The IPA ↔ SAMPA converter translates phonetic transcriptions between IPA Unicode symbols and the ASCII-only SAMPA / X-SAMPA encoding. SAMPA was created so phoneticians, speech scientists, and linguists could type and exchange transcriptions on systems that lacked IPA fonts — mapping each IPA symbol to one or more ordinary keyboard characters. This tool runs the mapping both ways, instantly and locally.
Why two systems exist
The International Phonetic Alphabet uses Unicode characters that were not available in the ASCII era and can still be awkward to type or embed in plain-text files, CSV data exports, or programming environments that expect ASCII. SAMPA (Speech Assessment Methods Phonetic Alphabet) was developed in the 1980s-90s to solve this problem for European languages by creating a machine-readable ASCII encoding. X-SAMPA extended the same idea to cover the entire IPA chart, including clicks, ejectives, tone marks, and the full vowel inventory.
Today the main uses are: legacy speech corpora and databases that store transcriptions in X-SAMPA, NLP pipelines that use plain text, and situations where copying IPA Unicode into a terminal, database field, or spreadsheet causes rendering problems.
How it works
The converter holds a two-column lookup table pairing each IPA symbol with its X-SAMPA equivalent. Some key mappings:
| IPA symbol | Meaning | X-SAMPA |
|---|---|---|
| ʃ | voiceless postalveolar fricative (“sh”) | S |
| ʒ | voiced postalveolar fricative (“zh”) | Z |
| ŋ | velar nasal (“ng”) | N |
| θ | voiceless dental fricative (“th” in “think”) | T |
| ð | voiced dental fricative (“th” in “the”) | D |
| ə | schwa | @ |
| ɪ | near-close front unrounded vowel | I |
| ʌ | open-mid back unrounded vowel | V |
| æ | near-open front unrounded vowel | { |
When converting IPA to SAMPA, the input is scanned character by character and each IPA glyph is replaced by its ASCII code. Going the other way, SAMPA is matched longest-token-first so multi-character codes (like r\ for the IPA trill r̈) are handled before single-character ones, preventing partial matches.
Any character with no table entry — ordinary Latin letters that are shared between systems (like p, t, k, b, d, m), stress marks ˈ, spaces or brackets — is passed through unchanged, so surrounding text stays intact.
Worked examples
The English word “ship” transcribed as IPA ʃɪp becomes X-SAMPA SIp. The word “thing” /θɪŋ/ becomes TIN. The article “the” as /ðə/ becomes D@. Going the reverse direction, the X-SAMPA string "hEpi represents IPA ˈhɛpi (“happy”), where " is the primary stress mark.
Because Latin letters like p, b, m, f, v, s, z, l, r, n are identical in both systems, words built from common consonants often look nearly the same in either notation. The divergence shows up with fricatives, nasals, vowels, and affricates that require special IPA characters.
Case sensitivity: a common gotcha
X-SAMPA uses both uppercase and lowercase letters with distinct meanings. S means ʃ (the “sh” sound) while s means s (the regular sibilant). N means ŋ (the velar nasal) while n means n. If you are typing X-SAMPA by hand, always use the correct case — a lowercase s where you meant S will produce a different phoneme entirely.
Notes
Most core IPA symbols round-trip cleanly through this converter. A few edge cases — rarely used diacritics, tone letters, or the most obscure extensions — may pass through unchanged if they have no defined X-SAMPA mapping. Everything runs locally; your transcription is never uploaded.