PART 1- BEFORE STARTING
1. I size my document according to how big I want the sprites. While 72 dots per inch (DPI) is good, and while Color Splash uses an unusual amount of DPI, I do 300 (the highest possible amount). That way, if I need to use less DPI for some reason, I fear no shenanigans.
2. I find an idle dummy for Paper Mario, for the sake of size comparison. That way, I always know how big the character is supposed to be. Of course, I use my own custom PM.
As for how I figured out the size for my Mario (and all future sprites in general), I used the CS sprite as a beginning point. Then, using the resize option, I increased t size by a certain amount (can't remember how much exactly), so my sprites would appear good/non-jaggy in close-ups on a 1920x1080 screen. Also, so my sprites would always be bigger than the Color Splash sprites in terms of resolution.
3. With that done, I increase the resolution of the document's canvas (not the image itself- the canvas) to make room for a number of sprites and disassembly.
PART 2A- NEW SPRITE
1. If it's a main series character with no Paper Mario sprites, I find pics on the internet for various incarnations of the character. If's it a completely new character, meanwhile, after thinking about the ideas for a bit, if I can't think up a good rough, I find pics of items and whatever general thing the character may be based on for guidance.
2. With the resource pics nearby, I keep drawing and roughs on layers labelled "rough #". I do it near the Mario dummy to ensure size consistency. Every time a rough doesn’t meet my standards, either I copy the layer and edit what I don’t like OR start from scratch in case it’s that far below acceptable.
3. Finally, I create the rough I want.
PART 2B- REMASTER SPRITE (EXISTING SOURCES)
1. I go to Spriter's resource to locate the original sprite.
2. I then copy and paste the sprite into a PhotoShop layer.
3. I resize it to match the new proportions. Off the top of my head, Paper Mario 64 is increased to 580%, while TTYD is 170%.
4. Finally, to give the remastered version more life, I saturate the source sprite layer by 15%.
PART 2C- REMASTER SPRITE (NO AVAILABLE SPRITESHEETS)
1. In case no sprites are available (such as with Zip Toad), I resort to more extreme measures. In this case, I find videos containing the character in question. I then pause the video when the character exhibits unique poses (near Mario if possible).
2. Using the Print Screen (prt sc) button of my computer, I copy and paste that part of the video into the PhotoShop document. Then, I either use the Marquee or Lasso tool to cut and paste the part of the video picture with the character and Mario into a new, smaller layer.
3. I resize this snippet until the Mario in the video pic and the dummy Mario are the same size.
PART 3- DRAWING AND ARRANGING THE CHARACTER
1. I create a folder/group labelled "parts”.
2. Using the various pictures as reference, I create and name part layers in the group based on what poses that character might make (“body”, “foot”, “arm”, “bent arm”, “hand”, “eyes”, “pupils”, “hit pupils”, “mouth”, “open mouth”, “happy mouth”, “happy mouth open”, and other exclusive parts.). This is usually in service of the character in their “idle 1” pose.
3. I trace over either the original sprite or the new roughs. If the outline is off, I use CTRL + Z (Back One Step) to quickly remove it. Alternatively, I use the eraser tool. Both the brush and eraser tool are equally important.
I use a Wacom Intuos drawing tablet with pressure sensitivity for my HD sprites. The brush size is usually 6, with Spacing of 1%. If the character is bigger, I increase the brush size to compensate. If pressure sensitivity is not available for some reason, using the Magic Wand tool and nudging it up and left, followed by deletion of what it encompasses, can create the illusion of dynamic lines. Be sure to use the eraser, too, to smooth out the lines.
4. SHORTCUT: When doing hands, draw the hand closed first. Then, copy the layer and draw it “open”, erasing and using the original hand as guidance. Finally, copy those two layers and redraw as needed for the other side of the hands. Ditto for angry and sad pupils being copied from normal pupils.
5. Before coloring anything in, copy and sort-of flip (-80%, rather than -100) those layers that have reversed parts (eyes, eyebrows, wings). This is so the edges won’t pixelate as much when rotated or resized later.
6. While I have the part outline on one layer, I have the actual color of the part on a layer directly beneath it, called “CLR”.
