Portuguese IPA Transcription - Help

← Back to Portuguese transcription ↑ All Languages

About This Implementation

This Portuguese transcription app uses the Wiktionary Portuguese Pronunciation Module to generate phonetic/phonemic transcriptions for Portuguese text.

The system is exceptionally powerful, capable of modeling numerous regional dialects and subtle phonetic variations. However, it requires specific user input for words with ambiguous pronunciation.

Dialects Supported in this App

This tool offers two main dialect choices that represent the major varieties of Portuguese.

• Brazil: This option generates a transcription based on the Rio de Janeiro (Carioca) accent. This is a widely recognized standard known for its palatalized "chiado" 's' sound.
• Portugal: This option generates a transcription based on the General / Standard European Portuguese accent, often based on the pronunciation of Lisbon.

Transcription Forms

You can choose between two levels of detail for the IPA transcription, depending on your needs.

• Phonetic: A detailed, narrow transcription that shows subtle, non-distinctive variations in pronunciation (allophones). It aims to represent how a word is actually said. For example, it will show the fricative pronunciation of 'b' between vowels in Portugal, as in saber [sɐˈβeɾ].
• Phonemic: A simplified, broad transcription that shows only the sounds that are essential for distinguishing meaning (phonemes). It represents the abstract sound system of the language. For example, saber would be shown as /sɐˈbeɾ/.

---

Important Note: Why You Might See an Error

This transcription engine is designed to be highly accurate. Because Portuguese spelling is not always perfectly phonetic, the engine will intentionally produce an error when it encounters an ambiguous word. This prevents it from guessing and generating an incorrect pronunciation.

To fix this, you need to provide a "respelling" in the input box to clarify the intended sound. Here are the most common reasons for errors:

1. Ambiguous Stressed Vowels ('e' and 'o')

This is the most common cause of errors. In Portuguese, a stressed 'e' or 'o' can be open or closed, but standard spelling often doesn't use an accent mark to show this.

Example: The word governo has stress on the 'e', but the script doesn't know if it's open or closed.

Solution: You must add an accent mark to the vowel in your input to specify the quality.

• For a closed sound, use a circumflex: govêrno.
• For an open sound, use an acute accent: férro.

The script will typically show an error like: "Stressed e or o... must be marked for quality using é/ê or ó/ô"

2. The Irregular Letter 'x'

The letter 'x' has four possible pronunciations, and the script cannot reliably guess which one to use.

Example: The script will fail on words like exame or táxi.

Solution: Replace the 'x' with its phonetic equivalent in the input.

• For the /z/ sound in exame, enter: ezame.
• For the /ks/ sound in táxi, enter: tácsi.
• For the /ʃ/ sound in caixa, enter: caisha.

The script will typically show an error like: "x must be respelled z, sh, cs, ss or similar"

---

Implementation Details

Underlying Dialect Profiles in the Engine

While this app provides two primary options for simplicity, the core engine is capable of modeling eight distinct dialect profiles. The available profiles are:

• Brazilian Profiles:
  • gbr: General Brazilian (a standardized variety).
  • rio: Rio de Janeiro (Carioca) - The dialect used for the "Brazil" option in this app.
  • sp: São Paulo (Paulistano).
  • sbr: Southern Brazil (Gáucho).
• European Profiles:
  • gpt: General Portuguese - The dialect used for the "Portugal" option in this app.
  • cpt: Central Portugal.
  • spt: Southern Portugal (Alentejo/Algarve).
  • npt: Northern Portugal (Porto/Minho).

Dialectal Differences (Between App Options)

This is a summary of the key differences between the two available options:

• Brazil (Rio de Janeiro Accent)
  • Coda 's' (Chiado): Pronounced [ʃ] or [ʒ] (e.g., pastas → [ˈpaʃ.tɐʃ]).
  • Unstressed 'e'/'o': Raised to [i] and [u] (e.g., ele → [ˈe.li]).
  • 't'/'d' before 'i': Palatalized to [tʃ] and [dʒ] (e.g., dia → [ˈdʒi.ɐ]).
  • Coda 'l': Vocalized to [w] (e.g., sol → [sɔw]).
• Portugal (General Accent)
  • Coda 's' (Chiado): Also pronounced [ʃ] or [ʒ], similar to Rio (e.g., pastas → [ˈpaʃ.tɐʃ]).
  • Unstressed 'e': Reduced to [ɨ] or deleted, not raised to [i] (e.g., ele → [ˈel(ɨ)]).
  • 't'/'d' before 'i': Not palatalized, remain [t] and [d] (e.g., dia → [ˈdi.ɐ]).
  • Coda 'l': Pronounced as a dark [ɫ] (e.g., sol → [sɔɫ]).

---

For technical issues or suggestions, please visit our GitHub repository.