ispeech swot analysis · 2013-07-22 · 2013 global public inclusive infrastructure (gpii)...
TRANSCRIPT
Interchangeable Modalities W3C Workshop on MultiModal Interaction
22-23 July 2013, New York
Background: iSpeech is a Text-to-Speech & Speech Recognition Company
Developer
s 15,000+ devs in
12 months, 2x
growth of
Enterprise Fast growing list of
Mobile, Auto, Home
and Publishing
Customers
Consumers 30+ million app downloads
Developer
Experience 25,000+ developers
Enterprise Mobile, Auto, Home, &
Publishing Customers
Consumer Experience 30+ million app downloads
Credibility: Developer Ecosystem
> 25K developers registered > 2 billion API calls serviced > 99.9% uptime .
Mobile Devs Mobile OEM/OS Auto Home Publishing
Speech: New Frontiers
2 Mobile/Nav 1 Entertainment 3 eLearning 4 Telephony
Growth Following
New Use Cases
Breakdown of Developers by Segment & Activity
Developers
API Usage
Challenges of Speech Technology
Many technologists have never ‘experienced’
working directly with speech technologies
Uncharted Technology Waters: Audio DSP, NLP,
Domain/Grammar/Lexicon, Multimodal UI
Speech mirrors humans; more like ‘wetware’
than ‘software’?
Life-cycle of continuous adaptations and QA
Consideration: Speech Technology Value Chain
ASR NLP TTS
Standards & HTLM5
• 25,000 Developers X 10 ways to package web services (APIs and SDKs) And that’s just
Cloud, dozens more embedded engines to account
• HTML5 adoption – audio playback of TTS ok, audio recording (ASR) not widely used
• Example: impact of mature standards on use of Speech Technology: VoiceXML, SSML, SRGS, MRCP
Talkz™ Case Study
• Talkz - successor to Drivesafe.ly • Interchangeable Multimodal App • Available today through iTunes
Section 508
1998
2000 Windows Narrator
VoiceOver Mac OS X
2005
2006 YouTube Captions
1st Accesible
Smart- phone 2009
2013 Global
Public Inclusive Infrastructure
(GPII)
Multi-Modal’s Silver Lining : Universal Accessibility
“The gap between usability and accessibility is narrowing and with it the digital divide between disabled and non-disabled people.” - Robin Christopherson, AbilityNet
Conclusions for Multi-Modal Developers
• Plan to Partner & Partner to Plan
• Multimodal is a CENTRAL UI pillar, not an after thought