Public Service and accessible live streaming

Every year the Swedish Public Service broadcast corporations and The Swedish Disability Federation (HSO) meet to share insights and needs. WestreamU have earlier had the honor to live stream these events in 2011 and 2013. But this year was special.

Handikappförbunden (HSO, our customer) wanted only one live stream (via that included live sign language interpretation and subtitles. In addition we faced three new challenges. First, the moderator spoke only sign language. Second, one presenter spoke English, but should be interpreted live to Swedish. Third, the organizer wanted subtitles on the stage too. It took some planning, but turned out pretty well. AB Stockholmstolkarn STAB provided sign language interpreters (for the stage/stream and the moderator) as well as Speech To Text Reporters (STTR).

Screen dump from

The event was held October 1, 2015 between 9:00 at 13:00 at Radiohuset (SR’s main building in Stockholm). This year’s theme was “En av oss?” (eng. One of Us?), see their web site The program included presentations from BBC, a discussion with the CEO’s of Sveriges Televison (SVT), Sveriges Radio (SR) and Utbildningsradion (UR), parallel workshops, as well as an award ceremony.

The plenary sessions took place in Studio 4. The organizers decided to live stream only from the workshop that was held in the same room. We knew from earlier experience that from a technical perspective live streaming from Radiohuset should not be a problem. The stage lights in a radio studio, at an event for a live audience, is perhaps not optimal for web TV. At least we attempted to ensure sufficient light at the sign language interpreter’s spot.

Basically we designed a triple-box layout for the video frame, somewhat inspired by the “EDSol format” by Edsol Producciones. The three boxes contained respectively the original camera/graphics mix (upper right), the sign language interpreter (left), and the subtitles (bottom).

We strived to find a suitable balance between the boxes’ sizes. Keeping the original box as large as possible, and maintaining the 16:9 ratio of the stage cameras was important. At the same time it was of course crucial to keep the sign language interpreter box large enough to support sign readibilty. Up in the left corner we got room for the event’s logo.

Actually, when the original video contained a double box (presenter torso + powerpoint slide), we ended up having a hexa-box (triple + double+ logo ;)

For the subtitling we used the beta version of Text-on-Top Video that you can read more about in our blog post from the Intersteno 2015 congress. The three videos embedded here (click to activate) describe how Text-on-Top Video works.

This solution allowed us to feed the text as subtitles into our live production, and to record the text for later conversion to closed captions. In addition the reported text could be used by additional Text-on-Top devices/applications. In theory we could have provided Text-on-Top dongles to the audionce so that they could in real time get the text on their laptops or Android phones/tablets – in up to 10 languages!

As normal in Sweden the text reporters/interpreters worked in pair. During the talks in English one of them listened to the presenter and translated into Swedish for his colleague, who in turn typed the words. The interpreters ingeniously created a small paper megaphone to overcome the problem with ambient sound that we did not think of during the planning.

A separate PC picked up the text produced by the reporters and projected it on to a large flat screen stage monitor. It was positioned to the far right in order to keep it out of the camera’s views. Given the circumstances the interpretation into written words had a very good sync. The delay was around 3-4 seconds for Swedish, and typically less than 10 seconds for English.

Watch the on demand videos

After the event we cut our local recordings into separate films and uploaded them to HSO’s YouTube channel. HSO decided to keep the original live subtitles when Swedish was spoken, as a way to exemplify how it actually looks live. For comparison we converted the recorded text to closed captions when English was spoken (BBC’s presentations). These captions was also re-timed to be in better sync. For the English films YouTube has auto-generated English subtitles/CC.

A look behind

Filmed with a smartphone during the production. It starts with a stage overview, moves over to the video mixer that adds sign language and subtitles to the original video mix. Then continues to the original video mixer, over to the speech to text interpreters, then a quick view of the Text-On-Top PC, and finally zooms into the text monitor at the stage. Note how one of the interpreters translates orally from English to Swedish, while the other writes in Swedish.