Tuesday, April 28, 2009

How to Correct an Image With the Levels and Shadow/Highlight Command

In this tutorial I will show how to correct the color and lighting of an image using the Levels and Shadow/Highlight command. As I will demonstrate, using these commands will give you greater control and better results than just the Auto Adjustments commands that come with Photoshop.

The image I have chosen for this tutorial is a photograph I took of a young porcupine several summers ago:


I had taken dozens of photographs of this porcupine and this one image was the perfect pose. However it was back lit from the sun bouncing off Gastineau channel in the background so the light is very harsh creating a lower contrast image and a greenish yellowish cast. I would like to correct this effect.

When you first start using Photoshop you might adjust your photos using one of the Auto commands. Here is the original photograph applying the Auto Level command - Image -->Adjustments -->Auto Levels or Ctrl + Shift + L :


As you can see, the Auto Levels does improve the contrast, but there is still is a yellowish greenish cast to the image. Let's see what Auto Contrast does.

Here is the original photograph applying the Auto Contrast Command - Image -->Adjustments-->Auto Contrast or Alt + Ctrl + Shift + L :


With Auto Contrast there doesn't seem to be any change at all from the original photograph. Next let's try Auto Color.

Here's the original photograph applying the Auto Color Command - Image -->Adjustments -->Auto Color or Ctrl + Shift + B :


Auto Color adjusts the color somewhat, but it still has a yellowish cast. What I would like to do is tone down the yellowish greenish cast, bring out the blues of the Gastineau channel in the background as well as some red tones. I would also like to add some contrast to this image.

With the exception of Auto Contrast, the Auto Commands did improve the quality of the image. You might say that it looks pretty good. However, you can get better results using the Levels command. The Levels command will let you edit individual color channels as well as shadows, midtones and highlights. To access the Levels command, select Image --> Adjustments --->Levels or press Ctrl + L:


Starting with my original image I am going to press the Auto button on the right hand side of the Levels window. This will give the image some auto correction on which we can do further correcting. The result is nearly identical to using the Auto Levels function. Next I'm going to reduce the greenish yellowish cast in the photo by selecting the Green Channel from the drop down Channel box. This will let me isolate the green values :

You adjust the shadows, midtones and highlights input values for each Channel using the black, grey and white slider triangles under the histogram box. You can drag the triangles using your mouse, nudge by selecting the number directly beneath the triangle and using the arrow keys or directly entering a value in the number box. For my image I raised the black (shadow) triangle to 11 and then reduced the grey (midtone) triangle down to .97, which reduced the greenish cast and brought out some red tones.

I Then selected the Blue Channel from the drop down box to adjust the blue tones. I adjusted the white Input triangle, which lowered the highlight input value to 206. This brought out some of the blue in the Gastineau channel in the lower left of the image. When I was finished, I clicked the OK box to save the changes. Here is the result:


With Levels, I am able to correct the color cast and bring out blue and red colors in the image. Here is a side by side comparison with just the Auto Levels command (click on image to enlarge):

As you can see the Levels command provides a better result than just the Auto Levels.

But I'm still not done with this image. I'd like to reduce the harsh backlight and lighten the face of the porcupine. This is where the Shadows/Highlight command becomes useful.

Go to Image --->Adjustments--->Shadow/Highlight to open the Shadows/Highlights command. A dialog box should appear. Click the Show More Options box at the bottom.

The Shadows amount will be set to 50%, which will lighten the image. Set the amount to 0% to display the original shadow amount. First I want to lighten the face of the porcupine a little. I don't want to lighten it too much so I adjusted the Tonal Width for Shadows down to 30%. The Tonal Width determines the range of values that is considered to be shadows or highlights. I then increased the shadow amount to 10% to lighten the shadows on the porcupine's face to bring its face out. Next I increased the Highlights amount to 70%, which greatly reduces the harsh backlight. Here is the end result:


Here is a side by side comparison with my original image (click on image to enlarge). What a difference!



Monday, April 20, 2009

Project 7 / Layer Masks

For my first image, I picked a black bear photo I took in Juneau last summer. Here is the original photograph:


Here is the finished image:


I pretty much followed the instructions John gave in the video and the handout for the layer mask project for creating the shadow and midtone layers and merging them. When I created a text layer I used a free font I had downloaded called 28 Days Later for the text. I decided to change the color of the text from black and white to yellow. I also moved the tonal mask to the top of the layers so the text would be behind the bear instead of over the bear image. I experimented with different gradient adjustment layers until I decided on picking a green color and then using a gradient fill.

