else is complaining about it, so I'm thinking it'd just be that Rick got a lemon (again, presuming it *isn't* CA, which is still a possibility). And as far as I'm aware, Alexa, Helium and SW/Dragon don't have realtime CA fixing, so it's not that. Could be a combo of little-to-no OLPF, higher than 3.2k resolution (smaller/<5.5u photosites), *and* the lens CA... Like a perfect storm that exacerbates the lens's artifacts (just odd that it's across all of Rick's glass).

Rick, are the other lenses as bad as that 50 1.4 FD? Cause I'd also think that if it's more or less on other glass, than it suggests it's lens performance/CA (not the sensor).
 
rand thompson said:
I think a better white balance helps. I don't own any FD lenses so I can't comment on how they handle C.A. throughout the Aperture range. However, as others have mentioned usually in areas where there is a high amount of contrast C.A will be exhibited. I'm not saying that the Komodo may be to some part contributing to this, I just think the Komodo may be less forgiving with slightly off white balances than usual.

Hey Rand, thanks for taking a look! I was able to to remove most of it via hsl and wb as well, but still can’t figure out what making this pop so much. I have tested all of my other lenses and while it’s more prominent on others it’s existing anytime a stick is against a white sky, very odd. It is very barely noticeable on my monitor while shooting even when magnifying.

I just can’t expect others to always Key this out of my footage in post.

I’m going to try shooting some tests in r3d HQ shortly to see if that helps.
Click to expand...
 
Rick Cook said:
Here is a link to one of the R3D's. Let me know what ya'll think. Worth noting this is R3D LT.

Its a wild a neon peaking effect. It's visible on edges against the sky in every backlit exterior shot Ive done regardless of stop.

https://drive.google.com/drive/folders/1nXX4UiKXiP--1aBUB-dkiAQM5ad_DRht?usp=sharing
Click to expand...

Confirmed. I put the clip into Resolve, opened up the Vector scope and watched it play. I adjusted color temp to 7000 and tint to +18 just to center the white blob, and with the Vector scope set to 2x zoom, you can really see the fringing come and go as focus and motion blur make certain features sharp or blurry. The last frame of the 2GB clip is one of the brightest in the vectorscope.

Using a qualifier to cut the purple saturation definitely removes the purple fringing, at the expense that one cannot really use purple in that shot without lots of extra secondary grading. Quite surprised at how strong the effect is.

Nevertheless, you can definitely see how much the lens can be to blame if you start at frame 209-214 (where purple fringing is prominent) and step through to 218-221 (where you can see some out-of-focus branches are clearly purple and others clearly green) you can see that the lens is truly splitting the wavelenghts for you, and that whatever you are choosing as your focus point is getting the red/blue stuff in focus and leaving the green stuff out of focus.
 
Rick,

I have been personally experimenting on how Red's change on how Kelvin and Tint reacts in the komodo from the "Planckian Locust" (1700K to 4000K range) to the "Daylight Locust" (5000K and above range) at the different Color temperatures ranges at different times of day. I think all of the image issues some are seeing in the Komodo can partially or completely be contributed to this , but that's just my personal opinion.

Screenshot-481.png
 
rand thompson said:
Rick,

I have been personally experimenting on how Red's change on how Kelvin and Tint reacts in the komodo from the "Planckian Locust" (1700K to 4000K range) to the "Daylight Locust" (5000K and above range) at the different Color temperatures ranges at different times of day. I think all of the image issues some are seeing in the Komodo can partially or completely be contributed to this , but that's just my personal opinion.

Screenshot-481.png
Click to expand...

Is that a hardware/sensor design change with Komodo, or a colour science/debayer/software-side change for all r3d footage?
 
At the root of it, the lens is introducing the color.

I just went out with three lenses at 50mm, this FD, CP2 and a Canon 24-70mm at 50mm at 2.8. The L series Zoom does a good job but I still get the neon purple hits. Its no doubt aberration causing it I'm just trying to work though what can help with it. As Mike P said I am feeling like its a perfect storm that exacerbates the lens's artifacts into something much more unpleasant. I'm more surprised others are not having the issue, which is the main reason I posted here. On most of my footage I can zoom in and find neon purple hits.
 
Mike,

I'm trying to figure that out myself. But if I had to make a guess, I would say Sensor. The reason, because all existing IPP2 Output Transform luts work fairly well with Komodo. Maybe IPP2 Transform Luts specifically aimed at how the Komodo reacts at certain color temperatures are needed.
 
You must log in or register to reply here.
Back
Top