Hide images and reveal effect with a red filter.

[Welsh Yellow Cheddar]

I found on the net a neat effect that I could use at school with my students. I would like to be able to reproduce it but I don't know the steps and I didn't find a plugin to do this either. As you can see, on the left side, two images have been merged in such a way that when a red filter is placed over them a hidden image will be revealed.

For example, I would like to combine Van Gogh and Sorolla's portraits and make the painter Sorolla appear when a red filter is used.

 

The combined image will be printed and red cellophane will be used as a filter.

 

Thanks for the help.

[AndrewDavid]

Hi @Welsh Yellow Cheddar

I seem to remember @BoltBait writing a plugin to do that.

Have a try and post any questions to the plugin thread.

Hope this helps

 

[Welsh Yellow Cheddar]

2 minutes ago, AndrewDavid said:

Hi @Welsh Yellow Cheddar

I seem to remember @BoltBait writing a plugin to do that.

Have a try and post any questions to the plugin thread.

Hope this helps

 

Hello AndrewDavid,

 

Thanks for the tip but I actually need to print the resulting images so the kids can walk around the classroom with a red filter in their hands and take notes.

[Reptillian]

I guess it would be possible to write a plugin/G'MIC script which import images, and then combine them per channel.

[MJW]

It can be done fairly easily without a special plugin.

• Convert the original image to black and white.
• Run Ed Henry's Extract Channel plugin. Set the channel to Red, and uncheck the Grayscale checkbox. The image will now be red.
• Add a layer, and set its Blend Mode to Additive.
• Fill the layer with some random black and white pattern. For example, run Clouds, perhaps followed by Dents or Crystalize.
• Run Extract Channel, this time setting the Channel to Blue.
• Add another layer, and set its Blend Mode to Additive.
• Fill the layer with some random black and white pattern.
• Run Extract Channel, this time setting the Channel to Green.
• Flatten the image.

The effect is best when the Green and Blue layers obscure the image in the Red layer.

 

EDIT: Though it's probably obvious, I should have mentioned that the Blue and Green images don't need to be random patterns -- they can be any black-and-white images, colored by applying Extract Channel. In the original examples, they're other portraits.

[Reptillian]

35 minutes ago, MJW said:

Run Ed Henry's Extract Channel plugin. Set the channel to Red, and uncheck the Grayscale checkbox. The image will now be red.

 

There's another Extract Channel plugin (which I had made).

 

Your step can be argued to be a bit too long, but works.

Edited September 2, 2022 by Reptillian

[MJW]

1 hour ago, Reptillian said:

Your step can be argued to be a bit too long, but works.

 

I think it's the perfect length.

[Welsh Yellow Cheddar]

On 9/2/2022 at 3:04 AM, Reptillian said:

I guess it would be possible to write a plugin/G'MIC script which import images, and then combine them per channel.

That would be great, unfortunately, scripting is way out of my capacities. ☺️

[MJW]

For those who for some reason don't want to use plugins, the coloring can be done without them. For the red layer, add a layer above the black-and-white image, set its Blend Mode to Multiply, fill it with red, then merge down. Likewise for green and blue.

[Welsh Yellow Cheddar]

20 hours ago, MJW said:

It can be done fairly easily without a special plugin.

• Convert the original image to black and white.
• Run Ed Henry's Extract Channel plugin. Set the channel to Red, and uncheck the Grayscale checkbox. The image will now be red.
• Add a layer, and set its Blend Mode to Additive.
• Fill the layer with some random black and white pattern. For example, run Clouds, perhaps followed by Dents or Crystalize.
• Run Extract Channel, this time setting the Channel to Blue.
• Add another layer, and set its Blend Mode to Additive.
• Fill the layer with some random black and white pattern.
• Run Extract Channel, this time setting the Channel to Green.
• Flatten the image.

The effect is best when the Green and Blue layers obscure the image in the Red layer.

 

EDIT: Though it's probably obvious, I should have mentioned that the Blue and Green images don't need to be random patterns -- they can be any black-and-white images, colored by applying Extract Channel. In the original examples, they're other portraits.

Thanks MJW. I followed your instructions. I don't know what the layers modes do but they do what I needed. 😊

[MJW]

4 hours ago, Welsh Yellow Cheddar said:

I don't know what the layers modes do but they do what I needed.

 

It's actually pretty simple. The Blend Mode of a layer determines how that layer is combined with the layers beneath it. Additive mode adds whatever's in the layer to the layers below. Because the Green and Blue layers use Additive mode, the combined image consists of the Red, Green, and Blue layers added together. Looking through a red filter removes all but the image in the Red layer.

 

Good job with the image you made, the portrait is nicely disguised by the top layers.

[Reptillian]

To do exactly what the author of the blog did, you'd need to invert the channels. I actually checked, here's the check:

 

 

You can tell that from the top left picture of each group are inverted. There may also be some control involved into each channels.

It's hard to see from the blue side, but you can tell that the pictures for other channels other than red don't have to be the same.

 

My example with the observation made:

 

Edited September 4, 2022 by Reptillian

[Welsh Yellow Cheddar]

9 minutes ago, Reptillian said:

To do exactly what the author of the blog did, you'd need to invert the channels. I actually checked, here's the check:

 

 

You can tell that from the top left picture of each group are inverted. There may also be some control involved into each channels.

It's hard to see from the blue side, but you can tell that the pictures for other channels other than red don't have to be the same.

Yes, at first I thought the images were the negatives of pictures and I could not blend them correctly till MJW mentioned channel extractions and the use of additive mode.  I find it interesting that there are several methods of doing this effect.

 

I used your Extract Channel plugin for this image.

 

[Welsh Yellow Cheddar]

[Reptillian]

One last observation, the red channel and blue channel are using the same image as base. It's only the green channel that should use a different image. Seems very doable as a script to be honest.

 

So basically:

Red has no change

Green is a negative, and has a different base

Blue is a negative, and is the same base as in red.

Edited September 4, 2022 by Reptillian

[Welsh Yellow Cheddar]

Thanks MJW and Reptillian. Both methods are great. If one day there is a script, I will be happy to use it.

 

 

[Reptillian]

@Welsh Yellow Cheddar

 

Now I'm developing a G'MIC script (which of course will be part of the G'MIC-QT plugin) to do this:

 

Now, I have added things like luminosity blending mode, and green control. I will find a way to allow users to tone the green down with human perception in mind. You could say this is more advanced than the original author intention.

Edited September 10, 2022 by Reptillian

[Reptillian]

Now, the GUI filter is in work:

 

[Welsh Yellow Cheddar]

@Reptillian

 

Every thing looks great. Even the name is interesting, it makes me think of red nail varnish. 🙃

[Reptillian]

I had just finished it. Took me a little longer as I have forgotten to account for alpha channel.

[anremo]

This looks great. do you already know when this will be added to the G'MIC release?