Jump to content
UltravioletPhotography

One-Image Aerochrome with Green and Orange filter


Christoph

Recommended Posts

I am aware of the channel swapping, just pointing it out for the gallery. :smile:

Yours looks good.

For example, one of the links Ulf posted above shows some examples I did using the typical yellow filter (Tiffen #12) with channel swapping plus some more processing to get the close simulation to Aerochrome,

thus the yellow tail lights.

https://www.ultravio...6088#entry26088

 

My favorite straight out of camera stack for red IR foliage with in camera white balance and 'see it in live view' is the Lee/Rosco #729 (polyester) + KG3 2mm stack,

red in live view, no processing needed. However, although that stack does very nice red foliage with no processing, it does not transpose red to yellow like swapping can.

Looks good though, with no swapping, and you can see the actual result in live view. In camera white balance on gray card (such a WhiBal card). Yes, even with a Nikon.

post-87-0-88667300-1610243982.jpg

Link to comment

 

 

Yes, because the blue serves as the IR channel. With Hoya and Orange you get blue/purple trees and green sky sooc

 

I get blue/purple trees and nuclear green yellow skies with an orange 530nm filter. What is the green one suposed to do?

Link to comment

I get blue/purple trees and nuclear green yellow skies with an orange 530nm filter. What is the green one suposed to do?

 

I dont know what the X1 does exactly, but from my understanding it cuts off and reduces parts of NIR ...and especially the 77mm seems to do that quite well. I'm a trial and error kind of guy, not a technics/optics guy at all... :grin:

 

What's the resulting image from the channel process? (Duplicate red and green channels, copy blue channel, paste into red channel, then copy the duplicated red channel, paste into green channel, then copy the duplicated green channel and paste it into the blue channel...)

Link to comment

I am aware of the channel swapping, just pointing it out for the gallery. :smile:

Yours looks good.

For example, one of the links Ulf posted above shows some examples I did using the typical yellow filter (Tiffen #12) with channel swapping plus some more processing to get the close simulation to Aerochrome,

thus the yellow tail lights.

https://www.ultravio...6088#entry26088

 

My favorite straight out of camera stack for red IR foliage with in camera white balance and 'see it in live view' is the Lee/Rosco #729 (polyester) + KG3 2mm stack,

red in live view, no processing needed. However, although that stack does very nice red foliage with no processing, it does not transpose red to yellow like swapping can.

Looks good though, with no swapping, and you can see the actual result in live view. In camera white balance on gray card (such a WhiBal card). Yes, even with a Nikon.

post-87-0-88667300-1610243982.jpg

 

That's a nice method too. I think it's simply a matter of taste. If I were only concerned with red/magenta vegetation, then I would clearly prefer the variants with Foveon sensor, especially with red filter or without filter. But it seems almost impossible to get the color changes (red->yellow, car taillights) with this sensor. So I tried to use my experience with SD1 and Foveon methods with already known methods for Bayer cameras. So I came to this result.

As I said, it's a matter of taste, but the common methods don't produce the images I want or require a lot of work.

 

IR Chrome: No color changes, no dark blue skies, rather orange and rather uniform vegetation.

 

550nm filter: Unnatural, quite uniform colors and unnatural general look, too much postprocessing needed.

Link to comment

 

 

I dont know what the X1 does exactly, but from my understanding it cuts off and reduces parts of NIR ...and especially the 77mm seems to do that quite well. I'm a trial and error kind of guy, not a technics/optics guy at all... :grin:

 

What's the resulting image from the channel process? (Duplicate red and green channels, copy blue channel, paste into red channel, then copy the duplicated red channel, paste into green channel, then copy the duplicated green channel and paste it into the blue channel...)

 

I just tried, problem is with this winter sun im not getting a lot of saturation.

So sooc i get blue/purple trees, green/yellow sky.

If i do a R-B swap i get pink trees and cyan sky.

If i do RGB -> BRG i get orange trees and blue skies.

Link to comment
Could you post the white-balanced sooc image? It seems as if nuances really make the difference... And I tend to believe that a little more nm than my tiffen 16 orange would even be better... I don‘t even have a transmission curve of this filter, so I don‘t know how much green gets transmitted. The solution seems to lie in a combination of x1 and a specific orange filter. Because if the filter is too yellow, the vegetation is too magenta (hence orange) and the sky too aqua green (hence magenta blue)...
Link to comment

Sure, so this is supposed to be a 530nm pass filter but i dont have the curve, got it from Ali.

 

oh Ive run out of upload space, see if this works:

Sooc white balanced:

 

wUUqKC8.jpg

 

RB swap

docqsfl.jpg

 

RBG swap

tUjDrah.jpg

Link to comment
Andrea can give you more space, but I looked at a few of the images you uploaded above and they are like 800K and more! You need to be saving them with a JPG compression of around 8 or so. Typical images on this forum should be around 1100px wide and less than ~300K if you can manage it, to keep bandwidth costs reasonable.
Link to comment

Andrea can give you more space, but I looked at a few of the images you uploaded above and they are like 800K and more! You need to be saving them with a JPG compression of around 8 or so. Typical images on this forum should be around 1100px wide and less than ~300K if you can manage it, to keep bandwidth costs reasonable.

