Does the voice make a pronunciation mistake that you don't want to hear ever again?
Would you like to expand an acronym into the desired spoken form?
The pronunciation correction feature lets you create substitutions to help the voice to get it right the next time!

For example, replace:
"AVE" by "avenue"
"Knot" by "not"

Case sensitivity
The input text is by default matched in a case insensitive way: in our first example, "AVE", "ave", "Ave", etc. will all be replaced by "avenue".
Tick the "Case Sensitive" option if you desire to create a case sensitive entry.

Spelling
Tick the "Spell it!" option if you desire the correction to be spelled: in our first example, the output "avenue" will be read as "a","v","e","n","u","e".

Saving
Add your correction to the dictionary with the "save" button.

Listening
To listen to a correction, click on the "play" button or tap once on a saved correction.

Editing and listening in context
Tap and hold on a saved correction to disable, edit, or delete it, or display it in a random context from the web.

Notes
The replacements will be valid for all voices in the same language set (e.g. for American English, Grace, Michael, Benny, Dark Lord and Ghost will share the same corrections).
Take special care with punctuation symbols in the input or output.
Punctuation symbols in the input will restrict your match. Punctuation symbols in the output will have an effect on what you hear: for example, a period in the output will introduce a long pause during text-to-speech.
You can EXPORT and IMPORT a dictionary, and share it with your friends by using the Android menu key.
Also, you can SEND your corrections to SVOX.
This is a great way to help the technology forward!

Known limitations
- The pronunciation corrections function is experimental. While we have worked hard to make the feature useful, certain symbols or corrections may stop the voice. In this case, please remove the problematic correction(s).
- The corrections are not app-specific. What is wanted in one app or context may be annoying in another. For example, replacing "No" by "Number" will have as a side effect that you won't hear "No" for an answer again.
- Corrections have to be specified using orthographic input. For you phoneticians out there it would be great to enter phoneme symbols such as IPA or XSAMPA; this is currently not supported. You have to find creative spellings to hear the output you want, which can sometimes be challenging!