Skip to content

Changelog

[3.2.1]

  • Minor license improvements.

[3.2.0]

  • New endpoint GET /api/v1/config to retrieve current service configuration.
  • New endpoint POST /api/v1/config to update service configuration without restarting the service.

[3.0.1]

  • Fix license issue to ensure compliance and smooth operation of the Voice Service.

[3.0.0]

  • Introduced the new verification algorithm, offering improved performance compared to previous versions. The old algorithm is still available for compatibility reasons.
  • Improved accuracy of the voice clones liveness detection algorithm.
  • Improved accuracy of the presentation (replay) attack liveness detection algorithm.
  • Fixed SNR Computer reporting TOO_NOISY error on files with lengthy silence.
  • Updated the OpenVINO backend from 2023.0 to 2024.4.

[2.9.0]

  • Update docker image version to ubuntu 24.04.
  • Number of threads in service configuration is 1 by default.
  • Added new verification algorithm delivering higher accuracy.
  • Improved accuracy of verification and liveness algorithms in noisy scenarios. The verification score might be slightly different from the previous release, but the voice templates remain compatible.
  • Recommended threshold for relative speech length quality check in verification and liveness use-cases is now 0 (any relative speech length is allowed now).
  • Updated algorithm for SNR estimation. The estimated values might differ from the previous product release, but overall algorithm accuracy remains the same.

[2.8.0]

  • New property check_quality in the enrollment and authentication endpoints to enable or disable the quality check.
  • New property check_liveness in the authentication endpoint to enable or disable the liveness check.

[2.7.2]

  • Improvements in the service health check.
  • Improvements in license management.

[2.7.1]

  • Concurrency improvements in base64 decoding
  • Some minor improvements in the code

[2.7.0]

  • Some refactors and performance improvements.
  • Concurrency improvements.
  • Improvements in the service health check.
  • New checks about license availability.

[2.6.0]

  • New Quality Check parameters:
    • minimum_speech_relative_length: Minimum required length of the speech in relation to the total audio length.
    • maximum_multiple_speakers_detector_score: Maximum numbers of speakers at the same time.
  • New response fields with the Quality Check detected values:
    • snr_db_detected: Signal-to-Noise Ratio detected.
    • speech_length_ms_detected: Speech length detected.
    • speech_relative_length_detected: Speech relative length detected.
    • multiple_speakers_score_detected: Multiple speakers score detected.

[2.5.0]

  • Support for new audio formats; AAC, PCM ulaw and mulaw, WMA, MP4, FLAC, ALAC and AIFF.
  • Support for audios with a sample rate below 16kHz.
  • Integrates tracking events for authentication.

[2.4.0]

  • Update voice-sdk to v2.4.0.
  • Multithreading in enrollment process.
  • Expose low level exceptions to the user.

[2.3.1]

  • Update voice-sdk to v2.3.1.
  • Improve exception and error messages.

[2.3.0]

  • Update voice-sdk to v2.3.0.
  • Update facephi-rest-manager to v2.4.2.
  • Add support for AAC audio format.
  • New configuration parameter client_max_body_size to limit the maximum size of the request body.

[2.2.3]

  • Update voice-sdk to v2.2.3.
  • Added operation_result field to Authentication and Enrollment result.
  • Added liveness_score field to Authentication and Enrollment result.
  • Added matching_score field to Enrollment result.
  • Added template_type field to Enrollment result.
  • Renamed probability field to Authentication result to matching_score.
  • Fixed template field in Enrollment result. The template field will be empty if the response is not valid.
  • Fixed template when multiple audios were passed to Enrollment but the minimum number of audios needed to generate a text-independent pattern was not met.

[2.2.2]

  • Update voice-sdk to v2.2.2.
  • Update facephi-rest-manager to v2.4.0.

[2.2.0]

  • Add support for voice-sdk v2.2.0.
  • Enpoints accept encrypted and not encrypted audios.
  • Quality check can be modulated in the request, with the new parameters minimum_snr_db and minimum_speech_duration_ms

[2.1.0]

  • Upgrade facephi-rest-manager to 2.1.3 version.

[2.0.0]

  • Add support for voice-sdk v2.0.0, with encrypted audios and templates.
  • Refactor of REST API implementation.
  • New message format for Bad Request errors and Internal Server errors.

[1.3.1]

  • Include http status code in json error response
  • Make all endpoints lowercase
  • Improve exception handling and exception messages
  • Add exceptions for voice-sdk errors
  • New endpoint api/v1/version
  • New endpoint api/v1/health

[1.3.0]

  • New licensing system.