AlexaDev Tuesday: Audio Resources For Foreign Language Alexa Skill Devs

Alexa’s available in many regions outside the US now (France was just announced today, and Italy will be joining us soon!), which presents both opportunities and challenges for English-speaking skill devs. Today I’m sharing some audio resources for foreign language Alexa skill devs, to help create foreign language audio files and get them into the proper format for Alexa.


Use Your Resources


MARY is a free, open-source text-to-speech conversion tool that ‘reads’ the text you provide, converts it to a speech audio file, and allows you to download the resulting file. MARY isn’t intended specifically for Alexa skill developers, but it’s very useful for converting written text—in multiple languages—into audio files. This can be a big help when creating foreign language versions of your skills.

First, it will give you some idea of what the end user will hear when Alexa speaks, which can help with optimizing your interaction model. Second, it can be used to generate your own foreign language speechcons, which you can then drop into your user interactions to add interest, humor or excitement to your skill.

It’s important to know the synthesized speech will not sound natural, so this tool cannot match the quality of foreign language text passages spoken by a human who is fluent in that language. Still, the tool definitely has its uses.

Click here for the MARY project homepage.

Click here for an in-browser MARY TTS conversion tool client.


Audacity – Convert Your MARY File to MP3
Unfortunately the MARY tool cannot output to MP3, but the free, open-source Audacity software can! Audacity is a fantastic tool to keep in your techie bag o’ tricks anyway, it’s feature-packed and makes audio remastering easy. The interface is not intuitive, but there are many YouTube videos and online tutorials available. Just Google Audacity + {feature or task name} to find the specific help you need.

Click here to go to the Audacity site.


JOVO Alexa MP3 File Converter
Now that you’ve got your audio in MP3 format, the JOVO Alexa MP3 Audio File Converter tool can convert it into the correct format to meet the Alexa SSML audio snippet spec. The JOVO site will also host the resulting audio file for 24 hours, to facilitate testing and quick changes during development.

Click here for the JOVO Alexa MP3 File Converter


Related Post

AlexaDev Tuesday: International Alexa Skills – this post addresses programmatic requirements for multi-language Alexa skills.

AlexaDev Tuesday: International Alexa Skills With No Code Changes – this post explains how to release your English-language skill in foreign regions where English is the primary spoken language for Alexa. It’s super-easy!