IR UV Cut filter - Can you spot the difference?

[Daniel Kunz]

There has been alot of concern on the interwebs about IR pollution of the MX Sensor (and any other CMOS sensor in that matter), so i went out an bought a 77mm IR/UV cut filter for my Canon Lenses.
I shot 2 examples with and without the filter, and i have a tough time telling which one is which. (after equalising them in RCX of course)

If you look at the native files they sure do look different, but once you got them equalised there is minimal difference, i suspect you could even get those slightly brownish shadows out with a tweak of the Lift color wheel.
Has anyone else done this experiment?

Here are the comparisons. Try to guess which one is with, and which one is without.

Exibit A1


Exibit A2


Exibit B1


Exibit B2

 

[jochem bakker]

just looking at the thumbnails, i'd say A2 and B1 are slightly reddish/brownish, so those without? Were these shot with ND or no? otherwise its not really an extreme example

 

[ericyoung]

IR pollution generally only becomes an issue with NDs of 0.9 and above. From the similarity of depth of field in your photos you don't seem to be using much, if any, Neutral Density filtering.

It also tends to show up particularly well on certain materials - eg dark cloth like black nylon.

 

[Jack Allred]

agree with eric above,

only need ir cutting at .9 nd and higher. ir tests on the mx are mute without nd filters.

appreciate the info nonetheless.

 

[Aaron Shirley]

just looking at the thumbnails, i'd say A2 and B1 are slightly reddish/brownish, so those without? Were these shot with ND or no? otherwise its not really an extreme example

Same here.

I'd like to see some uncorrected versions just to see how much work you had to do. I plan on shooting to an external recorder a lot so I'm trying to learn how to get everything right in-camera. I have an IR/UV cut and a .9ND, so I'll be testing this out myself as soon as I get adapter parts for my matte box in the mail.

 

[Scott Crawley]

For the purpose of clarity we should stop stating that it is the sensor that requires IR filtering. Any such filtering required by the sensor, is already happening inside the Epic and Scarlet. The need for hot mirrors arose with the R1 because it's in camera filter was found to be generally insufficient, but that has been corrected in the second generation cams.

ND filters above 0.9 require filtering in the IR band because the filter does not block light frequencies evenly across the spectrum and IR frequency light passes through at a higher rate than the other colors. Consequently, we have to add filtration to remove it. If you need to point a finger, aim it at the filter manufacturer. Perhaps they should not be making ND above 0.9 without additional IR blocking built in. I have a Tiffen ND 1.2 that is so red that it looks like a dark tobacco.

I don't know if this is an issue that is specific to digital imaging. If so, then I see a need for a set of "Digital ND" filters to reclaim a filtering stage in the ol' matte box and elimate filter placement errors related to stacking order.

 

[Satsuki Murashige]

If you need to point a finger, aim it at the filter manufacturer. Perhaps they should not be making ND above 0.9 without additional IR blocking built in. I have a Tiffen ND 1.2 that is so red that it looks like a dark tobacco.

I don't know if this is an issue that is specific to digital imaging. If so, then I see a need for a set of "Digital ND" filters to reclaim a filtering stage in the ol' matte box and elimate filter placement errors related to stacking order.

That's why IR ND filters exist... Some work better than others. That said, I can see using non IR NDs creatively for an IR "look" so I'm glad plain vanilla NDs are still available.

 

[Joel Arvidsson]

I have a "Tiffen IR ND1.2" which I guess have the ir built in.

 

[Scott Crawley]

The newer Schnieder Platinum IRNDs are also combination filters, but I have seen no grads available yet.