add stereo imaging and neural pan + fix error

app.py

@@ -7,10 +7,10 @@ import gradio as gr
 
 # Default presets for common genres
 PRESETS = {
-    "rock": {"gain": 6, "compress_threshold": -20, "compress_ratio": 4, "low_pass": 12000, "high_pass": 80},
-    "pop": {"gain": 4, "compress_threshold": -18, "compress_ratio": 3, "low_pass": 15000, "high_pass": 100},
-    "jazz": {"gain": 3, "compress_threshold": -22, "compress_ratio": 2, "low_pass": 14000, "high_pass": 60},
-    "electronic": {"gain": 8, "compress_threshold": -16, "compress_ratio": 5, "low_pass": 20000, "high_pass": 120},
+    "rock": {"gain": 6, "compress_threshold": -20, "compress_ratio": 4, "low_pass": 12000, "high_pass": 80, "stereo_width": 100, "pan": 0},
+    "pop": {"gain": 4, "compress_threshold": -18, "compress_ratio": 3, "low_pass": 15000, "high_pass": 100, "stereo_width": 100, "pan": 0},
+    "jazz": {"gain": 3, "compress_threshold": -22, "compress_ratio": 2, "low_pass": 14000, "high_pass": 60, "stereo_width": 90, "pan": 0},
+    "electronic": {"gain": 8, "compress_threshold": -16, "compress_ratio": 5, "low_pass": 20000, "high_pass": 120, "stereo_width": 120, "pan": 0},
 }
 
 # JSON export extension
@@ -27,33 +27,59 @@ def apply_mastering(audio: AudioSegment, params: dict) -> AudioSegment:
     # Gain
     processed = audio + params["gain"]
     # Compression
-    processed = compress_dynamic_range(processed,
+    processed = compress_dynamic_range(
+        processed,
         threshold=params["compress_threshold"],
-        ratio=params["compress_ratio"])
+        ratio=params["compress_ratio"]
+    )
     # EQ filters
     processed = high_pass_filter(processed, params["high_pass"])
     processed = low_pass_filter(processed, params["low_pass"])
+    # Stereo Imaging (Width)
+    processed = apply_stereo_width(processed, params.get("stereo_width", 100))
+    # Neural Pan
+    processed = apply_pan(processed, params.get("pan", 0))
     # Normalization
     processed = normalize(processed)
     return processed
 
 
-def process_and_export(audio_file, preset_name, gain, threshold, ratio, lp, hp):
-    # Load
-    audio = AudioSegment.from_file(audio_file.name)
-    # Determine params
+def apply_stereo_width(audio: AudioSegment, width: float) -> AudioSegment:
+    # Width in percent: 0 = mono, 100 = original, >100 = enhanced
+    left, right = audio.split_to_mono()
+    mid = (left + right) / 2
+    side = (left - right) / 2
+    # scale side for width control
+    side = side * (width / 100)
+    new_left = mid + side
+    new_right = mid - side
+    return AudioSegment.from_mono_audiosegments(new_left, new_right)
+
+
+def apply_pan(audio: AudioSegment, pan: float) -> AudioSegment:
+    # Pan from -100 (left) to 100 (right)
+    left, right = audio.split_to_mono()
+    pan = max(-100, min(100, pan)) / 100
+    if pan > 0:
+        new_left = left * (1 - pan)
+        new_right = right
+    else:
+        new_left = left
+        new_right = right * (1 + pan)
+    return AudioSegment.from_mono_audiosegments(new_left, new_right)
+
+
+def process_and_export(audio_file, preset_name, gain, threshold, ratio, lp, hp, width, pan):
+    audio = AudioSegment.from_file(audio_file)
     if preset_name != "custom":
         params = PRESETS[preset_name]
     else:
         params = {"gain": gain, "compress_threshold": threshold, "compress_ratio": ratio,
-                  "low_pass": lp, "high_pass": hp}
-    # Process
+                  "low_pass": lp, "high_pass": hp, "stereo_width": width, "pan": pan}
     out = apply_mastering(audio, params)
-    # Export audio
     buf = io.BytesIO()
     out.export(buf, format="wav")
     buf.seek(0)
-    # Create preset JSON
     preset = {"preset": preset_name, "params": params}
     preset_buf = io.BytesIO(json.dumps(preset, indent=2).encode())
     preset_buf.name = "preset" + EXPORT_EXT
@@ -65,40 +91,40 @@ def load_preset_json(file_obj):
     name = data.get("preset", "custom")
     params = data.get("params", {})
     return name, params.get("gain", 0), params.get("compress_threshold", 0), \
-           params.get("compress_ratio", 1), params.get("low_pass", 20000), params.get("high_pass", 20)
+           params.get("compress_ratio", 1), params.get("low_pass", 20000), \
+           params.get("high_pass", 20), params.get("stereo_width", 100), params.get("pan", 0)
 
 with gr.Blocks(css=CUSTOM_CSS) as demo:
     gr.Markdown("<div id='title'>Advanced Browser-based Audio Mastering Suite</div>")
     with gr.Row():
         with gr.Column():
-            audio_input = gr.Audio(label="Upload Audio", type="file")
+            # Use 'filepath' to get the file path for AudioSegment
+            audio_input = gr.Audio(label="Upload Audio", type="filepath")
             preset_dropdown = gr.Dropdown(choices=list(PRESETS.keys()) + ["custom"], value="rock", label="Preset")
             gain = gr.Slider(-10, 20, value=PRESETS["rock"]["gain"], label="Gain (dB)")
             threshold = gr.Slider(-60, 0, value=PRESETS["rock"]["compress_threshold"], label="Compress Threshold (dB)")
             ratio = gr.Slider(1, 10, value=PRESETS["rock"]["compress_ratio"], label="Compress Ratio")
             lp = gr.Slider(1000, 20000, value=PRESETS["rock"]["low_pass"], label="Low-pass Frequency (Hz)")
             hp = gr.Slider(20, 500, value=PRESETS["rock"]["high_pass"], label="High-pass Frequency (Hz)")
+            width = gr.Slider(0, 200, value=PRESETS["rock"]["stereo_width"], label="Stereo Width (%)")
+            pan = gr.Slider(-100, 100, value=PRESETS["rock"]["pan"], label="Pan (-100 Left to 100 Right)")
             load_preset = gr.File(label="Load .master Preset", file_types=[".master"])
             export_button = gr.Button("Process & Export")
         with gr.Column():
-            output_audio = gr.Audio(label="Processed Audio")
+            # Set output to filepath to serve downloadable file
+            output_audio = gr.Audio(label="Processed Audio", type="filepath")
             export_preset_file = gr.File(label="Download Preset (.master)")
-    
-    # Interactivity
     def sync_sliders(preset):
         if preset != "custom":
             p = PRESETS[preset]
-            return gr.update(value=p["gain"]), gr.update(value=p["compress_threshold"]), \
-                   gr.update(value=p["compress_ratio"]), gr.update(value=p["low_pass"]), \
-                   gr.update(value=p["high_pass"])
+            return (gr.update(value=p["gain"]), gr.update(value=p["compress_threshold"]), \
+                    gr.update(value=p["compress_ratio"]), gr.update(value=p["low_pass"]), \
+                    gr.update(value=p["high_pass"]), gr.update(value=p["stereo_width"]), \
+                    gr.update(value=p["pan"]))
         return None
-
-    preset_dropdown.change(sync_sliders, inputs=[preset_dropdown], outputs=[gain, threshold, ratio, lp, hp])
-    load_preset.upload(load_preset_json, inputs=[load_preset],
-                        outputs=[preset_dropdown, gain, threshold, ratio, lp, hp])
-    export_button.click(process_and_export,
-                         inputs=[audio_input, preset_dropdown, gain, threshold, ratio, lp, hp],
-                         outputs=[output_audio, export_preset_file])
+    preset_dropdown.change(sync_sliders, inputs=[preset_dropdown], outputs=[gain, threshold, ratio, lp, hp, width, pan])
+    load_preset.upload(load_preset_json, inputs=[load_preset], outputs=[preset_dropdown, gain, threshold, ratio, lp, hp, width, pan])
+    export_button.click(process_and_export, inputs=[audio_input, preset_dropdown, gain, threshold, ratio, lp, hp, width, pan], outputs=[output_audio, export_preset_file])
 
 if __name__ == "__main__":
     demo.launch()