guide/walk-through (ask the maker)
Requested at Ask the Maker 2.0 by likealight - thank you for asking something ♥
I apologize in advance for the lengthy novel I have written here, I don't even know if I answered this the way it was supposed to be answered, but well, I tried. :)
Also, I feel a little bad because it has taken me so long to get this up - I know I did say that it would be a while, but still, I'm sorry for making you wait for so long for this nonsense! :)
Ask the Maker 2.0 || Color & Contrast (a guide/walk-through/...thing)
So, the question was mainly about the use of contrast in my icons and what kind of techniques I'm partial towards. While I have not really settled yet for a particular way to create my icons, I am leaning more towards a certain kind of 'workflow' recently, using a small handful of tools that I feel comfortable with, so this is what I am going to describe here. Because I'm not good at making things short and up to the point this has basicly become some kind of walk-through of how I make my icons, minus the detailed coloring stuff and such; so yes, this is a lot longer than necessary. ;)
First, lets have a quick overview of what I do:
The techniques themselves aren't anything particular new, I'm afraid, I'm sure most people will know about them all, but I'm still going to try and explain them for anyone who maybe hasn't come across some of them yet.
A LITTLE MORE RAMBLING BEFORE WE START
When it comes to icons, the one major goal I run for is to make them bright and really vibrant, because that is what I want my icons to be. I want the colors to look strong and rich and striking, possibly without being painfully overwhelming, but I like to push the boundaries a little, if only to see how far I can go. Contrast on the other hand is something I feel I'm still struggling with, I often end up with something that is either too much or not enough and I'm still not sure how to find the right balance for every picture.
Anyway, as a result, almost everything that I do adds to this 'colorfulness' in a way. I always push contrast and colors around in several steps, often switching between adding color, saturation (as you will see I use lots of saturation) and contrast, back again to more color, some more contrast and even more saturation and back and forth and all over again as I go along. (This is one of the reasons why I usually end up with hundreds of layers in my icons. Ahem.)
So, coming back to my 'workflow' from above: I don't always use those steps in the exact order I've written them down here, I often go back and adjust the layers some more as I go along, sometimes even deleting some of them again, or changing the order they are in in the layer palette just to see where it will take me. So it's not so much a chronological and straightforward process when I make my icons, but a little more muddled. I usually start working on a larger sized screencap, fiddle with the different techniques until I have something that looks like a promising base, then I shrink down my image into a 100x100 pixel canvas and go further from there.
NOW LET'S GET TO THE DETAILS
Since I have the german version of Photoshop, the screencaptures of my settings are all in german, too; if you are unsure about anything, just ask and I'll look up the translation for you.
1. Fix basic contrast with Levels
Layer → New Adjustment Layer → Levels
Most images, especially screencaps, are rather flat and have only very little contrast to them, so the first thing I always do is make sure that the full tonal range in the histogram is used: I want the light parts of my image to be really bright and the dark parts really dark, but without losing any important details in the lights and shadows. You can do this with either Levels or Curves; personally, I prefer using Levels, just because it's a little more straightforward.
I pull the left and right sliders (the one in the middle I don't change) just to the point where the black mountain in the histogram begins to risen up, this will darken the shadows and lighten the highlights, just as I wanted to. Sometimes I go a little further than that*, but I always pay close attention to the light and dark areas in my image - especially when I have a close up from someone's face I'll want to make sure to not burn any details of the skin away by making the lights too light.
*In the first example, the first 1-pixel-large line of black pixels at the right side of the histogram represents the light of the candle and the window in the background - by dragging the slider further to the left those lights will burn out, but since I don't plan on using that area of the image for my icon I don't bother about that here.
1: levels 0 / 1,00 / 153 | 2: levels 0 / 1,00 / 215
2. Add saturation with Vibrance, Channel Mixer or Color Fill Layer
To bring out the colors that are somewhere hidden in the screencap I'm going to add a lot of saturation in the next step. There are various approaches of how you can do this; I'm going to describe the three techniques that I most frequently use, since I often combine several of them within an icon instead of using just one of them.
Vibrance
Layer → New Adjustment Layer → Vibrance
The quickest and easiest way to add saturation moderately is using a simple Vibrance layer - in earlier versions of Photoshop you can use Hue/Saturation instead. Vibrance tends to darken the image a little, whereas Hue/Saturation seems to make the colors lighter; Hue/Sat also has a stronger effect, so it needs to be used with a little more care.
I almost always add at least one Vibrance layer to my image, either now or later in the process, after I have cropped it down and am working on the coloring in my icon - sometimes I have even up to 3-4 different Vibrance layers set at various stages.
I usually start with about 50 for vibrance and go from there. Often I will pull it up even further, but too much of it can make some images look pixelated, so I push my luck only as far as I can without visibly decreasing the quality of my image. Sometimes I also move the saturation slider a bit, but a lot less than the vibrance one.
1: vibrance +70, saturation +22 | 2: vibrance +70, saturation +12
Channel Mixer
Layer → New Adjustment Layer → Channel Mixer
Another way of adding saturation to an image is using a Channel Mixer layer. It doesn't have the safety of a Vibrance layer (in terms of not oversaturating the image), but I like how it works with the colors and I often use it instead of the Vibrance variant while working on the larger sized image. It also doesn't have the tendency to make the image look as pixelated when used in greater amounts as the other two techniques do, which seems like another advantage to me.
Now, I won't try to explain how the Channel Mixer itself works (I'm not sure if I could, after all there's way too many scary-looking numbers for my taste); but basicly, what I do is increasing Red, Green and Blue in their individual color channels using equal numbers for them, so overall all the colors in my image become consistently stronger without introducing any color shifts. (...does that actually make any sense? I hope the picture above shows better what I mean than my poor English does...)
As I said, numbers scare me, so I don't bother about adjusting the settings every time, I just use the ones shown in the picture above, I even have saved them as a preset for me. Since the effect of this is very strong, I always lower the opacity of that layer afterwards, starting at around 50% and going further up or down from there, depending on my image. The blending mode of this layer is either set to saturation or left at normal, whichever seems to look better.
1: opacity 47%, blending mode: normal | 2: opacity 70%, blending mode: normal
Color Fill Layer
Layer → New Fill Layer → Solid Color
The third method of boosting up the saturation is actually one that I adopted from sallyna_smile, who has been generously sharing some of her amazing tips & tricks on Tumblr a while ago, and I hope she doesn't mind me sharing it here as well.
First I want to add that this technique ist the most 'dangerous' one of the three, it can end up making your picture look horribly pixelated very quickly, so you need to use it with great care, especially when working with low-quality images. When it does work, however, it can give a very nice touch to your image.
To use this method, I simple create a new Color Fill layer with a fully saturated color and set the blending mode of this layer to saturation. In a lot of cases, this will make certain areas of the image look pixelated, so I have to lower the opacity of this layer quite a lot, often as low as 10-20% or even less, until those ugly spots disappear (actually, they will never fully disappear, but they will at least become weaker). You can also try to smudge them out or otherwise try to get rid of them, e.g. by using a layer mask, but depending on what kind of image you are working on this can become a rather difficult task.
1: opacity 40% | 2: opacity 30% - you can still see pixelated areas in the background of both images
Combining these three techniques
As I said above, I often like to combine the different saturation methods - the way I go about this is usually that I create all of the layers and then switch them on and off and try different combinations of them to see what looks best on my image. Sometimes I will adjust them some more after the next step, too, as it will also influence the strength of saturation that I need - as I mentioned somewhere at the beginning I don't go very chronological with all of this.
In my examples here, I settled for the following:
Color Fill Layer set to saturation, opacity 7% + Channel Mixer, opacity 40%, set to normal
Color Fill Layer set to saturation, opacity 10% + Channel Mixer, opacity 70%, set to normal
3. Add some overall color with Curves
Layer → New Adjustment Layer → Curves
Now it's time to bring the colors out some more and lighten everything up a litte. Curves are a great tool for that, they make it very easy to infuse an image with lots of color.
