Recherche avancée

Sur d’autres sites (6639)

• How to handle differing .mp4 file types from different sources ?

  10 octobre 2017, par Dave502619

  If I take a .mp4 recorded on my mobile (Samsung S5) and pass it through FFmpeg with the below command, the output file (fileX.avi) is a greyscale bitmap uncompressed video file.

  • The offset values in fileX.avi (output from FFmpeg) to allow me to locate the video frame data are always 5680 bytes for the file header.

  • And 62 bytes for the inter frame header.

  • The data is uncompressed RGB24 so i can easily calculate the size of a video frame from height x width x 3.

  So my C# application can access the video frames in fileX.avi always at these above offsets.
  (This works great).

  My FFmpeg Command is :

  ffmpeg.exe -i source.mp4 -b 1150 -r 20.97 -g 120 -an -vf format=gray -f rawvideo -pixfmt gray -s 384x216 -vcodec rawvideo -y fileX.avi

  However... I recently took an .mp4 file from a different source (produced by Power Director 14 instead of direct from my mobile phone) and used this as the input source.mp4. But now the structure of fileX.avi differs as the offset values of 5680 + 62 bytes from the start in fileX.avi do not land me at the start of the video data frames.

  There seems to be different file formats for .mp4 - and obviously if there are my crude offset approach will not work for them all. I suspected at the time I wrote the code my method was all too easy a solution !

  So can anyone advise on the approach I should take now ? Should I check the original .mp4 or the output file (fileX.avi) to determine a "file type" to which I can determine the different offsets ?

  At the very least I need to be able to identify the "type" of .mp4 file that works so I can declare the type that will work with my software.

• lavc/h264dsp : optimise R-V V weight for shorter heights

  1er septembre 2024, par Rémi Denis-Courmont

  lavc/h264dsp : optimise R-V V weight for shorter heights
  The height is a power of two of up to 16 rows. The current code was
  
  optimised for large sample counts.
  T-Head C908 :
  
  h264_weight2_8_c :                                      211.7 ( 1.00x)
  
  h264_weight2_8_rvv_i32 :                   before       184.0 ( 1.15x)
  
  h264_weight2_8_rvv_i32 :                   after         54.2 ( 3.90x)
  
  h264_weight4_8_c :                                      285.7 ( 1.00x)
  
  h264_weight4_8_rvv_i32 :                   before       341.2 ( 0.86x)
  
  h264_weight4_8_rvv_i32 :                   after         82.2 ( 3.47x)
  
  h264_weight8_8_c :                                      498.7 ( 1.00x)
  
  h264_weight8_8_rvv_i32 :                   before       683.7 ( 0.73x)
  
  h264_weight8_8_rvv_i64 :                   after        128.5 ( 3.95x)
  
  h264_weight16_8_c :                                     878.2 ( 1.00x)
  
  h264_weight16_8_rvv_i32 :                  unchanged    239.5 ( 3.67x)
  SpacemiT X60 :
  
  h264_weight2_8_c :                                      207.2 ( 1.00x)
  
  h264_weight2_8_rvv_i32 :                   before       259.6 ( 0.80x)
  
  h264_weight2_8_rvv_i32 :                   after         82.2 ( 2.52x)
  
  h264_weight4_8_c :                                      290.8 ( 1.00x)
  
  h264_weight4_8_rvv_i32 :                   before       509.6 ( 0.57x)
  
  h264_weight4_8_rvv_i32 :                   after         61.5 ( 4.73x)
  
  h264_weight8_8_c :                                      498.8 ( 1.00x)
  
  h264_weight8_8_rvv_i32 :                   before      1019.8 ( 0.49x)
  
  h264_weight8_8_rvv_i64 :                   after         71.8 ( 6.95x)
  
  h264_weight16_8_c :                                     874.0 ( 1.00x)
  
  h264_weight16_8_rvv_i32 :                  unchanged    249.0 ( 3.51x)
  

  • [DH] libavcodec/riscv/h264dsp_init.c
  • [DH] libavcodec/riscv/h264dsp_rvv.S

• How to take snapshot with multiple web cameras at the same time using PHP in Centos ?

  13 novembre 2014, par galengodis

  I’m trying to take still photos / snapshots with multiple web cameras at the same time through php / shell_execute.

  This is what I use so I can the cameras in the background.

  shell_exec('ffmpeg -f video4linux2 -s 1280x960 -i /dev/video0 -q:v 0 -b:v 10000k -vcodec mjpeg -vframes 1 /var/www/html/cam1.jpg -y > /dev/null 2>/dev/null &');
  
  
  
  shell_exec('ffmpeg -f video4linux2 -s 1280x960 -i /dev/video1 -q:v 0 -b:v 10000k -vcodec mjpeg -vframes 1 /var/www/html/cam2.jpg -y > /dev/null 2>/dev/null &');

  It outputs only one image from the cameras. If i run them one at a time with the same code everything works. Fyi the "&" at the end makes the php run in background. Read more here about shell_exec background process : Is there a way to use shell_exec without waiting for the command to complete ?

  > [root@localhost ~]# lsusb 
  
  > Bus 001 Device 002: ID 8087:8000 Intel Corp.
  
  > Bus 002 Device 004: ID 1a40:0201 Terminus Technology Inc. FE 2.1
  
  > 7-port Hub Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root
  
  > hub Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus
  
  > 003 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub Bus 002
  
  > Device 006: ID 046d:0825 Logitech, Inc. Webcam C270 Bus 002 Device
  
  > 005: ID 0c45:6340 Microdia
  
  > 
  
  > [root@localhost ~]# find /dev/bus/ /dev/bus/ /dev/bus/usb
  
  > /dev/bus/usb/003 /dev/bus/usb/003/001 /dev/bus/usb/002
  
  > /dev/bus/usb/002/006 /dev/bus/usb/002/005 /dev/bus/usb/002/004
  
  > /dev/bus/usb/002/001 /dev/bus/usb/001 /dev/bus/usb/001/002
  
  > /dev/bus/usb/001/001

  I’ve tried with altering applications between ffmpeg and streamer. So the problem seems to be USB-related. The both cameras are plugged into a USB-hub (with an external power supply). The cameras are of different brands.

  I’m on Centos 7, 64bit.