|Traded as||NASDAQ: NUAN
Russell 1000 Component
|Headquarters||Burlington, Massachusetts, United States|
|Chairman and CEO: Mark Benjamin
Executive Vice President: Bruce Bowden
|Products||Productivity applications, OCR, speech synthesis, speech recognition, PDF, consulting, government contracts|
|Revenue||$1.949 Billion (2016)|
|$0.139 Billion (2016)|
|-$0.013 Billion (2016)|
|Total assets||$5.7 Billion (2016)|
|Total equity||$1.9 Billion (2016)|
Number of employees
Nuance is a U.S. based multinational computer software technology corporation, headquartered in Burlington, Massachusetts, United States on the outskirts of Boston, that provides speech recognition, and artificial intelligence. Current business products focus on server and embedded speech recognition, telephone call steering systems, automated telephone directory services, and medical transcriptionsoftware and systems. The company also maintains a small division which does software and system development for military and government agencies based in Westborough, Massachusetts, allegedly called Twined (Previously marked by a large wooden sign that has been removed).
Nuance merged with its competitor in the commercial large-scale speech application business, ScanSoft, in October 2005. ScanSoft was a Xerox spin-off that was bought in 1999 by Visioneer, a hardware and software scanner company, which adopted ScanSoft as the new merged company name. The original ScanSoft had its roots in Kurzweil Computer Products.
- 2List of countries with English speaking “Text-To-Speech” (TTS) voices
- 3List of other countries with TTS voices
- 5External links
In September 2005, ScanSoft Inc. acquired and merged with Nuance Communications, and the resulting company adopted the Nuance name. For a decade prior to that, the two companies competed in the commercial large-scale speech application business.
In 1974, Raymond Kurzweil founded the Kurzweil Computer Products, Inc. to develop the first omni-font optical character-recognition system—a computer program capable of recognizing text written in any normal font. In 1980, Kurzweil sold his company to Xerox. The company became known as Xerox Imaging Systems (XIS), and later ScanSoft.
In March 1992, a new company called Visioneer, Inc. was founded to develop scanner hardware and software products, such as a sheetfed scanner called PaperMax and the document management software PaperPort. Visioneer eventually sold its hardware division to Primax Electronics, Ltd. in January 1999. Two months later, in March, Visioneer acquired ScanSoft from Xerox to form a new public company with ScanSoft as the new company-wide name.
Prior to 2001, ScanSoft focused primarily on desktop imaging software such as TextBridge, PaperPort and OmniPage. Beginning with the December 2001 acquisition of Lernout & Hauspie, the company moved into the speech recognition business and began to compete with Nuance.
Nuance history prior to the 2005 merger with ScanSoft
Nuance was founded in 1994 as a spin-off of SRI International‘s Speech Technology and Research (STAR) Laboratory to commercialize the speaker-independent speech recognition technology developed for the US government at SRI. Based in Menlo Park, California, Nuance deployed their first commercial large-scale speech application in 1996. Their initial route to market was through call center automation. Call centers had just centralized the branch-office telephone handling function throughout many large companies. The highest cost of running call centers is the cost of staff. Early projects were completely developed by Nuance to prove the commercial practicality and benefits.
In 1994, Nuance was spun off from SRI’s STAR Lab. Two years later, Nuance deployed its first commercial speech application. On 13 April 2000, Nuance filed an initial public offering (IPO) on the NASDAQ under the symbol NUAN. On 15 November 2000, Nuance acquired SpeechFront voice instant messaging company for $10.5 million in cash and stock.
In simple terms, the technology produced allowed a computer to determine what a speaker was saying within a specific and limited vocabulary of phrases. Its key advantage over technologies such as ViaVoice was that the system did not need training for the specific speaker. This permitted the use of the system, so-called speaker-independent natural-language speech recognition (abbreviated as SI-NLSR or just NLSR), for call automation.
The limited vocabulary was typically a few thousand different variations of phrases. In complex systems this could be in the low millions. At the time, these systems were pushing the limits of computer processing power in commodity Intel x86 servers until the early 2000s.
During the late 1990s and into the 2000s, Nuance competed against other NLSR vendors including Philips SpeechPearl, SpeechWorks and other smaller players which were typically geographically focused such as Vocalis in the UK which used proprietary PCI cards with DSPs on board to improve the efficiency and density of the system.
Each speech-recognition engine provider had to determine how to convert written text into sounds. Determining how written text is spoken is a hugely challenging task in itself. Languages are “modeled”, samples of real spoken language is recorded and analyzed to create a language model. The higher the quality the language model the better the experience of the user, especially in complex interactions. Different language models were required for different dialects such as Flemish being a variant of Dutch, or Swiss Germanbeing a dialect of High German. Different models were also created for different qualities of telephone connection. Europe’s Philips had by far the largest language coverage, which included Flemish and Welsh.
Later, Nuance sold licenses (training and consulting) to their technology to third parties, including independent software vendors and interactive voice response (IVR) vendors who would build applications on top of an IVR platform. SpeechWorks on the other hand would typically deliver the application with the technology or with a group of key delivery partners. The technology was integrated into most of the leading IVR products from Avaya, Nortel Periphonics, Envox, Syntellect and many others. The requirements of telephony reliability meant many of these solutions ran on various versions of UNIX.
