This article lists all the features/capabilities of VOIP E2E testing
Botium text asserter can be used in test cases for voice bots. The text is automatically converted to speech before the test case is run with the voice bot.
Because this text is automatically converted to speech there can sometimes be errors thrown by a mismatch between the Expected Text and Actual Text. This is due to small organic errors introduced by the Text-to-Speech (TTS) engine.
For example the TTS converted the word assistant into ‘Office Tent’.
To overcome this issue we have added a ‘Word Error Rate’ asserter. This allows the user to set the acceptable error rate for a message in %. This in turn allows Botium to ignore any small errors introduced by the TTS Engine.
Word Error Rate Asserter with JOIN method
The silence duration between sentences in converting from Speech-to-Text (STT) can sometimes cause failures due to bot responses being incorrectly split into multiple bubbles.
To mitigate this issue, users can employ different message handling methods in the voice bot configuration, such as 'Original,' 'Join,' and 'Split,' within the Speech Recognition Bot.
- Original: The Original method keeps the original message from coming from the STT provider.
- SPLIT: The “SPLIT” method allows the user to split based on a specified character. This can be useful but can sometimes cause issues if the STT engine doesn't provide those specified characters.
- JOIN: When the JOIN method is selected, messages with a silence duration below the specified threshold are joined together. There is also a Delimiter which allows the user to set a custom way to join. For example, a user may choose to use a dot or a space to join these messages.
Join by Silence Duration Logic Hook
Dual-tone multi-frequency (DTMF) buttons
DTMF buttons are the sounds or tones generated by a telephone when the numbers are pressed. These tones are transmitted with the voice channel and often are used to allow users to make selections from a list of choices.
These can be tested in Botium using the typical button asserter. Simply set the button payload and text to the required DTMF tone and Botium will test it. It is also possible to set multiple digits within a single button asserter.
Transform to Voice
Botium automatically converts Text Asserters to Voice for VOIP E2E testing. However you can also transform the text to voice manually. In this way you can create a new test case where you can select the voice profile, the language, the voice along with any voice effects you require.
In this way you can try various voices, such as male and female voices along with different kinds of effects such as background noise or humanification.
Once this is applied you will see in the place of text there will be an audio file. You can check this audio file by playing it within Botium.
The transform to voice feature allows users to make selections on the test case level (along with voice effects).
Silence Duration Timeout
If there is a specific test case where the user would like to ignore the Silence Duration Timeout this can be achieved using the ‘VOIP - Ignore Silence Duration’ logic hook. This will ignore the silence between the step that it is applied to and the previous step. This can be useful for steps that don’t have a response - such as a fallback intent.