Intelligent Voice is pleased to announce the release of Intelligent Voice version 5.3.
This is a major new release containing new features, improvements and bug fixes
New features
VideoOCR
A major new component included in the default IV installation, VideoOCR extracts text from videos imported into Intelligent Voice. This text is made searchable and available through the Intelligent Voice API
Inverse Text Normalisation (ITN)
Inverse Text Normalisation is a major new transcript formatting technology which improves readability and the identification of semantically important punctuation in transcripts
New Import and Export options
Many new processing options available through the Create New Item and Import APIs, for example:
- Tagger on/off (default on)
- VideoOCR on/off (default off)
- Punctuator model version
For a full list see https://api-docs.intelligentvoice.com
Many new export options including options to select multiple export types and templates
Improvements
Improvements to the installers include:
- Script to install indexes for Tagger independently
- Option to install Azure autoscaling components such as telegraf for Red Hat installs
- PhonemeLoader made optional
- Option to select alternative Gearman server host during install
- New English ASR models included in the installer
- Single Sphinxsearch instance serving App and Proc
Improvements to JumpToWeb include:
- Support for new import and export options
- Export page uses the v2 API instead of legacy v1
- ASR model adaption page
- Optional Azure Storage Queue display
Improvements to the IV API include:
- New webservices to support VideoOCR (asr-models/ocr-language and asr-models/{modelId}/ocr-language)
- New webservice showing the status of model adaption jobs (/asr-models/building-jobs)
- Get Item Details API now returns processing options
The inclusion of LMBuilder 9.1 makes ASR model adaption available for latest NASR3.1 models
New V3 punctuator model is available which brings significant improvements to punctuation
waveform data is now available when installing on Red Hat
The configuration of container images has been improved make them easier to deploy in container orchestration environments
Bug fixes
INT-646 Fix for Import and Export are not being logged for the right User into the Audit Table
INT-3074 Enabling importItemEmailNotificationScheduler no longer sends email notifications for existing items
INT-3099 Fix for Tags constituting word(s) at the end of a sentence are not displayed in JumpTo Topics
INT-3655 Fix for duplicates in alternative words
INT-3927 Fix for Items without a "from" or "to" populated not added to the search index
INT-3739 Fix for items created from URL not being deleted when requested
INT-3828 JumpToWeb fix to display tags in Japanese, Russian, Korean, Arabic
INT-3928 Index rotation fixed for empty groups
INT-4235 Fix for multichannel files show same waveform on ST
INT-4886 Fix for issue with punctuting acronyms
INT-4824 Fix for issue searching with UTF-8 queries
INT-4987 Fix for issue importing items with payload
INT-4992 Fix for searches not finding items without a timestamp
INT-4556 Fix for installation failure when hostname is over 64 chars long
INT-4938 increase the maximum length of a word from 34 to 80 characters
INT-5014 iv-webapp installer now installs indexes for Tagger
INT-5085 Fix for Get Item Summary doesn't return allText when no transcript is produced
INT-5688 Fix for a few models producing unknown characters rendered as ??
Known Issues
We are aware that punctuation may not work as well with mono audio files when compared to multi-channel audio files. We have planned improvements coming in the forthcoming release of IV 5.4 in August 2022 with a view to a longer term fix later in the year.
INT-2059 Groups API and JumpToWeb groups page timeout with more than 1 million items.
INT-2396 Create new group in the front end does not work if there are no existing groups
INT-2430 Connections to gearman server are not encrypted (workaround available)
INT-2228 JumpToWeb Automatic Topics and People Map order changes on searching
INT-3677 "Data too long for column 'email' " message on creation of email items
INT-3749 Limited options for password characters
INT-4970 CrackerWorker not available in 5.3 release
INT-5015 Import Report page times out with lots of data
INT-5088 "ERROR: Unknown database 'languages'" at end of installation
INT-5920 CUDA Linux Repository Key Rotation issue (Apply this solution)
INT-5676 Red Hat installer cannot use existing repositories
INT-5697 Connections to Triton are not encrypted
INT-5698 Connections to the sphinxsearch server are not encrypted (workaround available)
Component versions
IV app server 5.3
JumpToWeb 5.3
SmartTranscript 2.0
LMBuilder 9.1
VAD 3.2
ASR 6.0
Diarization 4.3
Tagger 3.0
VideoOCR 1.0