Update app.py

app.py

@@ -2,77 +2,46 @@ import gradio as gr
 import edge_tts
 import asyncio
 import tempfile
-import re
-
-EBURON_VERSION = "1.8"
-
-# ---------------------------------
-# Emotion presets (ElevenLabs-ish)
-# ---------------------------------
-EMOTION_PRESETS = {
-    "Neutral": {
-        "rate_offset": 0,
-        "pitch_offset": 0,
-    },
-    "Comedy / Playful (Taglish)": {
-        "rate_offset": 14,   # faster
-        "pitch_offset": 6,   # brighter
-    },
-    "Storytelling / Warm": {
-        "rate_offset": -4,
-        "pitch_offset": 3,
-    },
-    "Emotional / Heartfelt": {
-        "rate_offset": -10,
-        "pitch_offset": 6,
-    },
-    "Angry / Rant": {
-        "rate_offset": 16,
-        "pitch_offset": 4,
-    },
-    "Sad / Dramatic": {
-        "rate_offset": -18,
-        "pitch_offset": -4,
-    },
-}
+import xml.sax.saxutils
+
+EBURON_VERSION = "2.0"
 
 # -----------------------------
-# Custom CSS (works with gradio==4.36.1)
+# Custom CSS – ElevenLabs-style
 # -----------------------------
-EBURON_CSS = """
-body {
+EBURON_CSS = f"""
+body {{
     background: radial-gradient(circle at top left, #020617 0, #020617 45%, #020617 100%);
     color: #e5e7eb;
     margin: 0;
     padding: 0;
-}
+}}
 
-* {
+* {{
     font-family: system-ui, -apple-system, BlinkMacSystemFont, "SF Pro Text", "Segoe UI", sans-serif;
     -webkit-font-smoothing: antialiased;
-}
+}}
 
-#eburon-root {
+#eburon-root {{
     max-width: 1100px;
     margin: 0 auto;
     padding: 20px 18px 32px 18px;
-}
+}}
 
-/* Top nav bar (fake, for look) */
-#eburon-top-nav {
+#eburon-top-nav {{
     display: flex;
     align-items: center;
     justify-content: space-between;
     margin-bottom: 18px;
-}
+}}
 
-#eburon-nav-left {
+#eburon-nav-left {{
     display: flex;
     align-items: center;
     gap: 14px;
-}
+}}
 
-#eburon-logo-circle {
+#eburon-logo-circle {{
     width: 32px;
     height: 32px;
     border-radius: 999px;
@@ -84,27 +53,27 @@ body {
     font-weight: 800;
     font-size: 17px;
     box-shadow: 0 0 22px rgba(59, 130, 246, 0.8);
-}
+}}
 
-#eburon-product-title {
+#eburon-product-title {{
     display: flex;
     flex-direction: column;
-}
+}}
 
-#eburon-product-title span:nth-child(1) {
+#eburon-product-title span:nth-child(1) {{
     font-size: 18px;
     font-weight: 700;
     letter-spacing: 0.08em;
     text-transform: uppercase;
     color: #e5e7eb;
-}
+}}
 
-#eburon-product-title span:nth-child(2) {
+#eburon-product-title span:nth-child(2) {{
     font-size: 11px;
     color: #9ca3af;
-}
+}}
 
-#eburon-nav-tabs {
+#eburon-nav-tabs {{
     display: inline-flex;
     align-items: center;
     gap: 4px;
@@ -113,96 +82,90 @@ body {
     background: rgba(15, 23, 42, 0.9);
     border: 1px solid rgba(55, 65, 81, 0.9);
     font-size: 11px;
-}
+}}
 
-.eburon-tab {
+.eburon-tab {{
     padding: 5px 10px;
     border-radius: 999px;
     cursor: default;
     color: #9ca3af;
-}
+}}
 
-.eburon-tab-active {
+.eburon-tab-active {{
     background: linear-gradient(135deg, #38bdf8, #6366f1);
     color: #020617;
     font-weight: 600;
-}
+}}
 
-#eburon-nav-right {
+#eburon-nav-right {{
     display: flex;
     align-items: center;
     gap: 8px;
     font-size: 11px;
     color: #9ca3af;
-}
+}}
 
-#eburon-pill-version {
+#eburon-pill-version {{
     padding: 4px 10px;
     border-radius: 999px;
     border: 1px solid rgba(148, 163, 184, 0.4);
     background: radial-gradient(circle at top, rgba(31, 41, 55, 1), rgba(15, 23, 42, 1));
-}
+}}
 
