* Use FFmpeg's MJPEG decoder instead of tinyjpeg. 
  * Each screen display has now its own capture device. Note that on Linux, multiple monitors are seen as one screen. JAWTRenderer on Linux will not display image > 2048x2048 due to driver limitation (XvCreateImage), so avoid dropping image if size mismatch and show what we can. 
  * Other recent changes in earlier versions:
  * - Attempts to fix a NullPointerException reported by Juanvi Martinez on the dev mailing list in the thread "Sip Communicator connects and disconnects very quickly on Windows 7 64 bits."
  * - Commit from translate.sip-communicator.org by user mandre.: 804 of 804 messages translated (0 fuzzy).
  * - Commit from translate.sip-communicator.org by user acastroy.: 799 of 799 messages translated (0 fuzzy).
  * - Commit from translate.sip-communicator.org by user acastroy.: 799 of 799 messages translated (0 fuzzy).
  * - Avoid using DNS resolver for an IP address.
  * - Fix possible deadlock when switching very quickly from media configuration form and another form.
  * - Updates otr4j to version 122.
  * - Fixes a couple of warnings.
  * - Fixes warning.
  * - Add support for V4L2's MJPEG and JPEG format.
  * - Support audio codecs that depend on native JMF libraries only for Windows 32-bit and Linux 32-bit.
  * - Add jspeex native library target. 
  * - Add possibility to initiate directly videocall and desktop streaming session in protocol service. Note that GUI need to be modified in order to propose this feature to users. 
  * - Fix NPE when removing group.
  * - Tries to fix the playback of the notification sounds with PortAudio on devices which cannot be opened with arbitrary sample rates. Reported by Werner Dittmann on the dev mailing list in the thread 'Wideband PortAudio'.
  * - fixed a few typo and fixed some brace placement
  * - Fix NPE when there is no audio capture device
  * - Increments alpha version to better reflect progress status and keep consistency with our roadmap. 
  * - Increments alpha version to better reflect progress status and keep consistency with our roadmap. 
  * - Increments alpha version to better reflect progress status and keep consistency with our roadmap. 
  * - Use native audio format settings to avoid resampling inside audio subsystem. Resampling may reduce the randomness of the audio data, thus use the native samples. Compute the correct number of bytes to gather based on the audio format.
  * - Fix race condition with jportaudio on Windows.
  * - Enhance GatherEntropy to use Javasound and PortAudio input. Use standard JMF mechanisms to determine the audio source and get the right DataSource and audio stream. Read a number of audio buffers and seed the Fortuna PRNG.