fx

app.py

@@ -464,6 +464,10 @@ def other_tts(text='Hallov worlds Far over the',
     return tmp_file
 
 
+def update_selected_voice(voice_filename):
+    return 'wav/' + voice_filename + '.wav'
+
+
 description = (
     "Estimate **age**, **gender**, and **expression** "
     "of the speaker contained in an audio file or microphone recording.  \n"
@@ -473,11 +477,7 @@ description = (
     f"whereas [{expression_model_name}]"
     f"(https://huggingface.co/{expression_model_name}) "
     "recognises the expression dimensions arousal, dominance, and valence. "
-    )
-
-def update_selected_voice(voice_filename, current_text):
-    _full = 'wav/' + voice_filename + '.wav'
-    return _full, gr.Textbox.update(label=f"Text for TTS: Vox=`{voice_filename}`", value=current_text)
+)
 
 css_buttons = """
             .cool-button {
@@ -504,30 +504,49 @@ css_buttons = """
 
 with gr.Blocks(theme='huggingface', css=css_buttons) as demo:
     with gr.Tab(label="other TTS"):
-        initial_voice = 'en_US_m-ailabs_mary_ann.wav'
-        selected_voice = gr.State(value=initial_voice)
+
+        selected_voice = gr.State(value='wav/en_US_m-ailabs_mary_ann.wav')
+
         with gr.Row():
-            text_input = gr.Textbox(label=f"Text for TTS: Vox=`{initial_voice}`",
-                                    placeholder="Enter Text for TTS:",
-                                    lines=4,
-                                    value="Farover the misty mountains cold too dungeons deep and caverns old.")
-            generate_button = gr.Button("Generate Audio",
-                                        variant="primary")
+            voice_info = gr.Markdown(f'`{selected_voice.value}`')
+
+        # Main input and output components
+        with gr.Row():
+            text_input = gr.Textbox(
+                label="Enter text for TTS:",
+                placeholder="Type your message here...",
+                lines=4,
+                value="Farover the misty mountains cold too dungeons deep and caverns old.",
+            )
+            generate_button = gr.Button("Generate Audio", variant="primary")
+
         output_audio = gr.Audio(label="TTS Output")
+
         with gr.Column():
             voice_buttons = []
             for i in range(0, len(VOICES), 7):
                 with gr.Row(elem_classes=["cool-row"]):
                     for voice_filename in VOICES[i:i+7]:
                         voice_filename = voice_filename[4:-4]  # drop wav/ for visibility
-                        button = gr.Button(voice_filename, elem_classes=["cool-button"])
-                        button.click(fn=lambda: update_selected_voice(voice_filename, text_input.value),
-                                     inputs=[],
-                                     outputs=[selected_voice, text_input])
+                        button = gr.Button(voice_filename,  elem_classes=["cool-button"])
+
+                        button.click(
+                            fn=update_selected_voice,
+                            inputs=[gr.Textbox(value=voice_filename, visible=False)],
+                            outputs=[selected_voice]
+                        )
+                        button.click(
+                            fn=lambda v=voice_filename: f'`{v}`',
+                            inputs=None,
+                            outputs=voice_info
+                        )
                         voice_buttons.append(button)
-        generate_button.click(fn=other_tts,
-                              inputs=[text_input, selected_voice],
-                              outputs=output_audio)
+
+        generate_button.click(
+            fn=other_tts,
+            inputs=[text_input, selected_voice],
+            outputs=output_audio
+        )
 
     with gr.Tab(label="Speech Analysis"):
         with gr.Row():
@@ -560,4 +579,4 @@ with gr.Blocks(theme='huggingface', css=css_buttons) as demo:
         submit_btn.click(recognize, input, outputs)
 
 
-demo.launch(debug=True)
+demo.launch(debug=True)