Recherche avancée

Sur d’autres sites (1333)

• Reducing a ffmpeg two-pass xml profile to a single pass profile

  14 octobre 2013, par M.frank

  I am to reduce the XML profile below to a single pass profile rather than a two-pass profile.
  The settings are (if possible) the same. The profile looks like this :

  <?xml version="1.0" encoding="utf-8"?>
  
    
  
      
  
        -i %infile% -vcodec libx264 -aspect 4:3 -s 480x360 -r 25.000 -vb 380000 -vprofile main -level 3.0 -pix_fmt yuv420p -pass 1 -sn -an -y %outfile%
  
      
  
      
  
        -i %infile% -vcodec libx264 -aspect 4:3 -s 480x360 -r 25.000 -vb 380000 -vprofile main -level 3.0 -pix_fmt yuv420p -pass 2 -sn -acodec libvo_aacenc -ab 48000 -ar 22050 -ac 2 -y %outfile%
  
      
  
    

  My idea was to just remove the two-pass cmdline, but I want to make sure that it is correct.

• How to save variables at start of script for use later if script needs to be re-run due to errors or bad user input

  28 novembre 2023, par slyfox1186

  I have a script that uses GitHub's API to get the latest version number of the repositories that I am trying to download and then compile.

  


  Due to the fact that without using a specialized token from GitHub you are only allowed 50 API calls a day vs the 5000 a day with the API user token.

  


  I want to be able to parse all of the repositories and grab the version numbers that my script will then import into the code up front so in case someone who accidentally cancels the build in the middle of it (for who knows what reasons) wont have to eat up their 50 day API call allowance.

  


  Essentially, store each repo's version number, if the user then needs to rerun the script and version numbers that have been saved so far will be skipped (thus eliminating an API call) and any numbers that are still needing to be sourced will be called and then stored for used in the script.

  


  I am kinda of lost for a method on how to go about this.

  


  Maybe some sort of external file can be generated ?

  


  So what my script does is it builds FFmpeg from source code and all of the external libraries that you can link to it are also built from their latest source code.

  


  The code calls the function git_ver_fn and passes arguments to it which are parsed inside the function and directed to another functions git_1_fn or git_2_fn which passed those parsed arguments that have been passed on to the CURL command which changes the URL based on the arguments passed. It uses the jq command to capture the GitHub version number and download link for the tar.gz file.

  


  It is the version number I am and trying to figure out the best way to store in case the script fails and has to be rerun, which will eat up all of the 50 APT limit that GitHub imposes without a token. I can't post my token in the script because GitHub deactivates it and thus the users will be SOL if they need to run the script more than once.

  


  curl_timeout='5'

git_1_fn()
{
    # SCRAPE GITHUB WEBSITE FOR LATEST REPO VERSION
    github_repo="$1"
    github_url="$2"

    if curl_cmd="$(curl -m "$curl_timeout" -sSL "https://api.github.com/repos/$github_repo/$github_url")"; then
        g_ver="$(echo "$curl_cmd" | jq -r '.[0].name')"
        g_ver="${g_ver#v}"
        g_ssl="$(echo "$curl_cmd" | jq -r '.[0].name')"
        g_ssl="${g_ssl#OpenSSL }"
        g_pkg="$(echo "$curl_cmd" | jq -r '.[0].name')"
        g_pkg="${g_pkg#pkg-config-}"
        g_url="$(echo "$curl_cmd" | jq -r '.[0].tarball_url')"
    fi
}

git_2_fn()
{
    videolan_repo="$1"
    videolan_url="$2"
    if curl_cmd="$(curl -m "$curl_timeout" -sSL "https://code.videolan.org/api/v4/projects/$videolan_repo/repository/$videolan_url")"; then
        g_ver="$(echo "$curl_cmd" | jq -r '.[0].commit.id')"
        g_sver="$(echo "$curl_cmd" | jq -r '.[0].commit.short_id')"
        g_ver1="$(echo "$curl_cmd" | jq -r '.[0].name')"
        g_ver1="${g_ver1#v}"
    fi
}

git_ver_fn()
{
    local v_flag v_tag url_tag

    v_url="$1"
    v_tag="$2"

    if [ -n "$3" ]; then v_flag="$3"; fi

    if [ "$v_flag" = 'B' ] && [  "$v_tag" = '2' ]; then
        url_tag='git_2_fn' gv_url='branches'
    fi

    if [ "$v_flag" = 'X' ] && [  "$v_tag" = '5' ]; then
        url_tag='git_5_fn'
    fi

    if [ "$v_flag" = 'T' ] && [  "$v_tag" = '1' ]; then
        url_tag='git_1_fn' gv_url='tags'
    elif [ "$v_flag" = 'T' ] && [  "$v_tag" = '2' ]; then
        url_tag='git_2_fn' gv_url='tags'
    fi

    if [ "$v_flag" = 'R' ] && [  "$v_tag" = '1' ]; then
        url_tag='git_1_fn'; gv_url='releases'
    elif [ "$v_flag" = 'R' ] && [  "$v_tag" = '2' ]; then
        url_tag='git_2_fn'; gv_url='releases'
    fi

    case "$v_tag" in
        2)          url_tag='git_2_fn';;
    esac

    "$url_tag" "$v_url" "$gv_url" 2>/dev/null
}

# begin source code building
git_ver_fn 'freedesktop/pkg-config' '1' 'T'
if build 'pkg-config' "$g_pkg"; then
    download "https://pkgconfig.freedesktop.org/releases/$g_ver.tar.gz" "$g_ver.tar.gz"
    execute ./configure --silent --prefix="$workspace" --with-pc-path="$workspace"/lib/pkgconfig/ --with-internal-glib
    execute make -j "$cpu_threads"
    execute make install
    build_done 'pkg-config' "$g_pkg"
fi

git_ver_fn 'yasm/yasm' '1' 'T'
if build 'yasm' "$g_ver"; then
    download "https://github.com/yasm/yasm/releases/download/v$g_ver/yasm-$g_ver.tar.gz" "yasm-$g_ver.tar.gz"
    execute ./configure --prefix="$workspace"
    execute make -j "$cpu_threads"
    execute make install
    build_done 'yasm' "$g_ver"
fi


  


• Revision 0ba1542f12 : Vidyo : Support for one-pass rc-enabled SVC encoder Adds support for one-pass rc

  7 novembre 2014, par Deb Mukherjee

  Changed Paths :
   Modify /examples/vp9_spatial_svc_encoder.c

  
   Modify /vp9/vp9_cx_iface.c

  
   Modify /vpx/src/svc_encodeframe.c

  
   Modify /vpx/svc_context.h

  
   Modify /vpx/vp8cx.h

  
   Modify /vpx/vpx_encoder.h

  
  
  Vidyo : Support for one-pass rc-enabled SVC encoder

  Adds support for one-pass rc-enabled SVC encoder with callbacks for
  getting per-layer packets.
   the callback function registration is implemented as an encoder
  control function.
   if the callback function is not registered, the old way of
  aggregating packets with superframe will take effect.
   one more control function “VP9E_GET_SVC_LAYER_ID” has been
  implemented to get the temporal/spatial id from the encoder
  within the callback. This can be used to get the ids to put on RTP
  packet.

  Change-Id : I1a90e00135dde65da128b758e6c00b57299a111a