sprite op• 데이터 종류: rgba → rgba
• 호출: import gfx2d; gfx2d.sprite_transform(sprite, angle_deg=0.0, scale=1.0, interp='bilinear', out_shape=None)(또는 opsgfx2d.get("sprite_transform"))
`rgba` 스프라이트를 중심 둘레로 회전·확대축소합니다.
> 아래 상세 설명은 원문입니다 —— 요약과 제목은 번역되어 있습니다.
A positive `angle_deg` turns the sprite clockwise on screen, because
the row axis points down. `interp is "nearest", "bilinear"` or
`"bicubic"` (a cubic spline; it overshoots, and the overshoot is clipped
back into the premultiplied range).
The resampling happens in premultiplied space and the result is
converted back. Interpolating straight colour instead mixes in the colour of
fully transparent pixels — usually black, and invisible in the alpha channel
— which shows up as a dark fringe. Measured on a white disc whose
transparent region is black, composited over mid-grey: the straight-space
result differs from this one by up to 0.203 and by 0.048 on average
over the pixels it changes (rotations of 13, 37 and 45 degrees; the maximum
was 0.197, 0.203, 0.156).
Exact where it can be: `angle_deg=0, scale=1` returns the input bit for
bit, and a multiple of 90 degrees with `interp="nearest"` permutes the
alpha channel exactly (colour matches to 1.2e-16, the cost of the
premultiply round trip). Everything else costs interpolation: the measured
round trip of `+37 then -37` degrees, bilinear, is a mean alpha error
of 0.0150 with a maximum of 0.8125 at the one pixel where the edge
crosses a sample.
`out_shape is (height, width)`; the default is the bounding box of the
transformed sprite, rounded up.
• 샘플 데이터 카탈로그(DL URL / 라이선스) —— 2-D 는 skimage.data(BSD/public)+ 합성, 3-D 는 실데이터 소스(Stanford/PDS 등)의 DL URL.
• 연산자의 내력·참고문헌 —— 이 연산자 족의 바탕이 된 연구/기법의 출처.
• 알고리즘의 정전(저자·연도)과 용도는 위의 패밀리 사용 가이드에 적혀 있습니다.
• gfx2d_scene — py -3.11 examples/gfx2d_scene.py
rgba 를 입력으로 받는 것)premultiply · alpha_composite · sprite_blit · sprite_sheet_slice · nine_slice
sprite)sprite_synthesize · sprite_blit · sprite_sheet_slice · nine_slice
*Provenance: gfx2d.py — GFX2D 연산자 레지스트리. 이 op 노트는 tools/opdocs.py md 가 자동 생성합니다(직접 편집하지 마세요).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.