Changes in version 0.1.6                        

  - list_voices() can now list voices for all language codes
    (@muschellij2, #6). All voices (for all language codes) are now
    shown by default. Use language = "en-US" to replicate the previous
    default.

  - list_voices() has new arguments engine (to filter on engine types)
    and include_additional_languages (to find bilingual voices with
    specific additional languages).

  - synthesize() and get_synthesis() have new arguments engine (to
    specify the engine type) and language (to specify the language code
    when using bilingual voices).

  - In synthesize() and get_synthesis(), arguments rate and lexicon are
    deprecated and replaced by sample_rate and lexicon_names.

  - In synthesize() and get_synthesis(), the default sample rate is no
    longer 22050 Hz. Instead, AWS will pick the appropriate default
    based on the voice, engine and output format.

  - Speech synthesis now fails with a useful error message when AWS
    returns an error.

                 Changes in version 0.1.5 (2020-03-11)                  

  - Released on CRAN 2020-03-11
  - New maintainer @antoine-sachet

                        Changes in version 0.1.4                        

  - Fixed a function not found bug in pollyHTTP(). (#3)

                        Changes in version 0.1.3                        

  - Requests to get_synthesis() or synthesize() now fail when
    nchar(text) > 1500. (#1, h/t Sean Kross)
  - Bumped aws.signature dependency to use new, more sophisticated
    credentials checking.

                 Changes in version 0.1.2 (2016-12-08)                  

  - Created separate synthesize() and get_synthesis() functions. The
    former as a convenience function wrapping the latter lower-level
    function.

                        Changes in version 0.1.1                        

  - Initial release.