I always work with the individual color channels first, mostly the red and green one, and often will not even touch the blue and the RGB channel. Sometimes I will adjust the RGB channel just a tiny bit, slightly dragging the curve up or down to lighten or darken my image a little or try a soft s-curve to add a bit of contrast in the midtones.
The main goal, for me at this stage is not to get all the colors 'right' here and now, but just to get some color into my image, so I have something to work with later.
Red channel
Unless my image is already very reddish from the start, I will pull the curve up here quite a bit, which will add some red color to the image. This will also 'open' up the midtones by making them brighter and just make everything look a lot more colored already (well, everything is just red yet, but we will change that in a minute).
Often, after having played with the other two color channels, I will come back here to adjust the red curve a little more, so it's always a bit of switching around between the channels until I'm satisfied with the result.
Green channel
In most cases, I will pull the curve up here as well, just a little to balance all the red out a bit. Sometimes I might also drag it down instead, in that case I will probably have to work with the blue channel afterwards, though. I try to take care not to make the image all-too green here, which can easily happen; pulling the green curve up will also brighten the image a lot which is not always what I want, so that's another reason why I am a lot more careful here.
Blue channel
This is the most difficult one for me; often I will just leave the curve completely untouched, or I try to make just very small adjustments here, trying to see if I can get something that I like. As opposed to the other two color channels I tend to pull the curve down here, though, to add a tiny bit of yellow instead of blue, unless my picture has already a good deal of blue in it that I want to enhance further.
RGB channel
When I feel that my image has become a lot too light by the adjustments in the color channels, I adjust the curve here a little to compensate for that, by slightly pulling it down - or up, if for some reason my image is still too dark.
1: see settings in the picture above | 2: see settings here
4. Highlight areas of the image with Dodge/Burn and Light blobs
Now it's time to decide where I want the focus in my image to be. The following techniques serve to lighten/darken and enhance certain parts of the image. I use the Dodge/Burn tool mostly while working on the large screencap and Light Blobs later to enhance smaller details in the final 100x100 pixel icon.
Dodge/Burn
Tools Palette → Dodge/Burn tool
(in CS5, the tool options will be open below the main menu bar after chosing the tool)
With the Dodge and Burn tools, you can work in the various tonal areas (ranges) of the image (highlights, midtones and shadows); I usually just stick to the midtones for Dodge and to either midtones or shadows for Burn, because I'm too lazy to change the options around all the time. Also, because I used that 'basic' levels layer in step 1, my highlights and dark shadows should be bright and dark enough already as they are.
Since I don't like to work directly on my precious image layer, I create a new layer filled with 50% grey ( Layer → New layer | Edit → Fill... → 50% grey ) and set it to overlay - this is the undestructive way of using Dodge and Burn.
Then I grab the Dodge tool, make sure that the tools options are set as seen in the picture above (the brush size can vary, of course), and place a few middlish-sized blobs on the image, just in the places that I want to lighten or enhance a bit. Afterwards, I adjust the opacity of this layer, depending on how strong I want the effect to be.
Then I do the same procedure with the Burn Tool, to darken some areas of the image if needed.
1: see dodge layer here | 2: see dodge layer here - I didn't use the burn tool in these examples
Now I am finally at the point where I'm satisfied with my result, at least for the moment, so I go and crop and resize my image down to 100x100 pixel - let's ignore the shabby cropping for now, it's certainly not one of my strengths:
Light blobs
I use light blobs in almost all of my icons and usually all throughout my coloring process, or just at the end when I'm more or less finished with my icon and just need a final touch to it. I know that most people will be already familiar with this technique (and also a lot better at using it than I am), but I'll still quickly explain how I do it anyway.
I create a new, empty layer on top and paint with white (or a very light color, but mostly I stick to pure white) and a small soft circle brush over the parts where there is already some light in the image or at certain spots that I want to emphasize, then I set the layer to soft light and adjust the opacity of it.
