As you may have noticed from previous articles, I’ve done a lot of research on which cameras allow infrared or far red to contaminate dark fabrics and change their color. There hasn’t been a satisfactory solution for the Sony EX1, EX3 and F35 cameras–until now.
Silicon is sensitive to infrared energy above all else, so camera manufacturers work hard to prevent their sensors from seeing anything but the visible spectrum. These three cameras don’t have a classic infrared contamination problem where they mistake heat energy, beyond the visible wavelengths of light, for actual visible light.
Sony installs very effective hot mirrors in their cameras to prevent any IR from reaching the sensor(s), but as these cameras see what Sony calls “broad spectrum color” they tend to be very sensitive to red. This isn’t a bug, it’s a feature!
This sensitivity to red can cause problems with dark fabrics that reflect not only infrared (which is cut very effectively in these cameras) but far red, which is red on the edge of the visible spectrum. Humans may have a little trouble seeing this hue of red but these cameras don’t, and sometimes it can be a bit much. The color red has, until recently, been the bastard stepchild of colorimetry because it’s very hard to reproduce properly–and thanks to Sony’s new broad spectrum color you’ll see beautiful hues of red you’ve never seen before. But there’s always a price to pay.
A while back I tested a prototype filter for Tiffen that worked brilliantly. It cut through far red like a knife. Previously the only filter that worked on any of these cameras was the Schneider Tru-Cut 680, which worked exceptionally well except for vignetting on wide lenses: the dichroic hot mirror was so thick that when viewed at an angle the filter turned cyan, so wide lenses yielded an image that was cyan around the edges. Tiffen’s filter, originally known as T1 for “Test 1,” used dyes alone to absorb far red, completely avoiding the risk of vignetting. Their reasoning was this: if the camera’s hot mirror works fine, and since we’re cutting visible light instead of heat energy, it’s clear that we can use a dye, which doesn’t vignette, over a hot mirror, which will. (The more a hot mirror cuts, the heavier the dichroic coating has to be. When cutting non-visible infrared the dichroic layer can be fairly light, but cutting visible far red requires a very heavy dichroic coating, which causes off-axis vignetting on wide lenses.)
While the T1 prototype worked marvelously, Tiffen wasn’t satisfied. The dyes used in the prototype weren’t stable and would fade over time. After trying a number of different formulas, all of which I’ve tested at one point or another, they settled on the current version which works exactly the way the original T1 did but without the original’s color instability. Having perfected it, Tiffen is now ready to release this filter into the wild.
If you want to see how it does, travel with me to page two and cast your gaze upon my homemade far red/infrared test chart…
Here’s my homemade far red/infrared fabric chart:
Read this article to learn how I developed it. (This test chart was shot with a Sony EX3 at Videofax.)
As you can see, the left two columns of fabrics reflect a lot of far red, while the third column doesn’t. The Macbeth chart is present to show whether there are any obvious color shifts when using the T1 filter.
Here’s the same chart with the T1 filter added:
No more far red.
How big a deal is far red on the EX1 and EX3 cameras? The answer is “really big”: I shot this test chart under tungsten light, period. No special camera settings, no neutral density filters. This is just the way it is. The F35 shows a lot less far red contamination but it’s still there–and the T1 filter cleans it up nicely.
My market has no F35’s currently available for testing, but as the original T1 worked fine on the F35 and this filter appears to be a perfect color match (I compared it to the original T1 that I still have in my possession) I have no qualms recommending it for the Sony F35 and the Panavision Genesis.
The black patch at the top of column two looks like it might still have some red in it, so I zoomed into the image in Final Cut Pro 7 and took a look. The top image is without the T1 filter, and the bottom image is with the T1 filter:
It looks like quite a difference, but eyes can be fooled. Let’s look at the corresponding vectorscope images, with some magnification applied:
The top vectorscope shows a definite skew towards red without the T1 filter. The bottom vectorscope shows that the T1 filter has restored black to perfect black.
The T1 filter is green, which requires white balancing when the filter is first attached, and there appears to be no issue with it skewing colors unfavorably. The left column shows a DSC chart shot with no filtration, and the right column shows a DSC chart shot through the T1 filter:
If you can see a difference then I’ll be forced to ask how strong a microscope you’re using. It’s a near perfect 99.99% match.
When I first wrote about the filter I referred to it as “T1” as that’s what it was labeled. I found out later that stood for “Test 1,” but the name stuck. Pre-order it from your local Tiffen dealer.
Production will begin October 1st, and it may take four weeks before all sizes are available. Standard sizes will include 72mm, 77mm, 82mm, 138mm, 4×4, 4×5.6 and 6.6x6x6. Other sizes will be available as a special order.
Tiffen let me keep the prototypes I’ve tested, and I’ve lent them out to friends to use on their EX1 and EX3 shoots. They’ve become quite desperate on the few occasions when I can’t get them a filter in time. I suspect Tiffen will be selling quite a lot of these. If you own or use Sony F35, EX1 or EX3 cameras, this filter is not an option: it’s a requirement.
Art Adams is a DP who is well red. His web site is at www.artadams.net.