In conversation with.. VRM
"Journalism can only fulfil its task if it is and remains relevant"
VRM is a regional media company in Hessen and Rheinland-Pfalz with its headquarters in Mainz. VRM reaches around 1.7 million unique users with its digital offerings. Jule Lumma is responsible for content development as a member of the editorial team and combines the areas of digital reporter storytelling, moving images, audio, documentation and editorial development.
Our conversation with Mario Geisenhanslüke, Deputy Head of Content Development at VRM:
What are your company's core competencies and which digital solutions for processing audiovisual content do you already use today?
"VRM's core competence is regional and local journalism. Our editors research the relevant topics for our readers and users every day - and prepare them in the best possible way on all channels: from our news portals and NewsApps to social media and ePaper to the print edition of the next day. Audio and moving images are a central part of VRM's editorial content strategy. In both audio and video, we rely on a modern system landscape in (post) production for the highest possible quality, as well as on a modern CMS in distribution, in which various assistance systems help us."
In your opinion, what forms of accessibility are urgently needed in the provision of media content?
"Journalism can only fulfil its task if it is and remains relevant - and we can only be relevant if we reach as many to all people in our circulation area as possible. That's why we try to offer our content as barrier-free as possible: from a read-aloud function in the ePaper to the subtitling of all videos or the provision of transcripts of all podcasts to a currently running experiment for the AI-supported provision of subtitles during ongoing livestreams."
Most companies have an ever-increasing amount of video and other digital content in their media archives. What do you think are the biggest challenges in managing, processing and using this content?"The biggest challenge for us is to (re)find all content in order to be able to use it meaningfully. For this purpose, we will rely on a central content management in which all editorial content - from texts and images to all storytelling elements to videos and audio files - can be found. Another challenge remains the meaningful addition of meta data per content."
How can AI-driven transcription solutions be used sensibly and what are the advantages?
"We already rely on the automatic generation of transcripts in the podcast area - both as a supplement to the meta data in our audio CMS and with regard to accessibility. The same applies to videos. We also use aiconix for preliminary work in the subtitling of longer videos as well as recorded spoken interviews. The advantage is clearly the time saved for what would otherwise be a very tedious and lengthy task."
What are currently still the biggest difficulties in using speech-to-text solutions?
"The quality - at least in the German language - is simply not good enough even under optimal conditions (one person speaking, without dialect, perfect sound quality). That's why STT solutions only ever do the preliminary work for our editors. Where this is not possible (for example, in the case of AI subtitling of livestreams), we explicitly point out that this text was created automatically."
For which fields of application do you see the greatest potential for the use of automated transcription?
"There is not this one field. The generation of searchable meta-data for content in the field of audio and video is incredibly valuable - and would be inconceivable for us without STT solutions. But it also makes work easier for reporters during interviews or in the area of video editing, for example."
How do you work with aiconix and how do you use aiconix solutions?
"We have a long partnership with aiconix. We have been using a tool for the automatic transcription of verbatim interviews as well as videos for a long time, whereby here the transcripts are "only" the basis for further work in the editorial office. In addition, the aiconix machine works together with our video CMS. In addition, an experiment of ours is running together with aiconix and Contentflow - our partner in livestreaming."
What solution enhancements / features would you like to see from aiconix in the future?
"For us, increasing the quality of STT remains crucial - especially in terms of multiple speakers, dialect and dependence on the sound quality of the recording."