Massively confused with Trichrome on a digital camera.

[estee.el]

Massively confused with Trichrome on a digital camera.

 

I was experimenting with trichrome photography off a 24MP bayer-layer-sensored camera and I have encountered some mind boggling confusions..

 

As instructed by literature for trichrome photography, I have used the 25 Red, 58 Green, and 47B Blue for each RGB channels and according to Tiffen website they have the filter correction for R-(3 stops), G-(2 2/3 stops) and B( 2 1/3 stops) respectively.

 

So here is what I did, I took a master shot( properly exposed with no filter) and then took RGB filter photos with corresponding filter factor compensated in Manual mode, by adjusting ISO values, so I don’t mess with low shutter speed and create shakes… All the photos are take with a 2s countdown timer so, again to avoid shakes.

 

The problem is that I realized immediately when adding the filter the under exposure value does not correspond to their filter factor compensation. For example, with the blue filter, the meter says I’m 4 stops underexposing it.

 

So our of curiosity, I took another RGB photo in Aperture priority mode, to let the camera auto correct to the proper exposures and thinking to compare the results later.

 

Then, when I moved all the files to Affinity photos, and mixed all the channels, I realized that

1) each RGB filter did not block out other colors. E.g for the photo I took with Blue filter, there are very much visible image show in its Red and Green channel. Which means still a lot of Red and Green light was passing through the 47B filter????

 

2) I mixed both the Aperture priority exposed and the manually exposed RGB copies, and the result looked nothing like the Master shot, they both have massive color shifts….

 

What did I do wrong???

 

I recorded the whole process in the YouTube video here: https://youtu.be/5aA4YD1gPjI

I also uploaded all the raw files in DNG as well as RGB mixed to here: https://we.tl/t-F7XlCC6Xmn

Edited November 22, 2022 by estee.el

[estee.el]

Here are the master, RGB mix for manaully adjusted for filter factor and aperture piority mode.

As you can see, they looked nothing like each other.... what on earth...Help

[Jochen_S]

Did you manually set your white balance to light source appropriate or did you let it go wild on auto?

[glen_h]

On 11/22/2022 at 3:44 PM, Jochen1664876637 said:

Did you manually set your white balance to light source appropriate or did you let it go wild on auto?

 

Filter factors were from the days with external (not through the lens) metering. 

It may still be necessary to be careful with metering, as TTL meters might not get it quite right with such filters.

They might not be exactly the same as the Bayer filter array, but should be close.

 

But yes, as above, with automatic light balance, it might try some strange things.

Even so, I suspect you need to do some balance when putting them back together.

[Niels - NHSN]

I will probably stand out as an idiot, as I have no experience doing this and may be misunderstanding.

My first thought: Don't you have to record the RGB images in BW for each filtered channel? and mix them together in their respective color channels? Or is that what you are doing?

An observation: Mike Johnston of "The Online Photographer" is currently experimenting with a digital camera converted to monochrome.
He mentioned in a note, that use of colorfilters in front of the monochrome sensor does not yield the expected results - as he would see with colorfilters and BW film or if playing with virtual filters in the software when converting a digital color image to BW.
I don't know anything about the subject, but it sounds like you shouldn't expect the same results with filters in front of digital sensors (for reasons I know nothing about).

[rodeo_joe1]

Why are you using filters on the lens at all? 

A digital camera already has RGB filters built into (or onto) the sensor. So all you need is an image editor capable of splitting the channels. 

A bit pointless if all you're going to do is recombine those channels with a still-life to get accurate colour again, but surely the Trichrome process comes into its own with moving subjects. 

Multiple images of, say, a crowd in a shopping mall, will show an accurate coloured background with multi-coloured 'ghosts' moving around in it. 

All that's needed for that effect are a series of 3 - unfiltered - shots in succession, then combining the R, G and B channels from separate shots into one image. 

The digital camera does the separation work automatically, and with no filter factors to figure in. 

Edited November 30, 2022 by rodeo_joe1

[rodeo_joe1]

On 11/26/2022 at 10:16 AM, NHSN said:

.. currently experimenting with a digital camera converted to monochrome.
He mentioned in a note, that use of colorfilters in front of the monochrome sensor does not yield the expected results

A bit of a sideline to this thread; but why? Why would you convert a camera to monochrome, only to expect to use filters over the lens to alter its tonal rendition? 

When the easiest way to get a filter effect in B&W is to shoot in colour and alter the tonal balance during a post conversion to monochrome. 

Example - original colour shot. 

And two 'filtered' variations got by altering the desaturate sliders in PS. 

Red filter simulation - 

Green filter simulation - 

Easy! And no expensive and irreversible camera mutilation to pay for. 😁

 

Edited November 30, 2022 by rodeo_joe1

[rodeo_joe1]

Further thought on the original question, and WRT white balance.

Just setting the camera WB to the ambient light is only half the story. 

If the light source is anything but standard daylight, then the given filter-factors won't be correct. 

For example: In artificial incandescent tungsten light, the red and blue filters will need different factors. With blue needing a higher factor (than daylight) and red needing a much lower factor.