NPR logo

Google Launches Closed Captioning For YouTube

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Google Launches Closed Captioning For YouTube


Google Launches Closed Captioning For YouTube

Google Launches Closed Captioning For YouTube

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

Google this week introduced closed captioning for the deaf on its YouTube video site. Ken Harrenstien, the lead engineer behind Google's automatic captioning technology, says that as a deaf person he lobbied his bosses for years to introduce the technology.


Many, if not most of us, have watched a video on YouTube. But have you ever read one? Well, now you can. Google, which owns YouTube, has begun offering closed captioning for the English language videos on its site. You simply go to a small box in the lower right hand corner of the screen and click your mouse.

The announcement came late last week and it's meant to open YouTube to the hearing impaired.

I'm joined now by Ken Harrenstien who is himself hearing impaired and who is the lead engineer behind Google's automatic captioning technology.

We're speaking to him through an interpreter from Google's headquarters in Mountain View, California.

Ken Harrenstien, welcome to the program.

Mr. KEN HARRENSTIEN: (Engineer, Google): (Through Translator) Hi, I'm happy to be here.

SIEGEL: And have deaf people been lobbying for Google to do this?

Mr. HARRENSTIEN: (Through Translator) No. I am a deaf person. I've been lobbying myself to do this.

SIEGEL: And to do this means to be translating into text on the fly, whatever speech is on a Google video, we see it spontaneously being turned into text?

Mr. HARRENSTIEN: (Through Translator) That would be nice. But in reality, it's a little more complex than that. I will group them into different kinds of captions: One prepared by a person, uploaded by the owner of a video. Now, the new stuff we call automatic captions, we use speech recognition to try our best to generate captions for the video that don't have them.

SIEGEL: The challenge, of course, is how good will the captioning be?

Mr. HARRENSTIEN: (Through Translator) Yes, it's a challenge for sure. I can't really say what percentage of videos have good captions or bad ones. I expect people will find that out for themselves. Several problems: First, we only support English right now; second, there could be a lot of noise in the background - music, for example. For some videos, we do pretty well.

SIEGEL: Yes, I looked at a few of them today. And I'll read to you, you know, some of the obvious pitfalls that I discovered in one. This was a description of a new AT&T phone. And it should have said: AT&T's first Android phone, long anticipated. It's the Motorola Backflip with Moto Blur. This will hit the stores March 7th.

It came out in the text: AT&T's first and read from long anticipated. It's the moral back flip with...

(Soundbite of laughter)

SIEGEL: ...with Moe Biller missile. Hit the stores March 7.

Let the record show, Moe Biller was the head of the Postal Workers Union for many years, and I guess his obituary is somewhere in the database that the...

(Soundbite of laughter)

SIEGEL: ...the computer is drawing upon. Obviously you have real problems with trade names and proper nouns.

Mr. HARRENSTIEN: (Through Translator) Oh, yes. In fact, YouTube itself comes out as you too.

(Soundbite of laughter)

SIEGEL: Well, mistakes and all, is it that valuable to people who can't hear, to have text that can be so far off?

Mr. HARRENSTIEN: (Through Translator) Okay, I have two answers to that. Basically, yes and yes. Yes for myself because just having some clue as to what's going on, I find it very helpful. But the second part of this answer is it's a huge benefit to people who create zip videos and want captions on them. But, again, maybe they don't have the time and the resources to do it.

What they can now do is download the transcript and they can use any editor to fix up the words that are wrong. Upload it again as a basically perfect caption video.

SIEGEL: Would you assume that five years from now, the transcription will be significantly better than it is today?

Mr. HARRENSTIEN: (Through Translator) Oh, yes. I'm a technological optimist. Some things just take longer than others.

SIEGEL: Well, Ken Harrenstien, thank you very much for talking with us about Google's new captioning.

Mr. HARRENSTIEN: (Through Translator) Sure.

SIEGEL: Ken Harrenstien, who is hearing impaired, is the lead engineer behind Google's automatic captioning technology. He spoke to us through an interpreter from Mountain View, California.

Copyright © 2010 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.