0.50

images_to_video_path.py

@@ -5,6 +5,7 @@ import tempfile
 import torch
 import numpy as np
 from PIL import Image
+import wave
 
 class ImagesListToVideo:
     @classmethod
@@ -13,6 +14,10 @@ class ImagesListToVideo:
             "required": {
                 "images": ("IMAGE",),
                 "frames_per_second": ("FLOAT", {"default": 30, "min": 1, "max": 120, "step": 1}),
+            },
+            "optional": {
+                "audio_path": ("STRING", {"default": "", "multiline": False}),
+                "audio": ("AUDIO", {"default": None}),
             }
         }
     
@@ -21,7 +26,7 @@ class ImagesListToVideo:
     FUNCTION = "images_to_video"
     CATEGORY = "Bjornulf"
 
-    def images_to_video(self, images, frames_per_second=30):
+    def images_to_video(self, images, frames_per_second=30, audio_path="", audio=None):
         # Create the output directory if it doesn't exist
         output_dir = os.path.join("Bjornulf", "images_to_video")
         os.makedirs(output_dir, exist_ok=True)
@@ -30,42 +35,85 @@ class ImagesListToVideo:
         video_filename = f"video_{uuid.uuid4().hex}.mp4"
         video_path = os.path.join(output_dir, video_filename)
 
-        # Create a temporary directory to store image files
+        # Create a temporary directory to store image files and audio
         with tempfile.TemporaryDirectory() as temp_dir:
             # Save each image as a PNG file in the temporary directory
             for i, img in enumerate(images):
-                # Convert the image to the correct format
                 img_np = self.convert_to_numpy(img)
-                
-                # Ensure the image is in RGB format
                 if img_np.shape[-1] != 3:
                     img_np = self.convert_to_rgb(img_np)
-                
-                # Convert to PIL Image
                 img_pil = Image.fromarray(img_np)
                 img_path = os.path.join(temp_dir, f"frame_{i:05d}.png")
                 img_pil.save(img_path)
 
-            # Use FFmpeg to create a video from the image sequence
+            # Prepare FFmpeg command
             ffmpeg_cmd = [
                 "ffmpeg",
                 "-framerate", str(frames_per_second),
                 "-i", os.path.join(temp_dir, "frame_%05d.png"),
                 "-c:v", "libx264",
                 "-pix_fmt", "yuv420p",
-                "-crf", "23",
-                "-y",  # Overwrite output file if it exists
-                video_path
+                "-crf", "19"
             ]
 
-            try:
-                subprocess.run(ffmpeg_cmd, check=True, capture_output=True, text=True)
-            except subprocess.CalledProcessError as e:
-                print(f"FFmpeg error: {e.stderr}")
-                return ("",)  # Return empty string if video creation fails
+            # Handle audio
+            temp_audio_path = None
+            if audio is not None and isinstance(audio, dict):
+                waveform = audio['waveform'].numpy().squeeze()
+                sample_rate = audio['sample_rate']
+                temp_audio_path = os.path.join(temp_dir, "temp_audio.wav")
+                self.write_wav(temp_audio_path, waveform, sample_rate)
+            elif audio_path and os.path.isfile(audio_path):
+                temp_audio_path = audio_path
+
+            if temp_audio_path:
+                # Create temporary video without audio first
+                temp_video = os.path.join(temp_dir, "temp_video.mp4")
+                temp_cmd = ffmpeg_cmd + ["-y", temp_video]
+                
+                try:
+                    # Create video without audio
+                    subprocess.run(temp_cmd, check=True, capture_output=True, text=True)
+                    
+                    # Add audio to the video
+                    audio_cmd = [
+                        "ffmpeg",
+                        "-i", temp_video,
+                        "-i", temp_audio_path,
+                        "-c:v", "copy",
+                        "-c:a", "aac",
+                        "-shortest",
+                        "-y",
+                        video_path
+                    ]
+                    subprocess.run(audio_cmd, check=True, capture_output=True, text=True)
+                except subprocess.CalledProcessError as e:
+                    print(f"FFmpeg error: {e.stderr}")
+                    return ("",)
+            else:
+                # No audio, just create the video directly
+                ffmpeg_cmd.append("-y")
+                ffmpeg_cmd.append(video_path)
+                try:
+                    subprocess.run(ffmpeg_cmd, check=True, capture_output=True, text=True)
+                except subprocess.CalledProcessError as e:
+                    print(f"FFmpeg error: {e.stderr}")
+                    return ("",)
 
         return (video_path,)
 
+    def write_wav(self, file_path, audio_data, sample_rate):
+        with wave.open(file_path, 'wb') as wav_file:
+            wav_file.setnchannels(1)  # Mono
+            wav_file.setsampwidth(2)  # 2 bytes per sample
+            wav_file.setframerate(sample_rate)
+            
+            # Normalize and convert to 16-bit PCM
+            audio_data = np.int16(audio_data * 32767)
+            
+            # Write audio data
+            wav_file.writeframes(audio_data.tobytes())
+
     def convert_to_numpy(self, img):
         if isinstance(img, torch.Tensor):
             img = img.cpu().numpy()