Products
HEAR ME - Speech Blog  |  Read more June 26, 2017 - How Hollywood gets biometrics wrong (and what it gets right)
Products

Quotables

  • The Ecobee4 is the new smart thermostat to beat (Nest who?)

    CNet
  • VoiceGenie users will be able to chat with Google Assistant, Siri, Cortana, or Alexa anywhere -- totally hands-free. You just speak the wake word.

    CNet
  • Ecobee4 Review: The First Smart Thermostat with Alexa

    Tom's Guide
  • VoiceGenie, a solution that allows most wireless headphones to utilize the power of Amazon’s Alexa on the go

    Android Headlines
  • VoiceGenie delivers low-powered speech recognition technology to wireless headphones and headsets.

    Wearable
  • VoiceGenie could change the way we interact with the devices already firmly embedded in our lives -- and it could make voice assistants just as integral.

    CNet
  • Amazon is relying on voice models that are key components of Sensory’s speech recognition suite.

    Business Intelligence
  • Sensory is enabling what it calls “wakeup words” — that is, terms and phrases that trigger the always-on listening that is a hallmark of assistants like Google Now and Siri — on devices that integrate Amazon’s Alexa.

    Digital Trends
  • Sensory, which focuses on voice recognition, teams up with Alexa's creators to make it easier for third-party developers to create voice-activated devices.

    CNET
  • Sensory won Best Innovation in Deep Learning at the Alconics awards; First Place in Mobile Security and Privacy at the CTIA Emerging Technology (E-Tech) Awards; and Star Performer at the Speech Industry Awards, where CEO Todd Mozer also won a Luminary Award.

    FindBiometrics
  • Samsung Partners With Sensory For Face/Voice Biometrics

    AndroidHeadlines.com
  • Say out loud 'GoPro, start recording' and the camera starts recording, no fuzzing with buttons.

    Geoffrey Fowler, Wall Street Journal
  • …it works well even at a distance. You just shout "okay Garmin," the lights on the camera will blink, and then you can tell it to do things like start or stop recording, or take a photo.

    The Verge
  • Huawei also created this cool feature to help you find your phone more quickly. It’s called voice wake up, and you can ask your phone “Where are you?” or some other phrase, and your phone will respond, saying, “I’m here,” and play music until you find it.

    Malarie Gokey, digitaltrends.com
  • The best upgrade to the controls, though, has to be voice control...The Hero5 Black understands seven languages.

    Brent Rose, Wired
  • Most notably though, this new shooter supports voice commands, letting you control it via phrases such as, ‘Okay Garmin, start recording.’

    Engadget
  • GoPro added voice controls for starting and stopping recordings, snapping photos and even tagging highlights in your clips by saying "That was sick!" and "Oh sh*t!" It works really well, and there are currently 10 commands and support for seven languages.

    Sean O'Kane, The Verge
  • So if the cloud’s not private, how can your TV respond to voice commands? Simple. Use speech-recognition services that are baked right into the TV – no cloud required.

    Ted Kritsonis, Digital Trends
  • Voice commands are where both Hero 5 cameras flex their smarts.

    Geoffrey Fowler, Wall Street Journal
  • One of the coolest features on the Garmin VIRB Ultra 30 is the ability to use your voice to control the camera.

    ZDNet
  • Of particular interest is the fact that TrulySecure is an on-device biometric identification system that does not rely on a connection to the cloud. Many users prefer this approach because they do not wish for their biometric data to be replicated and stored outside of their personal devices.

    Max Maxfield, EE Times
  • TrulySecure works by watching and listening as you repeat a passphrase a couple times. The system tracks the way your lips move and registers the unique attributes of your voice.

    Josh Ong, The Next Web
  • Given Qualcomm’s prominence as a mobile technology developer and the technological advancement on display in its latest offerings, the partnership reflects very well on the confidence the company has in Sensory’s technology.

    Alex Perala, Mobile ID World
  • With touchless control, Motorola and Google upped the ante.

