Voice Interfaces
Conversational technologies are in demand. Speech interfaces are a natural way to control and navigate devices with small
screens and keyboards for improved safety and ease of use. Using the Fonix VoiceIn SDK, developers can quickly and
easily integrate automatic speech recognition (ASR) into applications or onto devices. Most applications are likely to need a
voice interface soon. Fonix has it now!
Fonix VoiceIn 4.1
VoiceIn ASR is based on Fonix proprietary neural network technology. It provides accurate speaker-independent ASR in
noisy environments. This new release represents a significant leap forward in ASR technology. Fonix's newly developed
neural network phoneme localization technology employs significantly more phonetic, linguistic and prosodic information
about the speech signal, which improves recognition rates in noisy environments.
VoiceIn 4.1 Improvements
In addition to the operating systems previously supported, Fonix VoiceIn 4.1 now supports the new Microsoft Windows
Mobile 5.0. This includes support for the .NET Compact Framework and new programming interfaces C#.NET and VB.NET.
These new interfaces enable developers to quickly develop and deploy speech applications.
Previous versions of Fonix VoiceIn have performed well in noisy environments using a ‘push-to-talk’ paradigm. The new Fonix neural network phoneme localization technology utilizes additional prosodic and linguistic information provided by the neural net. It also processes prosodic and linguistic information in the speech to better frame speech onset in high-noise conditions to more accurately estimate phonetic probabilities. These advances provide higher recognition accuracies in very noisy environments.
New Features and Benefits
- Phonetics pronunciation analysis: Helps users learn correct language pronunciations
- Java support on Windows CE, Analog devices Blackfin: Additional platforms – more developer flexibility
- Improved phoneme estimates: Better recognition accuracy
- Improved speech detection and speech framing: Better recognition accuracy
- Improved rejection of non-speech sounds: Better rejection of background noises
An efficient, scalable, low-cost solution
With Fonix VoiceIn, the software components and architecture operate directly on the main processor, eliminating the need
for an additional CPU or DSP. The flexible, self-contained modules allow developers to design systems that operate within
application memory and MIPS constraints, and are highly portable across many processor and operating system platforms.
With multiple OS and hardware platform support, plus over 10 Languages, Fonix VoiceIn offers developers the best solution to meet their needs as well as the needs of end users. Fonix VoiceIn is a state-of-the-art ASR SDK that is flexible and simple to use, and provides the highest accuracy for applications to deploy on devices. This is truly a "write-once, deploy many times” solution, saving developers and manufacturers time and money.