Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -1,133 +1,66 @@
|
|
1 |
import gradio as gr
|
2 |
-
print(f"Gradio version: {gr.__version__}")
|
3 |
import moviepy.editor as mp
|
4 |
import numpy as np
|
5 |
import librosa
|
6 |
-
import librosa.display
|
7 |
import matplotlib.pyplot as plt
|
8 |
import io
|
9 |
-
import os
|
10 |
|
11 |
-
# Función principal para generar el video
|
12 |
def audio_to_video(audio_file, image_file, effect_type="waveform"):
|
13 |
-
"""
|
14 |
-
Genera un video a partir de un archivo de audio y una imagen, con un efecto visual sincronizado.
|
15 |
-
|
16 |
-
Args:
|
17 |
-
audio_file: Ruta al archivo de audio (wav o mp3).
|
18 |
-
image_file: Ruta al archivo de imagen (debe ser un formato soportado por MoviePy).
|
19 |
-
effect_type: Tipo de efecto visual a utilizar ("waveform" por defecto, otros tipos se pueden agregar).
|
20 |
-
|
21 |
-
Returns:
|
22 |
-
Ruta al archivo de video generado (mp4). Si falla, retorna un mensaje de error.
|
23 |
-
"""
|
24 |
try:
|
25 |
-
#
|
26 |
y, sr = librosa.load(audio_file)
|
27 |
duration = librosa.get_duration(y=y, sr=sr)
|
28 |
|
29 |
-
#
|
30 |
-
img_clip = mp.ImageClip(image_file)
|
31 |
-
img_clip = img_clip.set_duration(duration) # Asignar la duración del audio a la imagen
|
32 |
|
33 |
-
#
|
34 |
if effect_type == "waveform":
|
35 |
-
audio_envelope = np.abs(y) #
|
36 |
-
|
37 |
-
# Normalize audio envelope to image dimensions
|
38 |
-
audio_envelope = audio_envelope / np.max(audio_envelope)
|
39 |
-
audio_envelope = audio_envelope * img_clip.size[1] / 2 # Scale to half the image height
|
40 |
-
|
41 |
|
42 |
def make_frame(t):
|
43 |
-
|
44 |
-
fig, ax = plt.subplots(figsize=(img_clip.size[0]/100, img_clip.size[1]/100), dpi=100) # Adjust figsize for image dimensions
|
45 |
ax.set_xlim(0, duration)
|
46 |
-
ax.set_ylim(-img_clip.
|
47 |
-
ax.axis('off')
|
48 |
-
|
49 |
-
# Plot waveform
|
50 |
time_index = int(t * sr)
|
51 |
-
wave_slice = audio_envelope[max(0,time_index - sr//10):min(len(audio_envelope), time_index + sr//10)]
|
52 |
-
|
53 |
-
ax.plot(np.linspace(t-0.1,t+0.1,len(wave_slice)), wave_slice
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
# Convert the Matplotlib figure to an image
|
58 |
buf = io.BytesIO()
|
59 |
-
fig.
|
60 |
-
|
61 |
-
|
62 |
-
plt.close(fig) # Close the figure to prevent memory leaks
|
63 |
-
return img
|
64 |
-
|
65 |
-
|
66 |
-
audio_effect_clip = mp.VideoClip(make_frame, duration=duration)
|
67 |
-
audio_effect_clip = audio_effect_clip.set_fps(24) # Set a reasonable frame rate
|
68 |
|
|
|
|
|
69 |
else:
|
70 |
-
return "Error: Efecto
|
71 |
-
|
72 |
-
# 4. Overlay effect onto image
|
73 |
-
final_clip = mp.CompositeVideoClip([img_clip, audio_effect_clip.set_pos("center")])
|
74 |
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
final_clip =
|
79 |
-
|
80 |
-
|
81 |
-
# 6. Guardar el video
|
82 |
-
output_video_path = "output.mp4"
|
83 |
-
final_clip.write_videofile(output_video_path, fps=24, codec="libx264", audio_codec="aac") # Ajustar los parámetros de codificación según sea necesario
|
84 |
-
return output_video_path
|
85 |
|
86 |
except Exception as e:
|
87 |
return f"Error: {str(e)}"
|
88 |
|
89 |
-
|
90 |
-
# ----------------------------------
|
91 |
-
# Gradio Interface
|
92 |
-
# ----------------------------------
|
93 |
-
|
94 |
iface = gr.Interface(
|
95 |
fn=audio_to_video,
|
96 |
inputs=[
|
97 |
-
|
98 |
-
gr.Audio(type="filepath", label="Subir Archivo de Audio (WAV o MP3)"),
|
99 |
gr.Image(type="filepath", label="Subir Imagen"),
|
100 |
-
gr.Radio(["waveform"], value="waveform", label="
|
101 |
],
|
102 |
-
outputs=gr.Video(label="Video Generado"),
|
103 |
-
title="Audio
|
104 |
-
description="Sube un audio y una imagen para generar un video con efecto visual sincronizado."
|
105 |
-
examples=[["audio_example.wav", "image_example.jpg", "waveform"]]
|
106 |
)
|
107 |
|
108 |
-
|
109 |
-
# ----------------------------------
|
110 |
-
# Example files (optional). Create these files
|
111 |
-
# or remove the 'examples' line above.
|
112 |
-
# ----------------------------------
|
113 |
-
# Create dummy audio and image for example purposes if they don't exist
|
114 |
-
if not os.path.exists("audio_example.wav"):
|
115 |
-
sr = 22050
|
116 |
-
T = 5
|
117 |
-
t = np.linspace(0, T, int(T*sr), endpoint=False)
|
118 |
-
x = 0.5*np.sin(2*np.pi*440*t) # A4 frequency
|
119 |
-
librosa.output.write_wav("audio_example.wav", x, sr)
|
120 |
-
|
121 |
-
if not os.path.exists("image_example.jpg"):
|
122 |
-
# Create a simple placeholder image
|
123 |
-
import matplotlib.pyplot as plt
|
124 |
-
fig, ax = plt.subplots(figsize=(6,4))
|
125 |
-
ax.text(0.5, 0.5, "Placeholder Image", ha="center", va="center")
|
126 |
-
ax.axis("off")
|
127 |
-
fig.savefig("image_example.jpg")
|
128 |
-
plt.close(fig)
|
129 |
-
|
130 |
-
|
131 |
-
|
132 |
if __name__ == "__main__":
|
133 |
iface.queue().launch()
|
|
|
1 |
import gradio as gr
|
|
|
2 |
import moviepy.editor as mp
|
3 |
import numpy as np
|
4 |
import librosa
|
|
|
5 |
import matplotlib.pyplot as plt
|
6 |
import io
|
|
|
7 |
|
|
|
8 |
def audio_to_video(audio_file, image_file, effect_type="waveform"):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
try:
|
10 |
+
# Cargar audio
|
11 |
y, sr = librosa.load(audio_file)
|
12 |
duration = librosa.get_duration(y=y, sr=sr)
|
13 |
|
14 |
+
# Cargar imagen
|
15 |
+
img_clip = mp.ImageClip(image_file).set_duration(duration)
|
|
|
16 |
|
17 |
+
# Generar efecto visual (waveform)
|
18 |
if effect_type == "waveform":
|
19 |
+
audio_envelope = np.abs(y) # Envelope del audio
|
20 |
+
audio_envelope = (audio_envelope / np.max(audio_envelope)) * (img_clip.h / 2)
|
|
|
|
|
|
|
|
|
21 |
|
22 |
def make_frame(t):
|
23 |
+
fig, ax = plt.subplots(figsize=(img_clip.w/100, img_clip.h/100), dpi=100)
|
|
|
24 |
ax.set_xlim(0, duration)
|
25 |
+
ax.set_ylim(-img_clip.h/2, img_clip.h/2)
|
26 |
+
ax.axis('off')
|
27 |
+
|
|
|
28 |
time_index = int(t * sr)
|
29 |
+
wave_slice = audio_envelope[max(0, time_index - sr//10):min(len(audio_envelope), time_index + sr//10)]
|
30 |
+
ax.plot(np.linspace(t-0.1, t+0.1, len(wave_slice)), wave_slice - img_clip.h/4, color='red')
|
31 |
+
ax.plot(np.linspace(t-0.1, t+0.1, len(wave_slice)), -wave_slice + img_clip.h/4, color='red')
|
32 |
+
|
|
|
|
|
|
|
33 |
buf = io.BytesIO()
|
34 |
+
fig.savefig(buf, format='png', bbox_inches='tight', pad_inches=0)
|
35 |
+
plt.close(fig)
|
36 |
+
return np.array(Image.open(buf)) # Convertir a array de imagen
|
|
|
|
|
|
|
|
|
|
|
|
|
37 |
|
38 |
+
effect_clip = mp.VideoClip(make_frame, duration=duration).set_fps(24)
|
39 |
+
final_clip = mp.CompositeVideoClip([img_clip, effect_clip.set_pos("center")])
|
40 |
else:
|
41 |
+
return "Error: Efecto no soportado."
|
|
|
|
|
|
|
42 |
|
43 |
+
# Agregar audio al video
|
44 |
+
final_clip = final_clip.set_audio(mp.AudioFileClip(audio_file))
|
45 |
+
output_path = "output.mp4"
|
46 |
+
final_clip.write_videofile(output_path, fps=24, codec="libx264", audio_codec="aac")
|
47 |
+
return output_path
|
|
|
|
|
|
|
|
|
|
|
48 |
|
49 |
except Exception as e:
|
50 |
return f"Error: {str(e)}"
|
51 |
|
52 |
+
# Interfaz de Gradio
|
|
|
|
|
|
|
|
|
53 |
iface = gr.Interface(
|
54 |
fn=audio_to_video,
|
55 |
inputs=[
|
56 |
+
gr.Audio(type="filepath", label="Subir Audio (WAV/MP3)"),
|
|
|
57 |
gr.Image(type="filepath", label="Subir Imagen"),
|
58 |
+
gr.Radio(["waveform"], value="waveform", label="Efecto Visual")
|
59 |
],
|
60 |
+
outputs=gr.Video(label="Video Generado"),
|
61 |
+
title="Audio + Imagen → Video con Efecto Sincronizado",
|
62 |
+
description="Sube un audio y una imagen para generar un video con efecto visual sincronizado (waveform)."
|
|
|
63 |
)
|
64 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
65 |
if __name__ == "__main__":
|
66 |
iface.queue().launch()
|