- From: Satish Sampath <satish@google.com>
- Date: Tue, 16 Nov 2010 17:54:39 +0000
- To: Bjorn Bringert <bringert@google.com>
- Cc: Michael Bodell <mbodell@microsoft.com>, Dan Burnett <dburnett@voxeo.com>, "public-xg-htmlspeech@w3.org" <public-xg-htmlspeech@w3.org>
In the case of a trusted environment where the user thinks all the recorded audio is going to the recognizer they set in the UA, giving the web page/app access to the raw audio is a privacy issue. This is to prevent malicious web apps using the speech input API as a loophole to get raw audio from the user machine where other html APIs are more restrictive. We should perhaps reword this requirement to indicate that the web app should be given raw audio access only after explicit consent from the user. Cheers Satish On Fri, Nov 12, 2010 at 9:10 PM, Bjorn Bringert <bringert@google.com> wrote: > That seems like a fair use case. The original idea behind the > requirement was to disallow eavesdropping I think, but assuming that > the speech recognizer is good, you don't really need raw audio for > that. I'm ok with dropping this requirement altogether. > > /Bjorn > > On Fri, Nov 12, 2010 at 9:04 PM, Michael Bodell <mbodell@microsoft.com> wrote: >> Hmm, I'm not sure I agree. Even when using the default recognition service I can easily imagine scenarios that would require access to the audio via a record and recognize interaction. For a simple example consider a web application that allows a user to issue commands and/or dictate a SMS message and/or leave a voicemail. You may want to do that all in one user interaction. Or, it is often the case that you want to recognize speech and if it fails still have the audio. I.e., attempt to say an audio prompt that will be sent as text in an SMS message, but if the system fails to recognize your message very well you may opt to send it as an audio voice mail message instead. In this case you want the recognition result (to check with the user if it was ok, and if ok use that), but you also want the recording file in case recognition didn't work. >> >> What is the motivation (or what) that we are trying to capture and could we phrase it or address that without disallowing a user agent to use a more robust default speech service that allow a record and recognize scenario? >> >> -----Original Message----- >> From: public-xg-htmlspeech-request@w3.org [mailto:public-xg-htmlspeech-request@w3.org] On Behalf Of Bjorn Bringert >> Sent: Friday, November 12, 2010 1:22 AM >> To: Dan Burnett >> Cc: public-xg-htmlspeech@w3.org >> Subject: Re: R28. Web application must not be allowed access to raw audio. >> >> Since we already have a requirement that web apps should be allowed to specify their own speech recognizer, this requirement seems impossible to enforce. I think that we should restrict it to: >> >> "When using the user agent's default speech recognizer, web applications must not be given access to the captured audio." >> >> Together with "FPR10. If browser uses speech services other than the default one, it must inform the user which one(s) it is using." this means that captured audio will never be sent to a web-app specified location without the user being informed. >> >> /Bjorn >> >> On Fri, Nov 12, 2010 at 5:40 AM, Dan Burnett <dburnett@voxeo.com> wrote: >>> Group, >>> >>> This is the next of the requirements to discuss and prioritize based >>> on our ranking approach [1]. >>> >>> This email is the beginning of a thread for questions, discussion, and >>> opinions regarding our first draft of Requirement 28 [2]. >>> >>> Please discuss via email as we agreed at the Lyon f2f meeting. >>> Outstanding points of contention will be discussed live at an upcoming teleconference. >>> >>> -- dan >>> >>> [1] >>> http://lists.w3.org/Archives/Public/public-xg-htmlspeech/2010Oct/0024. >>> html >>> [2] >>> http://lists.w3.org/Archives/Public/public-xg-htmlspeech/2010Oct/att-0 >>> 001/speech.html#r28 >>> >>> >> >> >> >> -- >> Bjorn Bringert >> Google UK Limited, Registered Office: Belgrave House, 76 Buckingham Palace Road, London, SW1W 9TQ Registered in England Number: 3977902 >> >> >> > > > > -- > Bjorn Bringert > Google UK Limited, Registered Office: Belgrave House, 76 Buckingham > Palace Road, London, SW1W 9TQ > Registered in England Number: 3977902 > >
Received on Tuesday, 16 November 2010 17:55:09 UTC