Hi,
nice technique you show in the video. You’re right, the burn and dodge tool is best for fast changes on small parts of the picture. Even for that cases I use a copy of the layer. For changing big areas of the picture, I sometimes use a copy of the layer in a different blendmode and a layer mask, or several copys in different blendmodes, each for another area of the picture. But your technique with the three layers is very good. I will try that.
Greetings
Fornit

In addition you could use some technique of Charles Cramer;
Tonal Selection.

http://www.luminous-landscape.com/tutorials/Cramer_Selections.shtml?main_page=product_info&cPath=20&products_id=173

Before creating a layer with Highlights, midtones or lowlights you should examine the zone to dodge or burn using the curve tool.
Move the cursor over the image with CTRL-left click and look which colors are used

This is my first comment on your site, even though I have watched many of the videos. Great work which is very much appreciated. I agree that using layers is a more controllable method of dodging and burning, however, having seen your tutorial I thought i would experiment with the dodge and burn tool. It certainly is powerful and great for playing around with. I think I will use it in the same way as one of your previous posters, on a separate layer.

I think that the first example you gave of dodge and burn was towards the end in episode 10. The technique used there is similar to that shown here but differs in some respects and is much simpler but I suppose not so effective. Time moves on and so does one’s knowledge and experience and it is most interesting to do what I have just done use the ‘time machine’ and go back and look at the past.

There is an old saying – ‘There is more than one way of killing a cat’ ( apologies to animal lovers) – and often the simplest way is the best.

I discovered “Meet the Gimp” last month and have watched every episode until this one. Of course I plan to watch the others. They’re great! This one was, for me, the most interesting by far. It really tested my understanding of the underlying model.

Specifically, I am struggling to understand how the layer mode affects the resulting image.

Suppose the background image “midtone” area is a solid medium gray (808080) and that we have perfectly isolated this area with a white mask (the mask is black everywhere else). Now suppose we paint one white (dodge) and one black (burn) pixel in the midtone layer. Naturally these must be painted where the mask is white in order to affect the result. According to the formula for overlay mode given in the linuxtopia reference, the result should be an arithmetic average of screen mode and multiply mode since the “weights” B/255 and (1-B)/255 both equal 0.5.

Consider the white pixel. Screen mode would turn the result pixel white, and multiply would leave it unchanged. So the result will be the average of medium gray (808080) and white (FFFFFF), or A0A0A0, a light gray.

Consider the black pixel. Multiply mode would turn the result pixel black, and screen would leave it unchanged. So the result will be the average of medium gray and black, or 404040, a dark gray.

It took me an hour to figure this out, and I do not even have 100% confidence in my analysis. Perhaps someone will spot an error. With realistic background colors, like the ones in the image you used in this episode, screen and multiply are weighted unequally, and my mind simply cannot predict the result.

I guess that am looking for a non-mathematical mental model of layer modes that will enable me to predict results. Maybe such a model does not exist, but if you have one, it would be great if you devolted an episode to explain to us how your mind works. For example, how did you come up with this technique? Or is everything simply done by trial and error?

Thank you for providing a fantastic resource to the Gimp community.