Spaces:
Paused
Paused
+spectral gating filter
Browse files
app.py
CHANGED
@@ -35,7 +35,12 @@ def preprocess_audio(audio_tensor, original_sample_rate, apply_normalization):
|
|
35 |
audio_tensor = torchaudio.functional.resample(audio_tensor, orig_freq=original_sample_rate, new_freq=16000) # Resample
|
36 |
return audio_tensor
|
37 |
|
38 |
-
def
|
|
|
|
|
|
|
|
|
|
|
39 |
audio_data = audio_tensor.numpy()
|
40 |
filtered_audio = wiener(audio_data)
|
41 |
return torch.tensor(filtered_audio, dtype=audio_tensor.dtype)
|
@@ -43,13 +48,13 @@ def apply_wiener_filter(audio_tensor):
|
|
43 |
@spaces.GPU
|
44 |
def transcribe_speech(audio, progress=gr.Progress()):
|
45 |
if audio is None:
|
46 |
-
return "No audio received."
|
47 |
progress(0.5, desc="Transcribing audio...")
|
48 |
audio_np = audio.numpy().squeeze()
|
49 |
transcription = pipe(audio_np, chunk_length_s=10)['text']
|
50 |
-
return replace_symbols_back(transcription)
|
51 |
|
52 |
-
def transcribe_from_youtube(url,
|
53 |
progress(0, "Downloading YouTube audio...")
|
54 |
|
55 |
yt = YouTube(url)
|
@@ -62,16 +67,24 @@ def transcribe_from_youtube(url, apply_wiener, apply_normalization, progress=gr.
|
|
62 |
audio, original_sample_rate = torchaudio.load(audio_data)
|
63 |
audio = preprocess_audio(audio, original_sample_rate, apply_normalization)
|
64 |
|
65 |
-
if
|
66 |
progress(0.4, "Applying Wiener filter...")
|
67 |
-
audio =
|
|
|
|
|
|
|
|
|
68 |
|
69 |
-
transcription = transcribe_speech(audio)
|
|
|
|
|
|
|
|
|
70 |
|
71 |
except Exception as e:
|
72 |
-
return str(e)
|
73 |
|
74 |
-
return transcription
|
75 |
|
76 |
def populate_metadata(url):
|
77 |
yt = YouTube(url)
|
@@ -96,16 +109,18 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
|
96 |
mic_audio = gr.Audio(sources=['microphone','upload'], type="filepath", label="Record or upload an audio")
|
97 |
transcribe_button = gr.Button("Transcribe")
|
98 |
transcription_output = gr.Textbox(label="Transcription")
|
|
|
99 |
|
100 |
-
transcribe_button.click(fn=transcribe_speech, inputs=mic_audio, outputs=transcription_output)
|
101 |
|
102 |
with gr.Tab("YouTube URL"):
|
103 |
gr.Markdown("## Transcribe speech from YouTube video")
|
104 |
youtube_url = gr.Textbox(label="Enter YouTube video URL")
|
105 |
|
106 |
with gr.Accordion("Audio Improvements", open=False):
|
107 |
-
apply_wiener = gr.Checkbox(label="Reduce noise", info="Apply Wiener Filter", value=False)
|
108 |
apply_normalization = gr.Checkbox(label="Normalize audio volume", value=True)
|
|
|
|
|
109 |
|
110 |
with gr.Row():
|
111 |
img = gr.Image(label="Thumbnail", height=240, width=240, scale=1)
|
@@ -113,8 +128,9 @@ with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
|
113 |
|
114 |
transcribe_button = gr.Button("Transcribe")
|
115 |
transcription_output = gr.Textbox(label="Transcription", placeholder="Transcription Output", lines=10)
|
|
|
116 |
|
117 |
-
transcribe_button.click(fn=transcribe_from_youtube, inputs=[youtube_url, apply_wiener, apply_normalization], outputs=transcription_output)
|
118 |
youtube_url.change(populate_metadata, inputs=[youtube_url], outputs=[img, title])
|
119 |
|
120 |
demo.launch()
|
|
|
35 |
audio_tensor = torchaudio.functional.resample(audio_tensor, orig_freq=original_sample_rate, new_freq=16000) # Resample
|
36 |
return audio_tensor
|
37 |
|
38 |
+
def spectral_gating(audio_tensor):
|
39 |
+
audio_data = audio_tensor.numpy()
|
40 |
+
reduced_noise = nr.reduce_noise(y=audio_data, sr=16_000)
|
41 |
+
return torch.tensor(reduced_noise, dtype=audio_tensor.dtype)
|
42 |
+
|
43 |
+
def wiener_filter(audio_tensor):
|
44 |
audio_data = audio_tensor.numpy()
|
45 |
filtered_audio = wiener(audio_data)
|
46 |
return torch.tensor(filtered_audio, dtype=audio_tensor.dtype)
|
|
|
48 |
@spaces.GPU
|
49 |
def transcribe_speech(audio, progress=gr.Progress()):
|
50 |
if audio is None:
|
51 |
+
return "No audio received.", None
|
52 |
progress(0.5, desc="Transcribing audio...")
|
53 |
audio_np = audio.numpy().squeeze()
|
54 |
transcription = pipe(audio_np, chunk_length_s=10)['text']
|
55 |
+
return replace_symbols_back(transcription), audio
|
56 |
|
57 |
+
def transcribe_from_youtube(url, apply_wiener_filter, apply_normalization, apply_spectral_gating, progress=gr.Progress()):
|
58 |
progress(0, "Downloading YouTube audio...")
|
59 |
|
60 |
yt = YouTube(url)
|
|
|
67 |
audio, original_sample_rate = torchaudio.load(audio_data)
|
68 |
audio = preprocess_audio(audio, original_sample_rate, apply_normalization)
|
69 |
|
70 |
+
if apply_wiener_filter:
|
71 |
progress(0.4, "Applying Wiener filter...")
|
72 |
+
audio = wiener_filter(audio)
|
73 |
+
|
74 |
+
if apply_spectral_gating:
|
75 |
+
progress(0.4, "Applying Spectral Gating filter...")
|
76 |
+
audio = spectral_gating(audio)
|
77 |
|
78 |
+
transcription, processed_audio = transcribe_speech(audio)
|
79 |
+
audio_np = processed_audio.numpy().squeeze()
|
80 |
+
audio_output = BytesIO()
|
81 |
+
torchaudio.save(audio_output, torch.tensor(audio_np).unsqueeze(0), 16000)
|
82 |
+
audio_output.seek(0)
|
83 |
|
84 |
except Exception as e:
|
85 |
+
return str(e), None
|
86 |
|
87 |
+
return transcription, audio_output
|
88 |
|
89 |
def populate_metadata(url):
|
90 |
yt = YouTube(url)
|
|
|
109 |
mic_audio = gr.Audio(sources=['microphone','upload'], type="filepath", label="Record or upload an audio")
|
110 |
transcribe_button = gr.Button("Transcribe")
|
111 |
transcription_output = gr.Textbox(label="Transcription")
|
112 |
+
audio_output = gr.Audio(label="Processed Audio")
|
113 |
|
114 |
+
transcribe_button.click(fn=transcribe_speech, inputs=mic_audio, outputs=[transcription_output, audio_output])
|
115 |
|
116 |
with gr.Tab("YouTube URL"):
|
117 |
gr.Markdown("## Transcribe speech from YouTube video")
|
118 |
youtube_url = gr.Textbox(label="Enter YouTube video URL")
|
119 |
|
120 |
with gr.Accordion("Audio Improvements", open=False):
|
|
|
121 |
apply_normalization = gr.Checkbox(label="Normalize audio volume", value=True)
|
122 |
+
apply_spectral_gating = gr.Checkbox(label="Apply Spectral Gating filter", info="Noise reduction", value=True)
|
123 |
+
apply_wiener = gr.Checkbox(label="Apply Wiener filter", info="Noise reduction", value=False)
|
124 |
|
125 |
with gr.Row():
|
126 |
img = gr.Image(label="Thumbnail", height=240, width=240, scale=1)
|
|
|
128 |
|
129 |
transcribe_button = gr.Button("Transcribe")
|
130 |
transcription_output = gr.Textbox(label="Transcription", placeholder="Transcription Output", lines=10)
|
131 |
+
audio_output = gr.Audio(label="Processed Audio")
|
132 |
|
133 |
+
transcribe_button.click(fn=transcribe_from_youtube, inputs=[youtube_url, apply_wiener, apply_normalization, apply_spectral_gating], outputs=[transcription_output, audio_output])
|
134 |
youtube_url.change(populate_metadata, inputs=[youtube_url], outputs=[img, title])
|
135 |
|
136 |
demo.launch()
|