Nepali Phonetics and Character Mapping for Building Speech Technology

Script to Phoneme Mapping

Script	Phoneme (IPA)	Script	Phoneme (IPA)
अ	/ʌ/	आ	/aː/
इ	/i/	ई	/iː/
उ	/u/	ऊ	/uː/
ए	/e/	ऐ	/ẽ/
ओ	/o/	औ	/au/
अं	/ʌ̃/	अँ	/ã/
क	/k/	ख	/kʰ/
ग	/ɡ/	घ	/ɡʱ/
ङ	/ŋ/	च	/t͡s/
छ	/t͡sʰ/	ज	/d͡z/
झ	/d͡zʱ/	ञ	/ɲ/
ट	/ʈ/	ठ	/ʈʰ/
ड	/ɖ/	ढ	/ɖʱ/
ण	/ɳ/	त	/t̪/
थ	/tʰ/	द	/d̪/
ध	/dʱ/	न	/n/
प	/p/	फ	/pʰ/
ब	/b/	भ	/bʱ/
म	/m/	य	/j/
र	/r/	ल	/l/
व	/w/	स	/s/
ष	/ʂ/	ह	/ɦ/
श्र	/ʃr/	ज्ञ	/d͡zɲ/
क्ष	/kʃ/

JSON Representation

{
  "vowels": {
    "अ": "/ʌ/",
    "आ": "/aː/",
    "इ": "/i/",
    "ई": "/iː/",
    "उ": "/u/",
    "ऊ": "/uː/",
    "ऋ": "/r̩/",
    "ॠ": "/r̩ː/",
    "ए": "/e/",
    "ऐ": "/ẽ/",
    "ओ": "/o/",
    "औ": "/au/",
    "अं": "/ʌ̃/",
    "अँ": "/ã/",
    "अ:": "/ʌh/" 
  },
  "diacritics": {
    "ा": "/aː/",
    "ि": "/i/",
    "ी": "/iː/",
    "ु": "/u/",
    "ू": "/uː/",
    "ृ": "/r̩/",
    "ॆ": "/e/",
    "े": "/e/",
    "ै": "/ẽ/",
    "ॉ": "/o/",
    "ो": "/o/",
    "ौ": "/au/",
    "्": "virama",
    "ः": "/h/"  
  },
  "consonants": {
    "क": "/k/",
    "ख": "/kʰ/",
    "ग": "/ɡ/",
    "घ": "/ɡʱ/",
    "ङ": "/ŋ/",
    "च": "/t͡s/",
    "छ": "/t͡sʰ/",
    "ज": "/d͡z/",
    "झ": "/d͡zʱ/",
    "ञ": "/ɲ/",
    "ट": "/ʈ/",
    "ठ": "/ʈʰ/",
    "ड": "/ɖ/",
    "ढ": "/ɖʱ/",
    "ड़": "/ɽ/",
    "ढ़": "/ɽʱ/",
    "ण": "/ɳ/",
    "त": "/t̪/",
    "थ": "/tʰ/",
    "द": "/d̪/",
    "ध": "/dʱ/",
    "न": "/n/",
    "प": "/p/",
    "फ": "/pʰ/",
    "ब": "/b/",
    "भ": "/bʱ/",
    "म": "/m/",
    "य": "/j/",
    "र": "/r/",
    "ल": "/l/",
    "व": "/w/",
    "श": "/ʃ/",
    "ष": "/ʂ/",
    "स": "/s/",
    "ह": "/ɦ/",
    "श्र": "/ʃr/",
    "ज्ञ": "/d͡zɲ/",
    "क्ष": "/kʃ/"
  },
  "digits": {
    "०": "0",
    "१": "1",
    "२": "2",
    "३": "3",
    "४": "4",
    "५": "5",
    "६": "6",
    "७": "7",
    "८": "8",
    "९": "9"
  },
  "punctuation": {
    "!": "exclamation",
    "\"": "quotation",
    "'": "apostrophe",
    ",": "comma",
    ".": "period",
    ":": "colon",
    "?": "question_mark",
    "।": "danda",  
    "_": "underscore"
  }
}

  

What were the rules ?

Rules were inspired from wikipedia article: https://en.wikipedia.org/wiki/Nepali_phonology

Consonants

Spoken Nepali has 30 consonants in its native system, though some have tried to limit the number to 27.

Nepali consonant phonemes

			Bilabial	Dental	Alveolar	Retroflex	Dorsal	Glottal
Nasal			m (म)		n (न/ञ)	(ɳ (ण))	ŋ (ङ)
Plosive/ Affricate	Voiceless	Unaspirated	p (प)	t̪ (त)	t͡s (च)	ʈ (ट)	k (क)
		Aspirated	pʰ (फ)	tʰ (थ)	t͡sʰ (छ)	ʈʰ (ठ)	kʰ (ख)
	Voiced	Unaspirated	b (ब)	d̪ (द)	d͡z (ज)	ɖ (ड)	ɡ (ग)
		Aspirated	bʱ (भ)	dʱ (ध)	d͡zʱ (झ)	ɖʱ (ढ)	ɡʱ (घ)
Fricative					s (स/श/ष)			ɦ (ह)
Trill					r (र)
Approximant			(w (व))		l (ल)		(j (य))

Vowels

Nepali has 11 phonologically distinctive vowels, including 6 oral vowels and 5 nasal vowels. In some contexts, intervocalic "h" leads to breathy-voiced vowels.

Nepali vowel phonemes

	Front		Central		Back
	Oral	Nasal	Oral	Nasal	Oral	Nasal
Close	i (इ)	ĩ (ई)			u (उ)	ũ (ऊ)
Close-mid	e (ए)	ẽ (ऐ)			o (ओ)
Open-mid					ʌ (अ)	ʌ̃ (अँ)
Open			ä (आ)	ã (आँ)

Normalization Rules

1. Handling Halanta (्): A consonant with a halanta removes the inherent vowel "अ".
2. Combining Matras: Replace the inherent vowel "अ" with the vowel corresponding to the matra.
3. Clusters and Special Cases: Map clusters like क्ष, ज्ञ, श्र appropriately.