7. The Magic Wand Tool (in both Adobe Photoshop and PaintShop Pro) can help with the coloring process, whether by deleting excess color outside the part or keeping the color inside a selected area. Just pay attention make sure you’re coloring on the CLR layer beneath the part layer. Using the Magic Wand Tool alongside the Expand option (Select > Modify) also helps ensure no color seeps out of a part. The Eyedropper tool also helps with quickly picking the original extra-saturated sprite for it's color.
8. If there are special palette swaps of a character I’m making, I either copy the group a certain number of times and paint in Mask Layers (makes that layer only appear inside the layer it's atop of) of the color layers as needed, OR I just copy parts and CLR layers as needed. Either way, when done, I Merge Layers as needed. If I recolored in another group, I move those parts back into the main group, then delete the copied group once it’s unique parts are moved out.
9. I reorder the layers as one would see the character. Usually, it’s the arm and hands up front, then the face and all it’s features, then the nearest leg and foot, then body, then further leg and foot, and finally the furthest arm and hands. I also put recolored parts atop their original part and keep the order consistent through the group (NORMAL-GLOOMY-HYPER-DRY-GOLD foot, NORMAL-GLOOMY-HYPER-DRY-GOLD body, and so on and so forth).
PART 4- POSING THE CHARACTER
1. Keep copying the parts group for use in other poses (“idle/talk”, “walk”, emotion poses, action poses, and whatever else is absolutely necessary). If I believe I might accidentally use the original parts group instead of a copy (it’s happened), then I create a back-up copy of the group and place it at the very bottom of the layer order. To make things easier to manage, I write down the poses I need in a text layer, and do the poses one at a time.
2. In a given poses group, I delete what parts aren’t necessary for that pose (so the sad pupils go sayonara in a group for “happy”), while keeping parts for other versions of the pose (open mouths) and palette swap parts hidden. Even for idle, for easier management.
3. For emotions, if it’s not the idle/talk poses, I rotate parts for the first frame of that pose. Regardless, I then duplicate that pose's group at least 2 times- one for the second frame of the pose, and the other for the talking version.
4. I nudge the copied groups up or down with the arrow keys, the number and direction depending on the polarity of the mood. If it’s a neutral emotion, the second frame goes down while the talking frame goes up, both an equal amount. Positive goes down slightly less while going up slightly more. Negative is the opposite. I then go into each copied layer and nudge the legs and/or feet in the opposite direction an equal number of times.
5. Some smaller rotations are applied to arms and the head to give the character more life. The eyes can also be nudged slightly up or down where applicable to give the feeling the character is focused. The Group Visibility toggle is my friend for comparison.
6. If there’s multiple versions of a character, before Merging the Group (by rightclicking the visible group and selecting the option), I copy the groups a given number of times equal to the number of versions. Then, I go to each group and hide/show/even delete layers to switch the version of the character around. This is necessary, as just showing a layer atop another layer makes it look thicker than it should.
7. For special model mods like normal Goomba to Spiked Goomba, I keep the Spiked Helmet hidden on the normal version of the character. Then, before Merging the Group, I copy the three poses to “SPIKE” versions of them. Finally, I Show the Spiked Helmet Layers in the duplicates.
8. Once I have checked every pose group to ensure everything is correct, I Merge the Group to flatten it, saving on memory in the process. I then move onto the next pose.
9. This can be done at any point during Part 4, but I also duplicate the Parts group into a “disassembly” group. Here, I pull each part of the character apart, yet keep them as close as possible for lower file sizes. Again, multiple versions of a character lead to copying and showing/deleting layers as needed. Once everything is done, each group is Merged separately.
PART 5- ARRANGING AND SAVING THE SHEET
1. With all that done, there’s only the matter of spacing out the sprites and disassembly. Not only do I do this- I also include the Mario dummy for assistance in size comparison.
2. Once everything is all spaced out enough, I finally save the file as a non-Interlaced PNG with no background.
And that’s how I do it.
1. If there's a specific character of a certain species I want to create (Toad, Goomba, Koopa), I make the generic version of them first. That way, I can just reopen the original source file for future use and Save As a new file, rather than recreate the assets all over again.