V‑Blaze version 7.4 (October 2022)

V‑Blaze version 7.4.0-0, voci-webapi-3.1.0-1

New Features and Improvements

  1. Added automatic audio resampling using the resample tag. For more information on this parameter, refer to Adjusting for audio.

  2. Updated the warning field in the top level of JSON transcripts to include the highest priority warning if there are multiple warnings. Voci recommends logging the warning field for all ASR flows.

  3. Fixed a text processing issue where unrelated strings were pulled into earlier utterances when number merging was performed.

  4. Fixed an issue with the noise setting for the diarize parameter.

  5. Made improvements to model processing that decrease startup load time and improve ASR performance.

  6. Improved emotion=xa identification for faster speed and better accuracy.

  7. Added support for CUDA 11.6.

  8. Added capabilities for tracking audio hours processed by the engine. They are broken down by model, features, and billing tags. Additional tracked tags are configurable using the tracktags and tracksubs settings in Optional license configuration parameters.

  9. Added counter nsubs to /stats and /metrics in order to track the number of substitutions applied overall by the engine.

  10. (INTERNAL ONLY) Added a new feature that enables passing JSON transcription files to run post-transcription services, such as redaction and emotion, without having to re-transcribe the same audio. Refer to Other and advanced for more information.

  11. Several REST API changes including new usage endpoints, updated metrics, added AudioCodes support. Refer to V‑Blaze REST API version 3.1.0-1 (October 2022) for more details.