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 :)

Browny pollution when using ND filters

J

Jakub Burakiewicz

Well-known member
Joined
Dec 31, 2011
Messages
105
Reaction score
0
Points
16
Hey fellows!

I have a problem.
When I use ND filters from 0.9 up, the image from my Epic becomes brawny...
I'm conscious that it is connected to the infrared pollution.
I've bought a Schneider Tru-cut 750 filter to overcome this problem. It worked for the blacks but the color itself remains browny...

Any hints?!

I would be very grateful :)
Jakub.
 
The Tru Cut is said to be one of the best on Red cams. Some discoloration may remain even using them, but it is correctable. Some prefer the Schnieder Platinum IRND's. You might try those.
 
Tru-Cut worked 100% on Red One, NOT on Epic.

For Epic, it's Schneider Platinum, the last great design by the now sadly deceased Bob Zipka, who also created Classic Softs and many other beloved filters.
 
Happens to all cameras including 35mm film, but was not noticed much due to all film going through a colorist to correct before viewing... This is not a REDONE or EPIC issue.
What happens is you loose the color GREEN slowly as you put more filters on. Which you can put back in quite easily later.
I posted a thread on this some time back.

IR is influenced when loading up filters past the 4 stop barrier onwards...
IR (free) ND's still get it, but not as bad, you can push past 6 stops.

Some filters are worse than others.
 
This is not the same as what happened with film Mark.

Yes, some old NDs would mix out green before other colors, but IR pollution is due to sensors reading infrared light (basically heat, not really a light wave in the same way) along with the actual color red. It didn't happen with film because film was not very sensitiev to infrared light, but CMOS sensors are, very much so.

The hotter a scene is (in actual temperature i.e. 30 Celcius and above) and the more you are filming heat absorbing materials, like synthetic fabrics and foliage, the worse the effect will be. They are most apparent in dark fabrics and dark foliage, because these absorb a lot of heat, but emit little "normal" light, so the balance of infrared light vs what "should" be there is heavily skewed toward infrared - That is in fact the problem with NDs in general - they block out the visible spectrum, but not infrared to same degree, so one is left with a disproportionate amount of infrared light in the final mix of what is recorded. Even a CMOS recording a scene with no ND at all could have a problem with IR pollution, but the OLPF there by default will block out unwanted IR for ND =0- IRNDs attempt to fix but the effect, to some degree or other. If the OLPF were tuned for higher ND levels, it would start to carve away from the red channel when lower or no ND was used - as really CMOS has a problem telling the difference between infrared light and plain old Red light that films, especially modern emulsions, do not.

Because the amount of effect varies based on temperature, material type, etc, and NOT in a uniform way across the frame, to correct it completely sometimes requires creating detailed animated masks or pulling HSL secondaries that can accentuate noise. Even if you do these intensive corrections, infrared light is blurry (full of tiny versions of heat ripples) so you can be left with unwanted and non-sexy depth of field shifts within the image. You are also left with an image that was recorded with a color shift in the light itself - and fixing it is a bit like trying to white balance 8 bit footage - the camera simply was not fed light which it could hope to read correctly in terms of color.

Its' not something you can actually fix, in many cases, very easily at all.
 
Last edited:
IR contamination is usually not easy to fix. Some lenses, like Leica Apo Telyt-R, focus IR on exactly the same plane as visible spectrum. You may have a chance there. But, most lenses focus IR on a different plane and produce mushy diffused cast, not really possible to get rid of in post.

I use Pancro IR mirror ND filters up to ND 2.1 (custom made). No IR or colour shift detected. They work well on most cameras and lenses. Some users experienced front lens element reflections. I had no problems with my lenses and filters.
 
Rob Ruffo said:
This is not the same as what happened with film Mark.

Yes, some old NDs would mix out green before other colors, but IR pollution is due to sensors reading infrared light (basically heat, not really a light wave in the same way) along with the actual color red. It didn't happen with film because film was not very sensitiev to infrared light, but CMOS sensors are, very much so.

The hotter a scene is (in actual temperature i.e. 30 Celcius and above) and the more you are filming heat absorbing materials, like synthetic fabrics and foliage, the worse the effect will be. They are most apparent in dark fabrics and dark foliage, because these absorb a lot of heat, but emit little "normal" light, so the balance of infrared light vs what "should" be there is heavily skewed toward infrared - That is in fact the problem with NDs in general - they block out the visible spectrum, but not infrared to same degree, so one is left with a disproportionate amount of infrared light in the final mix of what is recorded. Even a CMOS recording a scene with no ND at all could have a problem with IR pollution, but the OLPF there by default will block out unwanted IR for ND =0- IRNDs attempt to fix but the effect, to some degree or other. If the OLPF were tuned for higher ND levels, it would start to carve away from the red channel when lower or no ND was used - as really CMOS has a problem telling the difference between infrared light and plain old Red light that films, especially modern emulsions, do not.

