It's not that I'm asking if it can be done, because I know it can, and I've done so before. What I'm asking is if there's a faster way to do it.
In particular I'm looking at games like Tsukihime or HoneyComing, where the stand CG is stored as either A) a single file with the actual image and a silhouette mask right next to it (Tsukihime) or B) two separate files, one of which is the actual image and one which is the silhouette mask (HoneyComing). B is faster simply by virtue of not having to splice the image into two pieces first, but it's still a rather involved process. This is my current workflow in Photoshop:
1) Duplicate the image layer and delete the background layer so that the image and the mask both have transparent backgrounds.
2) Delete the silhouette itself from the mask.
3) Copy the mask and paste it over top of the real image.
4) Disable the real image layer.
5) Select->Color Range and select the mask.
6) Re-enable the real image (and make sure it's highlighted in the layer selection box), and delete the area selected in step 5.
7) Delete the mask layer and save as PNG or PSD.
As that shows, it isn't very quick to do. Basic selection (by Color Range) of the solid color area fails because the color is often contained in the real image (and using the selection wand leaves ugly jagged edges behind in many cases because of edge anti-aliasing, not to mention the risk of selecting areas of the real image because the separating lines aren't defined well enough), and disabling the real image in step 4 is required because even if the mask is selected in step 5, it detects and will select areas of the real image just as if the images weren't on separate layers.
Since in cases like these there are a lot of images to do this to, is there a way of simplifying this, a program which can either fix all files within the directory selected or where the user specifies the real image and the mask image? While I doubt that approach would work with Tsukihime (as the silhouette is contained in the same image, just to the right - it'd still be necessary to splice them apart first), it would speed up the process substantially.
Constructing transparent images from masks?
- Qyot27
- Surreptitious fluffy bunny
- Joined: Fri Aug 30, 2002 12:08 pm
- Status: Creepin' between the bullfrogs
- Location: St. Pete, FL
- Contact:
- Qyot27
- Surreptitious fluffy bunny
- Joined: Fri Aug 30, 2002 12:08 pm
- Status: Creepin' between the bullfrogs
- Location: St. Pete, FL
- Contact:
- DriftRoot
- Joined: Mon Jun 09, 2003 7:18 pm
- Status: As important as any plug-in.
- Location: N.H.
Re: Constructing transparent images from masks?
What you're looking for does exist, it's Photoshop's Action scripting capability. I use this when I need to make masks for tons of frames. The general idea is that you work out a sequence of actions (which you have already done, it looks like) that can be applied across the board to every single file that needs editing. As long as there is no deviation from these steps, you basically can just record them in an action script, then play back that script on other files either in batch mode or individually. (Actually, there can be deviation, but that's a whole 'nother kettle of fish.)Qyot27 wrote: 1) Duplicate the image layer and delete the background layer so that the image and the mask both have transparent backgrounds.
2) Delete the silhouette itself from the mask.
3) Copy the mask and paste it over top of the real image.
4) Disable the real image layer.
5) Select->Color Range and select the mask.
6) Re-enable the real image (and make sure it's highlighted in the layer selection box), and delete the area selected in step 5.
7) Delete the mask layer and save as PNG or PSD.
It looks likes most, if not all, of your current process can be automated that way. I wrote a fairly detailed tutorial on how to action script that was specifically for masks, but it will get you started in the right direction. If you have questions or can't get the scripting to work, let me know and I'll be glad to see if I can help.