V-Blaze and V-Cloud Online Help

V‑Blaze REST API version 3.0.0 (May 2021)

V‑Blaze REST API version 3.0.0 is a major release with numerous new features and improvements.

  1. Added new V‑Blaze REST API Flows

    • Tag specification enabled via header or url parameters. Refer to Submitting Transcription Requests for more information.

    • Simple, single part POSTs. Refer to Submitting Transcription Requests for more information.

    • WebSockets to enable bidirectional streaming of data via upgraded connections. Refer to WebSocket API Interface for more information.

    • Simple results streaming back through the primary HTTP (POST) connection. Either JSON utterance data or scrubbed audio can be streamed back, but not both in the same connection. Refer to HTTP Results Streaming for more information.

  2. The vociwebapi service script now attempts to allow a clean exit. A stop or restart operation will block new /transcribe POSTS, but will allow existing /transcribe requests and any new GET requests to complete before exiting. Refer to V‑Blaze Service Commands for more information.

  3. Added new stats endpoint, and made updates to the status, statusfull, and sysinfo endpoints.

    • The /status and statusfull endpoints have changed to better reflect the current state of the ASR system. They are now lightweight calls that show summary information about current state of ASR processing.

      The following fields were added to /status and statusfull: efficiency, stream.audiosecs, stream.count, stream.errors, stream.speed, stream.warnings, utt.audiosecs, utt.count, utt.speed, utt.tat, utt.tat.max, utilization.idle, and shutdownin. The shutdownin field only shows up if a query is made during the shutdown waiting period, and indicates the number of seconds remaining before shutdown.

      The following fields were removed from /status and statusfull: lastactive, totalerrors, totalidle, totalstreams

    • The sysinfo endpoint shows the same product and server information with additional information included such as system sizing, licensing, and capacity.

      The following fields were added: aminfo, expires, gpudriver, gpumem, licenses, maxdecoders, maxstreams

    • The new /stats endpoint includes detailed performance-related metrics. Time-window averages allow users to determine how the system performed within the last 1, 5 and 15 minute windows. This allows clients to analyze the consistency of their machine by comparing measurements across different windows. The comprehensiveness of the /stats response provides users with a view of the system internals. Expert users may use this to fine tune their system and API implementation for optimal performance.

  4. The datahdr parameter now defaults to WAVE when not specified. Customers who require legacy behavior or a customized default can configure using the datahdr setting in webapi.cfg.

    For more information on datahdr, refer to Adjusting for Different Types of Input.

  5. Made multiple language ID (LID) improvements and additions.

    • Added the lideffort parameter, which specifies the required confidence level before lid will stop analyzing audio. This replaces the previous function of lidthreshold.

    • Changed lidthreshold to now adjust the confidence level required for the system to select the alternative language.

    • Optimized LID to automatically use different technologies based on language models, channels, per-stream vs per-utterance, and other characteristics.

    Refer to lid for more information on these changes.

  6. Minor bug fixes and general improvements.