Board of Command
Fri, 11-17-2006, 05:26 PM
I thought I lost this thing in the server rollback, but I saved a hard copy before I posted it that night because the server was acting really sluggish. I thought I deleted this file after I posted it, but apparently not. So luckily for you noobs, here is the guide I said I wrote.
This guide is not a sig tutorial. This guide will not teach you any Photoshop techniques. Instead, this is simply a guide and nothing more. Many beginners here have shown similar problems and I believe this will help to guide them in the right direction. If you want to learn Photoshop techniques, this guide is not for you. If you want to follow along and have a sig done by the end, this guide is not for you. But if you want to know what makes a sig good and where to start, please continue reading.
Progress:
Nov 14, 2006 - Sections 1 and 2
Nov 24, 2006 - Section 3 (I have school so bare with the time delays)
Nov 26, 2006 - Section 4
------------------------------------ Section 1: The Subject ------------------------------------
The first and foremost important thing is what you want to put on a sig. There are three general types of sigs: a) single subject, b) multiple subjects, and c) zero subjects.
a) Involves matching a subject (generally render) to a fitting background. This is the cookie-cutter sig that 99% of people use.
http://i82.photobucket.com/albums/j270/chacin_man/BoC/archer-1513246432.png
b) Involves putting two or more subjects together that mutually benefit each other as well and improves the overall presentation of the sig. These are rather rare because they are quite difficult to make properly.
(I'm not even gonna bother making one right now or go looking for one at the time being)
c) Generally a photo-manip, but it could be just an abstract background with some text on it. These are usually found at graphics forums and generally require some advanced skill to make. There's a big difference between making one and making a good one.
http://img.photobucket.com/albums/v700/blz123/vmadcopy3ww.png
http://img296.imageshack.us/img296/7264/cj19spadescloudssig2updatepc4.png
(not my sigs, just to get plagiarism out of the way...)
In your case, both are of type A (rather obvious), which is where most beginners naturally start at. For the rest I'll focus on type A sigs.
For type A sigs, the choice of subject is the most important aspect. Deblas chose a couple of very bad pictures to work with. The first one *might* be able to turn into a half-decent sig if you have some tremendous artistic ability and uber Photoshop skills, but the second one is a dud right from the start. What makes a good subject?
- Compatibility (can it fit within a box of said dimensions? if not, can you crop it without losing vital details?)
- Actions (is the subject "doing something?" does the action influence the ambiance?)
- Direction (is the subject "pointing somewhere?" see my current Black Lagoon sig for example)
All three aspects must be taken into account. Just because you like a picture does not mean you can make a sig with it. Believe me, I've had tons and tons of ideas for sigs but there are simply no pictures out there that would be suitable for sigs.
Here is a sig I attempted to make using an unsuitable subject, simply because I loved the movie and wanted to make a sig for it:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/c2165024.png
Here is a sig that contains a VERY suitable subject:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/UW_squall_edited.jpg
------------------------------------ Section 2: The Background ------------------------------------
The second part of type A sigs is making a suitable background for the subject. I cannot stress how it is unimportant to make a very "l33t" background for it to be good. A good background is one that fits the subject and the overall idea of the sig. Here is an example of such a background:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/48c3cfd0.png
To achieve the effect of this background is actually a lot more complex than it looks and takes a great deal of trial and error. However, the end product looks rather simple but also pleasing. Simple I understand, but why pleasing? Because it fits the subject which happens to be a silhouette. Since the subject is pure black, for the background to match it must mean that the background must also be dark but have high-brightness areas to reduce dullness. Since the subject is only one color, the background should be kept to only a few colors to reduce color-clashing.
I mentioned something about contrast in the previous paragraph. What I mean, in detail, is that the background should somewhat match the contrast of the subject itself. If the subject is really dull, the background should be dull too. If the subject is really high contrast, the background should be slightly less contrasted, but still rather high, to make the subject stand out. Mismatching of contrast is one of the most common problems in sigs.
I also mentioned colors earlier. What I mean, in detail, is that your background should contain relatively the same colors as the subject. If your subject is primarily red, don't make your background green. Instead, keep it to black or another similar color. If in doubt, start with black, add some some white, then color balance. Keep adjusting and you'll eventually end up with a decent result. Once you're comfortable working with different tools *cough*masks*cough*, you can get into the realm of multi-color balance, curves, and other adjustment layers. Invert is also *VERY* neat.
Next is blending. Well actually I'm not gonna get into blending because personally I'm not an expert at it. You can go read some other tutorials out there on different ways to blend subjects to backgrounds. I usually don't find a need to really "blend" the subject in. If the background fits well, it's not imperative that you do the "feather-3-px-press-delete-twice" technique, also known as "basic render blending." Don't ever use this technique unless you absolutely need to, and even if you do you must do some touch-ups afterwards.
Oh, and I also mentioned direction in the first half. If there's some direction in the subject, you should try to incorporate that into the background as well. This helps a lot with fitting the background and subject together. Take these for example:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/2a24f00d.png
http://i82.photobucket.com/albums/j270/chacin_man/BoC/t13kx.jpg
------------------------------------ Section 3: A Different Kind of Crop ------------------------------------
Preparation is key to success. Before doing anything, prepare your workspace. This does not include:
- Opening a new file in Photoshop and setting dimensions to 400x100 or whatever you "planned" on using.
- Pasting a picture into Photoshop and cropping out the chunk you want.
Never crop what you paste into Photoshop. Here is what you *should* be doing to select the part of the picture you want:
http://img293.imageshack.us/img293/4777/untitled1vt0.th.jpg (http://img293.imageshack.us/my.php?image=untitled1vt0.jpg)
1. Make a new layer above your picture or whatever.
2. Make a selection the size of your intended dimensions.
3. Fill selection with a very high contrast color relative to what's underneath. If your picture is predominantly red, draw a bright blue box. Etc etc.
4. Move the box around until you completely cover the desired area to be cut out.
5. Select the colored box layer and your picture and move both of then to one of the corners of your document. In this case, the top-left corner. To get it there with perfect precision, use your arrow keys to move 1 pixel at a time.
6. Image > Canvas Size > select top left corner like I did and then set the dimensions to the dimensions of the colored box. Press OK. Press PROCEED.
With this technique, you can get a better look at the big picture at hand before cutting out one section, as well as being able to dynamically adjust the overall dimensions of the sig.
But what's wrong with just cropping? If you crop a picture and later you want to edit the sig by moving the subject around or doing anything that requires the presence of some part of the original picture that isn't there anymore...well guess what? You're screwed because you cropped those parts off. That's why you shouldn't crop.
And to whoever that neg repped me for this, I bet you've never tried something like this. "Only the obvious," eh? The fact that you didn't bother to respond to my request for an elaboration on your comment is satisfactory enough for me.
This guide is not a sig tutorial. This guide will not teach you any Photoshop techniques. Instead, this is simply a guide and nothing more. Many beginners here have shown similar problems and I believe this will help to guide them in the right direction. If you want to learn Photoshop techniques, this guide is not for you. If you want to follow along and have a sig done by the end, this guide is not for you. But if you want to know what makes a sig good and where to start, please continue reading.
Progress:
Nov 14, 2006 - Sections 1 and 2
Nov 24, 2006 - Section 3 (I have school so bare with the time delays)
Nov 26, 2006 - Section 4
------------------------------------ Section 1: The Subject ------------------------------------
The first and foremost important thing is what you want to put on a sig. There are three general types of sigs: a) single subject, b) multiple subjects, and c) zero subjects.
a) Involves matching a subject (generally render) to a fitting background. This is the cookie-cutter sig that 99% of people use.
http://i82.photobucket.com/albums/j270/chacin_man/BoC/archer-1513246432.png
b) Involves putting two or more subjects together that mutually benefit each other as well and improves the overall presentation of the sig. These are rather rare because they are quite difficult to make properly.
(I'm not even gonna bother making one right now or go looking for one at the time being)
c) Generally a photo-manip, but it could be just an abstract background with some text on it. These are usually found at graphics forums and generally require some advanced skill to make. There's a big difference between making one and making a good one.
http://img.photobucket.com/albums/v700/blz123/vmadcopy3ww.png
http://img296.imageshack.us/img296/7264/cj19spadescloudssig2updatepc4.png
(not my sigs, just to get plagiarism out of the way...)
In your case, both are of type A (rather obvious), which is where most beginners naturally start at. For the rest I'll focus on type A sigs.
For type A sigs, the choice of subject is the most important aspect. Deblas chose a couple of very bad pictures to work with. The first one *might* be able to turn into a half-decent sig if you have some tremendous artistic ability and uber Photoshop skills, but the second one is a dud right from the start. What makes a good subject?
- Compatibility (can it fit within a box of said dimensions? if not, can you crop it without losing vital details?)
- Actions (is the subject "doing something?" does the action influence the ambiance?)
