Spaces:
Sleeping
Sleeping
| This folder contains the following three pronunciation customization files by default. | |
| • dont_translate_phrases.txt | |
| - You can add a list of phrases or words you do not want to be translated. | |
| - This will work for both Google Translate and DeepL | |
| • interpret-as.csv (Azure Only) | |
| - You can use SSML parameters to customize how specific words or phrases are pronounced | |
| - See this article for documentation: https://learn.microsoft.com/en-us/azure/cognitive-services/speech-service/speech-synthesis-markup-pronunciation#say-as-element | |
| - Note: The script will match the phrases in the TRANSLATED text. You may therefore wish to also add these phrases to 'dont_translate_phrases.txt'. | |
| - The first row contains the titles of each column - Do not change anything in the first row! | |
| - Descriptions of each column: | |
| • Text: The word or phrase that will be pronounced how you specify, if it is found in the text to be spoken | |
| • interpret-as Type: The way in which the word/phrase will be pronounced. See documentation link above. (Some examples include: characters, cardinal, ordinal) | |
| • Case Sensitive (True/False): Whether to only modify the pronunciation if the word/phrase matches exactly, being case sensitive | |
| • Format (Optional): Only applicable to some types, such as 'date', 'time', and others. Otherwise leave blank. See documentation link above for details | |
| - See 'Example - interpret-as.csv' for an example of how to use this file | |
| - This will only apply if using Azure TTS, not Google | |
| • aliases.csv (Azure Only) | |
| - Lets you effectively change what should be spoken instead of a certain word or phrase | |
| - Example: If the text to be spoken contains "BTW" you can have it say "by the way" | |
| -Note: It does NOT actually replace the text, but only changes how the voice will pronounce it | |
| - The first row contains the titles of each column - Do not change anything in the first row! | |
| - Description of each column: | |
| - Original Text: The original word or phrase to match | |
| - Alias: The word or phrase to speak instead of the original text | |
| - Case Sensitive (True/False): Whether it must be an exact match including capital/lower case. If nothing is entered, will default to False | |
| - This will only apply if using Azure TTS, not Google | |
| • Manual_Translations.csv | |
| - If you know you are going to use a word that gets incorrectly interpreted or translated, you can enter manual translations for any words for any languages | |
| - In Manual_Translations.csv, put the original text in the first column, your translation in the second, and the 2-letter language code for that entry into the 3rd column | |
| • url_list.txt | |
| - If you have any URLs in the original text, you can put them as a list in this file | |
| - This makes it so the URL will not be translated, and also improves the pronunciation in the TTS stage | |
| - It will really only work on basic URLs, such as "example.com/test". If it has anything other than slashes, periods, and colons, it won't work | |
| - See the notes at the top of the url_list.txt file for more details | |
| • Phoneme_Pronunciation.csv | |
| - Allows you to specify exact phonetic pronunciation of words or phrases | |
| - Note: This is different from 'aliases'. Using this requires using special phonetic alphabets (see links below) | |
| - See: https://learn.microsoft.com/en-us/azure/cognitive-services/speech-service/speech-ssml-phonetic-sets | |
| - See: https://learn.microsoft.com/en-us/azure/cognitive-services/speech-service/speech-synthesis-markup-pronunciation#phoneme-element | |