Sometimes people ask us about how the IPA phonetic translators work. Even for languages with relatively strict pronunciation rules, such as Spanish, the algorithm to obtain the phonetic transcription can be very complex. One reason for that is that the pronunciation of one letter can effect the pronunciation of the adjacent letters. This can happen inside a word or at word boundaries.
Here’s an overview of the algorithm of the phonetic translators on EasyPronunciation.com:
Cantonese and Chinese
2,500 lines of code. We use a database to split the text into individual words and show their phonetic transcription.
English
2,800 lines of code. And obviously we use a database combining several dictionaries. The number of words for each English dialect is indicated here.
French
3,400 lines of code. The pronunciation rules in French are extremely complex. We fist use a database containing the phonetic transcription of words (check the number of French words here). If a word is not found in the database, the pronunciation rules are used.
German
900 lines of code. The pronunciation rules are so complex in German that we use only the database to get the phonetic transcription of words.
Italian
1,200 lines of code. We use a database to get the position of stressed vowels and some other information. If the word is found is found in the database, we then apply the standard pronunciation rules in Italian to get its phonetic transcription.
Japanese
2,800 lines of code. And we use a database containing information about stress position in Japanese words and some other data.
Persian
100 lines of code. And the database with words’ transcription.
Portuguese
1,100 lines of code. We only use the pronunciation rules for Portuguese.
Russian
1,400 lines of code. We first use a database with stress position in Russian words. If the word is found in the database, we apply a complex set of pronunciation rules to get its phonetic transcription.
Spanish
1,200 lines of code. We only use the pronunciation rules to get the phonetic transcription of the Spanish text.