Filters

Our assignment was to take a new picture and apply filters to create 3 new images. We are to have a goal in mind and not just apply random filters. My goal was to turn a photo of a still life into a painting or sketch.

Some of the differences in these images are subtle. Click image to view in larger format for best viewing.



This is my original picture for this assignment. I was just in Dallas visiting family and took this picture in my sister's kitchen window where she displays a collection of antique purple glass bottles. The window was covered with dew and was a little frosty so the trees outside the window are obscured. I thought that if I painted, this would make a nice still life. So that is the effect I was hoping for.

I tried several different options to turn this into a "painting" only to discover that the frosty window turned into a very busy background. So my first step ended up to select the window with the quick selection tool and apply a Glassian Blur. Then I applied the Accented Edges filter to bring out the details of the bottles and compensate for a lack of contrast. I decided I liked that image. So I stopped there and saved it. I used it as the base layer for other images.

For the image above, I started off applying the Paint Daubs filter with a brush size of 1, sharpness of 0 and type of wide sharp. Then I applied a Box Blur (to the window) with a radius of 12 px. It didn't look enough like a painting, so I applied the Accented Edges with edge width of 9, edge brightness of 19 and smoothness of 14. Then I decided it needed to look more like a canvas, so I used the Texturizer filter for Canvas with a scale of 88, relief of 17 and light, top left. Then I made the mistake of trying to smooth out some of the splatter on the window. But I discovered that once you apply a texture filter, further adjustments really mess that up. So I backed off my last adjustment deciding I liked the window with a little splatter.

Then I decided to make the bottles into a still life sketch. This turned out to be the hardest effect to create. I tried several different filters and ended up with big areas of black (like on the big bottle to the right). This attempt was a combination of a Graphic Pen filter with stroke length of 15, light/dark balance of 40 and stroke direction of right diagonal and a Watercolor filter with brush detail of 6, shadow intensity of 1 and texture of 1. I was hoping this looked like a pen and ink sketch. It doesn't really. But I liked it anyway.


This was my final attempt to convert my purple bottles to a pen and ink sketch. I started with the second image above (Glassian Blur on the background and accented edges). Then I applied the Photocopy filter with detail of 24 and darkness of 50. I played around with this a lot and either have too much black in the center of the bottles or I lose the outline of the middle back bottle. I never found the right combination of filters to look like a pen and ink drawing. But I came up with some interesting variations of my original image.

So this is the last one worth putting out for display. I was going for a watercolor look. But I know the Watercolor filter in photoshop looks very little like a true watercolor. So I used the image with the Glassian Blur and this time Sharpen Edges. Then I used the Photocopy filter with a detail of 7 and darkness of 15. I over saturated the base layer and then reduced the opacity of the top (filtered) layer to 55%. There's not as much contrast as I wanted. But further work only messed it up. So now I really am done.

Thanks for viewing. All comments welcome.
Ann

Zac, a colorful cat

So for this assignment I was to download a drawing of the left side of Zac, complete the picture by duplicating it and then color it in using the techniques in Lesson 6 of our text. This seemed easy enough when I did it with the butterfly. But I ran into a snag that stumped me for about 30 minutes.

The first thing I did was download the halfZac tile. I followed the instructions in the text to expand the canvas, select and duplicate the half drawing, flip it and nudge it into place. I used the darken mode adjustment to line up my halfs to make a full Zac.

When I was ready to make my selection of the lines, I realized I had a RGB image, so I converted it to gray scale. I'm not sure if that really mattered, I think I could have made the selection using the RGB channel but I wasn't sure. I made my selection with the channel for gray, inversed my selection and created a new layer.

Then I filled my new layer selection with black and tried to delect the selection on the background layer so I could paint the selected lines. This is where I ran into trouble. I thought I had followed all the directions in the book just like with the butterfly. But when I deselected the background layer and filled it with white, it deselected the lines layer. Then when I tried to paint, nothing happened. I fiddled around with these two layers and the selection and the layer lock and the brushes. Everything seemed right - but no results when I painted. Finally I notice that my mode setting on my brushes tool bar was set to overlay instead of normal. I think the overlay was a left over from something I was doing in the book. Once I got that figured out, it was all fun with colors.

I used a standard brush for the outline of the face. Then switched to the scattering and color dynamics for the frame. Next I added my color fill layer and filled in the large sections. When I thought I was all done, I decided to go a little farther and add some texture to the face. This ended up a little more trouble than it was worth. I had to divide my color fill layer into 2 - the face and the frame. I added the pattern layer to the face layer using the ant farm pattern. I blended the layer with the soft light mode.

Then I flattened it all and posted it.

Doing the PS-CS3 Dance

Doing the PS-CS3 dance

So my assignment is to select a picture of myself, select the head, copy and paste it and superimpose it on one of the dancers. I have some issues with this assignment. The first is that I need to lose about 100 pounds before I take head shots. Otherwise, Rita or Ginger will look like The Incredible Dancing Woman With The Balloon Head. The other issue is with hair. Right now I have really short hair, kind of spiky on top. Both of these dancers have hair down their back or at least on their neck. How am I supposed to blend my spikes with their do’s?

I found some old pictures of myself from back in the day that I had long hair. Some of them are almost in the right position, but not the right lighting. So I first tried putting my head on Rita Hayworth’s body. It was a dismal failure. But I’m posting it anyway so everyone can have a good laugh.

My second attempt I used Ginger’s picture. But instead of putting my whole head on, I just selected my face and superimposed it over Ginger’s face. This solved the problems of the hair, the double chin and the size of my head. But I’m not sure it’s what John intended when he made the assignment.

So here’s a brief outline of my process.

I selected the part of my face to transfer to Ginger. Then I copied it onto a new layer in the head shot photo.

Before I copied and pasted the selection I tried to clean it up as best I could. The head shot has a white/black lace background which made it difficult to quick select or use the magic wand. I had to zoom in to almost a pixel level to clean it up. But it was much easier to do it on the head shot than on the dancer’s picture. And to help clean it up, I made an extra layer just behind my selection layer and painted the extra layer bright pink. This made it a lot easier to see what I was doing. But it’s kind of scary seeing ones face floating in a field of pink.

Once I had it cleaned up I transferred my selected face (still in color) to Ginger’s. Of course my face was as big as half her body. So I used the Transform and Free Transform functions to scale my face down and tilt it to match Ginger’s.

After I had it the right size and position, I changed it to black & white. Then I adjusted the levels to try to get the tones just right. Okay, they’re not perfect but as close as I could get them.

Next I reduced the opacity of my face so I could see Ginger’s hair line. I zoomed in and carefully erased any part of my face that was covering her hair. There is a little bit of a halo effect on her face around her hair (like in front of her ears). I could never get it quite right.

Then I used the smudge tool to blend the line between my face and Ginger’s neck.

Then I sat back and laughed.

Look! It's the Amazing Dancing Woman with the Big Balloon Head!!

This is the 3rd version of this one that I've posted. I keep trying to get the head smaller and look more proportioned to Rita's body. So now nobody needs to be embarrassed about theirs.

Selection Mania

I thought this was a pretty straightforward assignment - to start with a blank canvas, select the areas to duplicate John's image, and color in the selected area. Of course the trick to this assignment was doing all the selection at one time, then the painting all at once. I've done enough in Photoshop elements to know the basics of selections. But I found, as usual, that I didn't know as much as I thought I did. I had not used the shift and alt keys to add and subtract from selections. It was a smooth process once I figured out that to subtract the corners to make the cross, I had to start with the cursor outside the selection. I tried first to line my cursor up on the edge of the selection. But it didn't work. I had to be outside the selection. I also learned that when using the polygonal lasso tool, I didn't have to hold the mouse button down while I was drawing. I just had to left click at the corners. Also, I'm not so good at drawing with my mouse. So my circle and triangles are a little off in size and in position. (I even tried doing this a second time with the grid showing (view, show, grid). It didn't help much. But I got the idea of the procedure for selecting, adding and subtracting from a selection.

The most exciting part of this for me was finding out about the quick selection tool. I didn't use it for this assignment, but I will use it a lot in other projects.
Ann

New tricks for old pics

I'm procrastinating in other classes, so I thought I would do a little playing around in PS. I have about 3,000 photos on an external hard drive that could use some help. To start with, I selected a photo I had worked on before with another more limited software. I took this picture of a church in Mendocino, CA with a 35mm point and shoot. Not very impressive. But a friend of mine was doing a collection of black and white photos of old churches and she really wanted an improved copy of this.

I scanned it into my computer so it's not the best quality to begin with. The brightness/contrast is all wrong. Every time I tried to adjust it I lost details - either the cross on the top blew out to white, or the door and windows lost all the detail in a void of black. So instead of just trying the auto adjustments, I used levels and curves to mend the contrast without losing the details. My friend likes sepia, so I went a little brown instead of gray. I've still got to figure out what to do with the sky. But considering my previous attempts (which I'm too embarrassed to show) this is a definite improvement. Old first, then the new.

Project 2 Color

Bummer
I've redone this assignment at least 4 times over the last week. Always with the same result - a photo that looks over manipulated. But I think I'm getting closer. Here's what I did for my last attempt. I started with Auto Color correction because I thought it wouldn't hurt and it might help. From there I went to Variations and increased Red one time. Next I went to Levels by color and changed them to Red: 0, 0.37, 222 and Blue 1, 2.12, 159 and Green 12, 0.73, 214 and RGB 6, 0.72, 255. I also changed the output on RGB to 22, 255. And the output on Blue to 0, 233. All of these adjustments were made with trial and error - just playing around with the Level slide bars. I also tried using Curves - but I messed up more than I improved.

By this time the colors were looking better, but not great. The red bedspread isn't red enough, the blue curtain isn't blue enough and the wall is too dark. But I've noticed that the more I finagle the colors, the more the contrast, brightness and sharpness degrade. So I'm stopping here and will wait to learn how others in the class overcome the very yellow retro room.

Mine is on the top. John's on the bottom right.

Project 2 - Adjustments

I tackled this one first because I thought it would be easy. Of course I was wrong. I started off well, duplicating the color image so I have a working copy. I put that working copy and the goal photo on a split screen and off I went trying adjustments to Levels, Hue/Saturations, Variations, Curves. My first frustration was that I get in a hurry to try the next thing in my bag of tricks and I forget to write down what I'm doing. I know the goal here is to learn how to make adjustments, not just duplicate John's photo. So even if I get everything just right with a hit and miss approach, it does me no good if I can't remember how I got there. So I have to keep reminding myself to slow down, think about what I'm doing and write it down as I go. Thank God for the history window.

I started off adjusting the Levels - it looked heavy on blue. Then I adjusted Hue/Saturation and reduced the saturation to almost, but not quite, black and white. Next I spent a lot of time in the Variations window. I like the variations window and find it easy to work with because it gives me a preview of the changes with my options before I select them. But I'm frustrated that the window is so large that it blocks the goal photo that I'm trying to match. I played around with the midtones, adding red, too much red, adding cyan, overdoing darkening then lightening it up. Then I revisited Levels and Hues/Saturations. I ended up with Hue 34, Saturation 18, Lightness -17. By this time I thought the street looked pretty close but the sky was too dark. I used the Curves function, selected a point on the sky and lightened it up without messing up the rest of the work I had done. (My input was 136 and output 186 on the Curves)

I'm about 85% happy with the way the photo turned out. But I've learned a valuable lesson to slow down and document what I'm doing.

So here's my adjusted photo:


And here's the goal photo: