Navigation:  Managing a Site > Operator Management > Speech Profiles

 

Overview

 

A speech profile is a set of parameters (speech input/output, speech detection/recognition, etc.) to be used by operators/operator groups with Speech attributes.

This feature is only available if the site is associated to a subscription with speech services.

Speech Profiles can be created at account level (the profile becomes a global speech profile) or site level (the profile is a local speech profile. See Adding a Local Speech Profile).

The managing of speech profiles at account level (Global Speech Profile Management) is a default option but it is possible to allow each site to create speech profiles that will only be available within that site (Local Speech Profile Management). This can be done in the "Sites" module of the Admin Dashboard but also in the Site Dashboard's "Site Details" sub-module (see Managing a Site - Editing Site Details). Once a site is assigned the management of its own speech profiles, this overrides speech profile management at account level and a new sub-module will become available in the "Operators" module - the "Speech Profiles" sub-module.

To further help with overall recognition performance for a specific operator, it is possible to perform an adaptation. See Adaptation.

 

 

 

To access the Speech Profile Management page, click and, in the resulting row below, select .

 

 

The image below illustrates the page that is accessed when clicking .

 

 

This information can be displayed in a grid or a list view. By default, the speech profiles are displayed in a grid view but the view can be changed.

Click or to select the most appropriate view:

 

•Click for a grid view:

 

 

This view displays the following information:

 

 

The icon marks the current default local speech profile.

 

OR

 

•Click for a list view:

 

 

This view provides the following information:

 

 

 

The "Speech Profiles Management" page also displays the following options:

 

Use these buttons to control the number of displayed items per page. Define 10, 20, 50 or 100 items to be viewed in each page by clicking the corresponding button.

 

Use the search box to look for specific list items.

 

a. Enter what you want to search for.

b. Click or press <ENTER> in your PC keyboard to initiate the search.

c. To conclude the search operation, empty the search box and click or press <ENTER> in your PC keyboard.

 

Click this button to export a list of local speech profiles. Proceed as follows:

 

 

a. Click .

 

The speech profile list is exported with an Excel file format.

The conclusion of the export operation depends on the web browser used. You should be able to save the Excel file in your PC or open it directly after the download.

 

Click this button to refresh the page's information.

 

Use the buttons, included in this bar, to organize the listed items of the corresponding column:

•by ascending order.

• by descending order.

 

Click this button to select the listed items on display. In this case, the speech profiles.

 

Click the speech profile's name (displayed in light blue) to access the page with the corresponding details. See Local Speech Profile Details.

 

 

Available Operations

 

The "Speech Profile Management" page also provides direct access to specific operations:

 

To create a local speech profile, use . For more detailed information, see Adding a Local Speech Profile.

 

To delete a local speech profile, use . For more detailed information, see Deleting a Local Speech Profile.

 

To set a local speech profile as default, use . Proceed as follows:

 

a. Select the speech profile you want to define as default by checking the corresponding .

 

b. Click .

 

 

c. Click .

 

The selected speech profile is marked with an icon and becomes the default speech profile.

 

To apply a local speech profile, use . For more detailed information, see Applying a Local Speech Profile.

 

 

Adaptation

 

In order to improve speech recognition performance, the adaptation process will analyze the operator's pronunciation characteristics (accent) and store his deviation to the expected phonemes for a given language database. This adaptation is very useful for non-native speakers, speakers with accents or special voice characteristics because it improves overall recognition performance for that particular operator.

 

Within the MCL-Voice V4 environment, the adaptation process is managed by the Voice Engine, in the background, during operator execution of the application dialog.

 

Regarding MCL-Voice V3, the adaptation process starts when the operator logs in - the device communicates with the account to verify if no adaptation has already been performed and, if no adaptation data exists, an automatic procedure is set on the device to execute the adaptation process.

The adaptation performed in a V3 environment consists of data collection (the operator repeating words and digits that will be used to interact with the application in the device) and accent analysis (the resulting file stores that operator's pronunciation characteristics which are analyzed).

 

Adaptation can be enabled at speech profile level or operator profile level. In other words, you can set the adaptation as a default process for all the operators in an operator group (in the "Misc Speech Settings" tab of a speech profile - see Editing a Local Speech Profile - Misc Speech Settings Tab) or you can enable this process for specific operators (in the selected operator's "Operator Details" page, in the "Misc Speech Settings" tab - see Editing Operator Details - Misc Speech Settings Tab).

 

 

The operator profiles you create within an operator group after you have associated a speech profile with enabled adaptation to it, will inherit adaptation. Existing operators (meaning, operators already contained in the operator group and any operators you move into it in the meantime) do NOT inherit the feature from the speech profile. if you want an existing operator to go through the adaptation process, you MUST enable it in the corresponding operator profile.

 

 

This chapter describes the operations available regarding speech profiles: