Updated V‑Cloud to version 7.4.2-1 of the ASR engine.

Bug fix

1. Updated V‑Cloud to version 7.4.2-1 of the ASR engine to address an issue with transcribing diarized u-law, A-law and non-16-bit PCM audio. For full details, review the V‑Blaze 7.4.2-1 release notes here.

V‑Cloud 1.9.4-2023.01.09 updates to version 7.4.1-1 of the ASR engine and includes the following changes:

1. LID has been updated to version 2.1.0 and includes the following:

   1. Performance and accuracy updates for English-Spanish identification.

   2. Added language identification for English-French audio.

   3. Define the probability of alternative language speech using the lidprior parameter. Learn more: Additional LID Options

2. Fixes to text processing modules:

   1. Fixed an issue where setting punctuate to false resulted in mixed letter cases due to certain substitutions. Disabling punctuate now results in lowercase transcripts unless substitutions contain uppercase letters.

   2. Fixed an issue where substitution rules containing a combination of non-ASCII characters and a backreference in the same word would cause utterances to drop from results.

V‑Cloud 1.9.3-2022.09.14 updates to version 7.4 of the ASR engine and includes the following changes:

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. Improvements to the noise setting for the diarize parameter.

V‑Cloud 1.9.0-2022.06.02 includes the following changes:

1. Updates to the following language models for significantly increased accuracy and performance:

   1. English (North America) language models:

      1. eng-us:autodealership

      2. eng-us:financial

      3. eng-us:healthcare

      4. eng-us:insurance

      5. eng-us:largevocab

   2. French (Canada) language models:

      1. fre-ca:callcenter

      2. fre-ca:largevocab

Refer to Language model updates for more information on language model releases and updates.

V‑Cloud 1.8.1-2022.04.30 includes the following changes:

1. Updates to the eng-us:callcenter model. The update includes minor revisions to the language pack, additional training data, and a small configuration adjustment.

V‑Cloud 1.8.0-2022.03.30 includes the following changes:

1. Requests for deprecated eng1 models are now automatically remapped to use the improved eng-us models where available. Requests specifying any of the following eng1 models will default to the matching eng-us language model instead:

   1. English (North America) Automotive Industry — eng-us:autodealership has deprecated eng1:autodealership.

   2. English (North America) Call Center — eng-us:callcenter has deprecated eng1:callcenter.

   3. English (North America) Financial — eng-us:financial has deprecated eng1:financial.

   4. English (North America) Healthcare — eng-us:healthcare has deprecated eng1:healthcare.

   5. English (North America) Insurance — eng-us:insurance has deprecated eng1:insurance.

   6. English (North America) Large Vocabulary — eng-us:largevocab has deprecated eng1:largevocab.

2. V‑Cloud now includes the uttmaxgap parameter. Refer to Utterance Controls for more information.

V‑Cloud 1.7.2-2022.03.02 includes the following changes:

1. V‑Cloud has been updated to version 7.3.2-2 of the ASR engine.

2. V‑Cloud now includes the following language models:

3. English (North America) Language Models:

   1. eng-us:autodealership

   2. eng-us:financial

   3. eng-us:healthcare

   4. eng-us:insurance

   5. eng-us:largevocab

4. V‑Cloud now includes the filetype parameter, which can be used to manually specify the Content-Type of your audio. Refer to Common tags for more information.

1. V‑Cloud has been updated to version 7.3.2-1 of the ASR engine.

2. V‑Cloud includes updates to the following language models:

   1. English (North America) language models:

      1. eng-us:callcenter — Improved models for English (North America) Call Center domain including updates to the acoustic and language models for significantly increased accuracy and performance. The eng-us:callcenter model has deprecated eng1:callcenter, and is now the default when not otherwise specified.

      2. eng1:callcenter — Updated language model for increased accuracy and a small decrease in performance. This update was performed to provide some improvement for those who are unable to use the eng-us:callcenter model. The model has deprecated the eng1:callcenter model and is recommended for all uses.

3. V‑Cloud now includes the following language models:

   1. Portuguese (Brazil) language models:

      1. por-br:callcenter

      2. por-br:largevocab

V‑Cloud 1.6-2021.11.19 includes the following updates and fixes:

1. V‑Cloud has been updated to version 7.3 of the ASR engine. Refer to V-Blaze Version 7.3.0-1 Release Notes for more information.

2. V‑Cloud now includes the following language models:

   1. French (Canada) language models:

      1. fre-ca:callcenter

      2. fre-ca:largevocab

   2. Spanish (Argentina) language models:

      1. spa-ar:callcenter

      2. spa-ar:largevocab

   3. Spanish (Colombia) language models:

      1. spa-co:callcenter

      2. spa-co:largevocab

   4. Spanish (Mexico) language models:

      1. spa-mx:callcenter

      2. spa-mx:largevocab

   5. Spanish (Panama) language models:

      1. spa-pa:callcenter

      2. spa-pa:largevocab

   Refer to Language models for more information on available language models.

