Introduce trim frame by start and end

roop/core.py

@@ -39,6 +39,8 @@ def parse_args() -> None:
     program.add_argument('--reference-face-position', help='position of the reference face', dest='reference_face_position', type=int, default=0)
     program.add_argument('--reference-frame-number', help='number of the reference frame', dest='reference_frame_number', type=int, default=0)
     program.add_argument('--similar-face-distance', help='face distance used for recognition', dest='similar_face_distance', type=float, default=0.85)
+    program.add_argument('--trim-frame-start', help='start frame use for extraction', dest='trim_frame_start', type=int)
+    program.add_argument('--trim-frame-end', help='end frame use for extraction', dest='trim_frame_end', type=int)
     program.add_argument('--temp-frame-format', help='image format used for frame extraction', dest='temp_frame_format', default='png', choices=['jpg', 'png'])
     program.add_argument('--temp-frame-quality', help='image quality used for frame extraction', dest='temp_frame_quality', type=int, default=0, choices=range(101), metavar='[0-100]')
     program.add_argument('--output-video-encoder', help='encoder used for the output video', dest='output_video_encoder', default='libx264', choices=['libx264', 'libx265', 'libvpx-vp9', 'h264_nvenc', 'hevc_nvenc'])
@@ -64,6 +66,8 @@ def parse_args() -> None:
     roop.globals.reference_face_position = args.reference_face_position
     roop.globals.reference_frame_number = args.reference_frame_number
     roop.globals.similar_face_distance = args.similar_face_distance
+    roop.globals.trim_frame_start = args.trim_frame_start
+    roop.globals.trim_frame_end = args.trim_frame_end
     roop.globals.temp_frame_format = args.temp_frame_format
     roop.globals.temp_frame_quality = args.temp_frame_quality
     roop.globals.output_video_encoder = args.output_video_encoder

roop/globals.py

@@ -13,6 +13,8 @@
 reference_face_position: Optional[int] = None
 reference_frame_number: Optional[int] = None
 similar_face_distance: Optional[float] = None
+trim_frame_start: Optional[int] = None
+trim_frame_end: Optional[int] = None
 temp_frame_format: Optional[str] = None
 temp_frame_quality: Optional[int] = None
 output_video_encoder: Optional[str] = None

roop/uis/__components__/reference.py → roop/uis/__components__/face_selector.py

@@ -68,10 +68,10 @@ def update_face_reference_position(reference_face_position: int) -> Update:
     roop.globals.reference_face_position = reference_face_position
     if is_image(roop.globals.target_path):
         target_frame = cv2.imread(roop.globals.target_path)
-        maximum = get_faces_total(target_frame)
+        maximum = max(get_faces_total(target_frame) - 1, 0)
     if is_video(roop.globals.target_path):
         temp_frame = get_video_frame(roop.globals.target_path, roop.globals.reference_frame_number)
-        maximum = get_faces_total(temp_frame)
+        maximum = max(get_faces_total(temp_frame) - 1, 0)
     return gradio.update(value=reference_face_position, maximum=maximum)
 
 

roop/uis/__components__/preview.py

@@ -70,14 +70,14 @@ def update(frame_number: int = 0) -> Tuple[Update, Update]:
     if is_image(roop.globals.target_path):
         target_frame = cv2.imread(roop.globals.target_path)
         preview_frame = get_preview_frame(target_frame)
-        return gradio.update(value=ui.normalize_frame(preview_frame)), gradio.update(value=0, maximum=1, visible=False)
+        return gradio.update(value=ui.normalize_frame(preview_frame)), gradio.update(value=None, maximum=None, visible=False)
     if is_video(roop.globals.target_path):
         roop.globals.reference_frame_number = frame_number
         video_frame_total = get_video_frame_total(roop.globals.target_path)
         temp_frame = get_video_frame(roop.globals.target_path, roop.globals.reference_frame_number)
         preview_frame = get_preview_frame(temp_frame)
         return gradio.update(value=ui.normalize_frame(preview_frame)), gradio.update(maximum=video_frame_total, visible=True)
-    return gradio.update(value=None), gradio.update(value=0, maximum=1, visible=False)
+    return gradio.update(value=None), gradio.update(value=None, maximum=None, visible=False)
 
 
 def get_preview_frame(temp_frame: Frame) -> Frame:

