Services or capabilities described in Amazon Web Services documentation might vary by Region. To see the differences applicable to the China Regions, see Getting Started with Amazon Web Services in China (PDF).

Character sets for custom vocabularies and vocabulary filters

For each language Amazon Transcribe supports, there is a specific set of characters Amazon Transcribe can recognize. When you create a custom vocabulary or vocabulary filter, use only the characters listed in your language's character set. If you use unsupported characters, your custom vocabulary or vocabulary filter fails.

Many Unicode characters can appear identical in popular fonts, even if they use different code points. Only the code points listed in this guide are supported. For example, the French word déjà can be rendered using precomposed characters (where one Unicode value represents an accented character) or decomposed characters (where two Unicode values represent an accented character, one value for the base character and another for the accent).

Afrikaans character set

For Afrikaans custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

Arabic character set

For Arabic custom vocabularies, you can use the following Unicode characters in the Phrase and SoundsLike fields. You can also use the hyphen (-) character to separate words.

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

Chinese, Mandarin (Simplified) character set

For Chinese (Simplified) custom vocabularies, the Phrase field can use any of the characters listed in the following file:

The SoundsLike field can contain the pinyin syllables listed in the following file:

When you use pinyin syllables in the SoundsLike field, separate the syllables with a hyphen (-).

Amazon Transcribe represents the four tones in Chinese (Simplified) using numbers. The following table shows how tone marks are mapped for the word 'ma'.

Chinese (Simplified) custom vocabularies don't use the IPA field, but you must still include the IPA header in the custom vocabulary table.

The following example is an input file in text format. The example uses spaces to align the columns. Your input files should use TAB characters to separate the columns. Include spaces only in the DisplayAs column.

Phrase       SoundsLike               IPA     DisplayAs
康健          kang1-jian4 
谴责          qian3-ze2 
国防大臣        guo2-fang2-da4-chen2 
世界博览会      shi4-jie4-bo2-lan3-hui4          世博会

Chinese, Mandarin (Traditional) character set

For Chinese (Traditional) custom vocabularies, the Phrase field can use any of the characters listed in the following file:

The SoundsLike field can contain the zhuyin syllables listed in the following file:

When you use zhuyin syllables in the SoundsLike field, separate the syllables with a hyphen (-).

Amazon Transcribe represents the four tones in Chinese (Traditional) using numbers. The following table shows how tone marks are mapped for the word ㄇㄚ.

Chinese (Traditional) custom vocabularies don't use the IPA field, but you must still include the IPA header in the custom vocabulary table.

The following example is an input file in text format. The example uses spaces to align the columns. Your input files should use TAB characters to separate the columns. Include spaces only in the DisplayAs column.

Phrase        SoundsLike                     IPA        DisplayAs
健康          ㄐㄧㄢˋ-ㄎㄤ 
譴責          ㄑㄧㄢˇ-ㄗㄜˊ 
國防部長       ㄍㄨㄛˊ-ㄈㄤˊ-ㄅㄨˋ-ㄓㄤˇ 
世界博覽會     ㄕˋ-ㄐㄧㄝˋ-ㄅㄛˊ-ㄌㄢˇ-ㄏㄨㄟˋ             世博會

Danish character set

For Danish custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

Dutch character set

For Dutch custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

English character set

For English custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

Farsi character set

For Farsi custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields.

You can use the following International Phonetic Alphabet in the IPA field of your custom vocabulary file:

French character set

For French custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet in the IPA field of your custom vocabulary file:

German character set

For German custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

Hebrew character set

For Hebrew custom vocabularies, you can use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

Hindi character set

For Hindi custom vocabularies, you can use the following Unicode characters in the Phrase and SoundsLike fields:

Amazon Transcribe maps the following characters:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Indonesian character set

For Indonesian custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Italian character set

For Italian custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Japanese character set

For Japanese custom vocabularies, the DisplayAs field supports all hiragana, katakana, and kanji characters, and fullwidth romaji capital letters.

The Phrase field supports the characters listed in the following file:

The SoundsLike field supports lowercase romaji characters, as listed here:

Represent long vowels by doubling the vowel:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Korean character set

For Korean custom vocabularies, you can use any of the Hangul syllables in the Phrase and SoundsLike fields. For more information, see Hangul Syllables on Wikipedia.

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Malay character set

For Malay custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Portuguese character set

For Portuguese custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Russian character set

For Russian custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Spanish character set

For Spanish custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Swedish character set

For Swedish custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file:

Tamil character set

For Tamil custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Telugu character set

For Telugu custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Thai character set

For Thai custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Turkish character set

For Turkish custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of your input file:

Vietnamese character set

Amazon Transcribe represents the six tones in Vietnamese using numbers. The following table shows how tone marks are mapped for the word 'ma'.

For Vietnamese custom vocabularies, you can use the following characters in the Phrase and SoundsLike fields:

You can also use the following Unicode characters in the Phrase and SoundsLike fields:

You can use the following International Phonetic Alphabet characters in the IPA field of the custom vocabulary input file: