- Rapid Deployment: Nexidia's Suite of software development environments publishes to .NET, C++ and the Windows, Linux and OS/X operating systems, allowing for rapid deployment and easy integration, without requiring a detailed understanding of the underlying core processing engine.
- Future-Proof: By integrating through the published interfaces mentioned above, organizations can take advantage of improvements in Nexidia's system performance to evolve their products, ensuring backwards compatibility for all published APIs, with relatively little development overhead.
- Scalable: Nexidia Workbench is able to process very large volumes of audio content without the huge investment in hardware required by comparable solutions. Nexidia can index recorded content up to 207 times faster than real-time per CPU core. At these speeds, customers are able to fully index well over 20,000 hours of content per day on a single server, which provides significantly lower costs for organizations with large volumes of audio or other media content to review and analyze.
- Open-Standards Interfaces: Nexidia's suite of software development environments is designed to allow rapid integration of Nexidia's patented speech technologies into any Windows, Linux or OS/X environment.
- Nexidia Workbench provides a collection of published interfaces exposed in .NET, C and C++ which are well suited to rapidly incorporate voice analytics. Workbench can be used to develop customized solutions or to extend and enhance existing systems solutions.
- Powerful Real-Time Analysis: Nexidia Workbench simultaneously monitors and searches up to 1000 audio streams in real-time on a single server. As a result, high volume applications can be developed to capture and analyze audio directly from the audio source, e.g. telephony switch. A low-latency response of only 800 milliseconds provides rapid response to scanning and media sources.
- Automatic Audio Classification: A broad set of capabilities can identify specific characteristics that further classify and segment media, including the languages being spoken and other non-speech activity present:
- • Language ID automatically determines which languages are being spoken in a recording and identifies the segments of each language.
• Voice Activity identification classifiers detect the presence of "non-speech" such as music, silence, or other non-spoken activity, and differentiate these segments of audio from those with spoken language.
• Non-speech classification can help identify reasons for excessive call lengths, plus add valuable metadata to rich media content.
• Gender identification enables automatic detection of speaker gender by segmenting audio content into different categories based on the gender of the primary speakers.
• DTMF identification provides a means to classify segments of audio containing dual-tone multi-frequency and differentiate those segments from both speech and non-speech activity.
- Transcript Sync: Transcript Sync automatically locates and accurately time aligns transcripts with words spoken in audio and video recordings. Locating incomplete or poorly matching text within the audio, the resulting time-aligned text enables customers to leverage closed-captioning, pre-edit scripts and producer notes, providing jump-to navigation and text based editing within more content.
