W3C home > Mailing lists > Public > wai-xtech@w3.org > August 2010

RE: TTS from web content?

From: Richard Schwerdtfeger <schwer@us.ibm.com>
Date: Thu, 26 Aug 2010 08:59:12 -0500
To: "Sina Bahram" <sbahram@nc.rr.com>
Cc: "'David Bolter'" <david.bolter@gmail.com>, "'Léonie Watson'" <lwatson@nomensa.com>, "'Victor Tsaran'" <vtsaran@yahoo-inc.com>, wai-xtech@w3.org, wai-xtech-request@w3.org
Message-ID: <OFE56BC01C.27D48F45-ON8625778A.0073B239-8625778B.004CD6B6@us.ibm.com>
Hi Sina, 

Today, there are tools that access the web page in the cloud and read it. 
What they are currently doing is converting text to audio streams and 
sending those to the client. This is not very performant. If we did have 
access to TTS via the browser it would operate much faster. 

I should note that this solution would not be solely for screen reading. 
IBM has developed solutions like WebAdapt2Me that allows seniors to 
highlight text to be spoken. Another use case would be cognitive and/or 
reading impaired users (a much larger group than blind and low vision) 
which would benefit from the ability to highlight and read text on the 
page.

Could we create self voicing applications - sure. Although that may depend 
on the disposition of DOM Mutation events and a replacement down the road. 
We will have to see.

The big barrier to all this has been browser security. Unless you had a 
plug-in you could not get to these features. A standard script API would 
put things in the hands of the developer.

I think putting TTS API support in the browser is a win.

Rich

Rich Schwerdtfeger
CTO Accessibility Software Group



From:   "Sina Bahram" <sbahram@nc.rr.com>
To:     "'Léonie Watson'" <lwatson@nomensa.com>, "'Victor Tsaran'" 
<vtsaran@yahoo-inc.com>, "'David Bolter'" <david.bolter@gmail.com>, 
<wai-xtech@w3.org>
Date:   08/20/2010 09:52 AM
Subject:        RE: TTS from web content?
Sent by:        wai-xtech-request@w3.org



I could imagine a system where the JS access to the TTS is shared across a 
consistent priority model to that same TTS by the screen
reader, so that both can interact with it successfully; however, use 
concepts of priority, interrupt ability, blocking, and so on to
facilitate a seamless interaction.

This does however bring up the classic argument over self voicing versus 
screen reading; however, to that end it's imaginable that
one can have a hybrid approach whereby including a simple .js file from 
somewhere else on the web, suddenly enables your application
to speak, and that .js file could either contain self voicing capabilities 
or a screen reader of sorts.

Take care,
Sina

-----Original Message-----
From: wai-xtech-request@w3.org [mailto:wai-xtech-request@w3.org] On Behalf 
Of Léonie Watson
Sent: Friday, August 20, 2010 4:02 AM
To: Victor Tsaran; David Bolter; wai-xtech@w3.org
Subject: RE: TTS from web content?

"I think this would be a great addition to the realm of existing assistive 
technologies. Having Javascript API access to the OS's
built-in TTS would enable us to generate audio descriptions for videos 
directly from the web page in conjunction with ARIA live
regions, for example. Such a capability should be togglable by the user 
and/or disabled if a screen reader is detected (although a
shortcut key-based approach may be sufficient)."

                 Agreed, it really would open up some terrific 
possibilities. Handling the relationship with screen readers will 
certainly be
a challenge, as I can imagine times when both/neither would be desirable 
from the user's point of view.


Regards,
Léonie.

--
Nomensa - humanising technology

Léonie Watson            |  Director of Accessibility
t. +44 (0)117 929 7333 


-----Original Message-----
From: wai-xtech-request@w3.org [mailto:wai-xtech-request@w3.org] On Behalf 
Of Victor Tsaran
Sent: 19 August 2010 22:27
To: David Bolter; wai-xtech@w3.org
Subject: RE: TTS from web content?

Hi David,
I think this would be a great addition to the realm of existing assistive 
technologies. Having Javascript API access to the OS's
built-in TTS would enable us to generate audio descriptions for videos 
directly from the web page in conjunction with ARIA live
regions, for example. Such a capability should be togglable by the user 
and/or disabled if a screen reader is detected (although a
shortcut key-based approach may be sufficient). The power of this feature 
will depend on how many of OS's built-in TTS properties
will be exposed by the browser to the Javascript developer, eg voice rate, 
tone, pitch, voice, person, volume etc.

Is Firefox going to be one such browser? :) Thanks, Victor

-----Original Message-----
From: wai-xtech-request@w3.org [mailto:wai-xtech-request@w3.org] On Behalf 
Of David Bolter
Sent: Thursday, August 19, 2010 8:22 AM
To: wai-xtech@w3.org
Subject: TTS from web content?

  Hi all,

What do you think about having browsers provide built-in text-to-speech 
capability to web content? While I imagine a declarative
approach might be quite interesting I think we can go farther faster with 
a JavaScript API approach. The main two concerns I have
are:

1. We don't want to encourage unpolished aural interfaces.
2. We don't want to conflict with traditional screen readers.

The biggest potential I see is:

1. Innovation in Aural interfaces. The same kind of innovation we see 
happening in visual DHTML interfaces.
2. TTS solutions in places, and on devices where traditional screen 
readers are problematic. For example, perhaps on some mobile
devices that are currently not accessible.
3. The TTS can be done in the browser, on the native platform (e.g. 
Voice Over on OSX), or 'in the cloud'. We just need to get the API right.

Are we ready? Please speak up.

cheers,
David
Received on Thursday, 26 August 2010 13:59:51 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:16:09 GMT