July 2000
|
|
Cutting
The Cord BY
CAROL DRZEWIANOWSKI |
Go
Right To:
Corporate Solutions News
Speech Recognition On A Grand Scale
Speech Portals Bring The Web To
The Untethered World |
|
Like most people, I'm looking forward to spending a few lazy days of
relaxing without my fingers attached to a computer keyboard. That's right,
I actually want to be free of my PC for a while. But calm down! (I know
there are a few of you panicking right now -- how can I exisit without
being constantly connected?)
Well, even if I don't have a PC nearby, chances are I still have a regular
old telephone. And with the hubub about speech portals lately, I know that
I can get access to just about any information I need, as long as I have
dialtone.
The Kelsey Group, a leading e-commerce, voice, and Web technology analyst
firm, expects that 18 million consumers will use some kind of speech
recognition portal by 2005. With Internet-based e-commerce sales
increasing every year, emerging companies are seeing the many benefits of
extending their corporate information and e-commerce transactions to
anyone with a phone. As efforts increase to reach the burgeoning wireless
audience, The Kelsey Group believes that advertising revenues will reach
$17 billion in 2005, with local advertisers alone accounting for about $6
billion.
Lycos has teamed up with Quack.com to develop a voice-driven Web portal.
The free service will allow users to surf the Web using the spoken word
from any phone, thanks to a toll-free number. This service is not
positioned merely for Web users on the go; it also plans to promote the
service to non-Internet users so they can more rapidly access the
information they deem most relevant.
Natural MicroSystems is also jumping in the portal pool. They have
announced HearSay, an integrated, extensible product designed to drive the
rapid development and deployment voice voice-driven Web commerce. And
Intel and its Dialogic subsidiary are adding continuous speech processing
technology to its Dialogic family of voice processing products that
support speech-based Web access. This technology, developed in conjunction
with leading speech companies, allows developers to build speech
recognition applications, such as voice portals, at a lower cost using a
single-board solution.
So what makes a speech portal different from a cell-phone service that
deliver Web content? Some 600 million mobile phones will be in use by
2001, say IDC market researchers. Many of those phones will be able to
access Web content. Well, believe it or not, not everyone owns a mobile
phone, much less a Web-enabled mobile phone right now. (And as we all
know, wireless phones go in and out of service areas, causing
disconnections, or no signal at all). Under these circumstances, speech
portals are ideal solutions for quick, reliable access to information.
It will be interesting to see what happens with speech portals over the
next few months, but for now you can read about the latest announcements
from some of the leading vendors.
Corporate
Solutions News
WireNix.com To
Create Speech-Enabled Wireless Applications
WireNix.com, a new Dallas-based technology company, said it will create
speech-enabled wireless applications based on open systems and application
standards, including Wireless Application Protocol (WAP). WireNix.com
plans to develop speech-enabled applications, such as location finders,
contact managers, and other information retrieval services for wireless
providers, Internet service providers, and wireless Web portals.
No. 520, comsolmag.com/freeinfo
New Features For Locus Dialogue's Liaison Software
Locus Dialogue has announced a series of new features for version 5.0 of
its Liaison software, which provides speech-enabled solutions combining
the availability of the telephone with the simplicity of the spoken word.
Principal enhancements to Liaison software include the DigitalLinks
optional module and greatly expanded system capacity. The DigitalLinks
module enables seamless interoperation with both digital and analog
telephone lines. Digital integration accelerates call transfer time, which
in turn increases the overall efficiency of a telephony system. The
capacity of Liaison solutions ranges from 100 to 50,000 directory entries
and from 2 to 32 ports, making Liaison software a cost-efficient speech
recognition solution for small, medium, and large client applications.
Version 5.0 of Liaison software also delivers a number of additional
product enhancements that improve CallerCare service.
No. 521, comsolmag.com/freeinfo
Sensory Releases Voice Direct 364 Speech Rec Kit
Sensory has released the Voice Direct 364 Speech Recognition Kit, a
low-cost kit and development tool designed to easily enable hands free
speech recognition capability of consumer electronics. Voice Direct 364
can be used to voice-activate all types of ordinary consumer electronics
like set-top boxes, microwaves, cars, remote controls, garage door
openers, alarm clocks, telephones, lights, Internet appliances and many
other types of electronic products. Voice Direct 364 features continuous
listening technology that allows a device to be switched on or off with
just the sound of one key word or short phrase. "Our goal is to price
these development tools so inexpensively that we can enable tens of
thousands of product developers to create voice- activated products each
year. Many of these Voice Direct 364 developers will succeed and become
large volume chip customers for Sensory's RSC-364 IC," said Todd
Mozer, Sensory's president and CEO.
