pokemon/gen1recomp
1
0
Fork 0
mirror of https://github.com/bryanthaboi/gen1recomp.git synced 2026-08-15 07:41:21 +02:00
Code Issues Packages Projects Releases Wiki Activity

CLOSES #223, CLOSES #233, CLOSES #236, CLOSES #240, CLOSES #241, CLOSES #249, CLOSES #252, CLOSES #255, CLOSES #257, CLOSES #258, CLOSES #263, CLOSES #265, CLOSES #274, CLOSES #275, CLOSES #276, CLOSES #279, CLOSES #280, CLOSES #282, CLOSES #283, CLOSES #287, CLOSES #291, CLOSES #292, CLOSES #293, CLOSES #301, CLOSES #304, CLOSES #315, CLOSES #316, CLOSES #317, CLOSES #321, CLOSES #322, CLOSES #330

Browse Source
This commit is contained in:
bryanthaboi
2026-07-28 10:29:05 -04:00
parent 3b1032cc63
commit f9f38d161f
103 changed files with 11157 additions and 838 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/render/SpriteRenderer.lua
+26 -5
View File
@@ -18,10 +18,14 @@ local function getImage(path)
return imageCache[path]
end
-- RED++ overworld sprite OBJ-palette recolor (color/sprites.asm
-- ColorOverworldSprite), baked into an ImageData like BattleState's mon-pic
-- palette bake (src/battle/BattleState.lua getImage): CPU-remap the 4 DMG
-- shades to the resolved OBP colors, cached per (image path, group).
-- Overworld sprite OBJ-palette recolor, baked into an ImageData like
-- BattleState's mon-pic palette bake (src/battle/BattleState.lua getImage):
-- CPU-remap the 4 DMG shades to the resolved OBP colors, cached per
-- (image path, group). Every colour mode goes through it now (#301): RED++
-- resolves real per-sprite colours (color/sprites.asm ColorOverworldSprite),
-- OG RED the one boot-ROM object palette, and everything else the plain
-- rOBP0 = $D0 shade lift (PaletteFX.dmgObj) that leaves the sprite in DMG
-- shades for the zone shader to colour.
--
-- Sprite sheets carry no real alpha (every pixel, including the
-- background, is opaque -- confirmed by sampling the extracted PNGs): the
@@ -110,7 +114,13 @@ function SpriteRenderer:resolveImage()
-- collide in obpCache) -- see issue #155
return getObpImage(self.def.image, PaletteFX.ogObj())
end
return self.image
-- Every other mode (SGB and the mono/inverted novelties) leaves the sprite
-- in DMG shades so the zone shader colors it out of the map's own palette,
-- but still bakes rOBP0 = $D0 in and keys OBJ color 0 to alpha -- the two
-- things a raw sheet blit cannot express (#301, #150). The sheets carry no
-- real alpha (see getObpImage), so returning self.image here would put an
-- opaque white box behind every character a pipeline textures.
return getObpImage(self.def.image, PaletteFX.dmgObj())
end
-- facing: down/up/left/right; walkPhase: 0 stand, 1 walk; flip: alternate
@@ -152,6 +162,17 @@ function SpriteRenderer:draw(px, py, camX, camY, facing, walkPhase, stepFlip)
-- top.
image = getObpImage(self.def.image, PaletteFX.ogObj())
redraw = true
else
-- SGB and the mono/inverted modes (and OG RED's tilt upright pass, which
-- has no post-zone replay to restore a bake): the sprite stays in DMG
-- shades -- rOBP0 = $D0 baked in, OBJ color 0 keyed to alpha -- and the
-- whole-canvas zone shader colors it with the map's palette. That is the
-- only thing the Super Game Boy can do to an OBJ, since pokered never
-- sends the OBJ_TRN packet that would give sprites palettes of their own
-- (data/sgb/sgb_packets.asm defines ATTR_BLK / PAL_SET / PAL_TRN /
-- MLT_REQ / CHR_TRN / PCT_TRN and nothing else). No redraw is queued:
-- being colorized by the zone IS the point (#301).
image = getObpImage(self.def.image, PaletteFX.dmgObj())
end
-- single-frame sprites (item balls, fossils...) have one fixed pose;
-- still 3-frame sprites turn to face (the nurse at her machine,