-#eburon-pill-usage {
+#eburon-pill-usage {{
     padding: 4px 10px;
     border-radius: 999px;
     border: 1px solid rgba(59, 130, 246, 0.7);
     background: radial-gradient(circle at top, rgba(30, 64, 175, 0.85), rgba(15, 23, 42, 1));
-}
+}}
 
-/* Main cards */
-.eburon-main-card {
+.eburon-main-card {{
     border-radius: 20px;
     background: radial-gradient(circle at top left, #020617, #020617 60%);
     border: 1px solid rgba(51, 65, 85, 0.9);
     box-shadow: 0 24px 48px rgba(15, 23, 42, 0.95);
     padding: 16px 18px 18px 18px;
-}
+}}
 
-/* Headings inside cards */
-.eburon-section-header {
+.eburon-section-header {{
     display: flex;
     justify-content: space-between;
     align-items: center;
     margin-bottom: 8px;
-}
+}}
 
-.eburon-section-title {
+.eburon-section-title {{
     font-size: 14px;
     font-weight: 600;
     color: #e5e7eb;
-}
+}}
 
-.eburon-section-subtitle {
+.eburon-section-subtitle {{
     font-size: 11px;
     color: #9ca3af;
-}
+}}
 
-/* Script textarea */
-textarea {
+textarea {{
     background-color: #020617 !important;
     border-radius: 14px !important;
     border: 1px solid rgba(55, 65, 81, 0.9) !important;
     color: #e5e7eb !important;
     font-size: 13px !important;
-}
+}}
 
-/* Right panel controls */
-select, input[type="range"] {
+select, input[type="range"] {{
     background-color: #020617 !important;
     border-radius: 999px !important;
     border: 1px solid rgba(55, 65, 81, 0.9) !important;
     color: #e5e7eb !important;
-}
+}}
 
-/* Labels */
-label span, .gr-textbox label, .gr-slider label, .gr-dropdown label {
+label span, .gr-textbox label, .gr-slider label, .gr-dropdown label {{
     font-size: 11px !important;
     color: #9ca3af !important;
-}
+}}
 
-/* Generate button */
-#eburon-generate-btn button {
+#eburon-generate-btn button {{
     width: 100%;
     border-radius: 999px;
     font-weight: 600;
@@ -211,94 +174,58 @@ label span, .gr-textbox label, .gr-slider label, .gr-dropdown label {
     background: linear-gradient(135deg, #22c55e, #38bdf8);
     box-shadow: 0 12px 32px rgba(56, 189, 248, 0.75);
     border: none;
-}
+}}
 
-#eburon-generate-btn button:hover {
+#eburon-generate-btn button:hover {{
     transform: translateY(-1px);
     box-shadow: 0 18px 42px rgba(56, 189, 248, 0.95);
-}
+}}
 
-/* Audio player container */
-#eburon-audio-card {
+#eburon-audio-card {{
     border-radius: 18px;
     background: radial-gradient(circle at top right, #020617, #020617 65%);
     border: 1px solid rgba(55, 65, 81, 0.9);
     box-shadow: 0 18px 40px rgba(15, 23, 42, 0.95);
     padding: 12px 14px 14px 14px;
-}
+}}
 
-#eburon-audio-header {
+#eburon-audio-header {{
     display: flex;
     justify-content: space-between;
     align-items: center;
     margin-bottom: 4px;
-}
+}}
 
-#eburon-audio-title {
+#eburon-audio-title {{
     font-size: 12px;
     font-weight: 600;
     color: #e5e7eb;
-}
+}}
 
-#eburon-audio-subtitle {
+#eburon-audio-subtitle {{
     font-size: 11px;
     color: #9ca3af;
-}
+}}
 
-/* Warning styling (Markdown used as banner) */
-.svelte-1g805jl {
+.svelte-1g805jl {{
     border-radius: 999px !important;
-}
+}}
 
-/* Small badges */
-.eburon-mini-pill {
+.eburon-mini-pill {{
     padding: 2px 7px;
     border-radius: 999px;
     border: 1px solid rgba(75, 85, 99, 0.9);
     font-size: 10px;
     color: #9ca3af;
-}
+}}
 """
 
