File size: 20,878 Bytes
2cc7a05 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 | import threading
import queue
from typing import TYPE_CHECKING, Dict, Tuple
import time
import subprocess
from pathlib import Path
import os
import gc
from functools import partial
import cv2
import numpy
import torch
import pyvirtualcam
from PySide6.QtCore import QObject, QTimer, Signal, Slot
from PySide6.QtGui import QPixmap
from app.processors.workers.frame_worker import FrameWorker
from app.ui.widgets.actions import graphics_view_actions
from app.ui.widgets.actions import common_actions as common_widget_actions
from app.ui.widgets.actions import video_control_actions
from app.ui.widgets.actions import layout_actions
import app.helpers.miscellaneous as misc_helpers
if TYPE_CHECKING:
from app.ui.main_ui import MainWindow
class VideoProcessor(QObject):
frame_processed_signal = Signal(int, QPixmap, numpy.ndarray)
webcam_frame_processed_signal = Signal(QPixmap, numpy.ndarray)
single_frame_processed_signal = Signal(int, QPixmap, numpy.ndarray)
def __init__(self, main_window: 'MainWindow', num_threads=2):
super().__init__()
self.main_window = main_window
self.frame_queue = queue.Queue(maxsize=num_threads)
self.media_capture: cv2.VideoCapture|None = None
self.file_type = None
self.fps = 0
self.processing = False
self.current_frame_number = 0
self.max_frame_number = 0
self.media_path = None
self.num_threads = num_threads
self.threads: Dict[int, threading.Thread] = {}
self.current_frame: numpy.ndarray = []
self.recording = False
self.virtcam: pyvirtualcam.Camera|None = None
self.recording_sp: subprocess.Popen|None = None
self.temp_file = ''
#Used to calculate the total processing time
self.start_time = 0.0
self.end_time = 0.0
#Used to store the video start and enc seek time
self.play_start_time = 0.0
self.play_end_time = 0.0
# Timer to manage frame reading intervals
self.frame_read_timer = QTimer()
self.frame_read_timer.timeout.connect(self.process_next_frame)
self.next_frame_to_display = 0
self.frame_processed_signal.connect(self.store_frame_to_display)
self.frame_display_timer = QTimer()
self.frame_display_timer.timeout.connect(self.display_next_frame)
self.frames_to_display: Dict[int, Tuple[QPixmap, numpy.ndarray]] = {}
self.webcam_frame_processed_signal.connect(self.store_webcam_frame_to_display)
self.webcam_frames_to_display = queue.Queue()
# Timer to update the gpu memory usage progressbar
self.gpu_memory_update_timer = QTimer()
self.gpu_memory_update_timer.timeout.connect(partial(common_widget_actions.update_gpu_memory_progressbar, main_window))
self.single_frame_processed_signal.connect(self.display_current_frame)
Slot(int, QPixmap, numpy.ndarray)
def store_frame_to_display(self, frame_number, pixmap, frame):
# print("Called store_frame_to_display()")
self.frames_to_display[frame_number] = (pixmap, frame)
# Use a queue to store the webcam frames, since the order of frames is not that important (Unless there are too many threads)
Slot(QPixmap, numpy.ndarray)
def store_webcam_frame_to_display(self, pixmap, frame):
# print("Called store_webcam_frame_to_display()")
self.webcam_frames_to_display.put((pixmap, frame))
Slot(int, QPixmap, numpy.ndarray)
def display_current_frame(self, frame_number, pixmap, frame):
if self.main_window.loading_new_media:
graphics_view_actions.update_graphics_view(self.main_window, pixmap, frame_number, reset_fit=True)
self.main_window.loading_new_media = False
else:
graphics_view_actions.update_graphics_view(self.main_window, pixmap, frame_number,)
self.current_frame = frame
torch.cuda.empty_cache()
#Set GPU Memory Progressbar
common_widget_actions.update_gpu_memory_progressbar(self.main_window)
def display_next_frame(self):
if not self.processing or (self.next_frame_to_display > self.max_frame_number):
self.stop_processing()
if self.next_frame_to_display not in self.frames_to_display:
return
else:
pixmap, frame = self.frames_to_display.pop(self.next_frame_to_display)
self.current_frame = frame
# Check and send the frame to virtualcam, if the option is selected
self.send_frame_to_virtualcam(frame)
if self.recording:
self.recording_sp.stdin.write(frame.tobytes())
