IR Contamination..any further forward??

[Leonard Levy]

Bob, I assume that's the kind of work going on at Tiffen.

[Bob Grant]

Originally Posted by Leonard Levy

Bob, I assume that's the kind of work going on at Tiffen.

I'd sure hope that's what they and others are doing too.
If that's what they are doing though surely they should be publishing the spectral transmission curves of their filters. Then you and Art and anyone else testing these filters has baseline data to work from, you know that from the published data it should work and all you're doing is validating the manufacturers claims in real world situations.

If you scroll down the following page
diglloyd: Spectral Transmission Graphs for B+W, Hoya, XNite, Baader, Kenko filters
you can compare the data between the 489 and 486 filters. You can see why the 489 is of no use with our problem. Compare the 486 plot against the Kenko DR-655 and you can see a filter that might solve the far red problem but introduce a problem with color rendition.
This is the sort of hard data we should be working from.

I raised my concerns about having an IR problem with another camera (SI-2K). I'd asked if we should be using ND filters with IR cut and was told no way, that would for certain affect color rendition on a camera designed to provide a wider gamut than rec 709. Hence my concerns here. We need to know not only that the problem is solved but also that we're not giving up anything or if that's inevitable what we are giving up.

[Leonard Levy]

When I spoke to them at first they hadn't really done testing on the EX-1 specifically and were surprised that the filters that were developed for the S35 didn't work the same on the EX. They told me they were asking Sony for an EX to test with.
I can't say whether they'll publish spectral response curves or not though they usually do. You'd probably have to get Sony to publish spectral curves for EX response.

[Dave Morrison]

I posted a very brief clip showing my Schneider/B&W 486 filter in action:

IR contamination of Sony EX1 on Vimeo

All of the black fabric objects in the shot (except for one) showed the IR problem. I also took the filter out in the bright sun today and used the same tripod bag to test for IR. The sun made the bag just as brown/maroon as it appeared here and the filter cleaned it up again. The tiny bit of green tint introduced by the filter can be corrected in post MUCH easier than the all those black objects turning brown, but I still think it sucks that we have to buy $300 filters to fix this problem.

[Max Allen]

I tried our True-cut 750 4x5.65 yesterday with our EX1 in bright sunlight shooting a black Underarmour "loosegear" shirt which is made of 100% polyester. I white balanced without the filter and with the filter just to see if there's any difference. Brown. I saw a reduction with the 750 but to such an infinitesimal amount that most eyes would unquestionably say there was no difference. I'll be contacting Schneider.

The green vignetting of the 486 is unacceptable to me. I often hand off footage to editors and it just won't pass the muster to make excuses about a vignetted color cast with all my footage. Only thought I have now is to find out in advance if there are synthetic fabrics involved in the shoot and try to avoid them. Has anyone tried to mitigate this with the EX1's matrix controls aka "color correction" targeted to the black? I can only imagine the time you might have shooting a fashion job for example and being unaware of this problem until the client sees the footage.

[Steven Thomas]

Thanks for sharing the video Dave.
It clearly shows how well the 486 filter works.

[Dave Morrison]

I agree with the green vignette issue so I'd avoid using my 486 filter with any wideangle shots, but I'm still not happy about it....obviously. Also, you're going to have problems making any blanket statements about which fabrics will give you IR problems. In my test footage (above), there was a mixture of woven fabrics and the t-shirts (and possibly the hat) were 100% cotton. The tripod and light bags appear to have more of a synthetic sheen like a polyesther material. The only thing in that shot that did NOT reflect IR back to the camera was the office chair fabric. There was no indication on the bottom of the chair as to what the fabric covering was made from. A quick check of their web site showed that it might be something called Olefin or it might be a 55/45 cotton/poly blend. It's a 10+ year old chair, so who knows.

[Dave Morrison]

Originally Posted by Steven Thomas

Thanks for sharing the video Dave.
It clearly shows how well the 486 filter works.

Steven, the one thing I neglected to do was do another White Balance adjustment after mounting the filter just to negate the green shift. Maybe I'll repeat the test or somebody else can do it.

[Max Allen]

Good points Dave.

I skimmed over the possible upcoming Tiffen dye-based product. Personally I prefer Schneider filters over others when possible so I hope they come up with a solution, tested with EX. One problem I see with WB to cancel the green cast is the uneven pattern of a vignette. Obviously WB will take out a uniform level of green over the entire image so I'm wondering what results 486 users who WB with the filter on are getting at wide when there was more green around the edges than the center.

[Dave Morrison]

You're right, Max, and I'd probably avoid using it in extremly wide shots that had uniformly colored backgrounds. I know....it's a stupid way to have to work, but there you go. I think if you have a fairly complex or "busy" background, you might be able to hide the green effect. Frankly, I'm not going to use it unless I know that I have a LOT of black fabric subject matter in the shot but it sure seems like a waste of $300 to have it in the bag just waiting for those kinds of shots. I could have used that $300 for a lot of other stuff.

[Leonard Levy]

I hope to test the Tiffen prototype on Thursday.

[Steven Thomas]

Leonard, please let us know how the tiffen works out.

[Ronn Kilby]

Shot some interviews the other day with a group of firefighters. They were all wearing black t-shirts and black caps, with their station logo on them. All the clothing came out brown/purple, while the logos looked true. This was a bummer because these guys know what it should look like. Since there is no definitive cure to date, and I'm not spending $200-$500 for a filter that only kinda fixes it, I went back to informal testing. I grabbed 5 black t-shirts, a black hat, some black jockey shorts and a black bag and shot them using various PP and no PP. All but the jockey shorts showed IR contamination! Lighting was overcast daylight, manual white balance. ND made no difference. What I found was that the BBC Film PP showed the least contamination (IR still there, but in a pinch it's better than nothing). You can find it on the forum.