roop/uis/__components__/trim_frame.py

@@ -0,0 +1,59 @@
+from time import sleep
+from typing import Any, Dict, Tuple, Optional
+
+import gradio
+
+import roop.globals
+from roop.capturer import get_video_frame_total
+from roop.uis import core as ui
+from roop.uis.typing import Update
+from roop.utilities import is_video
+
+TRIM_FRAME_START_NUMBER: Optional[gradio.Number] = None
+TRIM_FRAME_END_NUMBER: Optional[gradio.Number] = None
+
+
+def render() -> None:
+    global TRIM_FRAME_START_NUMBER
+    global TRIM_FRAME_END_NUMBER
+
+    with gradio.Box():
+        trim_frame_start_number_args: Dict[str, Any] = {
+            'label': 'TRIM FRAME START',
+            'value': roop.globals.trim_frame_start,
+            'visible': False
+        }
+        trim_frame_end_number_args: Dict[str, Any] = {
+            'label': 'TRIM FRAME END',
+            'value': roop.globals.trim_frame_end,
+            'visible': False
+        }
+        if is_video(roop.globals.target_path):
+            trim_frame_start_number_args['visible'] = True
+            trim_frame_end_number_args['visible'] = True
+        with gradio.Row():
+            TRIM_FRAME_START_NUMBER = gradio.Number(**trim_frame_start_number_args)
+            TRIM_FRAME_END_NUMBER = gradio.Number(**trim_frame_end_number_args)
+
+
+def listen() -> None:
+    target_file = ui.get_component('target_file')
+    if target_file:
+        target_file.change(remote_update, outputs=[TRIM_FRAME_START_NUMBER, TRIM_FRAME_END_NUMBER])
+    TRIM_FRAME_START_NUMBER.change(lambda value: update_number('trim_frame_start', int(value)), inputs=TRIM_FRAME_START_NUMBER, outputs=TRIM_FRAME_START_NUMBER)
+    TRIM_FRAME_END_NUMBER.change(lambda value: update_number('trim_frame_end', int(value)), inputs=TRIM_FRAME_END_NUMBER, outputs=TRIM_FRAME_END_NUMBER)
+
+
+def remote_update() -> Tuple[Update, Update]:
+    sleep(0.5)
+    if is_video(roop.globals.target_path):
+        video_frame_total = get_video_frame_total(roop.globals.target_path)
+        roop.globals.trim_frame_start = 0
+        roop.globals.trim_frame_end = video_frame_total
+        return gradio.update(value=0, maximum=video_frame_total, visible=True), gradio.update(value=video_frame_total, maximum=video_frame_total, visible=True)
+    return gradio.update(value=None, maximum=None, visible=False), gradio.update(value=None, maximum=None, visible=False)
+
+
+def update_number(name: str, value: int) -> Update:
+    setattr(roop.globals, name, value)
+    return gradio.update(value=value)

roop/uis/__layouts__/default.py

@@ -1,6 +1,6 @@
 import gradio
 
-from roop.uis.__components__ import settings, source, target, preview, reference, output
+from roop.uis.__components__ import settings, source, target, preview, trim_frame, face_selector, output
 
 
 def render() -> gradio.Blocks:
@@ -13,7 +13,8 @@ def render() -> gradio.Blocks:
                 target.render()
             with gradio.Column(scale=3):
                 preview.render()
-                reference.render()
+                trim_frame.render()
+                face_selector.render()
         with gradio.Row():
             output.render()
     return layout
@@ -24,5 +25,6 @@ def listen() -> None:
     source.listen()
     target.listen()
     preview.listen()
-    reference.listen()
+    trim_frame.listen()
+    face_selector.listen()
     output.listen()

roop/uis/core.py

@@ -44,7 +44,7 @@ def get_theme() -> gradio.Theme:
     return gradio.themes.Soft(
         primary_hue=gradio.themes.colors.red,
         secondary_hue=gradio.themes.colors.gray,
-        font=gradio.themes.GoogleFont('Open Sans')
+        font=gradio.themes.GoogleFont('Inter')
     )
 
 

roop/utilities.py

@@ -43,7 +43,19 @@ def detect_fps(target_path: str) -> float:
 def extract_frames(target_path: str, fps: float = 30) -> bool:
     temp_directory_path = get_temp_directory_path(target_path)
     temp_frame_quality = roop.globals.temp_frame_quality * 31 // 100
-    return run_ffmpeg(['-hwaccel', 'auto', '-i', target_path, '-q:v', str(temp_frame_quality), '-pix_fmt', 'rgb24', '-vf', 'fps=' + str(fps), os.path.join(temp_directory_path, '%04d.' + roop.globals.temp_frame_format)])
+    trim_frame_start = roop.globals.trim_frame_start
+    trim_frame_end = roop.globals.trim_frame_end
+    commands = ['-hwaccel', 'auto', '-i', target_path, '-q:v', str(temp_frame_quality), '-pix_fmt', 'rgb24']
+    if trim_frame_start is not None and trim_frame_end is not None:
+        commands.extend(['-vf', f'trim=start_frame={trim_frame_start}:end_frame={trim_frame_end},fps={fps}'])
+    elif trim_frame_start is not None:
+        commands.extend(['-vf', f'trim=start_frame={trim_frame_start},fps={fps}'])
+    elif trim_frame_end is not None:
+        commands.extend(['-vf', f'trim=end_frame={trim_frame_end},fps={fps}'])
+    else:
+        commands.extend(['-vf', f'fps={fps}'])
+    commands.extend([os.path.join(temp_directory_path, '%04d.' + roop.globals.temp_frame_format)])
+    return run_ffmpeg(commands)
 
 
 def create_video(target_path: str, fps: float = 30) -> bool: