Asher Kelman
OPF Owner/Editor-in-Chief
Well, Ben are you going to do it? Or just an idea?
Asher
Asher
-
Please use real names.
Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed
Firstname Lastname
Register
We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know! -
Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!
Challenge: "Girls Choosing Leaves for Magic"
- Thread starter Asher Kelman
- Start date
Jerry P'Simer
New member
Ok, another intriguing challenge that plays...
...into something that I have been asked to research and experiment with.
I was asked to try and simulate a stereo pair of images (3D) from a single image that can be viewed on a monitor using the "cross eyed" method. I did some research on the web and found a vast amount of info. There was nothing that really explained how to do it, so I had to just experiment. I thought that Asher's image was as good a starting point as any, and since the idea here is to create something "MAGIC, I thought I would give it a shot.
To view the image takes a little patience and practice, so I have included two additional images for this purpose. The first image is used to practice with to allow your eyes to focus properly. It is a simple image. The idea is to get your left eye to focus on the right image and your right eye to focus on the left image. Once you have succeeded you should be able to see the green object in the image floating over it's background centerd between the pair of images which will then be out of focus. You can then just relax your eyes and you will see the image in near perfect 3D as if it were floating above the screen in front of you. Relax and you will then be able to look anywhere on the image while ignoring the backgroud completely.
It's best to view the image in Photoshop on a black screen if you have PS, although, your browser will also work fine since the images are placed on a black background. Download the image and open it in PS. Hit the letter f key 2 times to get the image onto a black screen. Hit the tab key to turn off the menus. Hold down the space bar and drag the image to an optimal place on the screen for viewing (PS CS). Sit back about three feet or so from your monitor. If your rulers are turned on then turn them off as well. Move your mouse curser to an area away from the images to prevent distraction.
If you are able to see the above image in 3D, then this next image will be startling. It is a true stereo pair that I found while researching. If you are able to focus properly and then relax your eyes you will swear you are looking at a perfect ViewMaster® image in full 3D. You should have no problem fully examining the image while remaining completely relaxed. This little guy should appear as though he is floating about a foot in front of your face.
Here is the simulated stereo pair that I created from Asher's image along with an explaination of how I accomplished it.
I created it after cropping the image and then opimizing it for the effect. I duplicated the image so I had to copies open at once. I took the right eye image (left image on screen) and duped the base layer. I then applied the emboss filter (Filter-Stylize-Emboss) using the following settings to the top layer. ANGLE = 0, Height = 4, Amount = 150. I then set the top layers blending mode to Soft Light. I then changed to the second copy (Left eye Image...right image on screen) and did the same steps as above but reversed the angle in the Emboss Filter to 180. I then set it's blend mode to Soft Light. I flattend each copy and placed them side by side on a new image with a black background making sure to place the images in the correct left and right placement.
The effect appeared to be reasonable at this point. To further advance the effect I did some dodging and burning on each image. I dodged in small amounts in several locations on the left image (seen with right eye) using various brush sizes and amounts with the brush set to highlight. I did some burning on the right image as above (seen with left eye) usng the midtone setting. The idea here was to add some depth to the image. Since most people are right eye dominate, I thought it best to add the highlights to the left image seen with the right eye. Since the effect tends to get lost as it touches the edges of it's bounderies, I darkend the edges using a soft mask.
Here is the result:
MAGIC?...You decide.
Jerry
Asher Kelman said:
...into something that I have been asked to research and experiment with.
I was asked to try and simulate a stereo pair of images (3D) from a single image that can be viewed on a monitor using the "cross eyed" method. I did some research on the web and found a vast amount of info. There was nothing that really explained how to do it, so I had to just experiment. I thought that Asher's image was as good a starting point as any, and since the idea here is to create something "MAGIC, I thought I would give it a shot.
To view the image takes a little patience and practice, so I have included two additional images for this purpose. The first image is used to practice with to allow your eyes to focus properly. It is a simple image. The idea is to get your left eye to focus on the right image and your right eye to focus on the left image. Once you have succeeded you should be able to see the green object in the image floating over it's background centerd between the pair of images which will then be out of focus. You can then just relax your eyes and you will see the image in near perfect 3D as if it were floating above the screen in front of you. Relax and you will then be able to look anywhere on the image while ignoring the backgroud completely.
It's best to view the image in Photoshop on a black screen if you have PS, although, your browser will also work fine since the images are placed on a black background. Download the image and open it in PS. Hit the letter f key 2 times to get the image onto a black screen. Hit the tab key to turn off the menus. Hold down the space bar and drag the image to an optimal place on the screen for viewing (PS CS). Sit back about three feet or so from your monitor. If your rulers are turned on then turn them off as well. Move your mouse curser to an area away from the images to prevent distraction.
If you are able to see the above image in 3D, then this next image will be startling. It is a true stereo pair that I found while researching. If you are able to focus properly and then relax your eyes you will swear you are looking at a perfect ViewMaster® image in full 3D. You should have no problem fully examining the image while remaining completely relaxed. This little guy should appear as though he is floating about a foot in front of your face.
Here is the simulated stereo pair that I created from Asher's image along with an explaination of how I accomplished it.
I created it after cropping the image and then opimizing it for the effect. I duplicated the image so I had to copies open at once. I took the right eye image (left image on screen) and duped the base layer. I then applied the emboss filter (Filter-Stylize-Emboss) using the following settings to the top layer. ANGLE = 0, Height = 4, Amount = 150. I then set the top layers blending mode to Soft Light. I then changed to the second copy (Left eye Image...right image on screen) and did the same steps as above but reversed the angle in the Emboss Filter to 180. I then set it's blend mode to Soft Light. I flattend each copy and placed them side by side on a new image with a black background making sure to place the images in the correct left and right placement.