# Update the widget values using parameters if it is not recording (The updation of actual parameters is already done inside the FrameWorker, this step is to make the changes appear in the widgets)
if not self.recording:
video_control_actions.update_widget_values_from_markers(self.main_window, self.next_frame_to_display)
graphics_view_actions.update_graphics_view(self.main_window, pixmap, self.next_frame_to_display)
self.threads.pop(self.next_frame_to_display)
self.next_frame_to_display += 1
def display_next_webcam_frame(self):
# print("Called display_next_webcam_frame()")
if not self.processing:
self.stop_processing()
if self.webcam_frames_to_display.empty():
# print("No Webcam frame found to display")
return
else:
pixmap, frame = self.webcam_frames_to_display.get()
self.current_frame = frame
self.send_frame_to_virtualcam(frame)
graphics_view_actions.update_graphics_view(self.main_window, pixmap, 0)
def send_frame_to_virtualcam(self, frame: numpy.ndarray):
if self.main_window.control['SendVirtCamFramesEnableToggle'] and self.virtcam:
# Check if the dimensions of the frame matches that of the Virtcam object
# If it doesn't match, reinstantiate the Virtcam object with new dimensions
height, width, _ = frame.shape
if self.virtcam.height!=height or self.virtcam.width!=width:
self.enable_virtualcam()
try:
self.virtcam.send(frame)
self.virtcam.sleep_until_next_frame()
except Exception as e:
print(e)
def set_number_of_threads(self, value):
self.stop_processing()
self.main_window.models_processor.set_number_of_threads(value)
self.num_threads = value
self.frame_queue = queue.Queue(maxsize=self.num_threads)
print(f"Max Threads set as {value} ")
def process_video(self):
"""Start video processing by reading frames and enqueueing them."""
if self.processing:
print("Processing already in progress. Ignoring start request.")
return
# Re-initialize the timers
self.frame_display_timer = QTimer()
self.frame_read_timer = QTimer()
if self.file_type == 'video':
self.frame_display_timer.timeout.connect(self.display_next_frame)
self.frame_read_timer.timeout.connect(self.process_next_frame)
if self.media_capture and self.media_capture.isOpened():
print("Starting video processing.")
if self.recording:
layout_actions.disable_all_parameters_and_control_widget(self.main_window)
self.start_time = time.perf_counter()
self.processing = True
self.frames_to_display.clear()
self.threads.clear()
if self.recording:
self.create_ffmpeg_subprocess()
self.play_start_time = float(self.media_capture.get(cv2.CAP_PROP_POS_FRAMES) / float(self.fps))
if self.main_window.control['VideoPlaybackCustomFpsToggle']:
fps = self.main_window.control['VideoPlaybackCustomFpsSlider']
else:
fps = self.media_capture.get(cv2.CAP_PROP_FPS)
interval = 1000 / fps if fps > 0 else 30
interval = int(interval * 0.8) #Process 20% faster to offset the frame loading & processing time so the video will be played close to the original fps
print(f"Starting frame_read_timer with an interval of {interval} ms.")
if self.recording:
self.frame_read_timer.start()
self.frame_display_timer.start()
else:
self.frame_read_timer.start(interval)
self.frame_display_timer.start()
self.gpu_memory_update_timer.start(5000) #Update GPU memory progressbar every 5 Seconds
else:
print("Error: Unable to open the video.")
self.processing = False
self.frame_read_timer.stop()
video_control_actions.set_play_button_icon_to_play(self.main_window)
#
elif self.file_type == 'webcam':
print("Calling process_video() on Webcam stream")
self.processing = True
self.frames_to_display.clear()
self.threads.clear()
fps = self.media_capture.get(cv2.CAP_PROP_FPS)
interval = 1000 / fps if fps > 0 else 30
interval = int(interval * 0.8) #Process 20% faster to offset the frame loading & processing time so the video will be played close to the original fps
self.frame_read_timer.timeout.connect(self.process_next_webcam_frame)
self.frame_read_timer.start(interval)
self.frame_display_timer.timeout.connect(self.display_next_webcam_frame)
self.frame_display_timer.start()
self.gpu_memory_update_timer.start(5000) #Update GPU memory progressbar every 5 Seconds
def process_next_frame(self):
"""Read the next frame and add it to the queue for processing."""
if self.current_frame_number > self.max_frame_number:
# print("Stopping frame_read_timer as all frames have been read!")
self.frame_read_timer.stop()
return
if self.frame_queue.qsize() >= self.num_threads:
# print(f"Queue is full ({self.frame_queue.qsize()} frames). Throttling frame reading.")
return
if self.file_type == 'video' and self.media_capture:
ret, frame = misc_helpers.read_frame(self.media_capture, preview_mode = not self.recording)
if ret:
frame = frame[..., ::-1] # Convert BGR to RGB
# print(f"Enqueuing frame {self.current_frame_number}")
self.frame_queue.put(self.current_frame_number)
self.start_frame_worker(self.current_frame_number, frame)
self.current_frame_number += 1
else:
print("Cannot read frame!", self.current_frame_number)
self.stop_processing()
self.main_window.display_messagebox_signal.emit('Error Reading Frame', f'Error Reading Frame {self.current_frame_number}.\n Stopped Processing...!', self.main_window)
def start_frame_worker(self, frame_number, frame, is_single_frame=False):
"""Start a FrameWorker to process the given frame."""
worker = FrameWorker(frame, self.main_window, frame_number, self.frame_queue, is_single_frame)
self.threads[frame_number] = worker
if is_single_frame:
worker.run()
else:
worker.start()
def process_current_frame(self):
# print("\nCalled process_current_frame()",self.current_frame_number)
# self.main_window.processed_frames.clear()
self.next_frame_to_display = self.current_frame_number
if self.file_type == 'video' and self.media_capture:
ret, frame = misc_helpers.read_frame(self.media_capture, preview_mode=False)
if ret:
frame = frame[..., ::-1] # Convert BGR to RGB
# print(f"Enqueuing frame {self.current_frame_number}")
self.frame_queue.put(self.current_frame_number)
self.start_frame_worker(self.current_frame_number, frame, is_single_frame=True)
self.media_capture.set(cv2.CAP_PROP_POS_FRAMES, self.current_frame_number)
else:
print("Cannot read frame!", self.current_frame_number)
self.main_window.display_messagebox_signal.emit('Error Reading Frame', f'Error Reading Frame {self.current_frame_number}.', self.main_window)
# """Process a single image frame directly without queuing."""
elif self.file_type == 'image':
frame = misc_helpers.read_image_file(self.media_path)
if frame is not None:
frame = frame[..., ::-1] # Convert BGR to RGB
self.frame_queue.put(self.current_frame_number)
# print("Processing current frame as image.")
self.start_frame_worker(self.current_frame_number, frame, is_single_frame=True)
else:
print("Error: Unable to read image file.")
# Handle webcam capture
elif self.file_type == 'webcam':
ret, frame = misc_helpers.read_frame(self.media_capture, preview_mode = False)
if ret:
frame = frame[..., ::-1] # Convert BGR to RGB
# print(f"Enqueuing frame {self.current_frame_number}")
self.frame_queue.put(self.current_frame_number)
self.start_frame_worker(self.current_frame_number, frame, is_single_frame=True)
else:
print("Unable to read Webcam frame!")
self.join_and_clear_threads()
def process_next_webcam_frame(self):
# print("Called process_next_webcam_frame()")
if self.frame_queue.qsize() >= self.num_threads:
# print(f"Queue is full ({self.frame_queue.qsize()} frames). Throttling frame reading.")
return
if self.file_type == 'webcam' and self.media_capture:
ret, frame = misc_helpers.read_frame(self.media_capture, preview_mode = False)
if ret:
frame = frame[..., ::-1] # Convert BGR to RGB
# print(f"Enqueuing frame {self.current_frame_number}")
self.frame_queue.put(self.current_frame_number)
self.start_frame_worker(self.current_frame_number, frame)
# @misc_helpers.benchmark
def stop_processing(self):
"""Stop video processing and signal completion."""
if not self.processing:
# print("Processing not active. No action to perform.")
video_control_actions.reset_media_buttons(self.main_window)
return False
print("Stopping video processing.")
