Closed GoogleCodeExporter closed 8 years ago
[deleted comment]
[deleted comment]
[deleted comment]
Notes :
Distort brush needs to be enabled both in boutons.c (for running from within
the
brush editing window) and moteur.c (for running with the shortcut)
Picture effects needs a new dialog (probably the same as the upper part of the
brush
editing) and should be enabled in init.c
Also, i noticed shortcuts for {NEXT|PREV}_USER_{FORE|BACK}_COLOR in moteur.c,
these
are not implemented and I don't know what they mean.
Original comment by pulkoma...@gmail.com
on 16 Dec 2008 at 5:58
These shortcuts are planned for a feature: next/previous buttons to browse not
all
colors, but a user-defined subset. The menu to choose them was planned to be
reachable from the alternate palette menu: there is a disabled button there
that
reads 'User's color series'
Original comment by yrizoud
on 16 Dec 2008 at 6:52
Original comment by pulkoma...@gmail.com
on 24 Dec 2008 at 3:48
Distort brushes : some ideas on implementation.
A) The brush is divided into a grid (squares about 16x16pixels each, maybe the
nearest to that we can get by successive divisions of the original brush size
by two)
B) Ask the user to draw the new shape of the brush as 4 edgelines (up down left
right). Each edgeline could be drawn like the 4-point splines in grafx2 (by
dragging
2 points) This would mean 8 moveable points on screen, can look a bit messy...
Maybe
use a different color for each edge ? or click on an edge to activate it ? It
would
be nice to have the horizontal edges as y=f(x) functions, and the vertical ones
as
x=f(y), and avoid (x,y)=f(t) functions
C) Split the new shape in quads (the same number as in the brushgrid), then
apply a
texturemapping to each quad using the small part of the brush that is linked to
it.
To do the splitting, we have to calculate some kind of weighted median between
the
edges. We have to calculate y=(y1+y2)/2 for each x (y1 and y2 are the values of
the
top- and bottom-edge functions for a given x) and the same thing along a
vertical
axis. Then start over with the 4 smaller parts we get.
Constraints of this approach :
-We have to find a compromise about the size of the squares at the end. It
seems
convenient to get them by splitting the brush in 2 parts again and again. Maybe
getting small quads (1x1 pixels) may get a perfect result, but will be heavy to
calculate. Maybe we can still use that on the final rendering, but not as a
preview.
-Using bigger quads will approximate the edgelines with straight lines. This
will
probably look ugly.
I think we already have a quad texture mapping used in the brush rotation
function.
Another thing that seems uneasy to handle : what about moving the brush corners
? it
may make the calculations a little more complex, but enables for more effects.
Bonus : it would be nice to have a way to :
-Force the edges to be straight lines
-Make the distortion symetrical (only edit up and left edges, the bottom and
right
will be the same)
Possible problems :
What happens if two lines are crossing inside the polygon ?
Original comment by pulkoma...@gmail.com
on 30 Dec 2008 at 10:55
[deleted comment]
[deleted comment]
[deleted comment]
WHAT WE STILL HAVE TO DO TO "COMPLETE" GRAFX 2.00:
- Distort brushes
- FX on the picture (flip, stretch, shrink, ...)
OTHER POSSIBLE IMPROVEMENTS:
- Add "Anti-aliasing" mode.
- Add more drawing modes.
- Add TIFf, TarGA and JPeG file supports (in our dreams ;)).
- Allow to load more types of truecolor pictures with Dithering or
Floyd/Steinberg color reduction.
- Add Pop-up menus.
Original comment by pulkoma...@gmail.com
on 4 Feb 2009 at 6:30
Isn't JPEG like totally useless for pixel art?
Original comment by ilija.melentijevic
on 4 Feb 2009 at 6:52
Yes it is :)
If we ever do it, it will be load-only.
May be useful in combination with the "animation" feature, or if we add
truecolor
modes.
Definitely not a priority for 2.0 release, but it may come later...
Original comment by pulkoma...@gmail.com
on 4 Feb 2009 at 7:00
I pity the fool who would make a truecolor program out of grafx2.
Original comment by yrizoud
on 4 Feb 2009 at 7:23
Working on brush distort operation.
I'm going to implement "planar" (2D) distortion.
If somebody can find the math for perspective distortion, I'm interested...
Original comment by yrizoud
on 22 Apr 2009 at 7:35
Brush distort Implemented in r763
Original comment by yrizoud
on 4 May 2009 at 7:43
Original issue reported on code.google.com by
pulkoma...@gmail.com
on 13 Oct 2008 at 6:21