Nuance 7 was launched in the late 1990s and was an efficient network-distributed NLSR speech recognition solution; it ran on Unix and Windows. Nuance 8 added Statistical Language Modeling, an adaption of technologies used in technologies, such as ViaVoice to improve the range of phrases that the system could recognize at the expense of greater implementation cost and complexity. Nuance 8.x series also introduced the W3C vocabulary definition language GrXML in addition to and eventual replacement of Nuance’s proprietary and very concise Grammar Specification Language, GSL.
Nuance 8.5 was the last point release before the takeover by ScanSoft. These systems were significantly different from the technology used in consumer speech recognition products such as ViaVoice, which is now also a Nuance product.
While Nuance marketed their technology at call-center exhibitions, they rarely delivered solutions directly. Instead, Nuance relied on telecommunications partners, such as Nortel Periphonics, Avaya and Syntellect. Nuance realized that designing and developing speech-based solutions requires a different skill set and to that of traditional DTMFsolutions.
For a couple of years prior to the takeover by ScanSoft, Nuance started selling products directly, including their call-routing product (“Call Steering”), which determined the “skill group” required for the call based on responses to reasonably open questions asked of the caller.
Nuance 9.0 is the first release (excluding service packs) of the recognizer product since the acquisition and is an amalgam of the technologies acquired from various companies including Philips Speech Pearl, Speechworks, Nuance Recognizer and others.
Partnership with the Siri and Apple Inc.
Siri is an application that combines speech recognition with advanced natural-language processing. The artificial intelligence, which required both advances in the underlying algorithms and leaps in processing power both on mobile devices and the servers that share the workload, allows software to understand words and their intentions.
Telephony application process
- User invokes the telephony application for call automation
- Application loads the phrases for the application and prompts the user to ask a question, opening a stream from the telephony input to the speech recognition software.
- User speaks and this is streamed to the recognizer.
- Recognizer returns a number of potential results with probability for each one that it is correct.
- Determines the start of speech input
- uses audio techniques to remove background noise
- slices the audio into small sections (10 to 100 ms in length)
- determines the sound in each slice
- matches the combination of sounds for the spoken phrase with the possible sound combinations provided by the possible phrases
A typical Nuance recognizer configuration required four or five applications to be started, often monitored by a sixth application.
- NLM: Nuance License Manager: kept a watch on the number of concurrent speech calls in use.
- Nrecclient: recognition client: it is the interface between the IVR speech path and the speech recognizing software, the recserver. The recclient can be developed into the IVR software.
- Nresource-manager: distributes the load over the recservers as required to balance load and to provide fault-tolerance.
- Nrecserver: where the speech is compared and processed against known vocabulary.
- Ngrammar-compiler: an application that dynamically adds words or phrases to an expected vocabulary for recognition.
- Nwatchdog: a Windows service or Unix daemon that monitors and maintains the above processes, restarting them if required.
Except for the watchdog which should be running on all the nuance speech servers, the other processes may be spread over a farm of servers, connected by an IP network with low latency and high-bandwidth, usually a dedicated LAN segment. The resource manager directs which resources it thinks are least utilized.
Prior to the 2005 merger, ScanSoft acquired other companies to expand its business. Unlike ScanSoft, Nuance did not actively acquire companies prior to their merger. After the merge, the company continued to grow through acquisition.
ScanSoft acquisitions prior to the merger
- March 2000 – Caere Corp., of Los Gatos, California – $145 million. Caere had developed OmniPage (scanner and OCR software.)
- December 2001 – Lernout & Hauspie, of Ieper, Belgium, Speech and Language division – $39.5 million
- This acquisition occurred following Lernout & Hauspie’s bankruptcy proceedings. Previously, Lernout & Hauspie had acquired these speech technology companies: BBS, Berkeley Speech Technologies (1996), Centigram Communications Corporation, Dragon Systems (2000), FDC, and Kurzweil Applied Intelligence (1998).
- January 30, 2003 — Royal Philips Electronics Speech Processing Telephony and Voice Control, Dialogue Systems — $35.4 million
- Philips had previously acquired Voice Control Systems, which had in turn had acquired Pure Speech, Scott Instruments and VPC.
- August 11, 2003 — SpeechWorks, Inc., of Boston, Massachusetts, — $132 million
- January 2004 — LocusDialog, of Montreal, Quebec
- May 2004 — Telelogue, Inc., of Iselin, New Jersey — $5.4 million
- November 2004 — ART Advanced Recognition Technologies, Ltd., of Tel Aviv, Israel – $21.5 million
- November 2004 — Rhetorical Systems Ltd., of Scotland — $6.7 million
- February 1, 2005 — Phonetic Systems, Ltd., of Burlington, Massachusetts and Israel — $35 million
- May 2005 — MedRemote Inc., of Westmont, Illinois — $6.2 million
List of countries with English speaking “Text-To-Speech” (TTS) voices
List of other countries with TTS voices
ScanSoft merges with Nuance; changes company-wide name to “Nuance Communications, Inc.”
From Wikipedia, the free encyclopedia (all content Belongs to Wikipedia)