- Direction (is the subject "pointing somewhere?" see my current Black Lagoon sig for example)
All three aspects must be taken into account. Just because you like a picture does not mean you can make a sig with it. Believe me, I've had tons and tons of ideas for sigs but there are simply no pictures out there that would be suitable for sigs.
Here is a sig I attempted to make using an unsuitable subject, simply because I loved the movie and wanted to make a sig for it:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/c2165024.png
Here is a sig that contains a VERY suitable subject:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/UW_squall_edited.jpg
------------------------------------ Section 2: The Background ------------------------------------
The second part of type A sigs is making a suitable background for the subject. I cannot stress how it is unimportant to make a very "l33t" background for it to be good. A good background is one that fits the subject and the overall idea of the sig. Here is an example of such a background:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/48c3cfd0.png
To achieve the effect of this background is actually a lot more complex than it looks and takes a great deal of trial and error. However, the end product looks rather simple but also pleasing. Simple I understand, but why pleasing? Because it fits the subject which happens to be a silhouette. Since the subject is pure black, for the background to match it must mean that the background must also be dark but have high-brightness areas to reduce dullness. Since the subject is only one color, the background should be kept to only a few colors to reduce color-clashing.
I mentioned something about contrast in the previous paragraph. What I mean, in detail, is that the background should somewhat match the contrast of the subject itself. If the subject is really dull, the background should be dull too. If the subject is really high contrast, the background should be slightly less contrasted, but still rather high, to make the subject stand out. Mismatching of contrast is one of the most common problems in sigs.
I also mentioned colors earlier. What I mean, in detail, is that your background should contain relatively the same colors as the subject. If your subject is primarily red, don't make your background green. Instead, keep it to black or another similar color. If in doubt, start with black, add some some white, then color balance. Keep adjusting and you'll eventually end up with a decent result. Once you're comfortable working with different tools *cough*masks*cough*, you can get into the realm of multi-color balance, curves, and other adjustment layers. Invert is also *VERY* neat.
Next is blending. Well actually I'm not gonna get into blending because personally I'm not an expert at it. You can go read some other tutorials out there on different ways to blend subjects to backgrounds. I usually don't find a need to really "blend" the subject in. If the background fits well, it's not imperative that you do the "feather-3-px-press-delete-twice" technique, also known as "basic render blending." Don't ever use this technique unless you absolutely need to, and even if you do you must do some touch-ups afterwards.
Oh, and I also mentioned direction in the first half. If there's some direction in the subject, you should try to incorporate that into the background as well. This helps a lot with fitting the background and subject together. Take these for example:
http://i82.photobucket.com/albums/j270/chacin_man/BoC/2a24f00d.png
http://i82.photobucket.com/albums/j270/chacin_man/BoC/t13kx.jpg
------------------------------------ Section 3: A Different Kind of Crop ------------------------------------
Preparation is key to success. Before doing anything, prepare your workspace. This does not include:
- Opening a new file in Photoshop and setting dimensions to 400x100 or whatever you "planned" on using.
- Pasting a picture into Photoshop and cropping out the chunk you want.
Never crop what you paste into Photoshop. Here is what you *should* be doing to select the part of the picture you want:
http://img293.imageshack.us/img293/4777/untitled1vt0.th.jpg (http://img293.imageshack.us/my.php?image=untitled1vt0.jpg)
1. Make a new layer above your picture or whatever.
2. Make a selection the size of your intended dimensions.
3. Fill selection with a very high contrast color relative to what's underneath. If your picture is predominantly red, draw a bright blue box. Etc etc.
4. Move the box around until you completely cover the desired area to be cut out.
5. Select the colored box layer and your picture and move both of then to one of the corners of your document. In this case, the top-left corner. To get it there with perfect precision, use your arrow keys to move 1 pixel at a time.
6. Image > Canvas Size > select top left corner like I did and then set the dimensions to the dimensions of the colored box. Press OK. Press PROCEED.
With this technique, you can get a better look at the big picture at hand before cutting out one section, as well as being able to dynamically adjust the overall dimensions of the sig.
But what's wrong with just cropping? If you crop a picture and later you want to edit the sig by moving the subject around or doing anything that requires the presence of some part of the original picture that isn't there anymore...well guess what? You're screwed because you cropped those parts off. That's why you shouldn't crop.
And to whoever that neg repped me for this, I bet you've never tried something like this. "Only the obvious," eh? The fact that you didn't bother to respond to my request for an elaboration on your comment is satisfactory enough for me.