I decided to use a photo of a buckbean flower I took at the Mendenhall Glacier a couple of summers ago for my other image. Here is the original photograph:


Here is the finished image with layer masks:

First I duplicated the photo and went into the channels palette. I wanted to isolate the white flowers from the rest of the image so I picked the channel that did that the best which was the blue channel. I then duplicated the blue channel. I selected the RGB image and then loaded the selection so that it would select the flowers. I then inverted the selection and deleted out the background.

I decided to add a pattern fill layer and experimented with the different pattern selections until I finally settled on a burlap type pattern. Then I chose a color fill layer and picked a purple color. I decreased the opacity for the color fill to bring the pattern through.

I didn't feel the white flowers stood out from the background as much as I would like. So I went into the layer blending option for the selected white flower layer and selected lighter color blending and also the drop shadow option. That made the white flowers stand out more.

I'm finding layer masks to be complicated and I definately will need to spend a lot more time to really understand all of the layering and masking options.

Wednesday, April 8, 2009

Project 6 - Photo-stitching and Text application

At first I didn't want to do a scenic panaroma for this assignment. I thought about doing a series of photographs where you have a moving object across a series of photographs. I was initially going to try to photograph a squirrel on the deck railing and then stitch it together to make it look like there were multiple squirrels on the railing. That sooned turned out to be much more complicated than I though because when you moved the camera across the deck the trees in the immediate background would still be in the same part of the background creating some issues with perspective. So I settled for a scenic photograph. Here are the original photos I took of the Gastineau Channel and the beach near where I live:



I basically followed the instructions that John gave for stitching the photographs together during the last class meeting. I modified the canvas size to accomodate 3 images. I copied and pasted the the second image next to the first creating a layer and setting it to 60% opacity. When I matched the pictures I set the opacity back to 100% and used the eraser tool to blend the picture lines together. I did the same with the third image. I found it difficult to match the pictures exactly, the lines on the beach and in the water did not completely match up. I also adjusted the RGB levels so that each photo would match in tone. One mistake I made was not putting more blue from the RGB levels in the first photo on the right.

I then flattened the image and cropped it and then made some adjustments for shadows and highlights to make it less dark. I also added some yellow saturation to bring out the beach grass on the beach.

The last photo is with text added. I found the text tool a little tricky to work with. I'm used to using Microsoft Publisher at work and because I wasn't working with the full sized image when I put on the text the large font sizes through me off. I played around with some different font types until I found a couple that I liked.

Wednesday, March 25, 2009

Project 5: Focus and Filters

One of the things I learned the most from reading Chapter 8 of the text was to use more than one filter for each image. When I have used filters before I've just tried applying one per image. Using more that one filter can add more depth to an image effect as well as creating a dizzying array of possibilities. I felt pretty overwhelmed at the possibilities.


Here is my original photograph:

This is a photo of part of the Mendenhall refuge in Juneau I had taken a couple years ago. It was taken during really rainy, stormy weather so it has sort of a painterly effect to begin with. I decided to try to apply three different artistic effects to the photograph. The first was to make it look like an oil or acrylic painting. The second was to make it look like a pastel drawing. The third was to make it look like a watercolor and ink painting. You really need to enlarge these images to really see the effects. They don't show up as well as these smaller images on this blog.


This first image I tried to make look like an oil or acrylic painting. I first choose Paint Daubs as a filter and adjusted the brush size and sharpness until I liked it. I also chose Wide Sharp for brush type. The result created an effect that looked like painting but it lacked depth. I created a new filter layer in the filter gallery and experimented around until I found the Dark Strokes filter. That added the darker color of paint strokes to the image. At first it was way too dark so I adjusted the black intensity filter down to 0, which still darkened the image but did not overwhelm it.


The second image I tried to make look like a pastel drawing. I chose the rough pastels filter from the filter gallery and then added a new filter layer. Then after experimenting with some different filters I also chose the sprayed strokes filter. Combined with the rough pastels filter it seemed to make the pastel effect look more natural.


The last effect I tried was to try to make the image look like a watercolor with ink drawing. I chose the watercolor effect from the filter gallery and then added another filter layer and choose ink outlines. I adjusted the ink outlines to minimize the amount of dark ink marks on the image.
I've found that experimenting with which filter you apply first does make a difference. I've also realized that the size of your final image is important in how you adjust these filters.

