I just replied to a thread where someone wanted to know about hard, soft, and lost edges and how to decide when or where to use them. I’m pasting it here too so that others can benefit from it or add to it.
---------From the original thread--------------- (http://forums.cgsociety.org/showthread.php?p=4342098#post4342098)
Marta - I got your PM, and I’ll reply here so that it might help others with similar frutrations, or those who’s got a handle on it can join in the discussion.
Edges was one of my problems too (and it still may be), and I had researched or asked the artists I respected, or read up on the subject in books and online. My current approach is this (and this is just my personal approach):
-Edges that are hard are either objects that need to appear hard-edged in order to look right or have the proper effect in the image (for example, the gleaming blade that caresses the vulnerable bare neck of a woman, and the blade itself is vital to the message of the image), or used in areas of interest where you want to draw the eyes.
-Soft edges are for areas that need the softness (for example, a teddy bear in a scene), or areas where you don’t want things to stand out too much (background objects), or used to help contrast areas that need to pop out with harder edges.
-Lost edges are for areas that you really want to push back into the image and not cause attention, and for lighting falloff where the light is no longer causing enough contrast in the scene.
The soft/lost edges can also be used to help gel foreground/background elements together so they don’t appear like paper cutouts in the scene–to harmonize the various elements in an image, so to speak. They can also be used to depict areas that are out of focus or further away, but there are two approaches to this. I’ll elaborate:
Approach 1 - To mimic what the camera sees–where the depth of field is locked in a given moment, and the focused/unfocused planes are very obvious and well separated.
Approach 2 - To convey what the human eye sees–which cannot be locked in a given moment, because our eyes can only see one tiny area clearly at a time, and the areas in our peripheral vision aren’t necessarily always in a different focus plane–they just aren’t in that “sweet spot” where our eyes can see clearly. If you really look at your surroundings, you’d be surprised by just how tiny of an area our eyes actually can focus on at any given moment–roughly an inch or so of space. But to do a painting with only a tiny area of clear detail would be frustrating (but it might be cool as an experiment). So what some artists do is to paint the overall image as if the eyes are looking at the entire scene in not an instant, but slightly longer period of time, stopping at various areas of interest. So this approach would have hard/soft/lost edges according to what areas the artist would want you to pay attention to.
But then again, there are artists who do not typically rely on using soft or lost edges to convey distance or areas of focus–they prefer to use values, colors, brush stroke, brush size, texture…etc to convey that information (Craig Mullins is a good example).
I hopes this helps. It’s one thing to know this stuff intellectually, but to apply them creatively is another story. I struggle with it too, just like the rest.