The Fraunhofer Institute for Integrated Circuits IIS, together with German public broadcasters, is conducting further tests to improve speech intelligibility in television.
Back in October, German public broadcaster Westdeutscher Rundfunk (WDR) joined with Fraunhofer IIS in conducting an online test where participants were able to try out and evaluate Dialog+. This technology, developed by Fraunhofer IIS, provides a solution for hard-to-understand speech or overly loud background noise on TV. By watching three short clips, participants could compare the original audio of various television productions with the technically processed version and then give their opinion.
The results of the survey are clear: Dialog+ and personalization are very popular with German television audiences. Overall, more than 80 percent of the participants liked the option of switching between audio tracks, with more than half of them rating Dialog+ as the better alternative.
WDR welcomed the positive feedback as an indication that it should continue with the series of experiments. On December 12 and 13, the broadcaster tested the quality and use of the additional audio signal in its linear television programming. All households that receive WDR’s Cologne, Bielefeld, Dortmund, Düsseldorf, Essen, Münster or Siegen HD channels via satellite had the option on selected programs of switching to the “Clear Speech” audio track, where background noise and music are quieter and the dialogue is easier to understand.
Bayerischer Rundfunk (BR), the Bavarian public broadcaster, also recently launched a pilot project to test Dialog+ as an alternative to regular TV audio. The first show where viewers could switch between the original soundtrack and Dialog+ aired on December 5. Another opportunity was available for the “World of Animals” documentary airing on December 12. Future tests will also include episodes of the documentary series “Natur exclusiv”. To be able to switch to the other audio track, viewers require a smart TV (HbbTV 2.0 capable) connected to the Internet, since the audio tracks are stored on a server and synchronized in real time with the broadcast.
The broadcasters see the test runs as an excellent opportunity to find out what the audience wants and needs, while the Fraunhofer IIS team can use the results to improve Dialog+ further. At the same time, the project illustrates the advantages of personalizing TV audio and can be seen as a major milestone towards enhancing broadcast and streaming services with new features and accessibility options. With its object-based approach,the MPEG-H Audio system will bring even more and unique personalization options.
Dialog+ is based on artificial intelligence and reduces the volume of music, sound effects or background noise to put spoken dialogue in the foreground. Developed by Fraunhofer IIS, the technology is easy to integrate into existing production processes. It does not replace the original audio mix, but instead offers an alternative that viewers can switch to if they so choose. Not only can Dialog+ improve new recordings, but it can also increase the speech intelligibility of existing material. Furthermore, Dialog+ perfectly complements the MPEG-H Audio object-based production and it can be used as part of the MPEG-H Audio production tools. Dialog+ enables better speech intelligibility for legacy services and advanced personalization for new MPEG-H Audio services at the same time. This allows for cost-effective end-to-end workflows that perfectly fit the requirements of broadcasters today. Embedded in the object-based workflow and transmission chain, personalization and interactivity offered by producers and broadcasters can be as manifold as the needs and individual taste of their audience.
Header image © Fraunhofer IIS
This post is also available in: Deutsch