Recherche avancée

Sur d’autres sites (11386)

• vulkan : temporarily disable threading for ASIC-based hwaccels

  9 août, par Lynne

  vulkan : temporarily disable threading for ASIC-based hwaccels
  The issue is that a race condition exists between threads locking
  
  frames, which results in cyclic loops and deadlocks.
  Compute-based implementations are not affected.
  Fixes #20169
  

  • [DH] libavcodec/vulkan_av1.c
  • [DH] libavcodec/vulkan_h264.c
  • [DH] libavcodec/vulkan_hevc.c
  • [DH] libavcodec/vulkan_vp9.c

• How to dynamically overlay images with ffmpeg

  23 mai, par Rorschy

  As a part of a bigger project, I'm trying to stream a live feed using ffmpeg through RTSP while also dynamically changing subtitles depending on the situation.

  


  As of now, I'm able to live stream with no issue. I also came across a solution for the subtitles by using a text file.

  


  However, I'd like to avoid having this text file in my project. I thought about creating a picture with the subtitles and overlaying it with the screen stream. However, with my current solution, the data is streamed only when I kill the running code (data streamed for a few seconds).

  


  Here is the current code for this problem :

  


  import subprocess
import threading
import string
import random
import time
import io
from PIL import Image, ImageDraw, ImageFont

RTSP_URL = "..."
ffmpeg = None

def generate_subtitle():
    width = 640
    height = 100
    font_size = 32
    while True:
        if ffmpeg:
            try:
                text = ''.join(random.choice(string.ascii_uppercase + string.digits) for _ in range(10))

                image = Image.new("RGBA", (width, height), (0, 0, 0, 128))
                draw = ImageDraw.Draw(image)

                try:
                    font = ImageFont.truetype("arial.ttf", font_size)
                except IOError:
                    font = ImageFont.load_default()

                bbox = draw.textbbox((0, 0), text, font=font)
                text_width = bbox[2] - bbox[0]
                text_height = bbox[3] - bbox[1]

                x = (width - text_width) // 2
                y = (height - text_height) // 2

                draw.text((x, y), text, font=font, fill=(255, 255, 255, 255))
                buffer = io.BytesIO()
                image.save(buffer, format="PNG")
                ffmpeg.stdin.write(buffer.getvalue())
                ffmpeg.stdin.flush()
                time.sleep(5)
            except Exception as e:
                print("Erreur d'envoi d'image :", e)
                break
        else:
            time.sleep(1)

def run_ffmpeg():
    global ffmpeg
    ffmpeg = subprocess.Popen([
        'ffmpeg',

        # Input 0: capture desktop
        "-f", "gdigrab",
        "-offset_x", "0",
        "-offset_y", "0",
        "-video_size", "1920x1080",
        "-i", "desktop",

        # Input 1: PNG overlay from stdin
        "-f", "image2pipe",
        "-vcodec", "png",
        "-i", "-",

        # Filter to overlay Input 1 on Input 0
        "-filter_complex", "[0:v][1:v]overlay=(main_w-overlay_w)/2:(main_h-overlay_h)-10",

        # Output settings
        "-vcodec", "libx264",
        "-preset", "ultrafast",
        "-tune", "zerolatency",
        "-g", "30",
        "-sc_threshold", "0",
        "-f", "rtsp",
        RTSP_URL
    ], stdin=subprocess.PIPE)

threading.Thread(target=run_ffmpeg, daemon=True).start()

threading.Thread(target=generate_subtitle, daemon=True).start()

while True:
    time.sleep(1)



  


  My question is how can I stream the data correctly ?
If there is another solution to change dynamically the subtitles without using a text file or a temporary file I'd be glad to hear it.

  


• avcodec/mpegvideo_dec : Move memcpy'ing ctx to mpeg4videodec.c

  29 avril, par Andreas Rheinhardt

  avcodec/mpegvideo_dec : Move memcpy'ing ctx to mpeg4videodec.c
  When the destination MpegEncContext in ff_mpeg_update_thread_context()
  
  is not initialized, the source MpegEncContext is simply copied
  
  over it before (potentially) calling ff_mpv_common_init().
  
  This leads to data races when this code is executed which is why
  
  it should be replaced with only copying the necessary fields
  
  (this is for future commits).
  Given that the RV30 and RV40 decoders always call said function
  
  with an already initialized MpegEncContext (they use context_reinit
  
  in case of frame size changes), they don't need this ugly
  
  initialization (and are therefore race-free). This means that
  
  this code can be moved to the only decoder that actually needs it :
  
  MPEG-4. This commit does so.
  Signed-off-by : Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
  

  • [DH] libavcodec/mpeg4videodec.c
  • [DH] libavcodec/mpegvideo_dec.c