LunaTranslator/docs/en/settings.md

Settings

Access the settings window by clicking the settings button in the toolbar or the tray icon.

Basic Settings

Choose your text output source. A pink ✓ indicates selection, while a gray × indicates inactive selection; only one can be selected at a time.

• When selecting OCR as the text input, you need to define the OCR region in the toolbar.

• Choosing Textractor (HOOK) as the text input will open a process selection window, followed by a text selection window after choosing a game.

• The default is clipboard mode, which automatically extracts and translates text from the clipboard.

Translation Settings

Configure various translation engines. The developer (HIllya51) hasn't separated different types of translators into categories.

Supported translation engines include:

  Offline Translation: Supports JBeijing7, Kingsoft FastAIT, and YiDianTong for offline translation.

  Free Online Translation: Supports Baidu, Bing, Google, Ali, Youdao, Caiyun, Sogou, DeepL, Kingsoft, iFlytek, Tencent, ByteDance, Volcano, Papago, and Yeekit.

  Registered Online Translation: Supports user-registered API keys for Baidu, Tencent, Youdao, Niutrans, Caiyun, Volcano, and DeepL.

  Pre-translation: Supports loading human translations and aggregated machine pre-translations.

You can select any number of translation engines without restriction.

• The buttons represent: Enable/Disable translator / Set translation text color / Settings

• Offline translation, API online translation, and pre-translation require setup before use.

• For Google Translate and DeepL, you may need to set up a proxy to access them.

• Pre-translation supports fuzzy matching (particularly effective in OCR mode).

HOOK设置

LocaleEmulator settings allow you to set the LocaleEmulator path (built-in for newer versions). Once set, you can launch games through LocaleEmulator.

• "Game Manager" stores previously hooked games for convenient launches (same as clicking "Open Game Manager" in the toolbar).

• "Record Translation File" saves extracted text to the "transkiroku" folder, outputting two files:

  1. game_md5_game_executable_name.sqlite: Records a single translation source output for generating "manual translation" files. Setting a preferred translation source prioritizes that source, and falls back to others if translation fails.

  2. game_md5_game_executable_name.premt_synthesize.sqlite: Used for "machine pre-translation" to record all valid translation results.

• "Export SQLite file to JSON" Exports to JSON for easy translation editing. Set the JSON file path as the "Manual Translation" file path to use manual translations.

• In clipboard and OCR modes, files are recorded with prefixes "0_copy" and "0_ocr" respectively.

➔ See HOOK instructions for detailed usage

OCR Settings

In OCR mode, select your preferred OCR source.

• The local OCR is a built-in engine that's easy to use.

• Baidu OCR, OCRSpace, and Docsumo require API keys.

• Youdao OCR and Youdao Image Translation are experimental interfaces that may be unstable.

• Windows OCR requires Japanese language components installed on your system.

• Setting "Perform OCR at regular intervals" and specifying a maximum interval forces OCR to occur every X seconds, regardless if the game scene has changed or not.

➔ See OCR instructions for detailed usage

Display Settings

• Opacity sets the window background opacity.

• When "Show Original Text" is enabled, you can set options to display furigana and word segmentation results.

• Font styles include four options (Normal, Hollow, Outline, Shadow). The latter three advanced styles can be adjusted using "Hollow Line Width," "Outline Width," and "Shadow Strength" settings.

• "Original Text Color" sets the color for the source text, while "Background Color" sets the window background color.

• "Fill Color" is used for advanced font styles.

• "Selectable Mode" allows content selection within the translation window.

The four font styles are shown below:

Furigana (phonetic guide) display example:

Tokenization (word segmentation) display example:

Voice Settings

• Windows TTS requires Japanese language components installed on your system.

• Azure TTS and Volcano TTS are online services that may become unavailable in the future.

• VoiceRoid2 is an offline TTS engine.

• VOICEVOX is an open-source TTS engine, but it's relatively slow at Text-to-Speech.

Translation Optimization

For text extracted via HOOK, you can set up simple processing operations to improve the content.

This includes common settings and some advanced options.

• "Simple Text Replacement" allows replacing or filtering extracted text.

• "Regular Expression Replacement" requires knowledge of Python's re.sub method (Regex).

• "Manual Translation for Proper Nouns" supports user-configured dictionaries for special terms (e.g., names, places).

• "Translation Result Correction" occurs after translation; forced replacement of translation results can be useful when noun translations fail with certain engines.

• This software partially supports using VNR shared dictionaries.

• Users familiar with Python can directly modify the LunaTranslator\LunaTranslator\postprocess\post.py file to implement custom processing.

Dictionary Settings

With a dictionary configured, LunaTranslaotr can help you in your Japanese learning:

• MeCab setup + "Show Word Segmentation": Displays word boundaries (tokenization)
• MeCab + "Show Furigana": Applies furigana to kanji
• MeCab + "Show Different Colors for Parts of Speech": Highlights grammatical elements
• "Quick Word Lookup": Enables click-to-translate in the translation window

Note: Without MeCab, a basic built-in tokenizer will be used, providing limited furigana and segmentation without part-of-speech distinction.

Resource Download and Update

Automatic updates and links to commonly used resources.

Hotkey Settings

Enable the use of hotkeys, where you can activate and configure specific hotkey settings as desired.