I like to use a new layer for almost every single light blob I do, just so I can adjust them individually, or even delete some of them again should I change my mind about them later, so I often end up with 3-5 or more of those light blob layers in my final icon, set to varying opacities.
Of course you can do the same with black (or dark colored) light blobs, to deepen the shadows in the image or make certain areas a bit darker. It's a great way to play with contrast in specific areas of your icon by combining white and black light blobs together.
+
=
*
+
+
=
*
*opacity of light blob layers adjusted individually
5. Coloring and other effects... Color Balance, Selective Color, Gradients, Textures...
Well, and the rest is up to coloring and all the fun stuff. I won't go all the way into detail here with this for now (I do remember this being about contrast in the first place... well, at some point it was supposed to be), so I'm just going to say that I use lots of Color Balance and Selective Coloring, as well as Textures and Gradients set to various blending modes. I always end up with lots and lots of layers here and I don't really have any recipe of how I do things, it just simply varies a lot.
For further contrast, I often just put a simple Brightness/Contrast layer at the end, playing either with the Brightness or Contrast value to find some kind of balance, or I might use a Soft Light layer or even some more Curves or Levels. Finishing off the contrast is actually the part where I'm often struggling, so I'm afraid I can't give a much better advice here.
+ Vibrance layer, Color Balance, Selective Color, Textures and Gradients set to different blending modes
+ Selective Color, Vibrance, Textures, several more Light Blobs and a Brightness/Contrast layer
And erm... with that I think I will leave you for now - just in case you are still here reading instead of having given up (p)ages ago already. ;)
---
So, this all has gotten a lot longer than I thought, and I'm not sure if I have talked too much about all the wrong things altogether (I have the feeling that I have), so please just let me know if I need to explain something else or if you have any more questions or if I should just shut up and take to other things rather than writing obvious stuff like that up. =)
Coming next: 3 more Ask the Maker tutorials, and then hopefully finally some icons soon :)
I apologize in advance for the lengthy novel I have written here, I don't even know if I answered this the way it was supposed to be answered, but well, I tried. :)
Also, I feel a little bad because it has taken me so long to get this up - I know I did say that it would be a while, but still, I'm sorry for making you wait for so long for this nonsense! :)
Ask the Maker 2.0 || Color & Contrast (a guide/walk-through/...thing)
So, the question was mainly about the use of contrast in my icons and what kind of techniques I'm partial towards. While I have not really settled yet for a particular way to create my icons, I am leaning more towards a certain kind of 'workflow' recently, using a small handful of tools that I feel comfortable with, so this is what I am going to describe here. Because I'm not good at making things short and up to the point this has basicly become some kind of walk-through of how I make my icons, minus the detailed coloring stuff and such; so yes, this is a lot longer than necessary. ;)
First, lets have a quick overview of what I do:
- Fix basic contrast with Levels
- Add saturation with Vibrance, Channel Mixer and/or Color Fill Layer
- Add some overall color with Curves
- Highlight certain areas with Dodge/Burn and/or Light Blobs
- Add further coloring and effects with Color Balance, Selective Color, Gradients, Textures,... etc.
The techniques themselves aren't anything particular new, I'm afraid, I'm sure most people will know about them all, but I'm still going to try and explain them for anyone who maybe hasn't come across some of them yet.
A LITTLE MORE RAMBLING BEFORE WE START
When it comes to icons, the one major goal I run for is to make them bright and really vibrant, because that is what I want my icons to be. I want the colors to look strong and rich and striking, possibly without being painfully overwhelming, but I like to push the boundaries a little, if only to see how far I can go. Contrast on the other hand is something I feel I'm still struggling with, I often end up with something that is either too much or not enough and I'm still not sure how to find the right balance for every picture.
Anyway, as a result, almost everything that I do adds to this 'colorfulness' in a way. I always push contrast and colors around in several steps, often switching between adding color, saturation (as you will see I use lots of saturation) and contrast, back again to more color, some more contrast and even more saturation and back and forth and all over again as I go along. (This is one of the reasons why I usually end up with hundreds of layers in my icons. Ahem.)
