Sensory launches new voice recognition platform SDK
Sensory Inc. announced it has launched TrulyHandsfree SDK, the new version of its embedded small-footprint voice user interface platform.
Designed especially with software developers in mind, TrulyHandsfree significantly simplifies the development of speech recognition applications. As such, it significantly reduces the amount of time required to implement a speech-controlled user experience.
TrulyHandsfree SDK features several new improvements, including support for fixed triggers, user enrolled triggers and commands (phrase spotting technology makes TrulyHandsfree highly robust to noise); speech trigger and commands can be added to an application with less than 20 lines of code; tasks’ details are managed with streamlined modular speech recognition functions; easier to digest documentation; all speech recognition models and task configurations are passed as one file to further simplify the API; and available JAVA API for Android ensures seamless integration with Android Studio (no JNI programming needed).
“Integrating always-on, always-listening speech triggers and speech command functionality into the user experience of software and apps used to require a deep understanding of how speech technologies work,” said Jacques de Villiers, engineer of technology deployment at Sensory. “With Sensory’s latest TrulyHandsfree SDK, we’ve made it simple for developers to integrate these must-have UX technologies into their products. We’ve worked closely with our partners to understand what they wanted and completely redesigned our SDK to incorporate dataflow and inversion-of-control techniques to greatly reduce the overhead and complexity of designing TrulyHandsfree technology into applications.”
Sensory’s TrulyHandsfree trigger technology has become the most widely adopted keyword voice wakeup technology in the speech industry, with over a billion products using TrulyHandsfree shipped over the past several years from the world’s largest CE manufacturers.
The TrulyHandsfree trigger is a highly noise robust, low power, and accurate, embedded phrase recognition technology that listens for a specific “wake up” phrase, while ignoring other conversations.
It can be used in many different ways, supporting various wakeup voice triggers including fixed triggers, user-enrolled fixed triggers, user-defined triggers and speaker verification passphrases.
The trigger also supports high accuracy noise immune command sets and works seamlessly with the applications processors of devices and cloud-based speech recognition technologies, which enables seamless trigger-to-search or trigger-to-cloud speech recognition functionality.
The technology can also be used with Sensory’s TrulySecure speaker verification security technology to allow for trigger-to-speaker-verification applications.
In terms of private security measures, TrulyHandsfree operates embedded on device so no cloud connection is needed, eliminating the chance of any shared personal data being stolen from the cloud.
Available for Android, iOS, Linux, QNX and Windows, TrulyHandsfree SDK currently supports a myriad of languages including US English, UK English, Arabic, Dutch, French, German, Italian, Japanese, Korean, Mandarin, Portuguese, Russian, Spanish, Swedish and Turkish.
In addition, ultra-low-power deeply embedded ports of TrulyHandsfree are available for DSP/MCU IP cores from ARM, Cadence, CEVA, NXP CoolFlux, Synopsys and Verisilicon, as well as for integrated circuits from Audience, Avnera, Cirrus Logic, Conexant, DSPG, Fortemedia, Intel, InvenSense, NXP, Qualcomm, QuickLogic, Realtek, STMicroelectronics, TI and Yamaha.