#Canada #Waterloo #Applied Brain Research #ABR SDK #Niagara ASR

Applied Brain Research Launches Cutting-Edge ABR SDK

Applied Brain Research (ABR) has recently unveiled its highly anticipated ABR SDK, a toolset designed for developers aiming to create real-time voice interfaces that operate entirely on edge devices. This marks a significant advancement in the field of voice technology, especially with growing demand for applications that function without reliance on cloud services.

The ABR SDK: A Game-Changer for Voice Interaction

The ABR SDK integrates both automatic speech recognition (ASR) and text-to-speech (TTS) functionalities through a single application programming interface (API). This streamlined approach allows developers to implement responsive voice interactions on embedded hardware, ensuring that speech processing happens locally without the need for network connectivity. Kevin Conley, CEO of ABR, emphasized that immediate response times are crucial for usability in voice interfaces, and the new SDK is built with this priority in mind.

With the SDK, speech input and output are processed with minimal latency, allowing users to engage with devices seamlessly, regardless of their internet connection status. This is a vital feature for applications in areas such as healthcare, automotive, and consumer electronics, where reliability and performance are non-negotiable.

Advanced Features for Multi-Language Support

The newly released SDK supports a variety of languages, including English, Spanish, Mandarin, Japanese, and Korean. Each language model is packaged as a self-contained unit, simplifying the integration process. Developers can easily switch languages or upgrade the model’s capabilities without tedious backend selections or complex configurations.

To cater to specific industry needs, the SDK includes voice cloning capabilities for TTS, allowing the generation of custom voice profiles based on a short audio sample, along with an extensive custom vocabulary feature for ASR and TTS. This ensures that brands can include domain-specific jargon and product names accurately in their applications.

Optimal Performance Across Platforms

ABR's smart models deliver notably quick performance, processing speech recognition in as little as 115 milliseconds and generating speech in just 147 milliseconds, even on basic embedded CPUs. This efficiency sets a new industry standard for real-time voice applications.

Currently, the SDK supports major platforms including Linux x86-64, Linux ARM64, and Android ARM64. Options for acceleration on specific partner platforms are also on the table, while support for additional hardware platforms, particularly those with Real-Time Operating Systems (RTOS), is expected to be released by year’s end.

Easy Access and Evaluation

For developers interested in exploring the functionality of the ABR SDK, comprehensive documentation and access to evaluation tools are available on the ABR developer portal. Through self-managed registration, users can evaluate the full suite of tools and models without upfront costs. Upon completion of testing, those looking to deploy the SDK commercially can do so through ABR’s licensing options.

ABR’s contributions not only position them as leaders in edge AI technology but also pave the way for exciting innovations in voice user interfaces. With a commitment to efficiency and user-friendly experiences, Applied Brain Research is empowering developers to harness the potential of voice technology in diverse applications.

About Applied Brain Research

Founded in Waterloo, Ontario, ABR specializes in creating efficient real-time speech and language AI solutions designed for edge devices. Through the development of their Niagara ASR and Nith TTS model families, ABR boasts a performance profile that allows applications to run at remarkably low latencies, positioning them as pioneers in cutting-edge voice technology.