The effect appeared to be reasonable at this point. To further advance the effect I did some dodging and burning on each image. I dodged in small amounts in several locations on the left image (seen with right eye) using various brush sizes and amounts with the brush set to highlight. I did some burning on the right image as above (seen with left eye) usng the midtone setting. The idea here was to add some depth to the image. Since most people are right eye dominate, I thought it best to add the highlights to the left image seen with the right eye. Since the effect tends to get lost as it touches the edges of it's bounderies, I darkend the edges using a soft mask.
Here is the result:
MAGIC?...You decide.
Jerry
Last edited:
Asher Kelman
OPF Owner/Editor-in-Chief
Jerry,
That's not magic, it withchcraft! I was impressed that the image in the center became so real and I was able to relax and view it. Now a hologram would be wizardry!
Thanks for using your special powers and sharing the new use of the "emboss" filter.
Asher
That's not magic, it withchcraft! I was impressed that the image in the center became so real and I was able to relax and view it. Now a hologram would be wizardry!
Thanks for using your special powers and sharing the new use of the "emboss" filter.
Asher
Jerry P'Simer
New member
You're welcome Asher
I'm glad you liked it.
Jerry
I'm glad you liked it.
Jerry
Hi Sean,Sean DeMerchant said:
Sorry for the reviving your post a little late, I've been on holiday and I'm trying to catch up...
You did a great job with those blown highlights, and I've not seen this particular technique before. Personally I'd really appreciate having a peek at the PSD file.
Could you explain the how the application of the calculation works. Why the red and green channels in particular?
Thanks,
Tim
Anita Saunders
New member
Closer square crop, gold blurred vignette, background lightened with boost in contrast, foreground darkened, centre leaf area sharpened.
Anita Saunders
New member
Retouching
Retouching really depends on a preconceived idea so it is difficult working on another artist's work. But it is useful to give an idea of how a different photographer may have approached the image capture.
For eg if I had taken the photo with the idea preconceived (as per my previous post), I would have used a wide aperture so the plane of focus only reached the hand/leaf area. I would also have altered my camera level so that the leaf itself was lit and contrasted against a dark shadowed area of the leaves behind. Obviously I would also frame closer rather than crop later.
Retouching really depends on a preconceived idea so it is difficult working on another artist's work. But it is useful to give an idea of how a different photographer may have approached the image capture.
For eg if I had taken the photo with the idea preconceived (as per my previous post), I would have used a wide aperture so the plane of focus only reached the hand/leaf area. I would also have altered my camera level so that the leaf itself was lit and contrasted against a dark shadowed area of the leaves behind. Obviously I would also frame closer rather than crop later.
Sean DeMerchant
Inactive
No worries Tim. As I often say, "all things happen in good time" and I do not worry about it past that. And online, I often delay responses by huge time frames due to other constraints on my time.Tim Armes said:
This technique is an extrapolation/interpretation of some of Dan Margulis' ideas. Go to http://www.ledet.com/margulis/articles.html and scroll down and read Plate Blending As Poetry to get an idea of where this comes from. Dan is the only writer on Photoshop who actually teaches the behavior of color in his writing as he actually talks about the mathematics of the color spaces and how to work with them. In mathematical terms, an ICC profile is basically equivalent to the field axioms while Dan writes about Fourier Analysis.Tim Armes said:
http://www.envisagement.com/opf/GirlsChoosingLeaves_crop.psdTim Armes said:
Please note my explanation was simplified. But in essence the image was created starting with the bottom layer and working upwards. So if you turn off the layers from top to bottom and then slowly turn them on moving upwards you will see a progression of changes.
Calculations and Apply Image are simply ways of applying ill-defined* mathematical operations to image channels. Calculations takes two channels and a mathematical operation (multiply, lighten, darken, ...) as input and does a pixel wise mathematical operation to generate a new channel (i.e., multiply the red and green channels to yield a new alpha channel). Apply Image takes a channel and an operation as input and applies it to the currently selected channel/s.Tim Armes said:
The mathematical operations are the layer blending modes. i.e.,
multiply -> multiply (on 0-1 not 0-255)
screen -> divide (on 0-1 not 0-255)
darken -> minimum
lighten-> maximum
...
Plus two special operations that are not in the blending modes:
add
subtract
These tools enable one to implement selection of various items in an image via mathematical algorithms. And by doing final adjustments on the selection with curves one filters the results through the human visual system to make it work as it is the image that matters (not getting the numbers right).
And that is the short answer.
Tim Armes said:
The goal was to select the blown out areas. After looking at the individual channels I noted that the blue channel's brightness did not correlate well with the blown detail. So I chose the red and green. I them multiplied them to change the lightness mapping from "linear" to "quadratic" which makes the blown parts stand out and shadows very dark. If only one channel had mapped well to the blown out details I would then have multiplied that channel by its own self.
Remembering that arithmetic here is on the interval from 0.0 to 1.0 and not 0-255 we can see that 0.1 is a shadow and 0.1*0.1=0.01 which is very close to zero. While a blown highlight is 1.0 and 1.0*1.0=1.0. So with the quadratic remapping of the data we relatively isolate blown detail from normal data. We than tweak this with curves.
Stepping back to Dan Margulis' ideas one can get that all highlights except specular highlights should have some ink on them. Hence, taking this selection and adding a small amount of noise (which is then blurred) creates the perception of the slightest texture which the human visual system then interprets as bright but not too bright rather than a lack of data that blown detail is.
enjoy,
Sean (who came at all this from the computer vision side and could not see the image for the numbers for several years)
* Adobe does not share the exact details and it is not worth the trouble to reverse engineer when it is the visual result that matters.