V‑Cloud 1.6-2021.09.29 includes the following updates and fixes:

1. Multiple improvements to the text processing modules to enhance presentation of words, numbers, and punctuation. This includes fixes for situations related to AM/PM, Q1-Q4, "O" as zero, words/events with spaces, and address-related ordinals.

2. Fixed an error with certain substitution patterns. Substitution patterns with left-hand-side unicode or slash-protected (from a previous rule) strings in multiple value sets/lists are now processed correctly.

3. Multiple language ID (LID) improvements and additions.

   • 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.

V‑Cloud 1.6-2021.05.20 includes the following updates and fixes:

1. Made multiple improvements to number translation and web URL formation.

   • Changed number translation behavior to improve transcript readability. The translation is now more conservative by considering more context for various situations.

2. Made multiple improvements to redaction functionality:

   • Added the scruboffset option when using redaction. Refer to Redaction for more information on this parameter.

   • Fixed an issue from V-Cloud 1.6.2021.04.20, released on April 24, 2021, where number words concatenation was not working properly. This issue caused all single numbers (0 through 9) to be printed as a word instead of a numeral; for example, 1 was transcribed as "one." Numbers 10 and up were correctly processed. This issue affected audio and text scrubbing configurations that depend on single digits.

   • Fixed a rare truncated audio issue when using redaction.

3. Improved error handling:

   • Fixed an issue that could result in ASR errors not being reported correctly.

   • JSON output now only includes the ended , model , nchannels , and audiosecs elements if the stream completed successfully. These elements do not display in JSON output if there was a problem processing audio.

V‑Cloud 1.6-2021.03.25 includes the following updates and fixes:

1. V‑Cloud now returns 502 or 504 status codes when a URL audio source fails.

   1. 502 Bad Gateway returns when a service required for the request is inaccessible.

   2. 504 Gateway Timeout returns when a request is unable to complete within a reasonable amount of time.

   3. Refer to Return codes for more information on V‑Cloud return codes.

2. Fixed an issue that caused URLs with a query component to break due to percent-encoding the URL for all URL audio sources before downloading.

V‑Cloud 1.6-2021.02.25 includes the following updates and fixes:

1. V‑Cloud has been updated to version 7.2 of the ASR engine.

2. Language identification (LID) has been updated to a new engine that significantly increases LID accuracy.

3. V‑Cloud now supports basic HTTP authentication for URL audio sources. When a URL audio source requires authentication, prepend your access credentials to the hostname in the URL as shown in the following example:

   curl -F token=your_token_here \
        -F url=http://username:password@hostname.com/sample-audio-file.wav \
         https://vcloud.vocitec.com/transcribe

4. V‑Cloud now supports HTTP authorization request headers for URL audio sources. When a URL audio source requires authorization, include the authorization header, type, and credentials in your request as shown in the following example:

   curl -F token=your_token_here \
        -F url=http://hostname.com/sample-audio-file.wav \
        -H 'Authorization: auth_type credentials' \
         https://vcloud.vocitec.com/transcribe

V‑Cloud 1.6-2020.12.10 includes the following updates:

1. V‑Cloud now includes requestid as a top-level field in JSON transcripts. requestid is a unique identifier which is automatically generated in all transcripts for tracking purposes.

2. V‑Cloud now includes the following language models:

   1. English (United Kingdom) language models:

      1. eng3:survey

      2. eng3:voicemail

   2. English (Europe) language models:

      1. eng4:survey

      2. eng4:voicemail

   3. French (Canada) language models:

      1. fre1:survey

3. Numerous updates to the following language models:

   1. English (Europe) language models:

      1. eng4:callcenter

      2. eng4:largevocab

   2. French (Canada) language models:

      1. fre1:callcenter

      2. fre1:largevocab

V‑Cloud 1.6-2020.11.05 includes the following updates and fixes:

1. Updated V‑Cloud to version 7.1 of the Automatic Speech Recognition (ASR) Engine.

2. Numerous updates to the following European and Mexican Spanish language models:

   1. European Spanish language models:

      1. spa2:callcenter

      2. spa2:largevocab

      3. spa2:utilities

   2. Mexican Spanish language models:

      1. spa3:travel

      2. spa3:food

3. Fixed an issue that prevented the use of the Content-MD5 HTTP header for data verification.

4. Added user IP address logging to front webserver logs.

5. Added billingSecs , srcBkt , and resBkt to front webserver report.

6. The default values for certain parameters can now be specified per individual token. The defaults of the following parameters are configurable:

   1. model

   2. output

   3. callback

   4. callbackfmt

   5. callbackurl

   6. keepsrc

V‑Cloud 1.6 includes several new features and parameters, including:

1. Language identification (LID) has been enhanced with new option parameters, JSON output elements, and other functionality improvements.

