It requires a whole new set of data: hundreds of hours of recorded audio and their associated transcriptions and trainings of our machine learning based engines to become available. Speech recognition is a totally different process than translation. In the process of adding a new language to the automatic translation system, the most important learning material is existing high-quality translations-the same text in two languages.Ī significant amount, typically 1+ million words of this type of text, is needed to build a reasonable-quality machine translation system for a particular language pair. More information on how we count characters for the Translator Text API can be found in our character count documentation.įor speech translation using Microsoft Speech services, see the Speech services pricing page. To estimate your monthly volume, take the total characters to translate, multiply it by the number of languages you want to have it translated into, then take the number and spread it over the maximum number of hours or days you are able to wait for completion.Īs an order of magnitude, this FAQ contains about 6,000 characters a 30-page document has around 17,000 characters the seven Harry Potter books comprise about 60 million characters. A space, tab, markup, and any kind of white space character.

A repeated translation, even if you have translated the same text previously.

Text you pass into the Translate, TranslateArray, TranslateArray2, GetTranslations, GetTranslationsArray, Dictionary Lookup, Dictionary Examples, Transliterate, and TransformText method.

The number of queries, words, bytes, or sentences is irrelevant. Each translation of a text to a new language counts as a separate translation.

Every Unicode code point counts as a character. For Translator text translation, the volume you are billed for is the number of characters in the input.

Https support skype com faq fa10238