First I did some "real world" tests but was unable to distinguish between with and without 486 filter (I can post frames if of any interest). So I tested with a chart (daylight illumination) and I believe the 486 improves things, but I'll let you judge for yourself. However I've forgotten how to place images here so I'll email them to you Bob.

edit: test conditions: max F, f/2.4 (max effective aperture at full zoom), 1/500, gain 0dB, gamma STD1, default settings. The chart is 600mm wide and shooting distance about 10 metres. The real world shots were done at effective infinity (3Km, anyway).

Thanks Serena,
the chart tests do show much the same as I saw, an improvement that's not overly dramatic. However in my not so precise tests I did notice objects further away than the test subject (which was 3M away) and hence out of focus did show a more significant reduction in red fringing.
Both our tests were done under daylight. The images I have from the stage show are under tungsten illumination where the lights on the subject are gelled red and the background is illuminated blue. The red / blue transition doesn't fare well in video which further compounds the problem. My instinct is that yes, the filter would reduce the initial problem and help prevent it getting worse as it's converted to SD and displayed on a CRT TV.
I'll try to setup a test closer to the original problem, I don't have much in the way of tungsten light sources at hand so I'll have to dig around a bit.

Bob.

I'll inspect the scenic shots I did first. They had a range of things at various distances, the closer stuff being way out of focus and I did note fringing on one of those. Don't know whether improved the 486 reduced that.

I also wonder about getting one of these filters. I hear that the IR problem is most obvious with black synthetic material especially under tungsten. Some bags or uniforms are like that. So dark things indoors will sometimes look brown. Being indistinguishable from actual brown, secondary color correction is risky. I haven't tested it but think I have observed something like that in practice.

So maybe more useful to test on a benchmark like that.

Bob, I have an EX3, but I do not own the 486. From everything I've read, and more importantly from what I've seen online, I cannot justify the expense, especially exchanging one problem for another (a reddish tint on certain blacks for a green vignette over everything) simply isn't acceptable.

I am sharing this with you for no other reason than to explain why I cannot participate in your experiments (don't own the 486). I sincerely hope you find your answers and I wish you the best of luck in your efforts!

Apit... what is " 'UM' OPTICS LAB"?

Bob,

I did use the 486 as well, but - since most of my serious projects involve multi-EX , and most EX's owners here don't care for, or cannot afford this filter - I sold it away. With multi camera projects, you either have the filter on all cameras, or on none of them!

Since I'm using a mattebox, anyway, I'm waiting for the promised Tiffen solution in the 4x5.65" form.

Back to your experiment: frankly, I never really complained about too much CA in my EX1 unit (and since I've a reputation of being a measurebator, I probably was lucky to get a really good lens). BUT, I have just compared 2 tele shots, with and without the 486 (taken when I was testing it) - and YES, it seems like the fringing is less pronounced with this filter on...

Piotr

Some people did get a poorly aligned lens but I believe Sony readily replaced those. In my tests the lens does show a small amount of CA at longer focal lengths, as the BBC assessment found. Their comment was that it was a good lens while not matching the performance of a HD zoom designed for a 2/3 inch sensor. Of course such lenses costs several times the price of an EX1, a point that seems to be overlooked by many who think Sony should have sold the camera with a better lens.

Have you seen this article?

It discusses some of the shortcomings of the EX cameras in this regard.

John

Nice find Craftech! Saves me some time, its a great "starting" page. I'll will toy around the lower band if I can find a camera we can dis-assemble and inspect the prism and optics. ;-)

Yeah, that's great news, but the loss of four stops really hurts!

Maybe this will be a starting point for someone--some filter maker--to find a better solution that won't cost four stops.

P.S. Apit, you didn't answered my question above. :o(

Read it long ago.
This is not really a shortcoming of the EX and other cameras, it's a deliberate design decision as far as I can see. To further cloud the issue a number of different although related issues get muddled up in the discussion.
The EX cameras are not unduly sensitive to IR, they have an extended senstivity to the far red part of the spectrum. At true IR wavelengths (1000nm) they seem no more sensitive than any other camera. Attempts to correct the problem using dye based filters do work except they negate the advantages of the EX cameras, you loose sensitivity at the visible red part of the spectrum. Dye based filters have a very broad cut so they're not just filtering IR, they're filtering far red and visible red.
The only type of filter that can provide the sharp cut at 850nm is a dichroic which is what the 486 filter provides. The problem with a dichroic filter is its response varies with the angle of incidence. So as noted at short focal lengths (wide shots) you do get a green vignette. This is relatively easy to correct in post if it's noticeable, something that I've never found necessary. We have 5 EX cameras all fitted with 486s and these have shot 1,000s of hours of footage in every scenario and so far no one has noticed the green vignette issue.

By chance I did find a local company (Francis Lord Optics) that know how to build a dichroic filter that would not have this problem, they did make a couple however the prototype that they had on display is around 12" in diameter! The process involves using something akin to an inkjet printer to deposit varying thickness layers of the metals used to build the dichroic filter. My offer to place an immediate order for 1,000 such filters did not spark any interest at all. I suspect we'd be looking at a per unit cost for such a filter of at least 10 times what the 486 costs i.e. around the cost of the camera!

In any case this has little to do directly with the issue at hand. The issue I'm trying to look at is the impact of the enhanced red response of the EX cameras on what looks like CA at the long end of the lens and in particular under tungsten lighting. A couple of days ago I bought a couple of PAR38 lamps, one with a built in red filter. I'll try to setup a test case as soon as I can.


One thing I did note yesterday while testing a couple of Z1 cameras and an EX1 is how much better the EX1 performs. With the Z1 I found it impossible to get rid of the edge enhancement artifacts without dialing Detail down to the point where the image was noticeably soft. By comparison the default Detail in the EX1 was pretty hard to notice (no ringing on hard edges), dialing it down to -30 produces an image that looks more 'filmic' rather than soft. All of which is in line with the BBC's analysis of the camera. All of those tests I did with an old Bravia HDTV. It's a horrid piece of technology but the way it exaggerates image defects does have it uses.

Bob.

Jay, I was going to "borrow" or, "talk into doing some extra work",some graduates from the University of MIchigan High Enegry laser optics dept and a couple from the aero space program testing for the scientific workup. UofM has a great video "inhouse" department because of Sports and theater, but was I leaning more engineering approach to the camera. Normally, this be would like "tomorrow soon enough" by the staff, but the alternate energy programs, climate studies and airborne drone research are hot issues, no money yet, but everyones running towards the $$$$, plus Dow Chemical has been talking up "ideals" for department Chairs to investigate.

"they did make a couple however the prototype that they had on display is around 12" in diameter! The process involves using something akin to an inkjet printer to deposit varying thickness layers of the metals used to build the dichroic filter." Farss, a lot of this high-end is done by high spinning and shaving, very similar to IC manufacturing. The "ink/dye" deposit technology is problematic because of the requirement of continuous thickness for "high-end" optics, but ion deposit is a hot technology. But even if the coating is perfect. the glass or subsurface still must be superior and not "scatter' any light.

The later article on the far red question needs to be read. http://provideocoalition.com/index.php/aadams/story/far_red_on_the_ex1_ex3_f35_its_a_feature_not_a_bug/later article[/link]

Serena, the article is fault because the F35 is a single CCD and does not use the prism-type splitter of the 3 ccd EX series. The filter info is correct, I think, but the F35 can not use the same optical sub assembly. ;-) I'm little surprise that an Sony engineer would suggest their similar phyiscally.

Thanks, Apit. It's really cool that you have access to such resources!

As I read the article it says only that the F35 and the EX1/3 all have extended red sensitivity, not that they have the same optical design. Both cut wavelengths longer than 700 nM. Sony are saying that the extended red sensitivity gives richer reds (it's a feature) but synthetic black fabrics have strong reflectivity in the far red and the cameras are seeing this (not so good).

"hot mirror" and "prism" comments had me leaning more towards assembly design but now, I think the article comment was more of a general reference, And your right, i'ts a good article explaining why the Exs handle the low end like they do. I had forgotten the large number of articles published on the EXs.

Just for general knowledge, Pure Silicon is less sensitive to blue vs red feq. but the Silicon goes thru a enhancement process called "doping" plus a few IP processes. ;-) Doping increases the speed of "freeing" the electrons, increasing sensitive. Companie like Canon, Nikon, Sony, IBM, TI and Toshiba, forgot Intel ;-), all have their "special" doping process, Some are patented, others guard theirs like the CoCoa formula. Usually this doping process is under constant change, especially in refinement and because of ic die changes, ie, 90nm -> 60nm-> 45 nm. Rarely will this process survive more than one size reduction.