Alternative Workflow Using Smart Objects and Partial Transparency
Applies at least to Photoshop CS3 and CS4
Here's an idea I nicked from my working methods with printed media. The idea doing the basic design of the sig in higher resolution than the final render. Why? It gives a whole lot of more room to maneuver in terms of tweaking the sig than making it in the final resolution. When you make a sig in say 400x150 pixels dimension how the this one turned out, you lose a lot of quality when you make adjustments to for example sizes or angles of the elements. Doing it this way allows you to retain far more detail and quality, because you're always scaling down when going to final size, never up. So you can for example retain high quality while saving the sig in multiple sizes.
Here's the end result. Obviously not Asuki safe since I got a bit carried away with transparencies.
Create larger image than the final size of the sig with the same proportions. In this case as I was making an image with dimensions of 400x150px, so I initially created an image with a size of 1600x600px. I recommend sizes with the side of each side multiplied by 2, 4, 8 or so on. This is purely due to how Photoshop is capable of scaling the preview image. In this case image zoomed out 25% gives me a fairly good idea how the sig is going to look in it's final size. If I would have used side dimensions 3 time larger than the final I would have had to use the zoom out of 33.3% with which the rendering quality of the preview frankly sucks.
Since I wanted to play around with transparencies and the method I use for those makes scaling them a bit tricky, I wanted to have small safety areas. So I placed guides on the edges of the image and then used Image -> Canvas size to re-size the sig to 1700x700px giving me extra 50px per side to play around with.
So you end up with something like this:
Beyond that you design the sig pretty much as usual just in larger format. I'm skipping the whole design process since this tutorial isn't about that.
This is what I came up with:
Now we get to partial transparency
There's actually very easy way to make transparencies to images with multiple layers using the "Knockout" and "Fill Opacity" options in Blending Options. (not in the way they were originally designed to be used but who cares) You could do this by using clipping masks or layer masks but I think this is way easier.
First you create a layer of the areas you want to make transparent and make it the topmost layer. Here represented by the bright red areas. The reason why you'd want to use that little extra space around the sig, the safety areas, is because when you resize the sig and you have a this transparency layer photoshop antialiases the outer edges as well and you end up with a little stroke around the edges of the image. Using safety areas moves this problem out of the visible area of the final sig.
Then you just set the "Knockout" to deep and "Fill Opacity" to 0%. This works just fine as long as you don't have a locked Background layer, which you can't have in the first place if you intend on using transparency in the sig.
Now you're done with the design.
And you should end up with something like this:
Create the smart object.
Select all the layers and then select from the menu "Layer->Smart Objects->Convert to Smart Object". Once you're done with that you can start editing the smart object. You can do so Smart Object by double clicking on the Smart Object icon in the layer panel.
In my case the Smart Object looks like this:
You can think of Smart Objects being PSDs within PSD since they work in fact in the exact same way. You'll find all the original layers now here. Now you can go back to the actual image by closing the Smart Object the changes are applied if you save the changes upon closing.
Now you can resize the image to the final size. I first used "Canvas Size" to get rid of the safety zones, so I resized the canvas to 1600x600px and then used the "Image Size" in the same "Image" menu to resize the image to 400x150px.
Now if you open the Smart Object you'll find that it's still in the original resolution.
Naturally you can use multiple smart objects as well for even more control.
Now you can use the "Save for Web & Devices" just as you would with any sig.
Here's a little demo.
Here's what the image would look like if you made it in 400x150px to begin with and then decided to make it 600x225px instead.
And here's the same thing done with Smart object method.
Much more like it. Well in this case you're not in fact resizing it from 400x150px to 600x225px but from 1600x600px to 600x225px so the result is pretty much given.
This is not all great, joyful and blissful though. It does add a little bit of time to the process but more importantly makes the PSDs HUGE.
Here's the 30Meg one I ended up with.
Edit: Almost forgot. If you add a stroke border you'll need to add that as a separate layer after
resizing the sig to it's final size, otherwise Photoshop will
antialias it and it will look crappy. And if you want to make it partially transparent as well you need to create the transparency layer as a separate smart object. When creating this you need to first convert it to smart object and then change the blending options and once again place it on top of the layers you want to add transparency to.