def save_mp4_from_vid_and_audio(self,
video_tensor,
audio_wav=None,
fps=25,
sr=16000,
outname=None,
extract_frames_hop=None):
"""
:param video_tensor: tchw
:param sr:
:return:
"""
from moviepy.audio.AudioClip import AudioArrayClip
from moviepy.video.VideoClip import VideoClip
video_tensor = video_tensor.transpose([0, 2, 3, 1]) # thwc
# that's to avoid error due to float precision
vid_dur = len(video_tensor) * (1. / fps) - 1e-6
v_clip = VideoClip(lambda t: video_tensor[int(np.round(t * 25))],
duration=vid_dur)
import tempfile
if outname:
outfile = os.path.join(self.savedir, outname)
if not outfile.endswith('.mp4'):
outfile += '.mp4'
else:
outfile = os.path.join(self.savedir, '%03d.mp4' % self.id)
if audio_wav is not None:
_, temp_audiofile = tempfile.mkstemp(dir='/dev/shm', suffix='.wav')
import torch
if isinstance(audio_wav, torch.Tensor):
audio_wav = audio_wav.numpy()
import scipy.io
scipy.io.wavfile.write(temp_audiofile, 16000, audio_wav)
self.id += 1
try:
os.makedirs(os.path.dirname(outfile))
except:
pass
_, temp_videofile = tempfile.mkstemp(dir='/dev/shm', suffix='.mp4')
v_clip.write_videofile(temp_videofile, fps=25, verbose=False)
if audio_wav is not None:
command = ("ffmpeg -threads 1 -loglevel error -y -i {} -i {} "
"-c:v copy -map 0:v:0 -map 1:a:0 -pix_fmt yuv420p "
"-shortest {}").format(temp_videofile, temp_audiofile,
outfile)
from subprocess import call
cmd = command.split(' ')
call(cmd)
else:
import shutil
shutil.move(temp_videofile, outfile)
v_clip.close()
import imageio
if extract_frames_hop: # extract the video as frames for paper
frames_dir = os.path.join(
os.path.dirname(outfile),
'frames_' + os.path.basename(outfile).replace('.mp4', ''))
os.makedirs(frames_dir, exist_ok=True)
import scipy.misc
for fr_id, frame in enumerate(video_tensor[::extract_frames_hop]):
scipy.misc.imsave(frames_dir + '/%04d.png' % fr_id,
frame[:, :-5, :])
pass
if args.image_resize == 'fill' and args.crop_output:
print(f"Cropping output video to {unfill_width}x{unfill_height}")
output_clip = movie_crop(output_clip,
x_center=256 // 2,
y_center=256 // 2,
width=unfill_width,
height=unfill_height)
print("Saving Video...")
output_clip.write_videofile(output_video_path,
logger=None,
verbose=False,
**codecs[args.codec])
print("Video saved to", output_video_path)
print()
if args.stack:
print("Saving Side by Side Video")
stack_clip = clips_array([[output_clip.margin(right=10),
source_video]])
stack_path = output_video_path.rsplit('.', 1)[0] + '_stacked.mp4'
stack_clip.write_videofile(stack_path,
logger=None,
verbose=False,
**codecs[args.codec])
print("Stacked Video saved to", stack_path)
stack_clip.close()
output_clip.close()
source_video.close()
