nodes/havvk-ComfyUI_AIIA
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

v1.9.153: Anchor Suppression & Nuclear Trigger - Fixed correctional feedback loop and added hard Pitch override

Browse Source
This commit is contained in:
Hawk Lee
2026-01-24 18:28:51 +08:00
parent 729f1261af
commit c8f5d75daa
2 changed files with 20 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
libs/Ditto/core/atomic_components/audio2motion.py
+19 -9
View File
@@ -231,6 +231,12 @@ class Audio2Motion:
if self.look_up_timer > 50:
# [v1.9.152] Tripelling restoration gravity to 60%
gravity_vec[0, 1:67] = 0.60
# [v1.9.153] Nuclear Trigger: 100% Hard Overwrite if stall > 4s
if self.look_up_timer > 100:
if self.look_up_timer % 10 == 0:
print(f"[Nuclear ACTIVE] Pitch stuck too long. Force-overwriting to 100% Baseline.")
gravity_vec[0, 1:67] = 1.0
# 6. Integration
next_pose = last_pose + noise
@@ -331,15 +337,19 @@ class Audio2Motion:
# Store for next batch
self.last_kp_frame = res_kp_seq[:, -1:]
# [v1.9.109] Dynamic Anchor:
# Slowly pull the brownian anchor towards the AI's final pose.
# This prevents the "Rubber Band" effect after long speech segments.
# Brownian position is the offset relative to s_kp_cond.
# res_kp_seq shape: [1, seq_frames, dim]
target_drift = (self.last_kp_frame - self.s_kp_cond).squeeze()
# Smoothly interpolate brownian_pos towards this target
# 0.1 factor per call block means it follows moderately fast during speech
self.brownian_pos = (self.brownian_pos * 0.9 + target_drift * 0.1).astype(np.float32)
# [v1.9.153] Anchor Suppression Logic:
# If we are in recovery, the "Stable Center" should NO LONGER follow the AI's drift.
# Instead, we should actively pull the anchor back towards the 0-point (The Original Photo).
if self.look_up_timer > 50:
# Force the anchor to decay back to neutral photo position
# This ensures recovery gravity pulls the head DOWN, not back to its CURRENT tilted position.
self.brownian_pos = (self.brownian_pos * 0.7).astype(np.float32)
if self.clip_idx % 5 == 0:
print(f"[Postural] Anchor Resetting... (Dist={np.abs(self.brownian_pos[0, 1:67]).mean():.4f})")
else:
# Normal speech persistence: anchor follows AI slowly to prevent rubber-banding
target_drift = (self.last_kp_frame - self.s_kp_cond).squeeze()
self.brownian_pos = (self.brownian_pos * 0.9 + target_drift * 0.1).astype(np.float32)
self.clip_idx += 1
pyproject.toml
+1 -1
View File
@@ -1,7 +1,7 @@
[project]
name = "aiia"
description = "The Ultimate AI Audio/Video toolkit for ComfyUI. Features an enhanced Ditto (with optimizations that outperform official demos and other SOTA talking head models in lip-sync accuracy and natural motion), EchoMimic V3 & FLOAT, VibeVoice & CosyVoice 3.0 (Zero-Shot Voice Cloning), Multi-Role Podcast Generation, and a powerful Media Browser."
version = "1.9.152"
version = "1.9.153"
license = {file = "LICENSE"}
readme = "README.md"
authors = [