So, coming back to my 'workflow' from above: I don't always use those steps in the exact order I've written them down here, I often go back and adjust the layers some more as I go along, sometimes even deleting some of them again, or changing the order they are in in the layer palette just to see where it will take me. So it's not so much a chronological and straightforward process when I make my icons, but a little more muddled. I usually start working on a larger sized screencap, fiddle with the different techniques until I have something that looks like a promising base, then I shrink down my image into a 100x100 pixel canvas and go further from there.
NOW LET'S GET TO THE DETAILS
Since I have the german version of Photoshop, the screencaptures of my settings are all in german, too; if you are unsure about anything, just ask and I'll look up the translation for you.
1. Fix basic contrast with Levels
Layer → New Adjustment Layer → Levels
Most images, especially screencaps, are rather flat and have only very little contrast to them, so the first thing I always do is make sure that the full tonal range in the histogram is used: I want the light parts of my image to be really bright and the dark parts really dark, but without losing any important details in the lights and shadows. You can do this with either Levels or Curves; personally, I prefer using Levels, just because it's a little more straightforward.
I pull the left and right sliders (the one in the middle I don't change) just to the point where the black mountain in the histogram begins to risen up, this will darken the shadows and lighten the highlights, just as I wanted to. Sometimes I go a little further than that*, but I always pay close attention to the light and dark areas in my image - especially when I have a close up from someone's face I'll want to make sure to not burn any details of the skin away by making the lights too light.
*In the first example, the first 1-pixel-large line of black pixels at the right side of the histogram represents the light of the candle and the window in the background - by dragging the slider further to the left those lights will burn out, but since I don't plan on using that area of the image for my icon I don't bother about that here.
1: levels 0 / 1,00 / 153 | 2: levels 0 / 1,00 / 215
2. Add saturation with Vibrance, Channel Mixer or Color Fill Layer
To bring out the colors that are somewhere hidden in the screencap I'm going to add a lot of saturation in the next step. There are various approaches of how you can do this; I'm going to describe the three techniques that I most frequently use, since I often combine several of them within an icon instead of using just one of them.
Vibrance
Layer → New Adjustment Layer → Vibrance
The quickest and easiest way to add saturation moderately is using a simple Vibrance layer - in earlier versions of Photoshop you can use Hue/Saturation instead. Vibrance tends to darken the image a little, whereas Hue/Saturation seems to make the colors lighter; Hue/Sat also has a stronger effect, so it needs to be used with a little more care.
I almost always add at least one Vibrance layer to my image, either now or later in the process, after I have cropped it down and am working on the coloring in my icon - sometimes I have even up to 3-4 different Vibrance layers set at various stages.
I usually start with about 50 for vibrance and go from there. Often I will pull it up even further, but too much of it can make some images look pixelated, so I push my luck only as far as I can without visibly decreasing the quality of my image. Sometimes I also move the saturation slider a bit, but a lot less than the vibrance one.
1: vibrance +70, saturation +22 | 2: vibrance +70, saturation +12
Channel Mixer
Layer → New Adjustment Layer → Channel Mixer
Another way of adding saturation to an image is using a Channel Mixer layer. It doesn't have the safety of a Vibrance layer (in terms of not oversaturating the image), but I like how it works with the colors and I often use it instead of the Vibrance variant while working on the larger sized image. It also doesn't have the tendency to make the image look as pixelated when used in greater amounts as the other two techniques do, which seems like another advantage to me.
Now, I won't try to explain how the Channel Mixer itself works (I'm not sure if I could, after all there's way too many scary-looking numbers for my taste); but basicly, what I do is increasing Red, Green and Blue in their individual color channels using equal numbers for them, so overall all the colors in my image become consistently stronger without introducing any color shifts. (...does that actually make any sense? I hope the picture above shows better what I mean than my poor English does...)
