feat: pass instruct in voice_cloning mode for accent/style influence

If instruct is set alongside ref_audio, it is now forwarded to
model.generate() — allowing accent/style transfer on top of the
cloned voice identity. Model may or may not honour both simultaneously.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
2026-04-05 20:07:18 +02:00
parent d2cb5c4249
commit d4bf7c825e
+9 -6
View File
@@ -142,16 +142,17 @@ class OmniVoiceGenerate:
"instruct": ("STRING", {
"default": "",
"tooltip": (
"Voice description for voice_design mode. Combine attributes freely.\n"
"Voice style description. Required for voice_design mode; optional in voice_cloning\n"
"mode to attempt accent/style transfer on top of the cloned voice.\n"
"Connect the OmniVoice Voice Design node for structured input.\n"
"\n"
"GENDER: male, female\n"
"AGE: child, teenager, young adult, middle-aged, elderly\n"
"PITCH: very low, low, moderate, high, very high\n"
"STYLE: whisper\n"
"PITCH: very low pitch, low pitch, moderate pitch, high pitch, very high pitch, whisper\n"
"\n"
"ENGLISH ACCENTS (text must be English):\n"
" american, british, australian, canadian,\n"
" indian, chinese, korean, japanese, portuguese, russian\n"
"ACCENT EXAMPLES:\n"
" british accent, american southern accent, indian accent,\n"
" australian accent, french accent, japanese accent ...\n"
"\n"
"EXAMPLE: female, high pitch, british accent"
),
@@ -213,6 +214,8 @@ class OmniVoiceGenerate:
kwargs["ref_audio"] = tmp_path
if ref_text:
kwargs["ref_text"] = ref_text
if instruct:
kwargs["instruct"] = instruct
audio_tensors = model.generate(**kwargs)
finally:
try: