About user dictionary¶

User dictionaries are editable Japanese dictionary. There are three types of user dictionaries: phrase dictionary, word dictionary, and symbol pause dictionary, which allow users to register phrases, words, and symbol pauses. The phrase and word dictionaries allow users to specify and register their own reading of phrases and words that appear in sentences when sentences are not read out as intended, so that when the sentences are synthesized, they are read in the pre-registered reading style. The symbol pause dictionary has a function to insert a pause anywhere in a sentence.

This section describes phrase dictionaries and word dictionaries, which have complex functions. For information on symbol pause dictionaries, see the “Symbol pause” section.

This section describes the Japanese user dictionary. For English word dictionaries, please refer to Word edit (English), and for Chinese word dictionaries, please refer to Word edit (Chinese).

Phrase dictionary¶

Phrase dictionary function¶

By registering the rhyme information of the adjusted phrase in the phrase dictionary, when the same phrase occurs, it can be read out preferentially in the registered phrase reading.

Definition of a phrase¶

AITalk 5 defines a phrase as part of a text (a string of characters) that follows these conditions.

Phrases cannot span multiple sentences.
Sentence endings (“.”, “?” “!”) at the end of a sentence can be applied.

Example 1) X

“お電話ありがとうございます。株式会社エーアイです。”

=> Because it straddles two sentences, it is not considered as a single phrase.

Example2) ○

“お元気ですか？”

=> A line containing a sentence-final symbol at the end is considered a phrase.

Phrases may not begin or end with a space or whitespace (space, tab).

Example 1) X

「別にあなたのためにやったのではないが、」

=> Not considered a phrase because it contains a comma at the end.

Example2) X

「　　　別にあなたのためにやったのではないが」

=> Not considered a phrase because it has a space at the beginning of the sentence.

Example3) ○

「別に　　あなたのために、やったのではないが」

Sentences that contain spaces or commas in the sentence is considered as a phrase.

Applicable conditions for phrases¶

If a phrase registered in the phrase dictionary appears appropriately in the sentence to be synthesized, the phrase will be read out loud in the registered reading style.

If a phrase that matches exactly a registered phrase and a heading appears in a sentence separated by a space, a comma, or a period, the phrase will be applied. If the phrase is not separated by a space, comma, a period, the phrase is not applied, even if the heading is an exact match.

Original text 「この任務が無事に終わったら、彼女にプロポーズしようと思います。」

-> Register phrase 「この任務が無事に終わったら」

Example1)

「この任務が無事に終わったら、旅に出ようと思います。」

=> It is separated by a comma, so the registered phrase applies.

Example2) ○

「この任務が無事に終わったら　旅に出ようと思います。」

=> It is separated by a space, so the registered phrase applies.

Example3) X

「この任務が無事に終わったら旅に出ようと思います。」

=> Registered phrases are not applied because it is not separated by a space, comma, or a period.

If multiple phrases overlap, only one phrase will be applied according to the following rules.
- If the starting point is different, the phrase with the earlier starting point is given priority.
- If more than one phrase can be applied at the start position, the phrase with the longest heading is given priority.

Word dictionary¶

Word dictionary function¶

Assigns unique readings and accents to words.

The final reading of a word is determined by the text-to-speech engine, taking into account the balance of other words, so in rare cases a word may not be read as registered in the word dictionary. If you want to be sure that the word is read as intended, use the sentence ending symbols (“.”, “?”, “!”) If you want to make sure that the phrase is read as intended, please register a phrase as a whole sentence including the sentence ending symbols.

Example) 「東京に出てきてから、もう１０年近く経つ。」

=> Want「東京」 to be read 「コッチ」.

Heading 「東京」 Reading 「コッチ」 Register as a 「Geographic name」

=> 「東京」 will only be read as 「トーキョー」.

Part of speech¶

The part of speech of a word determines the role of the word in the sentence, and depending on the part of speech selected, the word may not be read as registered.

Example) 「彼は、故郷である日向に帰った。」

=> Want 「日向」 to be read as 「ヒューガ」.

Heading 「日向」 Reading 「ヒューガ」 Register as a 「Person name」.

=> Reads 「日向」 as 「ヒナタ」.

Heading 「日向」 Reading 「ヒューガ」 Register as a 「Geographic name」.

=> Reads 「日向」 as 「ヒューガ」 as expected.

Priority¶

Words have a priority, and the higher the priority, the higher chance it will be selected.

Registering words with extremely high priority may cause an imbalance in priority between words, which may affect the process of Japanese text analysis. Unless a high priority is required to read a word as intended, the priority should be set to “Standard” or “Low”. If you have no choice but to set the priority to “High” or “Highest”, please be very careful, as this may cause problems when other text is read.

If you believe it is necessary to raise the priority, please consider whether the following methods can be used before raising the priority.

Register words in longer units.
Register phrase.

* Even if you set the priority to “Highest”, another reading may take precedence depending on the context. To ensure that the phrase is read the way it is intended to be read, please register the phrase.