Captioned Telephone

Captioned Telephone - Part Six

Editor: This article is part of our coverage of the 2007 TDI convention and is brought to you by the folks at NVRC. You do not need permission to share this information, but please be sure to credit NVRC.

Robert Engelke and Pam Holmes were the presenters. This is part six of six parts.

~~~~~~~~~~~~~~~~~

Questions and Answers (continued)

Q: I understand that this is being funded by the feds. The government is concerned about fraud through the Internet. Do you have ways to protect that particular thing through your Web CapTel?

Rob Engelke: It is a concern that everybody has with any of these technologies. The types of fraudulent use of IP Relay evidently center pretty much on people using the relay in a way where they are invisible. They can use a computer or other type device where no one hears their voice, no one knows where they're calling from, and since they're just typing, they're pretty much anonymous. Captioned telephone, because of its nature, does not really lend itself to that kind of fraud in nearly the same way. You actually hear the person's voice, and it's a conversation taking place just like any other telephone conversation would. The call assistant never is heard by either party, so they can't be used as a front. So we believe the incidence of that kind of fraud will be very low with captioned telephone in any form. In addition, the people who will be providing captioned telephone services over the Internet have the option to give out user names and passwords, and I think they are planning to do that at least in some instances

Kevin Colwell: One element that makes it less prone to that kind of misuse is that as the calls are connected, there is caller ID. In the world of the dialable phone, as you know from using relay services and CapTel today, it has true caller ID. So the caller ID moves down the line with the call. Your identity is arriving as the call to the called party is arriving. Unless you blocked that, it's delivering information about who you are.

Q: You now are getting to the point where it's putting all the pieces together and I'm really excited about it. Do you have data about the delays from when a person speaks the word and the time the actual word gets displayed?

Robert Engelke: It's actually pretty straightforward. The call assistant has to hear what the person says. They usually begin transcribing two or three or four words after the person begins to speak. According to my son, who is a linguist or claims to be anyway, it takes our brains a little while to sort of tune in to what people are saying before we actually are capable of understanding it, and then, in this case, saying it over again. So that's actually quite a bit of the delay that we see.

The next part of the delay is caused by the system itself. The system takes a few tenths of a second to transcribe the information. In some cases, if the words or the sentence have context that is difficult for the computer to understand, it will delay a few more words before it makes up its mind as to what you're really trying to tell it. This can add another second or two. Of course the call assistant has to watch what the computer came up with and decide whether it's okay or not. And that might take another second or so.

They send it out and you get it about 3 to 4 seconds after the word was spoken. If there is something wrong, then the correction usually takes about 3 or 4 seconds. There's also the natural delay in the system, which is about 3 seconds but can be up to 7 seconds. Longer delays do occur and we call that stack up. If you're the kind of person who likes to talk and talk, like me, and you don't ever pause, then any delays that are going on in the system effectively stack up on top of each other. If the CA is a little bit behind, they have to fix an error, all those delays stack up. So the next few words that I say are delayed by all the things that happened prior to that.

Conference call captioning is a very difficult task for the CA. We have seen in delays in captioning these calls get up into a couple of minutes, and it's all stack up. It's not that the CA isn't transcribing fast enough. It's just that they are trying to keep their text accurate. As soon as the calls pause for a little while, it gives the CA and the whole system a chance to catch up. For those of you who are CapTel users, there is a little known trick built into the system to help you catch up if you get behind, and I'll let Kevin describe it for you.

Kevin Colwell:
There are actually two ways to do it. If you have a large amount of stack up, one is to interrupt the party that is talking, briefly, to ask a question or make a statement. That causes the person speaking to stop momentarily and allows the captions to catch up. That is a conversational trick of CapTel users.

The second way, if you have a conventional CapTel phone and want the captions immediately to catch up to where you are, is to touch any key on the touch tone pad of the phone and the captions will catch up. You can use this if you're following along well enough and the captions are farther behind than you want.

Q: What is the major difference between Web CapTel, and IP Relay with the VCO? Can you help me understand the difference between those?