As I said, numbers scare me, so I don't bother about adjusting the settings every time, I just use the ones shown in the picture above, I even have saved them as a preset for me. Since the effect of this is very strong, I always lower the opacity of that layer afterwards, starting at around 50% and going further up or down from there, depending on my image. The blending mode of this layer is either set to saturation or left at normal, whichever seems to look better.
1: opacity 47%, blending mode: normal | 2: opacity 70%, blending mode: normal
Color Fill Layer
Layer → New Fill Layer → Solid Color
The third method of boosting up the saturation is actually one that I adopted from sallyna_smile, who has been generously sharing some of her amazing tips & tricks on Tumblr a while ago, and I hope she doesn't mind me sharing it here as well.
First I want to add that this technique ist the most 'dangerous' one of the three, it can end up making your picture look horribly pixelated very quickly, so you need to use it with great care, especially when working with low-quality images. When it does work, however, it can give a very nice touch to your image.
To use this method, I simple create a new Color Fill layer with a fully saturated color and set the blending mode of this layer to saturation. In a lot of cases, this will make certain areas of the image look pixelated, so I have to lower the opacity of this layer quite a lot, often as low as 10-20% or even less, until those ugly spots disappear (actually, they will never fully disappear, but they will at least become weaker). You can also try to smudge them out or otherwise try to get rid of them, e.g. by using a layer mask, but depending on what kind of image you are working on this can become a rather difficult task.
1: opacity 40% | 2: opacity 30% - you can still see pixelated areas in the background of both images
Combining these three techniques
As I said above, I often like to combine the different saturation methods - the way I go about this is usually that I create all of the layers and then switch them on and off and try different combinations of them to see what looks best on my image. Sometimes I will adjust them some more after the next step, too, as it will also influence the strength of saturation that I need - as I mentioned somewhere at the beginning I don't go very chronological with all of this.
In my examples here, I settled for the following:
Color Fill Layer set to saturation, opacity 7% + Channel Mixer, opacity 40%, set to normal
Color Fill Layer set to saturation, opacity 10% + Channel Mixer, opacity 70%, set to normal
3. Add some overall color with Curves
Layer → New Adjustment Layer → Curves
Now it's time to bring the colors out some more and lighten everything up a litte. Curves are a great tool for that, they make it very easy to infuse an image with lots of color.
I always work with the individual color channels first, mostly the red and green one, and often will not even touch the blue and the RGB channel. Sometimes I will adjust the RGB channel just a tiny bit, slightly dragging the curve up or down to lighten or darken my image a little or try a soft s-curve to add a bit of contrast in the midtones.
The main goal, for me at this stage is not to get all the colors 'right' here and now, but just to get some color into my image, so I have something to work with later.
Red channel
Unless my image is already very reddish from the start, I will pull the curve up here quite a bit, which will add some red color to the image. This will also 'open' up the midtones by making them brighter and just make everything look a lot more colored already (well, everything is just red yet, but we will change that in a minute).
Often, after having played with the other two color channels, I will come back here to adjust the red curve a little more, so it's always a bit of switching around between the channels until I'm satisfied with the result.
Green channel
In most cases, I will pull the curve up here as well, just a little to balance all the red out a bit. Sometimes I might also drag it down instead, in that case I will probably have to work with the blue channel afterwards, though. I try to take care not to make the image all-too green here, which can easily happen; pulling the green curve up will also brighten the image a lot which is not always what I want, so that's another reason why I am a lot more careful here.
Blue channel
This is the most difficult one for me; often I will just leave the curve completely untouched, or I try to make just very small adjustments here, trying to see if I can get something that I like. As opposed to the other two color channels I tend to pull the curve down here, though, to add a tiny bit of yellow instead of blue, unless my picture has already a good deal of blue in it that I want to enhance further.
RGB channel
When I feel that my image has become a lot too light by the adjustments in the color channels, I adjust the curve here a little to compensate for that, by slightly pulling it down - or up, if for some reason my image is still too dark.
