Embedded and Offline: Voice Control and Speech Synthesis without Eavesdroppers

Embedded solutions have by far not been replaced by cloud-based voice recognition, speech synthesis and NLU, as shown, for instance, by the announcement of the movi-module for the Arduino Shield. With this small hardware module, ASR and TTS can run piggyback on the Shield, and at low cost. movi is not the first module of this kind. EasyVR Shield 3.0 has been around for a while, allowing the integration of a limited voice recognition modality, which is restricted to 28 hotwords and 28 commands, both defined by the user, respectively.  movi, however, is able to understand whole sentences and respond via speech synthesis and. As long as English is being used, movi does not depend on a specific voice.

The advantages of these integrated solutions for voice recognition, processing and synthesis are quite obvious: they do not require an internet connection and can run anytime and anywhere as long as there is electricity or batteries available.  They do not save information on servers and are unmatched in terms of data security and safety. Furthermore, they do not require expensive backend systems like servers and the respective infrastructure, and they are independent from large suppliers like Google or Apple – who knows, after all, whether Google Voice will remain free, whether any technological changes will render your solution obsolete or whether Siri and Google Voice will still be there in 2020?

Aside from the insecure future and only partial integration capability of a range of suppliers, e.g., Siri, embedded solutions can be freely customized. The only constraints are memory space and computing power, which eventually determine the performance of the embedded system.

The combination of a powerful software with the highest expertise in the development and integration is hence especially important for embedded systems. For more than seven years, the SemVox team has been working intensively with solutions and systems for internet-independent voice dialog and voice control systems. The robustness and performance of these solutions continue to impress and convince customers.

Our ODP S3-based systems achieve the optimum for any solution, regardless of the specific target platform. And even more than that: the ODP Workbench and the entire SDK enable our customers to work on integrated, hybrid and, of course, cloud-based systems. Do not hesitate to get in touch – we will find the perfect solution for you!