Red Filter Study in B&W - Comments Please

01/09/2004 11:11:43 AM · #26

I think that the real problem duplicating the red filter effect on B+W film is that the film has significant sensitivity in infrared, the red filter cuts off most of the shorter wavelengths.
You can still cut off the shorter wavelengths on digital (with the red filter or in software), but the IR filter in most cams would preclude getting the same result as the B+W film.
It would be interesting to try this with an old Sony or other cam with no IR filter. I would bet the effect could then be replicated.

Sony 16-50mm f/3.5-5.6 E-mount

01/09/2004 11:12:48 AM · #27

Originally posted by jmsetzler:

Gordon, I believe that if you get a roll of b/w film and shoot a photo with a red filter installed, you won't be able to duplicate it with a digital camera with or without a red filter.

To my simple engineering brain, this doesn't make sense. The filter is just that, it blocks certain colours and allows others to pass. This may be a non-linear relationship, and vary for certain wavelengths, but it is just a transform, that can be equally applied digitally or with the physical filter.

It probably is not one number applied across the entire image, equally at all tonal values, but would require specific curve. That said, it isn't magic... just maths.

jmsetzler

Sony Alpha a6000

01/09/2004 11:14:53 AM · #28

I don't know the math. All I know is that I have been unable to duplicate, or even come close.

waterlilies

Nikon D70

Nikon AF-S DX Zoom-Nikkor 18-70mm f/3.5-4.5G IF-ED

01/09/2004 11:19:33 AM · #29

Does the red filter effect on black and white film work on C-41 film (developed for color processing), or must you use traditional black and white film?

jmsetzler

Sony Alpha a6000

Sony 16-50mm f/3.5-5.6 E-mount

01/09/2004 11:24:38 AM · #30

Originally posted by waterlilies:

Does the red filter effect on black and white film work on C-41 film (developed for color processing), or must you use traditional black and white film?

The c41 film will work fine.

soup

Canon EOS-30D

Canon EF 17-40mm f/4.0L USM

01/09/2004 11:26:50 AM · #31

couldnt you use channels, and maybe masking to alter the red channel
to simulate the effect of a red filter on a digital image changed to BW

Originally posted by awebsite:

Digitally recreating filter effects
Once again, the effect of any filter can be recreated electronically in a colour image by using the Channel mixer in Photoshop. It's one of the more advanced Photoshop options and with this you edit each of the red, green and blue channels that make up a colour image to lighten or darken each part of the spectrum and affect the brightness of each tone. Become familiar with this and you have the most versatile method of adjusting the tonal range of a black & white image.

01/09/2004 11:36:46 AM · #32

The main challenge/ problem is the transmissiveness of the bayer filters on a typical digital camera. Looking at some filter curves for these, it would seem that they aren't very ideal. So the 'Red' channel on an RGB image actually still has quite a lot of green and blue in it - this could be part of the issue. So to get a good red effect filter, you need to work the red channel a bit more heavily than you might at first think.

Looking at it, you can probably do a real good orange filter... but never quite get the red filter curves due to the cross contamination into the R, G & B channels. The blue channel does a great job of just being blue though.. (but noisy for other reasons)

However, in the end, its digital - it isn't film. You can do different things, in different ways - I'm not certain of the value of trying to 'copy' film exactly. Though, the upshot is, that anything you can do with a glass filter infront of a digital camera, you can do in photoshop (it might not look the same as the film response, as the sensors are different though)

What always bothers me more, and why B&W digital is currently rubbish is the limited tonal palette you end up with from an 8-bit greyscale image. I haven't found a good way to get to duotones or quadtones without going through a 256 level greyscale stage, which means my B&W digital results are dissappointingly limited. There doesn't seem to be a real reason for this 8-bit restriction, other than current software implementations, so hopefully it will improve in the near term. (and if anyone can point out a current solution, I'd love to hear it too)

Message edited by author 2004-01-09 11:51:40.

01/09/2004 12:18:34 PM · #33

Originally posted by Gordon:

...What always bothers me more, and why B&W digital is currently rubbish is the limited tonal palette you end up with from an 8-bit greyscale image. I haven't found a good way to get to duotones or quadtones without going through a 256 level greyscale stage, which means my B&W digital results are dissappointingly limited. There doesn't seem to be a real reason for this 8-bit restriction, other than current software implementations, so hopefully it will improve in the near term. (and if anyone can point out a current solution, I'd love to hear it too)

This is definitely a major issue. PS CS does not solve it, you can use 16-bit gray but when you try to convert to duotone, you can't.
One possible solution that I have not tried as yet would be as follows:

1.) copy each color channel to a separate layer
2.) delete the original RGB layer
3.) Change mode to gray (now 3 separate 8-bit gray layers)
4.) apply duotone (or tritone or quadtone) of choice to image
5.) use opacity to blend layers as desired
6.) flatten when satisfied

Thoughts?