No. 522, comsolmag.com/freeinfo
Interactive Telesis Joins VoiceXML Forum
Telesis has joined the VoiceXML Forum and is endorsing the VoiceXML
specification. Motorola, IBM, Lucent, and AT&T founded the VoiceXML
Forum to promote phone-enabled Internet access. With the backing and
technology contributions of these four founders, and the support of
leading Internet industry players, the VoiceXML Forum is uniquely
positioned to make speech-enabled applications on the Internet a near-term
reality. The Forum is focused on establishing a standard specification for
Voice Extensible Mark-up Language (VXML), a computer language used to
create Web content and services that can be accessed by telephone.
No. 523, comsolmag.com/freeinfo
Philips Speech Rec Technology Available On Lucent's Hands-Free Cellular
Phone Chip
Philips Speech Processing announced that its speaker-independent and
speaker-dependent discrete speech recognizers have been ported to the
Lucent Technologies Microelectronics Group DSP1629 DSP chip which also
hosts Lucent's adaptive acoustic echo and line echo cancellation, adaptive
noise suppression, simultaneous voice and data modem, and voice memo
recording for use in automotive hands-free applications. By adding speech
recognition to this device, users can now experience the safety and
convenience of voice-activated dialing when placing a cellular phone call
while walking, driving, or using their hands for other activities. Based
on its extensive experience in developing speech recognition solutions for
the automotive market, Philips has optimized its speech recognition engine
for high accuracy under harsh driving conditions in a variety of vehicles.
No. 524, comsolmag.com/freeinfo
Voyant's Voice Platform Supports
Video-On's Audio Conferencing System
Voyant Technologies, a platform provider for intelligent voice
applications, and Video-On, a leader in teleconferencing services, have
partnered to add Voyant's voice functionality to Video-On's recently
announced audio conferencing service called engageAudio. As part of the
agreement, Video-On has integrated Voyant's voice conferencing solution to
further enhance its video, audio, and document-based conferencing suite.
By providing robust, scalable, and reliable technology, Voyant's InnoVox
platform gives engageAudio customers full control over their conferences,
allowing them to initiate and conduct a call at anytime without
reservations or operator assistance.
No. 525, comsolmag.com/freeinfo
Lernout & Hauspie Ships L&H Voice
Xpress Version 5
Lernout & Hauspie (L&H) announced the shipment of L&H Voice
Xpress version 5, the newest offering in its family of award-winning
continuous speech recognition products. The new version includes
significantly improved accuracy and usability and support for e-mail,
Internet browsing, and chat applications. These features, and other
enhancements are designed to make the product more intuitive and to meet
general consumers' needs, helping to drive the use of continuous speech
recognition among a broader audience of consumers. As part of its accuracy
enhancements, L&H Voice Xpress version 5 includes Nothing But Speech (NBS)
Technology, a new disfluency filter that eliminates "ahh" and
"umm," sounds users often make while speaking that can increase
errors in dictation.
No. 526, comsolmag.com/freeinfo
[ return to the top ]
[ return
to the July 2000 table of contents ] |
|
|
Speech Recognition On A Grand Scale
BY BILL WOLFE
One-billion-plus a day. That's the consensus estimate of the combined
volume of e-mail and voice mail messaging traffic in the U.S. these days.
As inexorable changes in business continue, the traffic will only
increase.
In fact, the more active and mobile the user, the greater the need to keep
in touch, whether by voice mail, e-mail, fax, or other advanced calling
services. Their needs help drive the long-awaited advent of unified
messaging -- a crucial emerging service offering enabled by speech
recognition that is projected as one of the fastest growing
"convergent" applications (expected to grow to $12 billion in
market size in 2006, according to Dataquest).
Service providers are driven to creating a more personalized, long lasting
relationship with their customers providing a user friendly unified
messaging solution as a key way of adding profitable revenue to basic dial
tone.
The rapid growth of Internet applications and mobile users has created the
requirement for a next generation of solutions that combine voice and
Internet applications -- "unified communications." A unified
communications solution is a combination of real-time and message-based
solutions for both inbound and outbound communications based on an open,
IP architecture and offering a broad range of converged services including
voice, video, and data.
Using a simple, single account, subscribers can access their e-mail, voice
mail, and fax messages through a standard PC Internet message client, an
Internet browser or telephone -- regardless of location. They can create
voice messages using the telephone and send them to any other subscriber
on the service. They can respond to or forward existing e-mail messages
with voice annotations from any standard telephone.
