feat: Implement super-sampling for text rendering and adjust _RENDER_H from 16 to 8.

mainline.py

@@ -147,7 +147,8 @@ KATA = "ﾊﾐﾋｰｳｼﾅﾓﾆｻﾜﾂｵﾘｱﾎﾃﾏｹﾒｴｶｷﾑ
 _FONT_PATH = "/Users/genejohnson/Documents/CS Bishop Drawn/CSBishopDrawn-Italic.otf"
 _FONT_OBJ = None
 _FONT_SZ = 60
-_RENDER_H = 16         # terminal rows per rendered text line
+_RENDER_H = 8          # terminal rows per rendered text line
+_SSAA = 4              # super-sampling factor: render at _SSAA× then downsample
 
 # Non-Latin scripts → macOS system fonts
 _SCRIPT_FONTS = {
@@ -579,8 +580,11 @@ def _render_line(text, font=None):
     draw = ImageDraw.Draw(img)
     draw.text((-bbox[0] + pad, -bbox[1] + pad), text, fill=255, font=font)
     pix_h = _RENDER_H * 2
-    scale = pix_h / max(img_h, 1)
-    new_w = max(1, int(img_w * scale))
+    hi_h = pix_h * _SSAA
+    scale = hi_h / max(img_h, 1)
+    new_w_hi = max(1, int(img_w * scale))
+    img = img.resize((new_w_hi, hi_h), Image.Resampling.LANCZOS)
+    new_w = max(1, int(new_w_hi / _SSAA))
     img = img.resize((new_w, pix_h), Image.Resampling.LANCZOS)
     data = img.tobytes()
     thr = 80