phones.rb

# 'r' after open-mid central unrounded?

module Keats
  module Phones
    IPA_CONSONANTS = {
      "B"  => "b",          # b  - Lower-case B	    <be>      voiced bilabial plosive
      "CH" => "t\312\203",  #	tʃ - T-Esh ligature   <cheese>  voiceless postalveolar affricate 
      "D"  => "d",          # d  - Lower-case D     <dee>     voiced dental or alveolar plosive
      "DH" => "\303\260",   #	ð  - Eth              <thee>    voiced dental fricative
      "F"  => "f",          # f  - Lower-case F     <fee>     voiceless labiodental fricative
      "G"  => "g",          # g  - Lower-case G     <green>   voiced velar plosive
      "HH" => "h",          #	h  - Lower-case H     <he>      voiceless glottal fricative
      "JH" => "d\312\222",  #	dʒ - D-Yogh ligature  <gee>     voiced postalveolar affricate
      "K"  => "k",          # k  - Lower-case k     <key>     voiceless velar plosive
      "L"  => "l",          # l  - Lower-case L     <lee>     voice alveolar lateral approx.
      "M"  => "m",          # m  - Lower-case M     <me>      voiced bilabial nasal
      "N"  => "n",          # n  - Lower-case N     <knee>    voiced alveolar nasal
      "NG" => "\305\213",   #	ŋ  - Engma            <ping>    voiced velar nasal
      "P"  => "p",          # p  - Lower-case P     <pee>     voiceless bilabial plosive
      "R"  => "\311\271",   # ɹ  - Turned R         <read>    voiced alveolar approximate
      "S"  => "s",          # s  - Lower-case S     <sea>     voiceless alveolar fricative
      "SH" => "\312\203",   #	ʃ  - Esh              <she>     voiceless postvelar fricative
      "T"  => "t",          # t  - Lower-case T     <tea>     voiceless alveolar plosive
      "TH" => "\316\270",   #	θ  - Theta            <theta>   voiceless dental fricative
      "V"  => "v",          # v  - Lower-case V     <vee>     voiced labiodental fricative
      "W"  => "w",          # w  - Lower-case W     <we>      voiced labio-velar approximant
      "Y"  => "j",          # j  - Lower-case J     <yield>   voiced palatal approximant
      "Z"  => "z",          # z  - Lower-case Z     <zee>     voiced alveolar fricative
      "ZH" => "\312\222"    # ʒ	 - Yogh             <seizure> voiced postalveolar fricative
    }
    
    IPA_VOWELS = {
      "AA" => "\311\221",   # a  - Script A         <odd>     open back unrounded vowel
      "AE" => "\303\246",   # æ  - Ash              <at>      near-open front unrounded vowel 
      "AH" => "\312\214",   # ᴧ  - Turned V         <hut>     open-mid back unrounded vowel
      "AO" => "\311\224",   #	ɔ  - Open o           <ought>   open-mid back rounded vowel
      "EH" => "\311\233",   #	ɛ  - Epsilon          <Ed>      open-mid front unrounded vowel
      "ER" => "\311\234",   #	ɜ  - Reversed-Epsilon <hurt>    open-mid central unrounded vowel
      "IH" => "\311\252",   #	ɪ  - Small Capital I  <it>      near-close, front unrounded vowel
      "IY" => "i",          #	i  - Lower-case I     <eat>     close front unrounded vowel
      "UH" => "\312\212",   #	ʊ  - Horseshoe u      <hood>    near-close near-back vowel
      "UW" => "u",          # u  - Lower-case U     <two>     close back rounded vowel
      "OY" => "\311\224i",  #	ɔi                    <toy>     (diphthong)
      "OW" => "o\312\212",  # oʊ                    <oat>     (diphthong)
      "EY" => "e\311\252",  #	eɪ                    <ate>     (diphthong)
      "AY" => "a\311\266",  #	aɪ                    <hide>    (diphthong)
      "AW" => "a\312\212",  # aʊ   	                <cow>     (diphthong)
    }
    
    IPA_CHART = IPA_VOWELS.update(IPA_CONSONANTS)

    STRESS_CHART = {
      "1"  => "\313\210",   # ˈ  - Primary Stress
      "2"  => "\313\214",   # ˌ  - Secondary Stress
      "0"  => ""            #    - No Stress
    }
  end
end

class String
  include Keats::Phones
  
  def ipa_tokenize
    self.split.collect do |token| 
      phoneme, stress = token.scan(/(\w{1,2})(\d)?/).first
      pronounciation = "%s%s" % [STRESS_CHART[stress], IPA_CHART[phoneme]]
      pronounciation.gsub(/(      # Vowel elongation on stressed syllables
        (?: \313\210 | \313\214 ) # Stress marks
        (?: [iu] | \311\224 )     # Vowels to elongate
        (!? i | $ )?              # Dont elongate diphthong segments or at the end of words
      )/x, '\1:')
    end
  end

  def to_ipa
    "/%s/" % self.ipa_tokenize.join("")
  end
end

# Example
puts "AH0 B Z AO1 R B AH0 N S IY0".to_ipa