   1. Added new optional parameters for lid

      Table 1. New optional parameters for lid

      Name	Values	Description
      lidoffset= N	integer	Delay start of LID until specified (N) seconds into audio. If there is not enough audio left after offset, this will process preceding utterances in reverse.
      lidutt	true, false	Run LID on every utterance. The default is only once per stream or audio channel. This option is only available with V‑Blaze 7.1+. Note: This option has a significant performance impact and should only be used when necessary.

   2. Added a new lid parameter value for limiting utterance metadata.

      lid= language :notext - use this to not decode when a language is detected, just that a language was detected. If this option is specified, the JSON will not contain a model metadata element in each utterance. Ensure that the language is specified but the domain is not included. For example: lid=spa:notext is valid, but lid=eng1:callcenter:notext is not.

   3. Improved logic for LID decisions with low scores.

      When LID scoring is below the decision threshold, the ASR engine will transcribe the audio with the language model specified by the model tag (or the default model for the ASR configuration if model is not explicitly provided). The results are indicated by a lidinfo.langfinal element in the JSON output.

   4. Made additions to JSON output.

      • lidinfo - now provided at utterance-level when lidutt=true

      • langinfo - breakdown of language information that is added when there was more than one language detected.

      • langfinal - added when the language specified in LID is below threshold and not the default language.

   For more information on LID and using these parameters, refer to Receiving language identification information.

2. Added new debugging parameters and JSON elements to assist with improved warnings and logging when using substitutions.

   1. New debugging parameters

      Warning: These parameters are intended for debugging purposes only and should not be used in production.

      Table 2. New substitution debugging parameters

      Name	Values	Description
      subst	true, false (default), none	The subst parameter can be used to enable or disable automatic system- and model-level substitutions. subst=true Enables system- and model-level substitutions subst=false Disables system-level substitutions; model-level substitutions still apply subst=none Disables both system- and model-level substitutions
      substinfo	true, false (default)	Provides substitution details in JSON transcripts. Set substinfo to true to include a top-level JSON object that indicates the applied substitution rules and a number count for each rule. In addition to the top-level JSON object, substinfo includes another JSON object in the metadata that details each substitution's location, the substitution rule applied, and the substitution rule source.

      For more details on these and other parameters, refer to the Substitutions section of the V-Cloud API docs.

   2. Added a new JSON output element: nsubs shows a count of substitutions applied at both top-level and utterance levels. When substinfo=true , nsubs will also include numtrans counts within the substinfo array. Top-level nsubs does not include numtrans counts. The nsubs element will not appear if no substitutions were applied.

   For more information, refer to the V‑Cloud 1.6 Release Notes (July 2020).

3. Made quality improvements to eng2:largevocab and spa1:voicemail language models.

4. Hinting is now supported for eng1 version 7 models. Hinting for version 5 models is no longer supported.

   For more information on hinting support, refer to the English page of the Language Models Reference.

5. Made minor enhancements and fixes to speech-to-text output processing ( textproc ), including:

   • The system now preserves timestamps on backrefs in substitutions instead of interpolating.

   • Fixed inadvertent uppercase of English cased backrefs (for example, /\1/ ).

   • Eliminated unexpected behavior of pattern{min,max} when min=0 .

6. Made minor improvements to Spanish time formatting.

7. Corrected the scope of emotion scoring to always score individual utterances.

8. Eliminated rare edge case decode failures.

V‑Cloud 1.6 includes several new features and parameters, including:

1. V‑Cloud now allows users to submit audio data through a URL. This method uses the url parameter as an alternative to the file parameter. The provided URL must support HTTP GET and return a Content-Length header when queried. V‑Cloud will verify that the data was properly received from the URL if either a Content-MD5 or ETag header are provided in response to querying the URL.

2. V‑Cloud now allows users to specify a callback format with the callbackfmt parameter. There are three format options available:

   1. multi performs a multi-part POST to the callback with two parts specifying the request ID and the resulting file content.

   2. single performs a standard POST that only contains the result file content within the POST body.

   3. put is identical to single except the callback is performed using an HTTP PUT instead of a POST .

3. V‑Cloud now allows users to send the results of erroneous jobs to a specific callback server with the callbackerror parameter. If callbackerror isn't specified, erroneous job results are sent to the URL defined in the callback parameter instead.

4. V‑Cloud now allows users to enable or disable MD5 verification of audio data using the filemd5 parameter. filemd5 has 3 options:

   1. true is the default value and specifies that MD5 verification should be performed if possible.

   2. false specifies that MD5 verification should not be performed.

   3. User-specified 128-bit value - Specify a sequence of 32 hexadecimal digits to use for MD5 verification.

5. V‑Cloud now supports HTTP basic authentication for callbacks. When a callback server requires authentication information, prepend your access credentials to the hostname in the URL as shown in the following example:

   https://username:password@hostname.com