Intelligent Voice is delighted to announce the release of major new version 5.0 which brings many new advances, including:

Improved Speech Recognition

New, CTC-based Automatic Speech Recognition (nASR). Models using this offer improved accuracy over existing IV models.

Updated Diarization (Biometric Speaker Separation)

The previous i-vector based Diarization system is replaced by an x-vector based system for improved speaker separation accuracy across all languages.

Improved Punctuation

Further advances in punctuation, now including commas and question marks in English, together with full stops (periods) across multiple languages.

NVIDIA Triton Inference Server support

New support for running AI workloads with the addition of NVIDIA Triton Inference Server (https://developer.nvidia.com/nvidia-triton-inference-server).

Faster waveforms in SmartTranscripts (preview feature)

Waveforms in SmartTranscripts are now pre-generated, rather than created on-screen with a dependency on JavaScript. Only supported for audio files, not video and only when installed on Ubuntu 20.04 - itself in preview / beta – and not RedHat Enterprise Linux.