self.processing = False
if self.file_type=='video' or self.file_type=='webcam':
# print("Stopping Timers")
self.frame_read_timer.stop()
self.frame_display_timer.stop()
self.gpu_memory_update_timer.stop()
self.join_and_clear_threads()
# print("Clearing Threads and Queues")
self.threads.clear()
self.frames_to_display.clear()
self.webcam_frames_to_display.queue.clear()
with self.frame_queue.mutex:
self.frame_queue.queue.clear()
self.current_frame_number = self.main_window.videoSeekSlider.value()
self.media_capture.set(cv2.CAP_PROP_POS_FRAMES, self.current_frame_number)
if self.recording and self.file_type=='video':
self.recording_sp.stdin.close()
self.recording_sp.wait()
self.play_end_time = float(self.media_capture.get(cv2.CAP_PROP_POS_FRAMES) / float(self.fps))
if self.file_type=='video':
if self.recording:
final_file_path = misc_helpers.get_output_file_path(self.media_path, self.main_window.control['OutputMediaFolder'])
if Path(final_file_path).is_file():
os.remove(final_file_path)
print("Adding audio...")
args = ["ffmpeg",
'-hide_banner',
'-loglevel', 'error',
"-i", self.temp_file,
"-ss", str(self.play_start_time), "-to", str(self.play_end_time), "-i", self.media_path,
"-c", "copy", # may be c:v
"-map", "0:v:0", "-map", "1:a:0?",
"-shortest",
final_file_path]
subprocess.run(args, check=False) #Add Audio
os.remove(self.temp_file)
self.end_time = time.perf_counter()
processing_time = self.end_time - self.start_time
print(f"\nProcessing completed in {processing_time} seconds")
avg_fps = ((self.play_end_time - self.play_start_time) * self.fps) / processing_time
print(f'Average FPS: {avg_fps}\n')
if self.recording:
layout_actions.enable_all_parameters_and_control_widget(self.main_window)
self.recording = False #Set recording as False to make sure the next process_video() call doesnt not record the video, unless the user press the record button
print("Clearing Cache")
torch.cuda.empty_cache()
gc.collect()
video_control_actions.reset_media_buttons(self.main_window)
print("Successfully Stopped Processing")
return True
def join_and_clear_threads(self):
# print("Joining Threads")
for _, thread in self.threads.items():
if thread.is_alive():
thread.join()
# print('Clearing Threads')
self.threads.clear()
def create_ffmpeg_subprocess(self):
# Use Dimensions of the last processed frame as it could be different from the original frame due to restorers and frame enhancers
frame_height, frame_width, _ = self.current_frame.shape
self.temp_file = r'temp_output.mp4'
if Path(self.temp_file).is_file():
os.remove(self.temp_file)
args = [
"ffmpeg",
"-hide_banner",
"-loglevel", "error",
"-f", "rawvideo", # Specify raw video input
"-pix_fmt", "bgr24", # Pixel format of input frames
"-s", f"{frame_width}x{frame_height}", # Frame resolution
"-r", str(self.fps), # Frame rate
"-i", "pipe:", # Input from stdin
"-vf", f"pad=ceil(iw/2)*2:ceil(ih/2)*2,format=yuvj420p", # Padding and format conversion
"-c:v", "libx264", # H.264 codec
"-crf", "18", # Quality setting
self.temp_file # Output file
]
self.recording_sp = subprocess.Popen(args, stdin=subprocess.PIPE)
def enable_virtualcam(self, backend=False):
#Check if capture contains any cv2 stream or is it an empty list
if self.media_capture:
if isinstance(self.current_frame, numpy.ndarray):
frame_height, frame_width, _ = self.current_frame.shape
else:
frame_height = int(self.media_capture.get(cv2.CAP_PROP_FRAME_HEIGHT))
frame_width = int(self.media_capture.get(cv2.CAP_PROP_FRAME_WIDTH))
self.disable_virtualcam()
try:
backend = backend or self.main_window.control['VirtCamBackendSelection']
# self.virtcam = pyvirtualcam.Camera(width=vid_width, height=vid_height, fps=int(self.fps), backend='unitycapture', device='Unity Video Capture')
self.virtcam = pyvirtualcam.Camera(width=frame_width, height=frame_height, fps=int(self.fps), backend=backend, fmt=pyvirtualcam.PixelFormat.BGR)
except Exception as e:
print(e)
def disable_virtualcam(self):
if self.virtcam:
self.virtcam.close()
self.virtcam = None |