1: see settings in the picture above | 2: see settings here
4. Highlight areas of the image with Dodge/Burn and Light blobs
Now it's time to decide where I want the focus in my image to be. The following techniques serve to lighten/darken and enhance certain parts of the image. I use the Dodge/Burn tool mostly while working on the large screencap and Light Blobs later to enhance smaller details in the final 100x100 pixel icon.
Dodge/Burn
Tools Palette → Dodge/Burn tool
(in CS5, the tool options will be open below the main menu bar after chosing the tool)
With the Dodge and Burn tools, you can work in the various tonal areas (ranges) of the image (highlights, midtones and shadows); I usually just stick to the midtones for Dodge and to either midtones or shadows for Burn, because I'm too lazy to change the options around all the time. Also, because I used that 'basic' levels layer in step 1, my highlights and dark shadows should be bright and dark enough already as they are.
Since I don't like to work directly on my precious image layer, I create a new layer filled with 50% grey ( Layer → New layer | Edit → Fill... → 50% grey ) and set it to overlay - this is the undestructive way of using Dodge and Burn.
Then I grab the Dodge tool, make sure that the tools options are set as seen in the picture above (the brush size can vary, of course), and place a few middlish-sized blobs on the image, just in the places that I want to lighten or enhance a bit. Afterwards, I adjust the opacity of this layer, depending on how strong I want the effect to be.
Then I do the same procedure with the Burn Tool, to darken some areas of the image if needed.
1: see dodge layer here | 2: see dodge layer here - I didn't use the burn tool in these examples
Now I am finally at the point where I'm satisfied with my result, at least for the moment, so I go and crop and resize my image down to 100x100 pixel - let's ignore the shabby cropping for now, it's certainly not one of my strengths:
Light blobs
I use light blobs in almost all of my icons and usually all throughout my coloring process, or just at the end when I'm more or less finished with my icon and just need a final touch to it. I know that most people will be already familiar with this technique (and also a lot better at using it than I am), but I'll still quickly explain how I do it anyway.
I create a new, empty layer on top and paint with white (or a very light color, but mostly I stick to pure white) and a small soft circle brush over the parts where there is already some light in the image or at certain spots that I want to emphasize, then I set the layer to soft light and adjust the opacity of it.
I like to use a new layer for almost every single light blob I do, just so I can adjust them individually, or even delete some of them again should I change my mind about them later, so I often end up with 3-5 or more of those light blob layers in my final icon, set to varying opacities.
Of course you can do the same with black (or dark colored) light blobs, to deepen the shadows in the image or make certain areas a bit darker. It's a great way to play with contrast in specific areas of your icon by combining white and black light blobs together.
+
=
*
+
+
=
*
*opacity of light blob layers adjusted individually
5. Coloring and other effects... Color Balance, Selective Color, Gradients, Textures...
Well, and the rest is up to coloring and all the fun stuff. I won't go all the way into detail here with this for now (I do remember this being about contrast in the first place... well, at some point it was supposed to be), so I'm just going to say that I use lots of Color Balance and Selective Coloring, as well as Textures and Gradients set to various blending modes. I always end up with lots and lots of layers here and I don't really have any recipe of how I do things, it just simply varies a lot.
For further contrast, I often just put a simple Brightness/Contrast layer at the end, playing either with the Brightness or Contrast value to find some kind of balance, or I might use a Soft Light layer or even some more Curves or Levels. Finishing off the contrast is actually the part where I'm often struggling, so I'm afraid I can't give a much better advice here.
+ Vibrance layer, Color Balance, Selective Color, Textures and Gradients set to different blending modes
+ Selective Color, Vibrance, Textures, several more Light Blobs and a Brightness/Contrast layer
And erm... with that I think I will leave you for now - just in case you are still here reading instead of having given up (p)ages ago already. ;)
---
So, this all has gotten a lot longer than I thought, and I'm not sure if I have talked too much about all the wrong things altogether (I have the feeling that I have), so please just let me know if I need to explain something else or if you have any more questions or if I should just shut up and take to other things rather than writing obvious stuff like that up. =)
Coming next: 3 more Ask the Maker tutorials, and then hopefully finally some icons soon :)