Hot Mirror/IRND/Mirror ND Filter Matrix [Parts 1-4]

[Evin Grant]

[7/23/09]

All four parts have been merged into one thread. You can find Evin's original posts within this thread here:

Part 2
Part 3
Part 4

~Brook

----------------------------


I'm posting part one now because I think it's info everyone in the market for theses filters needs to now NOW! Unfortunately it's bad news for some.

Part 1: Hot Mirror off axis color shifting.
One of the first things you notice when picking up any of the Dichroic coated Hot Mirror filters is the strong magenta/cyan cast the filter makes when you turn it around a bit. This is how the filter cuts IR, by reflecting, but it also has some effect on the visible spectrum most notably a cyan color shift that gets stronger the further off axis you look through the filter.


This is visible on all the Hot Mirrors to some degree or another but the Schneider True Cut IR takes it too far and unfortunately creates a strong cyan cast that is clearly visible on wider lenses like 16mm, even in normal daylight environments.

(These frames are unprocessed other than a color boost in Red Alert! The port-holing you see is endemic of the Zeiss 16mm T2.1 lens)

Here you can see the subtle effects of the Tiffen and Rosco Hot Mirrors compared to the garish Schneider.

It's relatively easy to correct a vignette in post but a color vignette is nye impossible.

Here's how the pack lines up...
1st place: Tiffen, slight cast off axis, probably negligible to 12mm
2nd place: Rosco, moderate off axis cast, probably negligible to 14mm
3rd place: Schneider, strong off axis cast, avoid all wide angle lenses!

Note: Although the Pancro Mirrored ND filters have no off axis color cast there is a slight darkening of the frame at the extreme edges that might accentuate vignetting slightly on super wide lenses of 10mm or shorter.

Here is a real world test done with the Schneider Ture-Cut IR and a Nikon D3 and 17-35mm lens simulating a 16mm and 10mm lens on the Red One in 16x9 mode.
(I don't have anything wider than the 16mm available for the Red right now but even that shows the dramatic color effect.)


Our flat light test in action...

[Bruce Allen]

Good to know! Thanks for the 3AM post...

Bruce Allen
www.boacinema.com

[RivaiC]

I like Rosco and Pancro. Seems the right one to go. Thanks so much for the post Evin. Really need to know

[James T Mather]

Does anyone know where to get a tiffen or Rosco filter in 4x4 or 4x5.6 - or nay size that fits in a matte box? Thanks.

[Curran Giddens]

I'm awaiting shipment from Michael Morlan's Rosco group buy that ends today.

http://www.reduser.net/forum/showthread.php?t=12516

Or try Custom Supply....

....BJ at Custom Supply let me know that he's got the Rosco Hot Mirror, I believe, for $131.
Matthew

[Brigham Edgar]

So does this mean Pancro were not counted in the equation or was this test result just for Mirrors and not combined IR+ND?

[Evin Grant]

Pancros were tested, but since they have no off axis color shift they were'nt included in this section of the test.

[ayarbro]

A couple questions....

1- So these were the IR cut filters? Or were they the ND IR combos?

2- What exactly was the "color boost" in red alert? What was the amplitude of it?

3- Which ND's of the series were tested? Did you also test standard ND's, for a control group? Was there any color shift noticed when the color boost was applied to the regular ND's?

Don't mean to be a pain in the as*, just want the facts.

[Evin Grant]

This is just part one, so be patient. There are still three more parts left to post. as far as what filters were tested here is the break down...

Pancro:
Mirror ND .6 and Mirror IR+ND .9, 1.2 & 1.5

Rosco:
TruColor IR Hot Mirror

Schneider:
True-Cut IR Hot Mirror

Tiffen:
Hot Mirror, Hot Mirror/IRND 1.2 Combo, IRND .6 & 1.2

[Evin Grant]

What exactly was the "color boost" in red alert? What was the amplitude of it?

The saturation setting in Red Alert! has a default of 1 which is not very colorful I tend to set it at 1.3 for normal shooting. To accentuate the color shift happening with the filters I punched that up to 2, it's highest setting. The point being that it would be very hard to discern the differences in the character of these color shifts otherwise. This is not an "artificial" color boost, this is the Red debayer program being told to decode and express the colors strongly. So no chroma is actually being added.