Recherche avancée

Sur d’autres sites (6082)

• accelerate x264 encoding

  7 août 2012, par Saraswati

  i am making use of x264 to encode raw data captured from the iphone camera .. but the encoding is very slow .Can anyone help me accelerate the encoding speed.

  I have used following settings to build x264 lib :

  //for armv6

  CC=/Developer/Platforms/iPhoneOS.platform/Developer/usr/bin/gcc ./configure —host=arm-apple-darwin —sysroot=/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS4.3.sdk —prefix='dist' —extra-cflags='-arch armv6' —extra-ldflags=-L/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS4.3.sdk/usr/lib/system/ —extra-ldflags='-arch armv6' —enable-pic —disable-asm

  //for armv7

  CC=/Developer/Platforms/iPhoneOS.platform/Developer/usr/bin/gcc ./configure —host=arm-apple-darwin —sysroot=/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS4.3.sdk —prefix='dist' —extra-cflags='-arch armv7' —extra-ldflags=-L/Developer/Platforms/iPhoneOS.platform/Developer/SDKs/iPhoneOS4.3.sdk/usr/lib/system/ —extra-ldflags='-arch armv7' —enable-pic

  I am using default preset like this :

  x264_param_default_preset(param, "slow", "zerolatency");

  and setting few perameters :

  param->i_bframe = 0;
  
  param->analyse.i_me_method = X264_ME_HEX;
  
  param->analyse.i_subpel_refine = 2;
  
  param->i_frame_reference = 1;
  
  param->analyse.b_mixed_references = 0;
  
  param->analyse.i_trellis = 0;
  
  param->rc.b_mb_tree = 0;
  
  param->analyse.i_weighted_pred = X264_WEIGHTP_NONE;
  
  
  
  
  
  param->rc.i_bitrate = 180;
  
  param->rc.i_qp_min = 20; 
  
  param->rc.i_qp_max = 26; 
  
  
  
  param->i_keyint_max = 15; 
  
  param->i_keyint_min = 15;
  
  
  
  
  
  param->i_width = w;
  
  param->i_height = h;
  
  
  
  x264_param_apply_profile(param, "baseline");
  
  x264_picture_alloc( &(enc->pic), X264_CSP_I420, param->i_width, param->i_height );

• Use C# to converting apng to webm with ffmpeg from pipe input and output

  30 novembre 2022, par martin wang

  I was using ffmpeg to convert Line sticker from apng file to webm file.
And the result is weird, some of them was converted successed and some of them failed.
not sure what happend with these failed convert.

  


  Here is my c# code to convert Line sticker to webm,
and I use CliWrap to run ffmpeg command line.

  


  async Task Main()
{

    var downloadUrl = @"http://dl.stickershop.LINE.naver.jp/products/0/0/1/23303/iphone/stickerpack@2x.zip";
    var arg = @$"-i pipe:.png -vf scale=512:512:force_original_aspect_ratio=decrease:flags=lanczos -pix_fmt yuva420p -c:v libvpx-vp9 -cpu-used 5 -minrate 50k -b:v 350k -maxrate 450k -to 00:00:02.900 -an -y -f webm pipe:1";

    var errorCount = 0;
    try
    {
        using (var hc = new HttpClient())
        {
            var imgsZip = await hc.GetStreamAsync(downloadUrl);

            using (ZipArchive zipFile = new ZipArchive(imgsZip))
            {
                var files = zipFile.Entries.Where(entry => Regex.IsMatch(entry.FullName, @"animation@2x\/\d+\@2x.png"));
                foreach (var entry in files)
                {
                    try
                    {
                        using (var fileStream = File.Create(Path.Combine("D:", "Projects", "ffmpeg", "Temp", $"{Path.GetFileNameWithoutExtension(entry.Name)}.webm")))
                        using (var pngFileStream = File.Create(Path.Combine("D:", "Projects", "ffmpeg", "Temp", $"{entry.Name}")))
                        using (var entryStream = entry.Open())
                        using (MemoryStream ms = new MemoryStream())
                        {
                            entry.Open().CopyTo(pngFileStream);

                            var result = await Cli.Wrap("ffmpeg")
                                         .WithArguments(arg)
                                         .WithStandardInputPipe(PipeSource.FromStream(entryStream))
                                         .WithStandardOutputPipe(PipeTarget.ToStream(ms))
                                         .WithStandardErrorPipe(PipeTarget.ToFile(Path.Combine("D:", "Projects", "ffmpeg", "Temp", $"{Path.GetFileNameWithoutExtension(entry.Name)}Info.txt")))
                                         .WithValidation(CommandResultValidation.ZeroExitCode)
                                         .ExecuteAsync();
                            ms.Seek(0, SeekOrigin.Begin);
                            ms.WriteTo(fileStream);
                        }
                    }
                    catch (Exception ex)
                    {
                        entry.FullName.Dump();
                        ex.Dump();
                        errorCount++;
                    }
                }
            }

        }
    }
    catch (Exception ex)
    {
        ex.Dump();
    }
    $"Error Count:{errorCount.Dump()}".Dump();

}


  


  This is the failed convert file's error information from ffmpeg :

  


  

  


  And the successed convert file from ffmpeg infromation :


  


  It's strange when I was manually converted these failed convert file from command line, and it will be converted successed.


