A commenter on my previous post asked about how I make the borders on the images I post. I'll talk a bit about that here.
At first I thought that it was wasteful of screen real-estate to add a border. When printing a picture to hang on the wall, I don't add borders, but I do use colored matting to compliment the picture. That's what I've got in mind when adding borders to images that I post.
The real skill involved -- skill that I don't have much of -- is deciding what to do. How to do it is much easier.
I made a Photoshop action which does most of the work:
Set the image's color space to sRGB. Normally, I shoot in Adobe RGB.
sRGB is a Microsoft color-space standard, and like most of what comes out of Microsoft, it is technically inferior, but more popular, than other options. It's a del-facto standard on the web, and while some browsers can auto-adjust for a different embedded color profile, IE doesn't, so I've got to downgrade the color encoding for a general audience.
Make a duplicate copy of the background layer. This copy will become the image actually seen within the border about to be added.
Change the canvas size: add 30 pixels to both the width and the height.
Most raw images I work with are about 3,000 pixels on a side, so this adds just a small 1% border. I adjust the size later, manually, to suit the specific image.
Add a black, solid-color fill layer between the background and the copy. This renders the background layer irrelevant, with the copy now the image that's seen. The newly-added fill layer is seen only around the edges, that is, it's the border. I change the color later, manually, to suit the image.
Change the Layer Style of the background copy by adding a white “outer glow”. I sometimes change this to a “stroke,” especially when I intend to make small versions.
I take the trouble to do steps 2 and 4 because it allows me to change the border color later on, on the fly. You can set the color of the newly-expanded canvas when changing the canvas size, but once you do the color is not easy to change. So, I make a pre-enlarge copy of the image, which, unlike the background, does not expand when enlarging the canvas, and throw a solid-color fill layer behind the copy.
I then manually adjust the border color by adjusting the solid-color fill layer, usually by eye-dropping a color from the image and working from there. I futz around with it a lot to get something that just seems to work. I also adjust the color of the outer-glow (or stroke) around the main image, to get something subtle.
If I want to add a title, I'll make the border larger, especially on the bottom, then add the text. Title text usually gets a subtle outer glow or a drop shadow, depending on the colors involved, to highlight it a bit from the background. Yet, I often do the titles at less than 100% opacity, to keep them a bit more subtle. It's all a judgement call.
Even small borders can have a huge effect on an image's feeling. The problem I have as a geeky engineer devoid of even the most basic artistic sense, is in picking a border that has the right effect. It happens often that I'll revisit an image I was working on before dinner and be totally turned off by the border that I thought was so great an hour ago. I really have no clue what I'd doing.
Here's a copy of yesterday's main image with a different border:
I really have no idea which one is better.
Once the image is done, I use Photoshop's Save for the Web command to write resized copies in the sizes I intend to use. At this stage I doublecheck that the text is readable at the resized copy. This save-for-the-web step strips out most of the Exif data from the image (including embedded thumbnails, which are wasteful of space). I then run my own little command which reinserts the main Exif data (date, location, geo-data) back into the copies that I'd just saved.