Update app.py

app.py

@@ -34,13 +34,6 @@ def text_to_video(text, voice, rate, pitch, video_width, video_height, bg_image,
     # 字体文件路径
     font_path = os.path.abspath(text_font)
 
-    # 读取背景图像
-    if bg_image is not None:
-        bg_image_path = bg_image.name
-        bg_clip = ImageClip(bg_image_path).set_duration(0)  # Create a static clip
-    else:
-        bg_clip = ColorClip((video_width, video_height), color=(255, 255, 255))  # Default white background
-
     # 计算每页可以容纳的行数和每行可以容纳的字符数
     max_chars_per_line = video_width // (text_size // 2)  # 字体宽度假设为字体大小的一半
     max_lines_per_page = video_height // (text_size + 15)  # 10是行间距
@@ -68,6 +61,7 @@ def text_to_video(text, voice, rate, pitch, video_width, video_height, bg_image,
     # 为每页生成独立音频
     audio_clips = []
     video_clips = []
+    
     for i, page in enumerate(pages):
         # 将每页的文本连贯朗读生成一个音频文件
         audio_text = page.replace("\n", " ")  # 移除换行符以防止 TTS 停顿
@@ -98,21 +92,29 @@ def text_to_video(text, voice, rate, pitch, video_width, video_height, bg_image,
                 img.format = 'png'
                 img_path = os.path.join(tempfile.gettempdir(), f"page_{i}.png")
                 img.save(filename=img_path)
-
                 text_clip = ImageClip(img_path).set_duration(audio_clip.duration).set_audio(audio_clip)
-                text_clip = text_clip.set_position('center').set_duration(audio_clip.duration)
-                final_clip = CompositeVideoClip([bg_clip, text_clip])  # Overlay text on background
-                video_clips.append(final_clip)
+                video_clips.append(text_clip)
 
     # 合并所有视频片段
     final_video = concatenate_videoclips(video_clips)
+
+    # If a background image is provided, create a CompositeVideoClip
+    if bg_image:
+        bg_clip = ImageClip(bg_image).set_duration(final_video.duration).resize(newsize=(video_width, video_height))
+        final_video = CompositeVideoClip([bg_clip, final_video.set_position("center")])
+
     final_video_path = os.path.join(tempfile.gettempdir(), "output_video.mp4")
     final_video.write_videofile(final_video_path, fps=24, codec="libx264")
     return final_video_path, None
 
 # Gradio接口函数
 def tts_interface(text, voice, rate, pitch, video_width, video_height, bg_image, text_color, text_font, text_size):
-    video, warning = text_to_video(text, voice, rate, pitch, video_width, video_height, bg_image, text_color, text_font, text_size)
+    # Check if a background image is uploaded
+    if bg_image is not None:
+        video, warning = text_to_video(text, voice, rate, pitch, video_width, video_height, bg_image.name, text_color, text_font, text_size)
+    else:
+        # Pass None for the bg_image if not uploaded
+        video, warning = text_to_video(text, voice, rate, pitch, video_width, video_height, None, text_color, text_font, text_size)
     return None, video, warning
 
 # 创建Gradio应用
@@ -128,7 +130,7 @@ async def create_demo():
             gr.Slider(minimum=-20, maximum=20, value=0, label="Pitch Adjustment (Hz)", step=1),
             gr.Slider(minimum=640, maximum=1920, value=1080, label="Video Width", step=10),
             gr.Slider(minimum=480, maximum=1080, value=720, label="Video Height", step=10),
-            gr.File(label="Upload Background Image or Video", type="file", optional=True),
+            gr.File(label="Upload Background Image or Video", type="file"),
             gr.ColorPicker(value="#FFFFFF", label="Text Color"),
             gr.Textbox(label="Text Font", value="msyh.ttf"),  # 请确保字体文件路径正确
             gr.Slider(minimum=10, maximum=100, value=24, label="Text Size", step=1)