News and events - Tampere University of Technology

Kvazaar video encoding software wins in Amsterdam

A research group from TUT’s Department of Pervasive Computing won the open source software competition at the ACM Multimedia 2016 with its Kvazaar video encoder. The visualizer, showing the encoding in real time, is one of the innovative features.
Marko Viitanen (2nd from left), Assistant Professor Jarno Vanne, and Ari Koivula receiving the award.
Marko Viitanen (2nd from left), Assistant Professor Jarno Vanne, and Ari Koivula receiving the award.

The Ultra Video Group at TUT’s Department of Pervasive Computing won the Open Source Software Competition at the ACM Multimedia 2016 conference. The winning publication was Kvazaar: Open-Source HEVC/H.265 Encoder, a description of the Kvazaar HEVC video encoder developed at the Tampere University of Technology.

The winner was selected based on a publication submitted for the competition as well as a presentation at the event. Research Assistant Marko Viitanen, Assistant Professor Jarno Vanne, and Research Assistant Ari Koivula were there to accept the award at the conference.

‘Over the course of several years, our Kvazaar  has become so good for encoding videos that we thought it appropriate to prepare a publication explaining its essential features and comparing it to other similar software. The publication was written for the purpose of this competition at the ACM Multimedia conference,’ says Marko Viitanen.

The victorious software has been under development since 2012, and the most recent version is 1.0. HEVC, used by Kvazaar, is a new video encoding standard used with 4K content by Netflix as well as the DVB-T2 television standard, for example.The video produced by the software is compliant with the HEVC standard, and it can be utilised in any video application. Kvazaar also offers other researchers a solid foundation for developing their own video encoding tools.

In their publication as well as their live presentation, the research group presented the built-in visualizer, which shows the encoding as it takes place in real time. This is a unique feature in any video encoding software.

‘The visualizer has also been used in teaching at TUT, because it allows video encoding to be explained in practice.

The main focus of the ACM Multimedia event is on artificial intelligence and neural network research, so the win was quite a pleasant surprise for the research group.

Working with research, it is rare to even be able to participate in such a competition, and now we have won one! The award certificate will be placed on the office wall as soon as we find a suitable frame for it,’ Viitanen says.

‘The award was a result of systematic and ambitious work by the Ultra Video Group and I consider it a significant tribute to our persevering research. It is one sign of the fact that our research group is one of the foremost in its field in the academic world,’ Vanne highlights.

After initial qualification rounds, 16 teams participated in the competition, which was organised for the 13th time. The event took place at the Tuschinski Theatre in Amsterdam. Previous winners include the Caffe deep learning framework.

News submitted by: Sanna Kähkönen
Keywords: science and research