 # -----------------------------
-# Helper: map [pause], [laugh] etc to more natural text
-# -----------------------------
-def normalize_script_for_tts(text: str) -> str:
-    """
-    Convert expressive cues [pause], [laugh], [energetic intro] etc.
-    into punctuation and "ha ha ha" so Edge TTS doesn't read brackets
-    but still sounds more like a comedian.
-    """
-
-    def _repl(match: re.Match) -> str:
-        cue = match.group(1).strip().lower()
-
-        if "pause" in cue or "beat" in cue or "silence" in cue:
-            return "... "
-        if "laugh" in cue or "chuckle" in cue:
-            # let TTS actually say "ha ha ha" to mimic a laugh
-            return " ha ha ha, "
-        if "intro" in cue or "outro" in cue:
-            return " "
-        if "soft" in cue or "whisper" in cue:
-            return " "
-        if "energetic" in cue or "teasing" in cue or "conversational" in cue:
-            return " "
-        # default: just drop bracket cue
-        return " "
-
-    # Replace bracketed cues
-    out = re.sub(r"\[(.*?)\]", _repl, text)
-    # Compress any crazy spacing
-    out = re.sub(r"\s+", " ", out)
-    return out.strip()
-
-
-# -----------------------------
-# Core TTS helpers
+# Core TTS Logic with Emotions
 # -----------------------------
 async def get_voices():
     voices = await edge_tts.list_voices()
+    # Prioritize voices known for good emotional range (e.g., US, UK)
     voice_labels = [
         f"{v['ShortName']} - {v['Locale']} ({v['Gender']})"
         for v in voices
@@ -306,275 +233,200 @@ async def get_voices():
     voice_labels.sort()
     return voice_labels
 
-
-async def text_to_speech(text, voice, base_rate, base_pitch, emotion, expressiveness):
+async def text_to_speech(text, voice, rate, pitch, style, style_degree):
     if not text.strip():
-        return None, "Please enter some text to synthesize."
-
+        return None, "Please enter some text."
     if not voice:
         return None, "Please select a voice."
 
-    # Clean expressive brackets into comedy-friendly text
-    clean_text = normalize_script_for_tts(text)
-
     voice_short_name = voice.split(" - ")[0].strip()
 
-    # -----------------------------
-    # Emotion → rate/pitch shaping
-    # -----------------------------
-    preset = EMOTION_PRESETS.get(emotion or "Neutral", EMOTION_PRESETS["Neutral"])
-    # 0 = no emotion, 100 = base, 200 = 2x preset strength
-    factor = max(0.0, float(expressiveness) / 100.0)
-
-    rate_offset = int(preset["rate_offset"] * factor)
-    pitch_offset = int(preset["pitch_offset"] * factor)
-
-    eff_rate = int(base_rate + rate_offset)
-    eff_pitch = int(base_pitch + pitch_offset)
-
-    # Clamp into safe ranges
-    eff_rate = max(-50, min(50, eff_rate))
-    eff_pitch = max(-20, min(20, eff_pitch))
-
-    rate_str = f"{eff_rate:+d}%"
-    pitch_str = f"{eff_pitch:+d}Hz"
-
-    communicate = edge_tts.Communicate(
-        text=clean_text,
-        voice=voice_short_name,
-        rate=rate_str,
-        pitch=pitch_str,
+    # Format Rate and Pitch
+    rate_str = f"{rate:+d}%"
+    pitch_str = f"{pitch:+d}Hz"
+    
+    # Escape special characters for XML
+    safe_text = xml.sax.saxutils.escape(text)
+
+    # ---------------------------------------------------------
+    # Construct SSML for Emotional Output
+    # ---------------------------------------------------------
+    # If "General" is selected, we don't use the express-as tag.
+    # Otherwise, we wrap the content.
+    if style != "General":
+        ssml_content = (
+            f"<mstts:express-as style='{style.lower()}' styledegree='{style_degree}'>"
+            f"<prosody rate='{rate_str}' pitch='{pitch_str}'>{safe_text}</prosody>"
+            f"</mstts:express-as>"
+        )
+    else:
+        ssml_content = (
+            f"<prosody rate='{rate_str}' pitch='{pitch_str}'>{safe_text}</prosody>"
+        )
+
+    # Full SSML wrapper
+    ssml = (
+        f"<speak version='1.0' xmlns='http://www.w3.org/2001/10/synthesis' xmlns:mstts='https://www.w3.org/2001/mstts' xml:lang='en-US'>"
+        f"<voice name='{voice_short_name}'>"
+        f"{ssml_content}"
+        f"</voice>"
+        f"</speak>"
     )
 
