<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Pete,<br>
<br>
Real-time-text does not refer to human-speech-to-ASCII/UNICODE-text.
Rather, it refers to two people communicating via text, each typically
entering text from some flavor of keyboard, with their keystrokes being
transmitted "in real time" as they are entered. Rather than only being
transmitted after pressing a SEND key equivalent (e.g. <CR>).
This is seen as particularly important for E911 services, where someone
might not be in a position to say everything and then press SEND.<br>
<br>
You could certainly hook up a dictation system to this, but that would
be an additional layer on top of the ANPRM RTT requirement, not a part
of it. At least as I understand the ANPRM.<br>
<br>
<br>
Regards,<br>
<br>
Peter Korn<br>
Accessibility Principal<br>
Oracle<br>
<br>
<blockquote cite="mid:4BC38757.1070102@a11ysoft.com" type="cite">
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
Regarding the minutes, regarding the following:<br>
<br>
PK: ... The refresh adds a bunch of new rules. One of the notable
additions is real-time text for the deaf and a much more significant
effort on the deaf and on communication technologies<br>
<br>
PB: On the speech reco angle, did they talk about whether it was good
enough for what we want to do?<br>
<br>
PK: I don't remember any mention of speech recognition. The NRPM says
at a high level that someone without hands needs to be able to use your
app. Doesn't specifically mention speech recognition. Apps that support
audio and video chat must also support real-time text. <br>
<br>
This is an accurate representation of the exchange during the meeting,
but what I really was asking is, in the case where speech reco might be
used to transcribe speech, will the current (or near in) state of the
art of speech reco technology be able to provide acceptable real-time
text, considering the challenge of large vocabularies, speaker
independence, the variety of speakers, and conversational speech? Or
do the requirements allow for the use of real time human transcribers
(either local or remote)?<br>
<br>
<div class="moz-signature">-- <br>
<font color="#0060a0" face="Arial, sans-serif" size="1"><b>Pete Brunet</b></font><br>
<font color="#0060a0"><span style="text-decoration: line-through;">
</span></font><br>
<font color="#333333" face="Arial, sans serif" size="1">
a11ysoft - Accessibility Architecture and Development<br>
(512) 238-6967 (work), (512) 689-4155 (cell)<br>
Skype: pete.brunet<br>
IM: ptbrunet (AOL, Google), <a moz-do-not-send="true"
class="moz-txt-link-abbreviated" href="mailto:ptbrunet@live.com">ptbrunet@live.com</a>
(MSN)<br>
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://www.a11ysoft.com/about/">http://www.a11ysoft.com/about/</a><br>
Ionosphere: WS4G
<br>
</font>
</div>
<pre wrap="">
<hr size="4" width="90%">
_______________________________________________
Accessibility mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Accessibility@lists.linux-foundation.org">Accessibility@lists.linux-foundation.org</a>
<a class="moz-txt-link-freetext" href="https://lists.linux-foundation.org/mailman/listinfo/accessibility">https://lists.linux-foundation.org/mailman/listinfo/accessibility</a>
</pre>
</blockquote>
<br>
</body>
</html>