IBC Q&A: IBM Cloud Video’s Clevinger explains how Watson automates closed captioning at US Open
Over the past year, IBM has made building out its cloud-video technology platform a top priority, acquiring object-storage service Cleversafe in October 2015 (now known as IBM Cloud Object Storage), video-management startup ClearLeap in December 2015, and live-streaming outlet UStream in January of this year. In conjunction with the Ustream acquisition, IBM combined Ustream and Clearleap to create the IBC Cloud Video Unit, while Aspera and IBM Cloud Object Storage sit under the IBM Cloud and leverage the Cloud Video Unit’s technology. With the transition now complete, the IBM Cloud Video Unit is at IBC 2016 to highlight its new portfolio of solutions and how it is leveraging other IBM technology — like its Watson natural-language processing and machine-learning system — to advance its offerings.
SVG sat down with IBM Cloud Video Unit Senior Director of Product Management David Clevinger on the IBC show floor to discuss the M&E market’s reaction to IBM Cloud Video, how use of the cloud is evolving for live production and live streaming, and how Watson is being leveraged to provide automated closed-captioning for live-streaming of the US Open tennis tournament.
How do the four platforms of IBM Cloud Video — Ustream, Clearleap, Cleversafe, and Aspera — work together as a whole?
We don’t want to be just a repository for your content; we want to actually make it richer and make it smarter. The US Open is a great example of that.
Sports is different from the rest of the media-entertainment space. Sports needs a lot of low-latency solutions; it needs a lot of immediacy in terms of interaction with the consumer. What Ustream brings is that live-streaming expertise, that immediacy, and that interaction with consumers.
ClearLeap brings vast expertise in terms of ingesting VOD and live content, managing that content, applying metadata, and distributing it back out again to get it in front of users in the ways they want to receive it. And then managing the monetization of that for the client — whether it’s authenticated access or subscription product — we’ve got all those pieces. So, between those two things, we’ve got a really robust infrastructure for doing everything from a live event to VOD to playout to subscription model; we’ve got that full end-to-end service.
Cleversafe provides secure object storage; so it’s for storing really massive files. For example, let’s say you have a sports-event 4K-streaming asset that lasted six hours; you’re talking about several terabytes of information. You need a solution like Cleversafe to store that content as opposed to having it live someplace that’s less secure.
And then Aspera is an ultra-fast file-transfer service. If you’re a content provider and you’ve got large bodies of assets that you need to deliver to us, then we work with Aspera to deliver that content in a rapid fashion.
With those four platforms in place, we’re truly offering an end-to-end solution.
What has the market’s reaction been to the launch of IBM Cloud Video?
This is the first show where we are completely rebranded as IBM Cloud Video, and we have gotten a tremendous reception. I think people are very interested in the story that we’re telling and the components of the product that we’ve brought to market. There are a lot of niche services that do very specialized things, but we’ve got many different components of the ecosystem that fill many gaps for many clients. We can bring all of it or only the pieces that you need as a client, and I think that’s really important.
How is IBM Cloud Video leveraging the technical resources from the rest of IBM and interacting with other sectors of the company?
I think that’s been one of the most exciting elements for us. A perfect example is how we are using Watson for the US Open this week. We are demoing that live here at the booth as the Open is happening.
We are taking in a live feed and sending a derivative of that to a Watson API. That API is sending us back a text file, which is converted to a WebVTT [closed-caption/subtitle] file that we use as closed-captioning overlay for the US Open live stream. A media manager on the US Open client side is reviewing, not keying in the closed captioning, just simply reviewing based on accuracy scores. So we are able to see when Watson is 88% accurate about a line [of text] or 13% accurate because there is a proper name in it, in which case the media manager knows to check that line more closely. It’s streamlining the workflow on the media-manager side, and the more that media manager reviews it and provides feedback to Watson, the more advanced Watson will get on the closed captioning. Watson will understand better over time what it’s looking for.
Prior to the US Open, we fed [Watson] huge amounts of tennis articles and the history of tennis, like background on Arthur Ashe Stadium and Martina Navratilova, and so on. And it’s completely extensible: you could teach Watson to understand any other sport and any other vertical. It’s really cool when you see it in action.
How do you see the use of cloud-based workflows and systems evolving in the live-sports production and streaming sector?
The power of the cloud is that it gives you the flexibility to grow over time in a way that’s much more graceful than having something that’s local. The sports market is changing all the time and always innovating. As we add more services based on other IBM assets and integration with other IBM products, it’s going to be easy for us to work that in over time without disrupting our client’s workflows. We’re going to be able to help them grow as we grow.
In my view, the key to live sports for us is not just to deliver a quality live feed but also to enrich it for the client and for the consumer. That’s things like creating metadata automatically, creating closed captioning automatically like we’re doing with the US Open. And then there are other pieces that we can do downstream that we’ll be announcing in the future that take it several steps further.