+    # Note: When using SSML, we pass the SSML string as 'text' and don't use rate/pitch args in Communicate
+    communicate = edge_tts.Communicate(text=ssml, voice=voice_short_name)
+
     with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as tmp_file:
         tmp_path = tmp_file.name
         await communicate.save(tmp_path)
 
     return tmp_path, None
 
-
-async def tts_interface(text, voice, base_rate, base_pitch, emotion, expressiveness):
-    audio, warning = await text_to_speech(
-        text=text,
-        voice=voice,
-        base_rate=base_rate,
-        base_pitch=base_pitch,
-        emotion=emotion,
-        expressiveness=expressiveness,
-    )
-    if warning:
-        return audio, gr.update(value=f"⚠️ {warning}", visible=True)
-    return audio, gr.update(value="", visible=False)
-
+async def tts_interface(text, voice, rate, pitch, style, style_degree):
+    # Warning logic if user selects an emotion for a voice that likely doesn't support it
+    # (Simplified check: Non-Neural voices or specific locales might ignore it)
+    warning_msg = None
+    if style != "General" and "Neural" not in voice:
+        warning_msg = "Note: The selected voice might not support emotions. Neural voices work best."
+
+    audio, error = await text_to_speech(text, voice, rate, pitch, style, style_degree)
+    
+    if error:
+        return None, gr.Warning(error)
+    
+    if warning_msg:
+        return audio, gr.Warning(warning_msg)
+        
+    return audio, None
 
 # -----------------------------
-# Eburon Speech Studio v1.8 – ElevenLabs-ish expressive UI
+# Eburon Speech Studio v2.0 UI
 # -----------------------------
 async def create_demo():
     voices = await get_voices()
+    
+    # Common styles supported by Microsoft Azure/Edge Neural voices
+    styles = [
+        "General", "Cheerful", "Sad", "Angry", "Terrified", 
+        "Whispering", "Excited", "Friendly", "Unfriendly", 
+        "Shouting", "Hopeful"
+    ]
 
-    # Sample Taglish Alex Calleja–style comedy script
-    sample_script = (
-        "[energetic intro]\n"
-        "Magandang gabi sa inyong lahat! Ako nga pala si Alex… hindi Calleja, "
-        "pero pwede na rin sa murang kopya. [pause] Parang Shopee version ng Netflix special.\n\n"
-        "[conversational]\n"
-        "Alam n’yo, mahirap na maging adult ngayon. Nung bata tayo, gusto natin tumanda para "
-        "“walang mag-uutos”. Ngayon, tumanda tayo… at ang pinaka-maingay mag-utos: BILLS. [pause]\n"
-        "Kuryente, tubig, WiFi, GCash utang, BNPL… parang ex na hindi makamove on. "
-        "Laging bumabalik buwan-buwan.\n\n"
-        "[teasing tone]\n"
-        "Tapos ‘yung kuryente, grabe. Kahit wala ka sa bahay, mataas pa rin bill. "
-        "Parang Meralco, nag-a-assume: “Alam naming may iyak ka sa dilim, may load ‘yan sa emosyon.” [laugh]\n\n"
-        "[family observation]\n"
-        "Sa pamilya, laging may tita na human notification. Wala kang alam? Siya meron. "
-        "Pagdating mo sa handaan: “O, tumaba ka ah… wala ka pa ring jowa? Ano na plano mo sa buhay?” "
-        "Parang performance appraisal na walang increase. [pause]\n\n"
-        "[jeepney bit]\n"
-        "Sa jeep, may tatlong klaseng pasahero: una, ‘yung DJ ng boundary — siya taga-abot ng bayad, "
-        "taga-sabi ng “Bayad daw po sa likod!”, parang COO ng jeep; pangalawa, ‘yung ninja na ayaw mag-abot ng bayad, "
-        "kahit tumama na sa siko niya yung pera; pangatlo, ‘yung tulog na may superpower — biglang gigising: "
-        "“Sa kanto lang po!” sakto pa rin ang timing.\n\n"
-        "[closer, warm]\n"
-        "Pero kahit ganoon, solid pa rin tayo mga Pinoy. Kahit pagod, late, traffic, may punchline pa rin sa dulo. "
-        "Yun ang talent natin: kahit hirap sa buhay, may follow-up joke pa rin.\n"
-        "Maraming salamat, good night! [pause]"
-    )
-
-    with gr.Blocks(title="Eburon Speech Studio v1.8") as demo:
-        # Inject CSS
-        gr.HTML(f"<style>{EBURON_CSS}</style>", elem_id="eburon-style-inject")
-
+    with gr.Blocks(title="Eburon Speech Studio v2.0", css=EBURON_CSS) as demo:
         with gr.Column(elem_id="eburon-root"):
-            # Top nav
-            gr.HTML(
-                f"""
+            # HEADER
+            gr.HTML(f"""
                 <div id="eburon-top-nav">
                     <div id="eburon-nav-left">
                         <div id="eburon-logo-circle">E</div>
                         <div id="eburon-product-title">
-                            <span>EBURON SPEECH STUDIO</span>
-                            <span>Neural voice generation · v{EBURON_VERSION}</span>
+                            <span>EBURON EMOTION</span>
+                            <span>Neural Expression Engine · v{EBURON_VERSION}</span>
                         </div>
                         <div id="eburon-nav-tabs">
-                            <div class="eburon-tab eburon-tab-active">Speech</div>
+                            <div class="eburon-tab eburon-tab-active">Synthesis</div>
                             <div class="eburon-tab">Voice Lab</div>
-                            <div class="eburon-tab">Projects</div>
                         </div>
                     </div>
                     <div id="eburon-nav-right">
-                        <div id="eburon-pill-version">
-                            Studio {EBURON_VERSION}
-                        </div>
-                        <div id="eburon-pill-usage">
-                            edge_tts 7.2.0 · gradio 4.36.1
-                        </div>
+                        <div id="eburon-pill-version">Pro</div>
+                        <div id="eburon-pill-usage">SSML Enabled</div>
                     </div>
                 </div>
-                """
-            )
+            """)
 
