Use the offline mask blur ratio in the realtime blending path

[cuhawk]

Problem

get_image() (offline) and get_image_prepare_material() (realtime) build the blending mask the same way: face_seg over the expanded crop, crop to the face box, paste onto a black canvas, drop everything above upper_boundary_ratio, then Gaussian blur to soften the edge.

The blur is the only step where the two disagree. get_image() sizes the kernel at 0.05 of the crop width (blending.py:81) and get_image_prepare_material() at 0.1 (blending.py:134), so the realtime mask feathers over roughly twice the distance. Where the mask edge runs close to the mouth, the wider ramp mixes more of the original frame back over the generated pixels than the offline path would from the same inputs.

What the file history shows, because it cuts both ways

• Before v1.5 both functions used 0.1 and agreed.
• v1.5 (db20431) rewrote get_image only: expand 1.2 → 1.5, new mode and fp arguments, blur ratio 0.1 → 0.05. It left get_image_prepare_material on the older code. This commit created the divergence.
• 39ccf69 ("feat: real-time infer", feat: real-time infer #286) then ported that rewrite into get_image_prepare_material: expand 1.2 → 1.5, added fp and mode, passed both through to face_seg. It edited the function line by line and left the 0.1 in place.

So this is not a value that only one commit overlooked; a later commit went through the same function and kept it. Nothing in the code or the commit messages argues for it either way. The reason it still reads as a leftover is that 39ccf69 carried over every other difference the v1.5 rewrite introduced here, and the blur ratio is the only one it did not.

If the wider feather is deliberate for the realtime path, please close this — a comment recording why would be worth more than the change.

Measured

On a real prepared avatar frame (704x896, face box [250,123,431,365], crop 362x362), running both versions of the function on the same frame with the same face_seg output:

	kernel	partial-blend pixels	centre-column ramp
current (0.1)	37 px	16,428	64 px
this PR (0.05)	19 px	9,243	36 px

1.78x more partially-blended pixels on the current path; 12.5% of mask pixels change, max delta 67. The ratio is under 2x because the Gaussian tails round to 0/255.

Testing note

scripts/realtime_inference.py writes every computed mask to {avatar_path}/mask/*.png and pickles the crop boxes, and later runs read those back instead of recomputing them. get_image_prepare_material() is called only from prepare_material(), so this change is inert for an avatar that has already been prepared. Reproducing any difference needs a fresh avatar_id, or the existing avatar directory removed first.