New

• GPU & CPU

  • New languages: Irish (ga) and Maltese (mt) 

  • Audio Filtering: pre-process audio to remove low-volume background speech which might otherwise be detected and transcribed. Refer to documentation here to get started

  • Disfluency removal: automatically remove disfluencies from your transcript. Refer to documentation here to get started

Removed

• The legacy Speaker Change Detection feature is now obsolete. Any jobs using the speaker_change and channel_and_speaker_change parameters will be rejected

Improvements

• GPU

  • Ursa2 models released, giving a broad accuracy uplift across languages (Enhanced Operating Point only):

    • Uplift for all languages, including a major improvement for Arabic dialects

    • Updated vocabulary for English

  • Improvements to speaker diarization accuracy

  • Improved music detection accuracy in Audio Events

• GPU & CPU

  • Ursa2 models released, giving a broad accuracy uplift for below languages (Standard operating point only):

    • Uplift for Basque (eu), Estonian (et), Polish (pl), Swedish (sv), Tamil (ta), Turkish (tr), Uyghur (ug)

  • Improved transcription of music lyrics and a decrease in missing spoken words in transcripts

• Management API

  • Added health check endpoint /v2/management/health

Fixes

• GPU

  • Fix for occasional incorrect repetition of words in transcription output when Audio Events is enabled

• GPU & CPU

  • Written form for negative percentages in German transcription is now output as "%" instead of "Prozent"

  • Security fixes including an OS upgrade. A Software Bill of Materials (SBOM) is available for download from the Knowledge Base section of the Support Portal.

Known Issues

• Security

  • We are aware of vulnerabilities in linux-headers-5.15.0-122, linux-headers-5.15.0-122-generic, and linux-modules-5.15.0-122-generic. Despite upgrading to later versions, we cannot remove these due to dependencies. Analysis shows we do not load the cpu5wdt module, so we are unaffected by CVE-2024-38630. The two relevant CVEs are:

    • CVE-2024-27397

    • CVE-2024-38630

  • The vulnerability in the github.com/traefik/traefik/v2 Go module affects server configurations, but we only use client functionality, so we are not impacted. While we use Traefik separately as a proxy, it has been updated to the latest version. We cannot upgrade the Go module due to dependencies.

    • CVE-2024-45410

  • Vulnerabilities in Nvidia gpu-operator. We upgraded to the latest version, but some vulnerabilities remain unresolved by the vendor. These packages are inaccessible from outside the appliance.