Spaces:
Sleeping
Sleeping
File size: 3,506 Bytes
2f2406a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 |
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 |