refactor: simplify audio input by consolidating microphone and file uploads into a single Gradio Audio component

app.py

@@ -169,10 +169,7 @@ def demo_fn(
     atten_lim_db: float,
     wet_dry_mix: float,
     post_process: bool,
-    mic_input: Optional[str] = None,
 ):
-    if mic_input:
-        speech_upl = mic_input
 
     sr = config("sr", 48000, int, section="df")
     logger.info(f"Got parameters speech_upl: {speech_upl}, noise: {noise_type}, snr: {snr}, atten_lim_db: {atten_lim_db}, wet_dry_mix: {wet_dry_mix}, post_process: {post_process}")
@@ -251,8 +248,6 @@ def demo_fn(
     enh_im = spec_im(enhanced, sr=sr, figure=fig_enh, ax=ax_enh)
 
     filter = [speech_upl, noisy_mp3, enhanced_mp3]
-    if mic_input is not None and mic_input != "":
-        filter.append(mic_input)
     cleanup_tmp(filter)
 
     return noisy_mp3, noisy_im, enhanced_mp3, enh_im
@@ -369,15 +364,6 @@ def cleanup_tmp(filter: List[str] = [], hours_keep=2):
                     logger.info(f"Removed file {f}")
                 except Exception as e:
                     logger.warning(f"failed to remove file {f}: {e}")
-
-
-def toggle(choice):
-    if choice == "mic":
-        return gr.update(visible=True, value=None), gr.update(visible=False, value=None)
-    else:
-        return gr.update(visible=False, value=None), gr.update(visible=True, value=None)
-
-
 with gr.Blocks() as demo:
     with gr.Row():
         gr.Markdown(
@@ -389,11 +375,7 @@ with gr.Blocks() as demo:
         )
     with gr.Row():
         with gr.Column():
-            radio = gr.Radio(
-                ["mic", "file"], value="file", label="How would you like to upload your audio?"
-            )
-            mic_input = gr.Mic(label="Input", type="filepath", visible=False)
-            audio_file = gr.Audio(type="filepath", label="Input", visible=True)
+            audio_file = gr.Audio(sources=["upload", "microphone"], type="filepath", label="Input Audio")
             inputs = [
                 audio_file,
                 gr.Dropdown(
@@ -424,7 +406,6 @@ with gr.Blocks() as demo:
                     label="Post-Process (80Hz Low-Cut & Presence Boost)",
                     value=True,
                 ),
-                mic_input,
             ]
             btn = gr.Button("Generate")
         with gr.Column():
@@ -437,19 +418,18 @@ with gr.Blocks() as demo:
                 gr.Image(label="Enhanced spectrogram"),
             ]
     btn.click(fn=demo_fn, inputs=inputs, outputs=outputs, api_name='denoise')
-    radio.change(toggle, radio, [mic_input, audio_file])
     gr.Examples(
         [
-            ["./samples/p232_013_clean.wav", "Kitchen", "10", 15, 90, True, None],
-            ["./samples/p232_013_clean.wav", "Cafe", "10", 15, 90, True, None],
-            ["./samples/p232_019_clean.wav", "Cafe", "10", 15, 90, True, None],
-            ["./samples/p232_019_clean.wav", "River", "10", 15, 90, True, None],
+            ["./samples/p232_013_clean.wav", "Kitchen", "10", 15, 90, True],
+            ["./samples/p232_013_clean.wav", "Cafe", "10", 15, 90, True],
+            ["./samples/p232_019_clean.wav", "Cafe", "10", 15, 90, True],
+            ["./samples/p232_019_clean.wav", "River", "10", 15, 90, True],
         ],
         fn=demo_fn,
         inputs=inputs,
         outputs=outputs,
         cache_examples=True,
-    ),
+    )
     gr.Markdown(open("usage.md").read())
 
 cleanup_tmp()

requirements.txt

@@ -2,5 +2,5 @@ torch==1.13
 torchaudio==0.13
 deepfilternet==0.4.0
 matplotlib==3.6
-# gradio==3.17
+gradio==5.48.0
 Pillow==9.3