Sunday, March 1, 2009

Project 4 - Coloring Scanned Artwork: Variations

To get the baseline black and white image I followed the instructions in the text to double the width of the canvas, select, clone and use the Free Transform command. I then flipped the horizontal, nudged the selection into place and loaded the RGB channel as a selection. Again following the text I reversed the selection, made the line art layer and then filled the selection with black. I selected the background layer, then deselected it and filled it with white. I then locked the transparency of the line art layer. I made several copies of this image to try different things.

It was pretty overwhelming to look at all the brush presets because there was so many to choose from. For the first image I finally choose a basic brush at 60px and and a special effect brush called hypno lines that had some interesting patterns. I brushed over the image with the 60px basic brush with a darker color and then used the hypno lines brush with a lighter color to create some texture. This was pretty easy to do because I locked the transparency of the line art layer.

For the next image I first set the foreground color to a dark warm brown and with a basic brush of 48 px colored over all the black lines. I created a new layer to fill in with colors and textures and moved the layer backwards as in the text. I selected the paint bucket tool and clicked in the shapes to fill the shapes with various colors. I found it useful to select large thumbnails for the color swatches because the default swatches were so small. I also had to enlarge the image several times so that I could fill in the small shapes without flooding the image with the color.

I had a lot of fun trying out different filters in the filter gallery. I selected the color shapes I wanted to use a filter on with the magic wand tool and the shift key. That seemed to work fine for this image. I used the stained glass texture for the peach color on Zac's "forehead". It actually looks like snakeskin to me. The yellow rectangle in the middle of the forehead was filled in with a darker yellow and then I used the hypno lines brush again with a lighter yellow for the swirly pattern. I also used the texturizer, crosshatch (eyes), and rough pastels filters on some other shapes. I used the plastic wrap filter for the "nose".

For the last image instead of choosing a color to paint over the black lines I picked Styles and loaded the photographic effects. There was one effect called blue tone w/ shadow that made the lines look three dimensional with a bluish shadow. I used the paint bucket tool to fill in all the shapes with a light yellow. I selected the colored fills layer and then created a pattern layer and made it part of a clipping mask (not quite sure what that means) by checking the Use Previous Layer to Create Clipping Mask.

There were a lot of patterns to choose from but I eventually selected the gauze pattern. I choose Hard Light from the Layers palette and changed the opacity to 69%. That allowed the gauze pattern to show through the yellow colored fills layer.

The result was interesting. It looks like the lion sketch is made out of string and it is draped or tacked over a light yellow gauze canvas.

Monday, February 23, 2009

Project 3a: Redo

Apparently I did this assignment incorrectly, so here is the redone image. The first thing I did after creating a new document with size 8" x 8" ,72dpi and white background was to view the rulers on both the new document and the image to duplicate so I could get the approximate sizing and alignment.

I first selected the rectangular marquee to create the cross shape in the center and the four squares. I drew a rectangular bar down the center and with the shift key added the other horizontal bar. This created the cross. I then continued to hold the shift key and added the four squares around the perimeter. When I did this I would cross check with the completed image using the ruler to make sure I was getting the correct sizing.

Next I switched to the elliptical marquee tool to create the circle. I selected the subtract from selection option and drew the circle by holding down the alt key. I also used the spacebar to position the circle in the middle of the image. This seemed more tricky this time then when I did it the first time and I'm not sure why. I discovered the edit ->step backwards function which was helpful when I wanted to undo a selection mistake.

I then selected the polygonal lasso tool to draw the triangles. I started at one corner and drew one edge, clicked, drew the second edge, clicked, and drew the last edge and closed the triangle when the 0 appeared. I also would view the rulers while I was doing this to verify the sizing.

Lastly, I chose a foreground color, right clicked on the image and then filled the entire image with color.

Wednesday, February 18, 2009

Project 3b

The Rita Hayworth and Ginger Rogers dancing photos were not my cup of tea, so I tried to find a base image to superimpose my head on. I found one on Flickr that allowed reuse through a Creative Commons license, and proceeded to take a picture of myself. The results were absolutely hideous, and I almost gave up on this project. I really do not like doing self portrait type things. So you are getting a picture of a flower instead. The photographs are pictures that I took. This is what I ended up doing:

Here is the base image:



















Here is the selected image:



















Here is the end result:



















I first tried using the lasso tool to select the flower which really was a pain and didn't work very well. I then tried the magnetic lasso tool, which was wonderful. You could trace around the flower and it would easily attach itself to the edge. I dragged the flower selection to the picture of the rocks and pressed the ctrl key to move it into place.

I then went into match color adjustment and chose the flower image as the source and checked use selection source to calculate colors. I was then able to adjust the fade and color intensity settings so that it looked like the two images matched in terms of color.