Recherche avancée

Sur d’autres sites (28133)

• How should be the track ID encoded in to the media tracks / media content components available for MPEG - DASH Media Presentation ?

  2 avril 2014, par Ghanesh Babu J

  As per ISO/IEC 14496 Part 12 : ISOBMFF , the track Identifier specified in track header box "trak" and track fragment header "tfhd" box should be same and unique to each track available in an ISO file.

  Also as per ISO/IEC 23009 Part 1 , The track Identifier is same for all the representation of same content component type and unique to a media content component type( Audio / Video etc . . .) in an adaptation set.

  But , Most of the mpeg dash non multiplexed sample streams available in internet has used same track Identifier in all of the media tracks ( Audio , video etc. . . ) available for a media presentation. Say track ID for all content component type is 1.

  In this case, If a media player is supported with single mp4 demuxer means, it is not possible to demux the fragmented media sample from various track correctly for a media presentation.

  My question is,
  Having same track identifier for all media content component type in a dash media presentation is MPEG DASH Standard ?

• How to specify MPEG-DASH adaptation sets in FFmpeg, when using optional output audio and subtitle streams ?

  15 juillet 2021, par Sebsii

  I am using the script below to transform videos to MPEG-DASH compliant ones. I want also to be able to process video that have no audio and/or no subtitles. Because of that I marked the streams for audio and subtitles with a ? at the end of the stream specifier, f.e. -map a:0?.

  


  Now when I want to specify the adaptation sets I can set it for the video and audio streams with id=0,streams=v and id=1,streams=a, but I can not select the subtitle streams, like I would want to with id=2,streams=s, as it is stated in the docs ("To map all video (or audio) streams to an AdaptationSet, "v" (or "a") can be used as stream identifier instead of IDs."). The only thing I could do is to specify an ID, like id=2,streams=8, but for that to work, an audio and subtitle stream would be mandatory, because otherwise there is no stream with the id 8 and FFmpeg would fail. I thought of putting the -map s:0? at the first position, then I could be sure, that it is always the stream with id 0, but if the input did not have a subtitle stream, then there would be an additional adaptation set, with the first version of video. I also thought about using FFprobe to get info about the available streams, but did not look too much into depth.

  


  Does anybody know how I could tell FFmpeg, that an adaptation set should be created, when there's a stream with subtitles present independent of there being audio streams or not ?

  


  Any help is much appreciated.

  


  #!/bin/bash

INPUT_PATH=$1
OUTPUT_PATH=$2

FPS=25
GOP_SIZE=100
PRESET_P=veryfast
V_SIZE_1=256x144
V_SIZE_2=426x240
V_SIZE_3=640x360
V_SIZE_4=854x480
V_SIZE_5=1280x720
V_SIZE_6=1920x1080

ffmpeg -i "$INPUT_PATH" \
    -preset $PRESET_P -keyint_min $GOP_SIZE -g $GOP_SIZE -sc_threshold 0 \
    -r $FPS -c:v libx264 -pix_fmt yuv420p -c:a aac -c:s copy \
    -map v:0 -s:0 $V_SIZE_1 -b:v:0 160k -maxrate:0 160k -bufsize:0 320k \
    -map v:0 -s:1 $V_SIZE_2 -b:v:1 450k -maxrate:1 450k -bufsize:1 900k \
    -map v:0 -s:2 $V_SIZE_3 -b:v:2 1M -maxrate:2 1M -bufsize:2 2M \
    -map v:0 -s:3 $V_SIZE_4 -b:v:3 2.5M -maxrate:3 2.5M -bufsize:3 5M \
    -map v:0 -s:4 $V_SIZE_5 -b:v:4 5M -maxrate:4 5M -bufsize:4 10M \
    -map v:0 -s:5 $V_SIZE_6 -b:v:5 8M -maxrate:5 8M -bufsize:5 16M \
    -map a:0? -b:a:0 128k -ac:a:0 1 \
    -map a:0? -b:a:1 384k -ac:a:1 2 \
    -map s:0? \
    -use_template 1 -use_timeline 1  \
    -seg_duration 4 -adaptation_sets "id=0,streams=v id=1,streams=a id=2,streams=s" \
    -f dash "$OUTPUT_PATH"


  


• Serving static video content directly vs. via adaptive streaming protocols (HLS, DASH)

  7 mai 2020, par astralmaster

  Is there an advantage of serving static video content (not a live stream) via adaptive streaming protocols such as HLS or DASH over serving them directly as files using HTTP server in terms of speed ?

  



  Example case is when you have a 500MB mp4 h264+AAC video that you have to serve on a website via HTML5 video element. Would you rather serve it directly, since most popular browsers implement functions such as seek without downloading the whole file first ; or would you rather use ffmpeg or similar solution to create HLS chunks from the mp4 file and instead provide .m3u8 playlist source to the HTML5 video element. Is there a real advantage in terms of speed of doing this ?

  



  Which one would you implement if you had hundreds of video files all served as static content ?