Voice search "in Russian"
Hello. Technologies are developing very quickly, and it is very nice that not only abroad.
Yandex presented its development, namely a multi-platform library that gives developers access to their own speech recognition technology - SpeechKit.
More details more ...
This library only works on the Android and iOS platforms. While she is able to "understand" only the Russian language, but there is much to develop!
As for the quality of recognition, it is:
- 84% for general requests made by Internet users;
- 94% for geographic queries.
In its work, Yandex SpeechKit relies on acoustic and language models, for which Yandex needed to record and mark out 300 hours of male, female and children's speech. Language models were formed on the basis of search queries and specially prepared texts.
The voice search query recognition rate is approximately 1 second.
One of the conditions for using the service is as follows: Yandex SpeechKit can be used for free if the number of requests does not exceed 10,000 per day.
Of course, Yandex is not alone in this matter. Other companies have their own technologies. For example, Google’s Android Speech API from Google, which, however, only works under Android, and some others.
Actually
brief description of the technology: api.yandex.ru/speechkit
documentation for Android: api.yandex.ru/speechkit/generated/android/html/index.html
documentation for iOS: api.yandex.ru/speechkit/generated/ios/html/ index.html
You can download the libraries on the Yandex Technologies portal: api.yandex.ru/speechkit/downloads
Yandex presented its development, namely a multi-platform library that gives developers access to their own speech recognition technology - SpeechKit.
More details more ...
This library only works on the Android and iOS platforms. While she is able to "understand" only the Russian language, but there is much to develop!
As for the quality of recognition, it is:
- 84% for general requests made by Internet users;
- 94% for geographic queries.
In its work, Yandex SpeechKit relies on acoustic and language models, for which Yandex needed to record and mark out 300 hours of male, female and children's speech. Language models were formed on the basis of search queries and specially prepared texts.
The voice search query recognition rate is approximately 1 second.
One of the conditions for using the service is as follows: Yandex SpeechKit can be used for free if the number of requests does not exceed 10,000 per day.
Of course, Yandex is not alone in this matter. Other companies have their own technologies. For example, Google’s Android Speech API from Google, which, however, only works under Android, and some others.
Actually
brief description of the technology: api.yandex.ru/speechkit
documentation for Android: api.yandex.ru/speechkit/generated/android/html/index.html
documentation for iOS: api.yandex.ru/speechkit/generated/ios/html/ index.html
You can download the libraries on the Yandex Technologies portal: api.yandex.ru/speechkit/downloads