前言
最近两周调研了一下字幕特效的底层程序逻辑，因为工作内容的原因，就分享几个自己找的链接具体细节就不分享了，CSDN也是我的个人笔记，只记录一些简单的内容用于后续自己方便查询，顺便帮助一下正在苦苦查阅资料入门的路人。总体来说先利用Aegisub学习ASS文件的字幕特效标签，然后借助脚本编写复杂的特效脚本，例如Python的pyonFX,以及c上位机python脚本的TCAX。Aegisub的lua脚本。如果是剪辑人员可以采用Pr+Ae以及剪映。本文章是程序化的字幕特效实现。
GITHUB pyonFX
Aegisub - 大神之路
关键字
字体添加背景
from pyonfx import *
import numpy as np
from PIL import Image
import math
#https://stackoverflow.com/questions/50297272/how-to-draw-text-with-image-in-background
io = Ass("in.ass")
meta, styles, lines = io.get_data()
img_path = "d:\Download\CgdkT.png"
fill_image = Image.open(img_path)
fill_image = fill_image.convert('RGBA')
fill_np = np.array(fill_image)
fill_width, fill_height = fill_image.size
def romaji(line, l, fill_np, fill_width, fill_height):
p_sh = Shape.rectangle()
for syl in Utils.all_non_empty(line.syls):
l.start_time = line.start_time
l.end_time = line.end_time
syl_width = syl.width
syl_height = syl.height
scale_x = fill_width / syl_width
scale_y = fill_height / syl_height
for pixel in Convert.text_to_pixels(syl):
x, y = math.floor(syl.left) + pixel.x, math.floor(syl.top) + pixel.y
print(x,y)
x2, y2 = x, y
alpha = ("\\alpha" + Convert.alpha_dec_to_ass(pixel.alpha) if pixel.alpha != 0 else "")
img_x = int(pixel.x * scale_x) % fill_width
img_y = int(pixel.y * scale_y) % fill_height
color = fill_np[img_y, img_x]
color_bgr = "&H{:02X}{:02X}{:02X}&".format(color[2], color[1], color[0])
l.text = f"{{\\p1\\an7\\1c{color_bgr}\\pos({x},{y}){alpha}}}{p_sh}"
io.write_line(l)
for line in lines:
if not line.comment:
romaji(line, line.copy(), fill_np, fill_width, fill_height)
io.save()
io.open_aegisub()
#ALKOON
