adjustments

app.py

@@ -642,7 +642,7 @@ with gr.Blocks(title="VEVO DEMO") as demo:
         timbre_button.click(vevo_timbre, inputs=[timbre_content, timbre_reference], outputs=timbre_output)
     
     with gr.Tab("Vevo-Voice"):
-        gr.Markdown("### Vevo-Voice: Transfer both style and timbre with separate references")
+        gr.Markdown("### Vevo-Voice: Transfers both style and timbre with separate references")
         with gr.Row():
             with gr.Column():
                 voice_content = gr.Audio(label="Content Audio", type="numpy")
@@ -665,17 +665,14 @@ with gr.Blocks(title="VEVO DEMO") as demo:
         style_button.click(vevo_style, inputs=[style_content, style_reference], outputs=style_output)
     
     with gr.Tab("Vevo-TTS"):
-        gr.Markdown("### Vevo-TTS: Text-to-speech with controllable style and timbre")
+        gr.Markdown("### Vevo-TTS: Text-to-speech with separate style and timbre references")
         with gr.Row():
             with gr.Column():
                 tts_text = gr.Textbox(label="Input Text", placeholder="Enter text to synthesize...", lines=3)
                 tts_src_language = gr.Dropdown(["en", "zh", "de", "fr", "ja", "ko"], label="Text Language", value="en")
                 tts_reference = gr.Audio(label="Style Reference", type="numpy")
+                tts_timbre_reference = gr.Audio(label="Timbre Reference", type="numpy")
                 tts_ref_language = gr.Dropdown(["en", "zh", "de", "fr", "ja", "ko"], label="Reference Audio Language", value="en")
-                
-                with gr.Accordion("Advanced Options", open=False):
-                    tts_timbre_reference = gr.Audio(label="Timbre Reference (Optional)", type="numpy")
-                
                 tts_button = gr.Button("Generate")
             with gr.Column():
                 tts_output = gr.Audio(label="Result")
@@ -691,8 +688,8 @@ with gr.Blocks(title="VEVO DEMO") as demo:
     VEVO is a versatile voice synthesis and conversion model that offers four main functionalities:
     1. **Vevo-Style**: Maintains timbre but transfers style (accent, emotion, etc.)
     2. **Vevo-Timbre**: Maintains style but transfers timbre
-    3. **Vevo-Voice**: Transfers both style and timbre simultaneously
-    4. **Vevo-TTS**: Text-to-speech with controllable style and timbre
+    3. **Vevo-Voice**: Transfers both style and timbre with separate references
+    4. **Vevo-TTS**: Text-to-speech with separate style and timbre references
     
     For more information, visit the [Amphion project](https://github.com/open-mmlab/Amphion)
     """)