Infrared Challenge: HPX300 vs. EX1

[Barry_Green]

While up at the Panasonic HQ, I took the opportunity to test the two cameras for infrared (IR) performance. It's a well-known issue that the EX1 sees into the infrared portion of the spectrum, which has the effect of turning black objects into brown or purple. Adam Wilt did an excellent infrared comparison (http://provideocoalition.com/index.php/awilt/story/schneider_ir_filter_tests/)among various cameras to try to test the performance of infrared-cut filters like the 486 or "hot mirror" filters.

So, with IR being a "hot topic" (ooh, a minor pun there), I wanted to see how the HPX300 performs as compared to the EX1. Being Panasonic's first CMOS camcorder, I expected it might show some IR contamination.

Accordingly, I had my wife find every stitch of black fabric we could locate. I took up her dresses and my socks and a tie and some of her quilting fabrics and, basically, anything and everything black. I couldn't find anything quite as scintillating as Tim Blackmore's shirt (http://provideocoalition.com/index.php/awilt/story/schneider_ir_filter_tests/P1/), but I certainly found enough to test with.

Both cameras were set to the same exposure, gamma, white balance, and color. All the objects in these photos were black. Both shots were taken simultaneously, under identical lighting (tungsten light from a Lowel DP light).

Here's how the EX1 rendered them:
http://dvxuser.com/barry/EX1-Infrared.jpg

And, in comparison, here's how the HPX300 rendered the exact same scene:
http://dvxuser.com/barry/HPX300-Infrared.jpg

The results show the HPX300 to be substantially superior in IR rejection and shouldn't be needing any Infrared filters to properly render black colors. The HPX300's rendition looks exactly like it did in real life; the EX1's rendition is way off.

These frames are direct from the footage, no post processing involved (other than to shrink them to 50% size). The white balance was manual; the exposure was set by a waveform monitor.

[Cees Mutsaers]

looks mighty impressive :-))))))))))))

[CovenantPictures]

Hmm. Interesting. What situation/circumstances cause this phenominon to occur? Because certainly, I've seen the EX1 render blacks beautifully. What about this test was different?

I've heard that using more than two stops of ND can cause IR polution in the footage. Why is this Barry?

[Sumfun]

Barry,

It looks like the red color cast in the EX1 picture is also on the white Lowell box, which leads me to believe that it's NOT caused by IR contamination. I took the picture into Photoshop and did an auto color to remove the red cast from the white box, and the blacks look fairly normal.

I don't doubt that the EX1 has an IR problem, but maybe the black fabrics that you chose just aren't "fluorescing" in the IR spectrum.

[Barry_Green]

Hmm. Interesting. What situation/circumstances cause this phenominon to occur? Because certainly, I've seen the EX1 render blacks beautifully. What about this test was different?
It's in the amount of infrared present, and in the types of fabrics. If you're shooting under daylight, with no ND, the blacks will usually look really good. If you're shooting synthetic fabrics under tungsten, infrared contamination can be pretty significant.

Tungsten light spews a lot of infrared out. Daylight, not so much. Tungsten lighting is closer to the red end of the spectrum and tungsten sprays a lot of infrared.

The type of fabric is also important; Tim Blackmore's shirt is pretty much the perfect IR reflector. It usually comes down to the dyes that are used in the fabric; some reflect a lot more IR than others.


I've heard that using more than two stops of ND can cause IR polution in the footage. Why is this Barry?
Because an ND filter will cut visible light, but it doesn't cut infrared. So you're raising the relative proportion of infrared that's passing through the filter.

IR is present in daylight, yes, but not in the same proportions that it is in tungsten. IR gets reflected from foliage, for example.

[Barry_Green]

Barry,
It looks like the red color cast in the EX1 picture is also on the white Lowell box, which leads me to believe that it's NOT caused by IR contamination.
Yeah, what's with that? Both cameras were manually white-balanced... hmmm...


I took the picture into Photoshop and did an auto color to remove the red cast from the white box, and the blacks look fairly normal.
I don't know, because there were lots of other fabrics that we tried that showed up black as black can be. Some cotton stuff, for example. I had too much stuff to use it all, so I pulled out anything that wasn't showing different. Maybe I should have left at least one there for comparison...


I don't doubt that the EX1 has an IR problem, but maybe the black fabrics that you chose just aren't "fluorescing" in the IR spectrum.
Definitely not as fluorescent as Tim's shirt... that's like the ultimate IR test fabric!

[Justyn]

Man.. that's pretty telling stuff there and would add quite a bit of work to correct this. This would be a nightmare if its a clothing shoot or something and the client is there on monitor looking at the muddy image..