    Eric Mack, CNET
  • Speech recognition company Sensory is expanding into the computer vision space with a new smartphone security client that uses both voice and face recognition to lock down your phone.

    Kevin Fitchard, Gigaom
  • MotoX is a fantastic phone with many great features. My favorite is Touchless Control… Ask it the weather, to call a friend or do a Google search, and it'll just do it, and you never have to touch the phone.

    Pete Pachal, Mashable
  • Sensory is continuing to exhibit leadership in handsfree control by allowing a secure multimodal biometric that doesn’t require touching devices to make them work.

    Dan Miller, Opus Research
  • Touchless mode...is the most useful feature [on Moto X].

    David Pogue, NY Times
  • The defining feature of the Moto X is it’s a virtual ear, always straining to hear its owner’s voice say three magic words that will rouse it to action: "Okay, Google Now."

    Steven Levy, WIRED
  • The phone [Moto X] has all the standard features expected of today’s top smartphone, with a twist: the ability to control the phone by talking to it, without lifting a finger.

    The New York Times
  • The voice-response system, called BlueGenie is surprisingly accurate for such a small device. It's better than the voice system in my Blackberry phone.

    US News and World Report
  • [BlueGenie is] an intuitive voice control system...the finest voice recognition user interface we've seen.

    Good Gear Guide
  • Sensory is trying to revolutionize voice and speech recognition by creating TrulyHandsfree, which looks to evolve our interactions with our smart devices.

    Talk Android
  • It may not seem like much, but that little detail of getting the phone to wake up via a voice command - which Sensory calls ‘TrulyHandsfree’ - is one of the trickiest.

    Mashable
  • With its dual biometric factors, AppLock comes closer to the security-and-convenience ideal than I've ever seen.

    Mike Feibus, usatoday

TrulyNatural

TN-FSE-72dpiSTM15AWARD_starperformTrulyNatural™ is the first embedded large vocabulary continuous speech recognizer system which can provide the same state-of-the-art performance as cloud based technologies, but on the device. Many applications don’t need, or don’t want to rely on, a cloud based connection to do high performance speech recognition. Connections can frequently fail or cause fatal delays, and recent concerns have been expressed about sending personal data to the clouds where it could be stolen or used for purposes undesirable to the owner.

Sometimes a client/cloud distributed system with optimal performance is desired or a standalone embedded solution that can perform complex or natural language tasks without being too slow or lacking accuracy. TrulyNatural is the solution for these needs.

Sensory uses a unique form of a neural network with deep learning to achieve acoustic models an order of magnitude smaller than the present state-of-the-art. These neural networks employ the most recent breakthroughs in speech feature extraction to produce superior accuracy in real world noise. Combining this with a small Finite State Transducer (FST) enables storage and accurate processing of multiple large search domains as required in today’s market.

FluentSoft Data Sheet pg1 copy

TrulyNatural is:

  1. Scalable:
    • Includes a micro FST and tiny models for an embedded recognizer of up to 1000 words in under 1MB of memory
    • Includes sophisticated FSTs that can be deployed with document retrieval and robust parsing to enable natural language interfaces for limited domains in under 100MB of memory
  2. Most accurate, flexible, and fastest search retrieval on a client:
    • Benchmarked at <8% word error rate (WER) on a music retrieval task for 1,000,000 entries
    • “Bag of Words” approach for data retrieval that works in background noise and with broken speech or extra words and sounds
  3. Compatible with existing and readily available tools and standards:
    • Compatible with open source language model tools such as OpenFST and Kaldi
    • Compatible with standards such as the ATT FSM
    • Data access in the client if available, or in the cloud if a connection exists
  4. Available for a variety of natural languages:
    • Available for US English, UK English, Mandarin Chinese, Korean, French, Italian, German, Portuguese, Spanish and Russian
    • New language models constantly under development
    • Applications using TrulyNatural include Voice Dial

Download a full copy of our TrulyNatural Data Sheet.

Video Library

TrulyNatural Coffee Demo
TrulyNatural Coffee Demo
TrulyNatural Apps and Contacts
TrulyNatural Apps and Contacts