V-Blaze Online Help

Substitutions

Substitution rules contain lists of “target:replacement” mappings, which enable you to correct consistent and frequent transcription errors that result from out-of-vocabulary words, excess noise in the audio, poor enunciation, or strong accents. They can also be also used to correct word combinations that rarely occur in general speech but occur frequently within a specific domain or company.

For example, useful substitutions in the insurance domain could include “giant whale : giant hail” and “hit a beer : hit a deer”.

Table 1. Substitutions Parameters

Name

Availability

Type

Values

Description

subst_list

V‑Blaze, V‑Spark

string

filename

The subst_list deployment method involves placing the substitution file in the /opt/voci/state/substitutions/ directory on all ASR servers being used for transcription.

Once the substitution file has been placed, set the value of the subst_list parameter to the name of the substitution file as shown in the example below.

subst_list=file.sub          

Note

The subst_list method of deployment should only be implemented if multiple clients are utilizing the same ASR servers. Placing the substitution file on the ASR server ensures substitution rules are synchronized across multiple clients.

subst_rules

V‑Cloud, V‑Blaze v5.6+, V‑Spark

string

string/filename

Specifies a newline-delimited string that contains substitution rules to be applied during transcription. Rules can be specified directly as a string, however, the recommended approach is to auto-populate the string with the contents of a local substitution file into the subst_rules field by using the redirection operator <. This approach makes it easier to edit and maintain these rules over time.

Use the subst_rules parameter as shown in the example below.

subst_rules=</path/to/file.sub