Budget:	$ 300-1000
Status:	Closed for Bidding
Project Creator:	jvdsandt
	Rating: (No Feedback Yet)
Required Skills:	C++ / C, Programming, Python
Attached Files:	(None)

NEW	SEND	DISCUSS	CONTACT	OPTIONS

Description
Assignment description VoIP and Internet Radio integration

We have a platform where users can dial in to listen to podcasts and other audio content. The platform uses VoIP and SIP/RTP technology to offer this functionality in an efficient way.

We would like to offer our users to possibility to listen to Internet radio stations via their mobile phones. Another feature that we would like to offer is the reverse: Let one or more users on the phone broadcast their conversation on the Internet as an audio stream. This is similar to a personal Internet radio station.

Currently both features are implemented as proof of concepts. We think that custom development is required to make them robust and scalable. Below are the technical details of both features. We are looking for a Linux C (and/or Python) programmer who can help us to make these features production ready.


Listening to Internet radio stations via VoIP

Functionality: A user dials our access number and selects a radio station by typing in (DTMF) its code, for example #644 for “Lounge Radio”.

Currently this feature is implemented using the Music-on-Hold functionality (MoH) of Asterisk, see for example http://nerdvittles.com/index.php?p=92 This approach works but has a number of disadvantages:

The transcoding from wma or mp3 to the G711 codec has to be performed on the Asterisk server. Transcoding is a very CPU intensive process limiting the scalability of the PBX.

When no one is listening to a particular radio station it’s very complex to stop the streaming. So often a lot of bandwidth is wasted streaming audio that nobody listens to.

We think that a better solution is to develop a specialized VoIP/SIP user agent that is responsible for reading the audio stream from the Internet, transcoding it and stream it via SIP/RTP to the phone listener. In this case the Asterisk PBX server only has to forward the caller to the correctly configured user agent, after this step the media will be transferred via a direct connection between the two user agents (or via a PSTN gateway).

The PJSIP library (http://www.pjsip.org/) already offers a lot of the functionality required, for example the PJSUA, a command line SIP client, can play a wav file when it receives a call. By using a program like ffmpeg the wma/mp3 Internet radio stream can be transcoded to wav on the fly.

On top of the PJSIP library it should be possible to develop a Linux executable that can handle the following requirements:

At startup it must be possible to specify the SIP server where the user-agent must register itself. Also the URL of the Internet Radio station must be a startup parameter.

When it receives the first call it should start the streaming and transcoding process. A small delay for buffering to prevent hiccups is allowed.

The executable must be able to handle multiple simultaneous calls, the single “in” stream must be transcoded and passed to all the callers.

When the last caller leaves the executable must stop reading the Internet radio stream to save bandwidth.

The executable must be able to handle DTMF input and possibly transfer a caller based on the input given.


Make a VoIP conversation accessible via the Internet

Functionality: A user dials a special access number and then selects the option (via DTMF) to start broadcasting. From that moment everything he/she says must be accessible as an Internet audio stream similar to an Internet radio station. Another option is that the conversation of a conference call (multiple callers) is made avaialble as an audio stream via the Internet.

Currently this functionality is implemented by a specialized SIP user agent that outputs the conversation as “raw” WAV to a named pipe. Programs like ffmpeg and VideoLan can than be used to transcode this audio and make it available as a correctly configured audio stream that users can listen to using Windows Media Player, Apple Quicktime or other media player software. The “specialized” SIP user agent is just the PJSUA command line client with the --rec-file --auto-rec parameters. See: http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/2008-February/001767.html

We require a custom PJSIP based user agent that can receive calls and output the audio as a correctly formatted Internet audio stream. It must support different formats like MP3 and WMA. When no call is active is must be capable to just output silence or repeat a special prerecorded message until a call is received.

Summary

To help develop the programs described above we are looking for an experienced Linux C and/or Python programmer. Preferably the person should have experience with one or more of the following products/libraries:

* VoIP / SIP / PJSIP
* ffmpeg or gstreamer or mplayer

Reminder
You may not start working in this and any project before your bid is accepted. Any user who violates this policy may have their account permanently suspended.

Bids Received (6)   Shortlist (0)   Declined Bids (0)    Average bid amount:   $5,225.00   Average delivery time:   45 Day(s)

Place Bid | Post Similar Project | Send Project | Message Board(1) | Contact jvdsandt Order by: Rating Bid sum Bid date

Remember that contacting the other party outside the site (by email, phone, etc.) on all business projects (before the project is awarded) is a violation of our terms of use. We supervise all site activity for such infringements and can immediately expel transgressors on the spot, so we thank you in advance for your cooperation. If you notice a violation please help out the site and report it. Thank you for your help.

nbsp;