From 17e7df43d19bde49efa46a32b89f5153b9cb0ded Mon Sep 17 00:00:00 2001
From: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com>
Date: Fri, 6 Feb 2026 21:02:11 -0800
Subject: [PATCH] Pad ace step 1.5 ref audio if not long enough. (#12341)

---
 comfy/model_base.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/comfy/model_base.py b/comfy/model_base.py
index 3aa345254..858789b30 100644
--- a/comfy/model_base.py
+++ b/comfy/model_base.py
@@ -1577,6 +1577,10 @@ class ACEStep15(BaseModel):
             else:
                 out['is_covers'] = comfy.conds.CONDConstant(False)
 
+        if refer_audio.shape[2] < noise.shape[2]:
+            pad = comfy.ldm.ace.ace_step15.get_silence_latent(noise.shape[2], device)
+            refer_audio = torch.cat([refer_audio.to(pad), pad[:, :, refer_audio.shape[2]:]], dim=2)
+
         out['refer_audio'] = comfy.conds.CONDRegular(refer_audio)
         return out