I do:
PLAY rtsp://addr/track1
I get ok response and I send another one but for track2. One is for audio and another one is for video, question is: How can I TEARDOWN only one of them only? let's say: TEARDOWN rtsp://addr/track1. Is this even possible or should I just TEARDOWN rtsp://addr/ and play again only one track?
Yes, the RTSP specification definitely allows this but in the end I guess this depends on server implementation/support. Easiest thing would be to write a quick script to test if your server supports it.
The "Media on Demand (Unicast)" example section shows such a scenario.
Related
I`m using Asterisk::AGI and I need to stream a music file and interrupt streaming according to some background condition (like checking data in DB, if changed smth -> interrupt streaming).
Can somebody advice/point me where should I look for a solution?
Thanks a lot
Pavel
You can use AsyncAGI command, which will play musiconhold class unless get some other action
After that you have use Asterisk::AMI to transfer that channel to other context/dialplan
No, there are no any way do it using just AGI interface without AMI.
I need to send a stream of data to Play server. The length of the stream is unknown and I need to get a response every line break \n or for every several lines. Rather then wait for the whole data to be sent.
Think of the following usecase:
lets say i'm intended to write a console application, that when launched, connects to my web server, and all the user input are being sent to play on every line break, and gets responded asynchronously. All above should be performed on a single connection, i.e. I don't want to open a new connection on every request I send to Play (a good analog would be 2 processes communicating through 2 pipes).
What is the best way to achieve this?
And is it possible to achieve with a client that communicates with the server only via http (with a single http connection)?
EDIT:
my current thoughts on how to approach this are as follows:
i can define a new BodyParser[Future[String]] which is basically an Iteratee[Array[Byte],Future[String]]. while the parsing takes place, i can compute the result asynchronously and the action can return the result as ChunkedResult in the future's onComplete method.
does this sound like the right approach?
any suggestions on how to achieve this?
Maybe you should look at websockets.
Java: http://www.playframework.com/documentation/2.1-RC3/JavaWebSockets
Scala: http://www.playframework.com/documentation/2.0/ScalaWebSockets
I want to initiate calls to two numbers using a 3rd party API. I need to make sure that the devices ring when they get the calls, media starts to transmit when the call is answered, and the call is terminated when hungup.
Ideally I would like to do this from JUnit tests so that I can automate this whole process. But any other tool will also be fine.
So this is what I want to do programatically,
1) Configure two SIP soft phones to answer on 2 different numbers using some credentials provided by the test.
2) Make a call using the API
3) Assert that two phones are in ringing state
4) Answer the call
5) Assert that RTP media is being transmitted among them
6) Hang up
7) Assert that the call is now successfully disconnected
I am quite new to telephony, so would appreciate any pointers on any tool or SDK that will help me accomplish this.
Please check out SipUnit https://code.google.com/p/commtesting/wiki/SipUnit.
It can do all above except check out the media is flowing yet.
Check out the KitCAT framework. It's based on JUnit and can support all of your requirements. It supports multiple user agents, which can all be coordinated within a test case. It also provides coordination with other protocols (e.g., RTP, HTTP).
Sample code:
SIPAgent callee1 = createAgent("callee1"); // sip:callee1#host
SIPAgent callee2 = createAgent("callee2"); // sip:callee2#host
// invoke your API here
invoke3rdPartyAPI(callee1.getSipURI(), callee2.getSipURI());
pause(2000);
assertThat(callee1, is(invited()));
assertThat(callee2, is(invited()));
callee1.answer();
callee2.answer();
pause(500);
assertThat(callee1, is(connectedTo(callee2))); // check for SDP match
callee1.playAudio(audioFileName);
pause(500);
assertThat(callee2, has(incomingMedia());
callee1.disconnect();
pause(500);
assertThat(callee2, is(disconnected()));
Check out the complete API here.
Live555 lib has a nice example testOnDemandRTSPServer.cpp This example just stream "one" given file. I want to stream more than one file. Does Live555 has playlist concept or how to stream more than one file in Live555?
Best Wishes
PS: I try to add more than one subsession, in that case Live555 just stream the last session file...
There is one more application that comes with the live555 code. Live555Media server is present inside the source code's mediaServer directory. This does the job. It uses the dynamicRTSP server class. You give it the folder with all your media files and access them as rtsp://ip/filename.
My 0.02 cents:
I'm not sure if that makes sense: how would you ensure that they are all encoded in the same format which is a requirement if you want to stream them in the same session. RTSP describe gets a media session description of the file and this is used to setup the streaming sessions so it is crucial that all files encoded similarly.
RTSP does not make any provision for playlists. Usually playlists are not transferred via RTSP, but say via HTTP. IMO if the playlist resides on the client it would make more sense to await the RTCP bye packet (at the eof) and then to do a SETUP and PLAY for the next file/RTSP URI in the playlist.
If you just want to stream a sequence of files (playlist is on the server) where the RTSP client just initiates one session, of course nothing prevents you from creating a custom file source in the live555 library that does what you want...
Recently I had to do similar task and with similar functionality:
Here what you can do for video H264 stream files to play in the row like playlist (of course if they are same resolution, encoding profile,etc)
You would have to modify ByteStreamFileSource::doGetNextFrame method.
There is code like feof(fFid)
if (feof(fFid))
{
CloseInputFile(fFid);
fFid = OpenInputFile(envir(), "test.264");
//fileName
}
else ....
Of course if you still need LGPL compliance you there will be more work to do... You will have to copy/rename this class outside library and do the same with H264VideoFileServerMediaSubsession and modify method createNewStreamSource that it would use you rewritten class of ByteStreamFileSource.
I've not found a answer to this question anywhere, but this seems like a typical problem:
I would like to send some POST-Requests (with ASIHTTPRequest, what I already do), but if something goes wrong, ther user can decide to "Try Later", that means, the task should be put on a queue and this queue should be read next time the application starts. So, that's my question: how to "save" the queue, so that the app can read it next time it starts? Is it possible to "read" the queue and try sending this POST-Request again, let's say, 10 min later, even if the application is not running?
What kind of documentation should I read in order to be able to do this?
I would be very glad to hear any answers. Thanks in advance.
P.S.: Another Idea I have: as I just have to Upload Photos, I could have a folder with all the Photos that still need to be uploaded, and when the App starts, the app looks at this folder and try to send all the photos in this folder. Does it make sense?
My approach for this issue would be like this:
Whenever you fail to send details - write content of the array to a file using '[NSArray writeToFile:]' you can use serialization if array contain any data which is custom defined (if your array contain standard cocoa objects(NSString,NSData etc) they already implemented with serialization )
When app launches; load the content from file directly to an array object ('[NSArray arrayWithContentsOfFile:]')
then construct http request and try sending. In application the data(in your case array) is stored/serialized not the request, you need to reconstruct the http request when you want to try one more time.(don't try serializing ASIHTTPRequest, you have reconstruct it)
I'm going to assume you've already looked at NSOperationQueue and NSOperation. AFAIK there is no built-in support for serializing NSOperation, but you could very easily write your own serialization mechanism for an NSOperation subclass that you use for posting data and write the an NSOperationQueue's operations to disk if something goes wrong.
Without knowing too many details it's hard to give a precise answer. There are many ways to write data to disk and load it again later, the direction you take will be largely dependent on your situation.