Recherche avancée

Sur d’autres sites (5069)

• FFmpeg vidstab generates jerky wiggling video

  27 mars 2022, par praet0ri4n

  I run Kdenlive 21.12.3 on Ubuntu 20.04.
Recently I stared editing footage from my Nikon D3100 DSLR. Source videos are 1920:1080 24fps in MOV format (camera set to PAL system).

  


  After trying to stabilize a video where I am walking I noticed that the result is kind of stabilized but the image is distorted and contains some strange wiggling effect. Changing the stabilization settings doesn't have any impact on this effect. I haven't seen this before. Previously I stabilized footage from other cameras without problems. This is the first time I use that DSLR.

  


  To me it looks like stabilization is doing some strange interpolation and it is stretching and skewing the image or pixel aspect ratio.
For convenience I tried the vidstab filter with ffmpeg, played around with the interpol option too, but the result ist exactly the same.

  


  Here the source MOV and stabilized mp4 file for comparison
https://nc.p1p0.eu/index.php/s/tTQgPj9rDbmReYn

  


  Thank you !

  


• How to save last 30 seconds of video in py

  5 juin 2024, par Mateus Coelho

  I want the last 30 seconds to be recorded every time I click enter and sent to the cloud. for example, if I click at 00:10:30, I want a video that records from 00:10:00 to 00:10:30 and if I click in sequence at 00:10:32, I need another different video that in its content is recorded from 00:10:02 to 00:10:32.

  


  I think I have a problem where I will always end up recovering from the same buffer in the last 30 seconds. Is there any approach so that whenever I click enter I retrieve a unique video ? Is my current approach the best for the problem ? Or should I use something else ?

  


  import subprocess
import os
import keyboard
from datetime import datetime
from google.cloud import storage

# Configuration
STATE = "mg"
CITY = "belohorizonte"
COURT = "duna"
RTSP_URL = "rtsp://Apertai:130355va@192.168.0.2/stream1"
BUCKET_NAME = "apertai-cloud"
CREDENTIALS_PATH = "C:/Users/Abidu/ApertAI/key.json"

def start_buffer_stream():
    # Command for continuous buffer that overwrites itself every 30 seconds
    buffer_command = [
        'ffmpeg',
        '-i', RTSP_URL,
        '-map', '0',
        '-c', 'copy',
        '-f', 'segment',
        '-segment_time', '30',  # Duration of each segment
        '-segment_wrap', '2',  # Number of segments to wrap around
        '-reset_timestamps', '1',  # Reset timestamps at the start of each segment
        'buffer-%03d.ts'  # Save segments with a numbering pattern
    ]
    return subprocess.Popen(buffer_command, stdout=subprocess.PIPE, stderr=subprocess.PIPE)

def save_last_30_seconds_from_buffer(buffer_file):
    datetime_now = datetime.now()
    datetime_now_formatted = f"{datetime_now.day:02}{datetime_now.month:02}{datetime_now.year}-{datetime_now.hour:02}{datetime_now.minute:02}"
    output_file_name = os.path.abspath(f"{STATE}-{CITY}-{COURT}-{datetime_now_formatted}.mp4")

    # Copy the most recent buffer segment to the output file
    save_command = [
        'ffmpeg',
        '-i', buffer_file,
        '-c', 'copy',
        output_file_name
    ]
    subprocess.run(save_command, check=True)
    print(f"Saved last 30 seconds: {output_file_name}")
    return output_file_name

def upload_to_google_cloud(file_name):
    client = storage.Client.from_service_account_json(CREDENTIALS_PATH)
    bucket = client.bucket(BUCKET_NAME)
    blob = bucket.blob(os.path.basename(file_name).replace("-", "/"))
    blob.upload_from_filename(file_name, content_type='application/octet-stream')
    print(f"Uploaded {file_name} to {BUCKET_NAME}")
    os.remove(file_name)  # Clean up the local file

def main():
    print("Starting continuous buffer for RTSP stream...")
    start_time = datetime.now()
    buffer_process = start_buffer_stream()
    print("Press 'Enter' to save the last 30 seconds of video...")

    while True:
        # Verify if 30 seconds has passed since start
        if keyboard.is_pressed('enter'):
            print("Saving last 30 seconds of video...")
            elapsed_time = (datetime.now() - start_time).total_seconds()
            # Determine which buffer segment to save
            if elapsed_time % 60 < 30:
                buffer_file = 'buffer-000.ts'
            else:
                buffer_file = 'buffer-001.ts'
            final_video = save_last_30_seconds_from_buffer(buffer_file)
            upload_to_google_cloud(final_video)

if _name_ == "_main_":
    main()




  


• avutil/threadmessage : split the pthread condition in two

  1er décembre 2015, par Clément Bœsch

  avutil/threadmessage : split the pthread condition in two
  Fix a dead lock under certain conditions. Let’s assume we have a queue of 1
  
  message max, 2 senders, and 1 receiver.
  Scenario (real record obtained with debug added) :
  
      [...]
  
      SENDER #0 : acquired lock
  
      SENDER #0 : queue is full, wait
  
      SENDER #1 : acquired lock
  
      SENDER #1 : queue is full, wait
  
      RECEIVER : acquired lock
  
      RECEIVER : reading a msg from the queue
  
      RECEIVER : signal the cond
  
      RECEIVER : acquired lock
  
      RECEIVER : queue is empty, wait
  
      SENDER #0 : writing a msg the queue
  
      SENDER #0 : signal the cond
  
      SENDER #0 : acquired lock
  
      SENDER #0 : queue is full, wait
  
      SENDER #1 : queue is full, wait
  Translated :
  
   - initially the queue contains 1/1 message with 2 senders blocking on
  
     it, waiting to push another message.
  
   - Meanwhile the receiver is obtaining the lock, read the message,
  
     signal & release the lock. For some reason it is able to acquire the
  
     lock again before the signal wakes up one of the sender. Since it
  
     just emptied the queue, the reader waits for the queue to fill up
  
     again.
  
   - The signal finally reaches one of the sender, which writes a message
  
     and then signal the condition. Unfortunately, instead of waking up
  
     the reader, it actually wakes up the other worker (signal = notify
  
     the condition just for 1 waiter), who can’t push another message in
  
     the queue because it’s full.
  
   - Meanwhile, the receiver is still waiting. Deadlock.
  This scenario can be triggered with for example :
  
      tests/api/api-threadmessage-test 1 2 100 100 1 1000 1000
  One working solution is to make av_thread_message_queue_send,recv()
  
  call pthread_cond_broadcast() instead of pthread_cond_signal() so both
  
  senders and receivers are unlocked when work is done (be it reading or
  
  writing).
  This second solution replaces the condition with two : one to notify the
  
  senders, and one to notify the receivers. This prevents senders from
  
  notifying other senders instead of a reader, and the other way around.
  
  It also avoid broadcasting to everyone like the first solution, and is,
  
  as a result in theory more optimized.
  

  • [DH] libavutil/threadmessage.c