[alpi69]

IR is not a EX-1 phenomenon. It is a SONY thing. I had huge problems a few years ago when shooting a series of backpacks that had black insides. Shot with a SONY VX1000. Has nothing to do with CMOS. Must be in their prosumer algorithms, because on Betacam SP (which we used eventually to shoot that particular project) and newer broadcast cams it isn´t a factor.
When I got my DVX100 soon after, this was my first test...and it passed easily -using Panasonic ever since :D

[shapna]

Hi Barry,

i have a question about your test, the difference they hace in colors are huge, but for what i know, this "color" that produces the EX1 can be corrected in the 1 on 1 color correction before shooting. Is this ok?

Is this "color" problem easy to solve, with filters o anything you can recomend??

In avdance i appologize for my horrible use of english, and also i ask this because like any other guy im deciding in what i want to buy for my first Cam

Thank you for the time.

Damian

[Buck Forester]

Brown is the new black.

[Barry_Green]

The color results from the camera being sensitive to light in the infrared portion of the spectrum. It's not easily correctable in post. You have to use an Infrared filter to cut out the offending portion of the spectrum. Look at the link (http://provideocoalition.com/index.php/awilt/story/schneider_ir_filter_tests/)I posted to Adam Wilt's infrared filter comparison and you'll see what needs to be done.

[shapna]

Barry,

Terrific link, 1 doubt, thou. This means that if y buy the EX1 and a 486 filter i wont have any problem in getting the same image ( in terms of color) with the hpx 300?

sorry i bother with this questions. Thanks for the time

Damian

[CovenantPictures]

Thank you Barry for these terriffic tests.

[Barry_Green]

Barry,

Terrific link, 1 doubt, thou. This means that if y buy the EX1 and a 486 filter i wont have any problem in getting the same image ( in terms of color) with the hpx 300?

sorry i bother with this questions. Thanks for the time

Damian
The 486 filter will remove the infrared contamination. The guys in the Sony forum, specifically over on DVInfo, have been going through all sorts of discussions of the pros and cons of various types of IR filters. Most of what they tried early on led to side effects such as green vignetting in the corners. You'd probably be best served to check over there to find out the current status of recommendations for IR filters.

[Shipsides]

ProVideo Coalition posted a good article on IR filtration for the EX1/EX3. Check it out here - http://provideocoalition.com/index.php/aadams/story/ex1_ex3_ir_filter_shoot_out/

HPX300 seems to be free of any of these issues. Great job Barry!

[Bassman2003]

Barry, maybe you could re-shoot the test so the EX-1 white balance could be correct?

Just to give it a fair shake?

[Jan_Crittenden]

Barry, maybe you could re-shoot the test so the EX-1 white balance could be correct?

Just to give it a fair shake?

They were white balanced. I think the red splash on the white box is from the shiny black fabric.

There is no one more meticulous than Barry. If you have never watched him work, he is a little boring, sorry Barry, but thorough.

Best,

Jan

[Barry_Green]

I don't have access to the cameras. I did that while I was up at Panasonic headquarters for the Skew-Off. Maybe Adam Wilt could do a comparative test with that super-reflective shirt that his friend Tim has...

[Barry_Green]

There is no one more meticulous than Barry. If you have never watched him work, he is a little boring, sorry Barry, but thorough.
Heh. :)

[asphaltplanet]

Really happy to have discovered this post. I'm prepping a feature that will be shot with an EX1 and an EX3 and this would have been a bad thing to find out about after shooting had commenced. Thanks for the great information.

[David Saraceno]

I'd wager that when the lions, elephants, and rhinos were charging in Africa, he worked real fast.

[Lez]

Here is an example of the EX3 (wide shot) and ... HVX 200 (tight shot)...

Cheers

PS.. The EX3 looked great on set through the viewfinder and flip out screen... totally flat once in the edit suite...

[Jan_Crittenden]

PS.. The EX3 looked great on set through the viewfinder and flip out screen... totally flat once in the edit suite...

That might just be the difference in the fact that you were monitoring over HD/SDI which is a 4:2:2 output and the recording is 4:2:0.

Just a thought,

Jan

[Lez]

No idea why the EX3 was so flat... I was operating my HVX200 and the other operator set up his EX3... on set the HVX flip out screen looked out of focus compared to the EX3 (I kept on using the focus assist just to make sure I was in focus when comparing to the EX3)...

But once back in the edit suite the Panna footage looked crisp and color rendition was true... the EX3 looked dead and flat...

Cheers

[Bassman2003]

Well I guess this proves that EX-3 needs to be properly setup!?