Recherche avancée

Sur d’autres sites (5803)

• Wrapping an existing RTP stream with RTSP

  12 janvier 2015, par Neverbirth

  I have a RTP multicast stream from ffmpeg that requires users to use external SDP files, as this sometimes becomes a problem I was thinking of wrapping the stream using RTSP.

  I have an application that already makes some management for the RTP stream, so it would be useful to have the RTSP wrapper inside it as well, although I don’t want the wrapper to do anything beyond delivering the SDP information. Right now I’d say I’m correctly delivering the SDP information (otherwise it’s stuck in that request), however I’m not getting any PLAY command, the player is stuck after the SETUP request, so I guess my reply is wrong, but after testing many different values for the Transport header I haven’t been able to get anything beyond this point.

  This is the SDP :

  v=0
o=- 0 0 IN IP4 127.0.0.1
s=No Name
c=IN IP4 236.0.0.1
t=0 0
a=tool:libavformat 56.7.102
m=video 2000 RTP/AVP 96
b=AS:500
a=rtpmap:96 H264/90000
a=fmtp:96 packetization-mode=1

  This is a sample SETUP request from VLC :

  SETUP rtsp://127.0.0.1:34343/live.sdp/ RTSP/1.0
CSeq: 4
User-Agent: LibVLC/2.1.5 (LIVE555 Streaming Media v2014.05.27)
Transport: RTP/AVP;multicast;client_port=2000-2001

  And one of many replies I’ve tried :

  RTSP/1.0 200 OK
CSeq: 4
Transport: RTP/AVP;multicast;client_port=2000-2001;source=236.0.0.1;port=2000-2001
Session: 0456804596

  I’ve tried using destination instead of source, server_port instead of port, also, although the RTP multicast port is 2000, I’ve seen the streaming is also using other ports like 57927 and 57928, so I’ve tried those ones as well, etc.

• Anomalie #3375 (Nouveau) : Espace Privé > Recherche > Cas d’une recherche numérique

  19 janvier 2015, par xdjuj -

  La révision 20305 semble induire une erreur sur la recherche :
  - https://core.spip.net/projects/spip/repository/revisions/20305

  La boucle conditionnelle ajoutée considère que si le ENV recherche est numérique, alors il doit forcément rechercher cette valeur dans les tables objets comme étant un id_objet : table articles avec id_article=#ENVid / table rubriques avec id_rubrique=#ENVid, etc.

  Si l’un de mes champs (normaux ou extra) possède un chiffre (par exemple un ISBN ou même pour d’autres besoins, un ID quelconque) alors il n’est pas remonté par la recherche car bloqué dans la boucle conditionnelle.

  Je pense que le comportement attendu devrait être :
  - si j’identifie un numeric => je cherche dans les id_objets PUIS dans les différents champs / champs extra => j’affiche le tout

  Actuellement, on est bloqué dans la recherche dans les id_objet.

• AudioContext HTML5 Player

  15 février 2015, par Plyto

  So I’ve been playing with the Web Audio API and have the following issue.

  I am making a project in which I call an external library’s API with Ajax and get audio back (arraybuffer).

  • I send them the text and get audio back.
  • This is not necessarily a GET request (can be POST, etc.)
  • If text is too large, I split it into smaller chunks and send multiple requests

  So far so good, now comes the issue of how to play the multiple audios that I got back.

  Since users do not care that I have split the text and actually have multiple audio tracks, I need somehow to make it look like a single track or as a playlist.

  So I have tried to :

  • merge arraybuffer (apparently it does not work like that and most likely I need ffmpeg or simiar tools to do the merging, which is hard to do on client-side ? (like there is ffmpeg for browsers, but I don’t know how good is it to burden a client with it). If it’s not so, maybe you can suggest something here)
  • load it as a playlist, but so far cannot find a library that accepts multiple audiobuffers/audiocontexts and/or gives a playlist with it back.

  The easiest solution that I see so far is to create my own small library that accepts AudioBuffers/arraybuffers and go either with the playlist approach or play the ’chunked’ audios one by one and make scrubber that jumps between audio contexts.

  Is there a library/easier approach ?

  Will be thankful for any suggestions :]