Recherche avancée

Sur d’autres sites (12276)

• Yes or no, will ffmpeg api do hardware decoding on iOS ?

  15 janvier 2019, par Fattie

  There seems to be conflicting information on this.

  https://trac.ffmpeg.org/wiki/HWAccelIntro

  notice the first diagram, it firmly marks iOS as “Y” on VideoToolbox

  

  however in the comments down the bottom it says

  VideoToolbox. ​VideoToolbox, only supported on macOS. H.264 decoding is available in FFmpeg/libavcodec.

  And in the confusing second diagram it says "Standalone" is not done for VideoToolbox.

  We have found that using ffmpeg compiled in to iOS .... it seems to not use hardware decoding, which is really a pain.

  1. With avcodec_get_hw_config() we get AV_PIX_FMT_VIDEOTOOLBOX, AV_HWDEVICE_TYPE_VIDEOTOOLBOX which is seemingly correct.

  2. But usage and framerates clearly shows everything is being done in CPU. The code is in ff_hevc_hls_residual_coding all the time. (That’s fffmpeg’s software decoder.)

  3. This very diff very long git.videolan.org URL here seems to suggest again it should all be working.

  4. Have tried every iPhone etc. of course

• Python subprocess or os.system not working with ffmpeg

  13 février 2024, par confused

  I've been trying to get this darn programming run since last night and cannot seem to get anything to work. I want to first trim the video and then resize the video. I'm reading which video and what to give the final name from a text file, over 780 lines long, quite a few videos.

  


  Thus far with every idea under the sun I have tried, subprocess and os.system, I can't get anything more than error statements or right now all I get is no file creation of any kind. How the heck do I get this to work correctly ?

  


  import ffmpeg
import subprocess
import os

os.chdir('/home/Downloads/SRs/')
a = open('SRt.txt', 'r')
b = a.readlines()
a.close()
for c in range(0, len(b)-1):
    words = list(b[c].split(" "))
    d = len(words)
    e = words[d-1]
    f = b[c].replace(e, 'FR' + str(c) + '.mp4')
    words[d-1] = 'FR' + str(c) + '.mp4'
    print(f)
    subprocess.call(f, shell=True)
    subprocess.call([
        'ffmpeg',
        '-i',
        "'FR' + str(c) + '.mp4'",
        '-vf scale=320:240',
        words[d-1],
        ])


  


  Here are some examples of what the original file would look like :

  


   ffmpeg -i SR.mp4 -ss 00:00:00 -to 00:01:22 -c:v copy -a copy CPH.mp4
 ffmpeg -i SR.mp4 -ss 00:01:24 -to 00:02:58 -c:v copy -a copy CG.mp4
 ffmpeg -i SR.mp4 -ss 00:02:59 -to 00:05:41 -c:v copy -a copy CSGP.mp4


  


  Nothing fancy just separating video in its own individual segments and then resaving it before resizing it.

  


  I tried :

  


  z=subprocess.call(f, shell=True, stdout=subprocess.PIPE)
print(z)


  


  But all I get is '1'.

  


  When I changed it to :

  


  z=subprocess.call(f, shell=True, stderr=subprocess.PIPE)
print(z)


  


  All I get is '1'.

  


  Maybe I'm doing something wrong.

  


• Converting MSB padding to LSB padding

  23 novembre 2020, par Oier Lauzirika Zarrabeitia

  I am using FFmpeg and Vulkan for a video related project. The problem is that FFmpeg uses MSB padding for the 9, 10 and 12bit formats, whilst Vulkan accepts LSB padded pixel formats.

  


  I am using the following code to convert from one another, but the performance is terrible (1 second per frame), which cannot be accepted for video playback.

  


  if(paddingBits > 0) {
    //As padding bits are set to 0, using CPU words for the shifting should not affect the result
    assert(dstBuffers[plane].size() % sizeof(size_t) == 0);
    for(size_t i = 0; i < dstBuffers[plane].size(); i += sizeof(size_t)) {
        reinterpret_cast(dstBuffers[plane][i]) <<= paddingBits;
    }
}


  


  Note that this code will be executed 1-4 times per frame (once per plane) and dstBuffers[plane].size() will be in the order of a couple of MegaBytes. dstBuffers[plane]'s data has an alignment greater than size_t's, so I am not performing unaligned read/writes. Using a smaller type such as uint16_t makes it perform worse.

  


  My question is if there is any standard function (or inside the FFmpeg library) which implements the same behavior in a fancy way so that the performance is not a concern. If not, is there a more efficient way to implement it ?

  


  Edit :dstBuffers[plane] stores std::byte-s and although is not of type std::vector, (it is a custom class) it is contiguous.