  


  The question is the resource of images are all the same apng file,
so I just can't understan why some of files will convert failed from my c# code
but also when I manually use command line will be converted successed ?

  


  ---

  


  I have written same exampe from C# to Python...
and here is python code :

  


  from io import BytesIO
import os
import re
import subprocess
import zipfile

import requests


downloadUrl = "http://dl.stickershop.LINE.naver.jp/products/0/0/1/23303/iphone/stickerpack@2x.zip"
args = [
    'ffmpeg',
    '-i', 'pipe:',
    '-vf', 'scale=512:512:force_original_aspect_ratio=decrease:flags=lanczos',
    '-pix_fmt', 'yuva420p',
    '-c:v', 'libvpx-vp9',
    '-cpu-used', '5',
    '-minrate', '50k',
    '-b:v', '350k',
    '-maxrate', '450k', '-to', '00:00:02.900', '-an', '-y', '-f', 'webm', 'pipe:1'
]


imgsZip = requests.get(downloadUrl)
with zipfile.ZipFile(BytesIO(imgsZip.content)) as archive:
    files = [file for file in archive.infolist() if re.match(
        "animation@2x\/\d+\@2x.png", file.filename)]
    for entry in files:
        fileName = entry.filename.replace(
            "animation@2x/", "").replace(".png", "")
        rootPath = 'D:\\' + os.path.join("Projects", "ffmpeg", "Temp")
        # original file
        apngFile = os.path.join(rootPath, fileName+'.png')
        # output file
        webmFile = os.path.join(rootPath, fileName+'.webm')
        # output info
        infoFile = os.path.join(rootPath, fileName+'info.txt')

        with archive.open(entry) as file, open(apngFile, 'wb') as output_apng, open(webmFile, 'wb') as output_webm, open(infoFile, 'wb') as output_info:
            p = subprocess.Popen(args, stdin=subprocess.PIPE,
                                 stdout=subprocess.PIPE, stderr=output_info)
            outputBytes = p.communicate(input=file.read())[0]

            output_webm.write(outputBytes)
            file.seek(0)
            output_apng.write(file.read())



  


  And you can try it,the result will be the as same as C#.

  


• iOS SDK avcodec_decode_video Optimization

  6 août 2013, par Johny Cage

  I've recently started a project that relies on streaming FLV directly to an iOS device. As most famous i went with ffmpeg (and an iOS wrapper - kxmovie). To my surprise iPhone 4 is incapable of playing even SD low-bitrate FLV videos. The current implementation i'm using is decoding the video/audio/sub frames in dispatch_async while loop and copies the YUV frame data to a object, where the object is parsed to 3 textures - Y/U/V (in case of RGB color space - just parse the data) and rendered on screen. After much trial and error, i've decided to kill the whole rendering pipeline and leave only the avcodec_decode_video2 function to run. Surprisingly the FPS did not improve and videos are still unplayable.

  My question is : What can i do to improve the performance of avcodec_decode_video2 ?

  Note :
  I've tried a few commercial apps and they play the same file perfectly fine with no more than 50-60% cpu usage.

  The library is based off the 1.2 branch and this is are the build args :

  '--arch=arm',
  
  '--cpu=cortex-a8',
  
  '--enable-pic',
  
  "--extra-cflags='-arch armv7'",
  
  "--extra-ldflags='-arch armv7'",
  
  "--extra-cflags='-mfpu=neon -mfloat-abi=softfp -mvectorize-with-neon-quad'",
  
  '--enable-neon',
  
  '--enable-optimizations',
  
  '--disable-debug',
  
  '--disable-armv5te',
  
  '--disable-armv6',
  
  '--disable-armv6t2',
  
  '--enable-small',
  
  '--disable-ffmpeg',
  
  '--disable-ffplay',
  
  '--disable-ffserver',
  
  '--disable-ffprobe',
  
  '--disable-doc',
  
  '--disable-bzlib',
  
  '--target-os=darwin',
  
  '--enable-cross-compile',
  
  #'--enable-nonfree',
  
  '--enable-gpl',
  
  '--enable-version3',

  And according to Instruments the following functions take about 30% CPU usage each :

  Running Time    Self        Symbol Name
  
  37023.9ms   32.3%   13874,8                   ff_h264_decode_mb_cabac
  
  34626.2ms   30.2%   9194,7                    loop_filter
  
  29430.0ms   25.6%   173,8                     ff_h264_hl_decode_mb