Message edited by author 2004-01-09 12:30:23.

01/09/2004 12:28:59 PM · #34

Originally posted by kirbic:

This is definitely a major issue. PS CS does not solve it, you can use 16-bit gray but when you try to convert to duotone, you can't.

Exactly - now I know I can't print more than 8bit greyscale, but if I can get to a duotone with more than 256 levels, you can do more interesting things - I suppose potentially you could work on 3 channels independantly to manufacture a duotone without the greyscale step, potentially. Would be painful though...

01/09/2004 12:32:34 PM · #35

Originally posted by Gordon:

Originally posted by kirbic:

This is definitely a major issue. PS CS does not solve it, you can use 16-bit gray but when you try to convert to duotone, you can't.

Exactly - now I know I can't print more than 8bit greyscale, but if I can get to a duotone with more than 256 levels, you can do more interesting things - I suppose potentially you could work on 3 channels independantly to manufacture a duotone without the greyscale step, potentially. Would be painful though...

We're thinking in the same direction; see my final edit above! It's really not that painful, I just ran one as a test. The trick is in getting the channels separated to layers; there may be a faster way than I am currently using, and in any case it should be "actionable".

01/09/2004 01:57:35 PM · #36

Originally posted by kirbic:

There is always the 'separate channels' command, though that puts them in unique documents, but it sounds do-able and you are right, an action would be easy enough to set up. With CS you can keep it 16 bit right up to the duotoning step I guess. Might be possible to do it all in 16 bit, by doing the duotones more directly, rather than through the automated approach too

//www.outbackphoto.com/workflow/wf_29/essay.html

01/09/2004 05:18:10 PM · #37

cbeller

Nikon D300

Nikon AF-S DX Zoom-Nikkor 16-85 f/3.5-5.6G ED VR

01/09/2004 06:24:14 PM · #38

Thanks for all the helpful info, folks. I was shocked when I logged in here and saw how busy this thread had been today.

ronners

Nikon D100

01/09/2004 08:58:57 PM · #39

As far as I can tell, B&W filters won't work with a digital camera under any cirumstances to produce the same results as with B&W film. This is because B&W mode on your camera takes the shot 'in color' then does a simple desturation on it. The sensors aren't converted into B&W sensors - the conversion happens in the camera.

The only way to get a B&W effect is in PS (or whatever you use) by adjusting levels. It's not exactly the same as using a red filter, or green filter, or yellow filter, but will allow you to replicate some of the key effects.

This one is a good example (I didn't even use a polariser here - all enhancements to the sky and clouds were done later).

ursula

Fujifilm X-T2

Fujifilm XF 55-200mm f/3.5-4.8 R LM OIS

01/09/2004 09:30:22 PM · #40

Just levels? After converting to b/w? (assuming the picture was in colour originally).

faidoi

Pentax K10D

Pentax DA 18-55mm f/3.5-5.6 AF

01/09/2004 09:37:48 PM · #41

.

Message edited by author 2004-01-09 21:41:17.

ronners

Nikon D100

01/10/2004 05:38:58 PM · #42

Originally posted by ursula:

Just levels? After converting to b/w? (assuming the picture was in colour originally).

Sorry, I did indeed do a B&W conversion first of all using a channel mixer layer.

soup

Canon EOS-30D

Canon EF 17-40mm f/4.0L USM

01/19/2004 03:04:15 PM · #43

when coverting to duotone, how exactly does the mode work. for instance when adding the colors to the duotone pallette how is PS applying the colors. i was thinking about the 16bit -vs the need to convert to gray scale to then convert to duotone steps, and in so doing losing the 16bit pallette.

if you were to use the channel mixer to convert to grayscale ( so to speak ) would there be a way to use overlayed fill layers at less than 100% opaque, and then using curves on that layer to simulate a duotone effect. i have done this in the past with some decent results, but not sure overlay, applies the colors the same as the duotone mode... this would however keep the image in 16bit mode for the entire process and is pretty simple to do,, ( not too many steps )...

soup