Pronunciation

The AppArmor voice mechanisms is able to convert the text of your notifications to voice using a “text-to-speech” engine. Under normal circumstances, the text-to-speech engine uses the context to determine the correct pronunciation of all words, and is very effective.

Limitations of Text-to-Speech

In some situations, the text-to-speech engine is unable to properly pronounce some words. Typically proper names with less common pronunciation can be challenging for the text-to-speech engine.

To address this problem, AppArmor provides a pronunciation feature. It allows you to replace words with alternatives to "trick" the text-to-speech engine into the correct pronunciation.

Example

As an example, there is a street here in Toronto called "Strachan Avenue". However, the pronunciation is "Straun Avenue". Proper names are good examples of where the pronunciation tool can be used effectively. For this situation, we’ll provide “Strachan” as the literal, and “Straun” as the alternate. Here is the configuration:

Challenging words

There are certain words that are challenging for text-to-speech engines to pronounce if no clear context is available. For example, the word "resume" can refer both to a “professional resume” or the verb “to resume”. This particular word is difficult word for text-to-speech engines. The engine tries to use the context of the word to figure out the pronunciation, but that doesn't always work.

Limitations of the pronunciation feature

Our pronunciation tool won't be much help for words like “resume”. The word "resume" is also a verb, "... classes will resume on Monday...". So using the pronunciation tool to provide an alternate will replace all occurence of the the word "resume". This will likely mean incorrect pronunciations when the "resume" is used as a verb.

Using synonyms for challenging words

You might consider using synonyms such as "CV", "curriculum vitae", "biography", etc, as an alternative to “resume”.

Evolution of Text-to-Speech

AppArmor is contantly improving our systems. Hopefully the TTS technology will evolve in the future and these types of issues will be less frequent.