 

Cheers yes ive been compressing them to half or less but obviously not enough. I dont know if i can edit the old ones though?

Link to comment

@microbat52: Thanks for posting them! That's what I expected. The blues aren't "blue enough" and too pale - I guess, the x1 somehow darkens and accentuates them... It's nuances, really.

So here's a result from a quick test I did some hours ago in the evening light. I tested yellow K2 (with Hoya X1), added cyan filter et cetera. No improvement. Then I tried a Cokin sepia filter because I had bought it in the past for trying out experiments with blue vegetation.

So here's the comparison. They are sooc (white-balanced) with one exception: After opening in PS raw I shifted the blue hues more towards blue: excatly -30, color preset in camera was standard. Then I did the channel stuff, auto contrast, and nothing else... First just orange and green:

 

post-309-0-48345000-1610303564.jpg

 

Then with the additional sepia. So what it did: It gave more magenta to the plants and more blue to the sky.

 

post-309-0-94288900-1610303324.jpg

Link to comment

The just orange and green filter combo may be superior to the method with additional sepia, because it seems to provide more variation in the plants...

Here's just orange and green, sooc (except -30 in the blue hues) and then channel shifting...

 

post-309-0-56862000-1610370717.jpg

Link to comment

The "original" colour dye scheme of Infrared Ektachrome was,

 

IR -> R

R -> G

G -> B

 

when the addition you needed to keep blue (and UV) at bay by suitable filters.

 

Based on Birna's post earlier in this thread, would the Midopt filter which captures red, green and IR be of use here? Link to the filter - https://midopt.com/filters/tb550660850/

Link to comment
Maybe. I'm unsure about how much red is transmitted by the X1. How would (more) red impact the white-balanced image sooc?
Link to comment

Whoops, you are right: I don’t think the other dual band existed when I tried that so I didn’t read your post carefully enough. I just glanced at the spectrum and figured it was the same one.

 

The one you linked cuts all the red light though so I’m not sure how that would work here.

Link to comment

 

The one you linked cuts all the red light though so I’m not sure how that would work here.

 

That's the big question... is (visible) red needed? And why? :cool:

Link to comment

Christoph, the red light goes into the green channel. If you cut most of it, you will not get yellows from objects that reflect infrared+red light (e.g. stop signs).

 

Jonathan's suggestion of this one is much better:

https://midopt.com/f...rs/tb550660850/

 

Don't forget you would still need to subtract the blue channel from the other two because the IR gets distributed equally among the three channels.

Link to comment

Tell me, Andy, since you know a lot about this:

1. Is there a way to detect the IR in the separate color channels? What would be the difference between a, let's say, red channel image with IR contamination and one without? Would leaves be brighter with IR contamination?

2. Another thing that I've been wondering about: Why is the transmission curve of the X1 at max only by around 60 percents? And the Midopt at max around 90% (green). Is this on purpose? And what's that purpose?

Link to comment

I have no idea about point number 2 -- I don't know what that X1 was originally intended for even. In the MidOpt case they obviously want to transmit as much as possible since the sensor gain starts to get lower out in the 800nm's.

 

With regard to number 1: You know that at 850nm+, the IR goes into all three channels equally (this has been explicitly measured by Jonathan on this board even) pre-white balance. (You need access to the RAW to see this -- pre-WB components are not the same as white balance on PTFE or something white.) Since none of the R or G ends up in the blue channel, that allows you to simply subtract off the blue from the other two as I described in the link I sent before to the other thread. Since the IR goes into all three channels equally if you don't do the subtraction, the effect would be to lower the saturation in regions where infrared is high.

Link to comment

I have no idea about point number 2 -- I don't know what that X1 was originally intended for even. In the MidOpt case they obviously want to transmit as much as possible since the sensor gain starts to get lower out in the 800nm's.

 

With regard to number 1: You know that at 850nm+, the IR goes into all three channels equally (this has been explicitly measured by Jonathan on this board even) pre-white balance. (You need access to the RAW to see this -- pre-WB components are not the same as white balance on PTFE or something white.) Since none of the R or G ends up in the blue channel, that allows you to simply subtract off the blue from the other two as I described in the link I sent before to the other thread. Since the IR goes into all three channels equally if you don't do the subtraction, the effect would be to lower the saturation in regions where infrared is high.

 

Thanks for your answers! :-)

Link to comment

Normally a proper Aerochrome image renders red like car tail lights as green.

.

 

Green only if LED (no IR.) Yellow if incandescent, and red plastic lenses also go yellow.

 

Pulling this off without subtraction looks difficult to me.

Link to comment
I don't know why people are so squeamish about subtracting channels or so set on getting things straight-out-of-camera. I always edit all my images anyhow, what's one more step? And you can always make a photoshop action or whatever.
Link to comment

What would be the difference between a, let's say, red channel image with IR contamination and one without? .

 

I'll take a stab at that. Foliage in red is dark due to chlorophyll absorbing out to about 720 nm. Throw in an IR signal and the same leaves go quite pale. Some dark blue surfaces will behave similarly.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...