Because the amount of effect varies based on temperature, material type, etc, and NOT in a uniform way across the frame, to correct it completely sometimes requires creating detailed animated masks or pulling HSL secondaries that can accentuate noise. Even if you do these intensive corrections, infrared light is blurry (full of tiny versions of heat ripples) so you can be left with unwanted and non-sexy depth of field shifts within the image. You are also left with an image that was recorded with a color shift in the light itself - and fixing it is a bit like trying to white balance 8 bit footage - the camera simply was fed light which it could hope to read correctly in terms of color.

Its' not something you can actually fix, in many cases, very easily at all.
Click to expand...

Understood, Thanks Rob.

But this problem is not isolated to RED cameras... that's for sure.
 
Shot 2 years with Panchro's, though they were good in eliminating the IR, the strange reflections that came with it in by example sun reflections in cars forced me to quit with those. I ordered a set of Schneider Platinum in december, and received only the 1.5 and 1.8 in march, the rest was postponed till june.. Too late for me for the movie I was shooting, so I rented a set which was all over the place in terms of green; between 09 and 18 and all different. Got a second set after a pissed off email from the rental house to Schneider, and that one had teh same problem. Apparantly it should be better, but when I used my neutral Platinum 1.8 (the only 2 neutral filters that I got ) on a sunny day on a guy in a black suit, it turned purple anyhow.
So, no good experiences with the Schneiders, unfortunately.
 
Thank you for yours responses!

Has anyone else had the same problem as Mick with Schneider Platinium IRND?
Do you thing it's a good idea to buy only a 0.9 IRND filter? If I need 1.5 for example, I would mix it with a standard 0.6 ND. The pollution shouldn't be that harmful than. Am I right? I just don't have budget yet for a complete set of IRND as I've just bought a new ND set :-/

The interesting thing is that it seems that the Epic is much more susceptible to IR pollution than RED MX....weird.

Jakub.
 
I have the 1.5 and 1.8 platinum IRND and I am a little disappointed. I used them on one shoot and had magenta browny pollution in the dark foliage. Next shoot we pulled them out and pointed them at my assistants nylon jacket where you could easily see the IR contamination. I have not returned them yet as I have not compared them and don't know if any of these heavy ND ilters do better? The only filter that worked perfectly for me was the mirror pancros but not the easiest filter to work with.

Reds sensors with their no gain design can be a real pain in full sun where clients want shallow depth. If Dragon is faster again I really hope Red helps out with their clip on internal IRND filter that was mooted last year.
 
It's been posted a few times here, but if you are looking for a solid IR solution Formatt Hot Mirror/ND combo filters are a great way to go. It's what I use. It's what many others use. They actually cut IR light instead of simple filtration.
 
Formatt works for us just the same on our current Epic like on the R1 before.
 
Mick van Rossum said:
Shot 2 years with Panchro's, though they were good in eliminating the IR, the strange reflections that came with it in by example sun reflections in cars forced me to quit with those. .
Click to expand...

I use Pancro IR+ND and find that I can get rid of unwanted reflections by using a strip of gaffer tape between the edge of the matte box and the camera to slightly tweak the angle of the filter relative to lens front element.
 
Mark Toia said:
Understood, Thanks Rob.

But this problem is not isolated to RED cameras... that's for sure.
Click to expand...

No, they all have it - CCDs less than CMOS, and film much less than both. Alexa, ALSO CMOS, is even just a touch worse.

Canons seem to address the problem by using a very aggressive IR filter on the OLPF, and then adding back back red digitally - but I find the color Red with Canon to be a bot overblown and funky, probably for this reason.
 
Brigham Edgar said:
I also use Pancro IRND's up to 2.1 and they are great...custom made, took a while and not cheap! But I have seen no color shifting from them at all.
Click to expand...

Problem with Panchros is the ludicrously shiny surfaces they have. Make ghosts like crazy, and you cannot stack them with diffusion.
 
Tom Gleeson said:
I have the 1.5 and 1.8 platinum IRND and I am a little disappointed. I used them on one shoot and had magenta browny pollution in the dark foliage. Next shoot we pulled them out and pointed them at my assistants nylon jacket where you could easily see the IR contamination. I have not returned them yet as I have not compared them and don't know if any of these heavy ND ilters do better? The only filter that worked perfectly for me was the mirror pancros but not the easiest filter to work with.

Reds sensors with their no gain design can be a real pain in full sun where clients want shallow depth. If Dragon is faster again I really hope Red helps out with their clip on internal IRND filter that was mooted last year.
Click to expand...

Try them with a Rossco IR (Rossco behind). Works like a charm. Also make sure they are facing the right way (text to front)

How hot was it outside?
 
We use Formatt which have worked well on Epic, even the separate IR cut but the combined ND IR's are better. I think you can specify the IR cut frequency depending no the camera that are intended for. These things are not cheap where ever you buy them from.

There s a really good thread on this board comparing the IR filters, would be great if some one has time to update this for Epic.

http://www.reduser.net/forum/showth...D-Filter-Matrix-Parts-1-4&highlight=IRND+test
 
Hi Phil Dave and Uli
Just spoken to Format here in the UK and they do a 680 and 720 Hot Mirror/ND combo. Any idea which frequency works best on the Epic? They don/t know themselves which I find a bit odd!
Thanks
Dewald Aukema
 
You must log in or register to reply here.
Back
Top