REAL - TIME COMMUNICATIONS SERVICES
Building upon the IP-based messaging platform and the IP network
infrastructure, a unified communications platform lets subscribers take
advantage of a range of real-time communications services that improve
access to information and communications while providing complete control
over information delivery time and priority. These include:
- Call management and "follow-me" roaming capabilities, which
use speech recognition to identify and screen inbound calls.
- Enterprise application integration, to provide a voice interface to
applications such as customer relationship management, order entry,
customer databases, inventory management, and others.
- WAP (Wireless Application Protocol) message management, to enable mobile
subscribers to prioritize and route e-mails, voice mail, and fax messages.
- Voice-enabled Web servers, to create an integrated customer experience.
- On-demand conferencing, to improve the productivity of distributed
workgroups.
SCALABILITY
A unified communications solutions' IP-based architecture can scale
upwards at the network layer or the application layer. At the network
layer, standard network design engineering methods add capacity in the IP
cloud. Installing more voice or RAS (remote access server) boards in
gateway devices, or adding access servers at the point of presence, can
boost POP capacity. Application and message processing scalability is a
matter of increasing the number of servers. To add more directory capacity
(to store more subscriber profiles), simply add more directory servers and
use the directory vendor's partitioning or scaling features to maximize
performance. To add more message store capacity, add more message disk
drives.
Speech recognition will empower sophisticated network applications with a
highly natural and truly intuitive interface -- replacing rigid touch-tone
menus. But that interface must be built on a platform that's flexible,
open, cost-effective, and scalable.
Bill Wolfe is the general manager of Cisco Systems' Unified
Communications Business Unit, based in Richmond, Virgina. For more
information, please visit www.cisco.com.
[ return to the top ]
[ return
to the July 2000 table of contents ]
|
|
Speech Portals Bring The Web To The Untethered World
BY LEAH LESSER
W ho hasn't been stuck in traffic, left only to wonder if the flight you're
desperately trying to make will
be on time, or wondering where the latest stock du jour closed, only to be
left wondering because your teenager is glued to the family computer? None
of us are tethered to a computer all of the time, and incidents like these
are inevitable. But with the recent advent of speech portals, consumers
have untethered, anytime, anywhere access to popular Internet information.
Speech portals offer consumer news and information that anyone can
navigate and access from any phone, using speech recognition technology.
The increasing popularity of speech portals is obvious when you examine
the evidence:
- There are over 1.7 billion phones in the world.
- 90 percent of households in the Western hemisphere own phones.
- The Gartner Group measures the rate of wireless phone adoption as three
times the rate of new Internet subscribers.
- Consumers' expectations for instant information are already established
by the Web.
WHY NOW?
But, why is it now that speech portals are coming to market? Thanks to
great technological and user interface advances in telephony-based speech
recognition, speech portals are as easy to use as speaking into any phone.
One of the great advantages to speech portals versus getting Web content
on devices that have tiny screens, such as Internet-enabled cell phones,
is sheer convenience. "I would much rather talk into a phone to
receive information I can listen to rather than have to squint at a screen
I can hardly see and try to finagle my way into navigating around tiny
content," said one user at a recent speech portal focus group.
Not to mention when consumers are walking down the street, sitting in
their cars or otherwise multi-tasking, navigating screens becomes a very
dangerous proposition. This is where the power of speech recognition comes
in very handy.
FREE SPEECH FOR THE CONSUMER
Speech portals are following in the historical footsteps of the Web. Just
as Web sites offer free content and generate revenue from advertising,
many of the speech portals offer free services that are subsidized with
interactive, speech-based ads. These ads are permissive, relevant 5-8
second audio files consumers hear when interacting with a speech portal.
For example, a caller to a restaurant service who just made a dinner
reservation might hear, "Would you like to see a movie near that
restaurant?" The caller could then say, "Skip it" or
"Tell me more." If the caller responds "Tell me more,"
an extended speech ad plays offering a new set of interactive options to
the caller.
In a recent focus group led by a company deploying a speech service to
complement its online offerings, all participants preferred hearing an ad
to paying a service fee for using the system. Not only do these ads help
make the service free for consumers, they also help companies reach and
capture a new, larger audience -- anyone with access to a telephone.
So, next time you find yourself untethered, pick up any phone, use your
voice, and get your answers -- it's as simple as that!
Leah Lesser is marketing manager for SpeechWorks International, Inc.
For more information, visit the company's Web site at speechworks.com.
[ return to the top ]
[ return
to the July 2000 table of contents ]
|
|