-            # Main body
             with gr.Row():
-                # Left: Script
+                # LEFT: Script Input
                 with gr.Column(scale=2, min_width=460):
                     with gr.Group(elem_classes="eburon-main-card"):
-                        gr.HTML(
-                            """
+                        gr.HTML("""
                             <div class="eburon-section-header">
                                 <div>
                                     <div class="eburon-section-title">Script</div>
-                                    <div class="eburon-section-subtitle">
-                                        Taglish Alex Calleja–style skit with cues like [pause], [laugh], [energetic intro].
-                                    </div>
-                                </div>
-                                <div class="eburon-mini-pill">
-                                    Cues auto-converted to pauses & “ha ha ha”
+                                    <div class="eburon-section-subtitle">Type your text. Use standard punctuation for best pause handling.</div>
                                 </div>
+                                <div class="eburon-mini-pill">Unlimited</div>
                             </div>
-                            """
-                        )
+                        """)
                         text_input = gr.Textbox(
-                            label="",
-                            value=sample_script,
-                            lines=16,
+                            label="", 
+                            placeholder="Enter text here...", 
+                            lines=14,
+                            value="I can't believe you did that! That is absolutely amazing."
                         )
 
-                # Right: Voice & emotion
+                # RIGHT: Voice & Emotion Controls
                 with gr.Column(scale=1, min_width=340):
                     with gr.Group(elem_classes="eburon-main-card"):
-                        gr.HTML(
-                            """
+                        gr.HTML("""
                             <div class="eburon-section-header">
                                 <div>
-                                    <div class="eburon-section-title">Voice & Delivery</div>
-                                    <div class="eburon-section-subtitle">
-                                        Emotion preset + intensity + fine speed & pitch (ElevenLabs-style).
-                                    </div>
-                                </div>
-                                <div class="eburon-mini-pill">
-                                    Emotion-shaped rate & pitch
+                                    <div class="eburon-section-title">Voice & Emotion</div>
+                                    <div class="eburon-section-subtitle">Select neural voice and emotional overlay.</div>
                                 </div>
                             </div>
-                            """
-                        )
-
+                        """)
+                        
                         voice_dropdown = gr.Dropdown(
-                            choices=[""] + voices,
-                            label="Voice",
-                            value="",
-                            info="Tip: pick a lively EN voice (e.g. male) for stand-up style.",
+                            choices=[""] + voices, 
+                            label="Voice Model", 
+                            value="en-US-AriaNeural - en-US (Female)" if any("AriaNeural" in v for v in voices) else "",
+                            info="Select 'Neural' voices for best results."
                         )
-
-                        base_rate_slider = gr.Slider(
-                            minimum=-50,
-                            maximum=50,
-                            value=5,  # slightly faster by default
-                            label="Base Speed",
-                            step=1,
-                            info="Baseline speaking speed. Emotion will adjust on top.",
+                        
+                        style_dropdown = gr.Dropdown(
+                            choices=styles,
+                            label="Expressive Style",
+                            value="General",
+                            info="Applies emotional tone to the voice."
                         )
-
-                        base_pitch_slider = gr.Slider(
-                            minimum=-20,
-                            maximum=20,
-                            value=2,  # slightly brighter by default
-                            label="Base Pitch",
-                            step=1,
-                            info="Baseline pitch. Emotion will adjust on top.",
+                        
+                        style_degree = gr.Slider(
+                            minimum=0.1, maximum=2.0, value=1.0, step=0.1,
+                            label="Emotion Intensity",
+                            info="< 1 is subtle, > 1 is exaggerated."
                         )
 
-                        emotion_dropdown = gr.Dropdown(
-                            label="Emotion preset",
-                            choices=list(EMOTION_PRESETS.keys()),
-                            value="Comedy / Playful (Taglish)",
-                            info="High-level emotional profile (approximate, using rate+pitch).",
+                        gr.HTML("<div style='margin-top:15px; margin-bottom:5px; border-top:1px solid #334155;'></div>")
+                        
+                        rate_slider = gr.Slider(
+                            minimum=-50, maximum=50, value=0, label="Speed", step=1
                         )
-
-                        expressiveness_slider = gr.Slider(
-                            minimum=0,
-                            maximum=200,
-                            value=130,  # a bit stronger than normal
-                            step=5,
-                            label="Expressiveness (intensity)",
-                            info="0 = off, 100 = normal, 200 = stronger emotion.",
+                        pitch_slider = gr.Slider(
+                            minimum=-20, maximum=20, value=0, label="Pitch", step=1
                         )
 
-            # Bottom: Generate + audio
+            # BOTTOM: Generate & Player
             with gr.Row():
                 with gr.Column(scale=1, min_width=260):
-                    generate_btn = gr.Button(
-                        "Generate",
-                        variant="primary",
-                        elem_id="eburon-generate-btn",
-                    )
-                    warning_md = gr.Markdown("", visible=False)
+                    generate_btn = gr.Button("Generate Audio", variant="primary", elem_id="eburon-generate-btn")
+                    status_msg = gr.Markdown(visible=False)
 
                 with gr.Column(scale=2, min_width=460):
                     with gr.Group(elem_id="eburon-audio-card"):
-                        gr.HTML(
-                            """
+                        gr.HTML("""
                             <div id="eburon-audio-header">
                                 <div>
-                                    <div id="eburon-audio-title">Latest generation</div>
-                                    <div id="eburon-audio-subtitle">
-                                        Auto-plays after each render. Browser must allow audio playback.
-                                    </div>
-                                </div>
-                                <div class="eburon-mini-pill">
-                                    MP3 · edge_tts neural
+                                    <div id="eburon-audio-title">Output</div>
+                                    <div id="eburon-audio-subtitle">Generated result</div>
                                 </div>
+                                <div class="eburon-mini-pill">MP3</div>
                             </div>
-                            """
-                        )
-                        audio_output = gr.Audio(
-                            label="",
-                            type="filepath",
-                            autoplay=True,
-                            interactive=False,
-                        )
+                        """)
+                        audio_output = gr.Audio(label="", type="filepath", autoplay=True, interactive=False)
 
             generate_btn.click(
                 fn=tts_interface,
-                inputs=[
-                    text_input,
-                    voice_dropdown,
-                    base_rate_slider,
-                    base_pitch_slider,
-                    emotion_dropdown,
-                    expressiveness_slider,
-                ],
-                outputs=[audio_output, warning_md],
+                inputs=[text_input, voice_dropdown, rate_slider, pitch_slider, style_dropdown, style_degree],
+                outputs=[audio_output, status_msg]
             )
 
     return demo
 
-
 async def main():
     demo = await create_demo()
-    demo.queue(default_concurrency_limit=50)
+    demo.queue(default_concurrency_limit=20)
     demo.launch()
 
-
 if __name__ == "__main__":
-    asyncio.run(main())
+    asyncio.run(main())