NAVER CLOUD PLATFORM

Customer Success Stories

These customer case studies introduce successful companies that provide safe and efficient services via the NAVER CLOUD PLATFORM.


Although we’d tried a few speech-recognition APIs, there was no service like Clova Speech Recognition, which is user-friendly and uses only simple code.
The greatest advantage of the NAVER CLOUD PLATFORM's Clova Speech Recognition feature, apart from its speed, was being able to choose from five levels of speech analysis. Another benefit was being able to use different Naver artificial intelligence features in addition to speech recognition. We are planning to "cont"inue upgrading our app using Clova Face Recognition, also part of the NAVER CLOUD PLATFORM.
Interview
Ji Woo Shin, Ji Eun Lee, Ji Soo Lim
3G Development Team
Please introduce your company.
The PPD ('pronunciation practice device') app helps hearing-impaired adolescents practice pronunciation.
We are classmates at Duksung Women's University and our club, Neulpurunsori, performs volunteer service in the area of IT education. Two years ago, during a volunteer visit to Chungeum Hall, a charity organization that helps hearing-impaired people, we got to know some people with cochlear implants who needed steady practice to correct their pronunciation. That’s when the idea of developing a pronunciation practice app first occurred to us. We learned that there were very few speech therapy institutions in the country and even those few were concentrated in metropolitan areas. Knowing how difficult it was for our friends to get the services they needed, we decided to create an app to help them practice their pronunciation conveniently without leaving home. Fortunately, we were just in time to participate in a government ICT human resources development project called Hanium ICT Mentoring. It was a good opportunity to organize a team called 3G and work with a mentor to develop this app. Hearing-impaired people can use it for independent practice, listening and checking their accuracy on their own, or they can use the 'homework' function and get feedback from a trained speech therapist.
How and when did you start using the NAVER CLOUD PLATFORM?
Since September 2017 we've been using Clova Speech Recognition, which is part of the NAVER CLOUD PLATFORM.
Since our app requires an STT function to convert speech to text, we'd been looking around for open-source STT software. Although we’d used a foreign cloud company’s STT service for a while, it hadn't worked properly: the conversion speed gradually got slower. While looking for other options, we discovered Clova Speech Recognition and ended up using it. We didn’t encounter the same problem, and the best thing was that we could choose from five levels of output quality. Other companies’ STT services didn’t meet our needs because they offered only one level of quality: the auto-complete function automatically corrected for mispronunciation, rendering the service rather unhelpful for our purposes. Clova, on the other hand, was much more satisfactory because we could program it to reflect users’ pronunciation realistically.
What was your reason for choosing the NAVER CLOUD PLATFORM?
We selected this product because it offered a high recognition rate for the Korean language, as well as for its ease of use and reasonable price.
Clova has an outstanding rate of success in recognizing the Korean language. We needed our app to reflect users’ pronunciation as precisely as possible, and Clova Speech Recognition proved to be a good choice in that regard. Perhaps that’s because it was developed by a Korean company, Naver. Furthermore, it is easy to use. Even beginners can use it, since the sample source is full of helpful explanations that anyone can understand. Error codes are also explained in detail on the website, so we could find out why a given error was occurring at each stage of the development process. Another big advantage, especially for students like us, was the price. Not only is Clova more reasonable than other services of its kind, but so are most of the NAVER CLOUD PLATFORM’s services. Affordability played a part in our decision.
How are you using the NAVER CLOUD PLATFORM, and what improvements have you experienced since you began using it?
The conversion speed is much faster, the codes are noticeably shorter, and the accuracy of our speech-recognition function has improved as well.
When we were using another company’s speech-recognition API, the slow speed was always an obstacle. Using the service became increasingly difficult because converting voice to text took longer as the sentences got longer. All those inconveniences disappeared after we switched to Clova Speech Recognition. The codes also became noticeably shorter, making maintenance much easier. Plus it was great to be able to look at the converted text and select the text we wanted. Those improvements allowed us to objectively reproduce the sound of a user’s voice and correct pronunciation in a much more helpful way.
Lastly, please give us an opinion that would be helpful to people who are still deciding on whether to use the NAVER CLOUD PLATFORM.
Although we’d tried a few speech-recognition APIs, there was no service like Clova, which is user-friendly and uses only simple code.
The greatest advantage of the NAVER CLOUD PLATFORM's Clova Speech Recognition feature, apart from its speed, was being able to choose from five levels of speech analysis. Another benefit was being able to use different Naver artificial intelligence features in addition to speech recognition. We are planning to continue upgrading our app using Clova Face Recognition, also part of the NAVER CLOUD PLATFORM. That's because we think we'll be able to correct pronunciation more effectively if we can also analyze the way users move their mouths. In the future, we're planning to use NAVER CLOUD PLATFORM technology to help more of our friends get the speech therapy they need. Way to go, PPD! And way to go, NAVER CLOUD PLATFORM!
Please introduce your company.
The PPD ('pronunciation practice device') app helps hearing-impaired adolescents practice pronunciation.
We are classmates at Duksung Women's University and our club, Neulpurunsori, performs volunteer service in the area of IT education. Two years ago, during a volunteer visit to Chungeum Hall, a charity organization that helps hearing-impaired people, we got to know some people with cochlear implants who needed steady practice to correct their pronunciation. That’s when the idea of developing a pronunciation practice app first occurred to us. We learned that there were very few speech therapy institutions in the country and even those few were concentrated in metropolitan areas. Knowing how difficult it was for our friends to get the services they needed, we decided to create an app to help them practice their pronunciation conveniently without leaving home. Fortunately, we were just in time to participate in a government ICT human resources development project called Hanium ICT Mentoring. It was a good opportunity to organize a team called 3G and work with a mentor to develop this app. Hearing-impaired people can use it for independent practice, listening and checking their accuracy on their own, or they can use the 'homework' function and get feedback from a trained speech therapist.
How and when did you start using the NAVER CLOUD PLATFORM?
Since September 2017 we've been using Clova Speech Recognition, which is part of the NAVER CLOUD PLATFORM.
Since our app requires an STT function to convert speech to text, we'd been looking around for open-source STT software. Although we’d used a foreign cloud company’s STT service for a while, it hadn't worked properly: the conversion speed gradually got slower. While looking for other options, we discovered Clova Speech Recognition and ended up using it. We didn’t encounter the same problem, and the best thing was that we could choose from five levels of output quality. Other companies’ STT services didn’t meet our needs because they offered only one level of quality: the auto-complete function automatically corrected for mispronunciation, rendering the service rather unhelpful for our purposes. Clova, on the other hand, was much more satisfactory because we could program it to reflect users’ pronunciation realistically.
What was your reason for choosing the NAVER CLOUD PLATFORM?
We selected this product because it offered a high recognition rate for the Korean language, as well as for its ease of use and reasonable price.
Clova has an outstanding rate of success in recognizing the Korean language. We needed our app to reflect users’ pronunciation as precisely as possible, and Clova Speech Recognition proved to be a good choice in that regard. Perhaps that’s because it was developed by a Korean company, Naver. Furthermore, it is easy to use. Even beginners can use it, since the sample source is full of helpful explanations that anyone can understand. Error codes are also explained in detail on the website, so we could find out why a given error was occurring at each stage of the development process. Another big advantage, especially for students like us, was the price. Not only is Clova more reasonable than other services of its kind, but so are most of the NAVER CLOUD PLATFORM’s services. Affordability played a part in our decision.
How are you using the NAVER CLOUD PLATFORM, and what improvements have you experienced since you began using it?
The conversion speed is much faster, the codes are noticeably shorter, and the accuracy of our speech-recognition function has improved as well.
When we were using another company’s speech-recognition API, the slow speed was always an obstacle. Using the service became increasingly difficult because converting voice to text took longer as the sentences got longer. All those inconveniences disappeared after we switched to Clova Speech Recognition. The codes also became noticeably shorter, making maintenance much easier. Plus it was great to be able to look at the converted text and select the text we wanted. Those improvements allowed us to objectively reproduce the sound of a user’s voice and correct pronunciation in a much more helpful way.
Lastly, please give us an opinion that would be helpful to people who are still deciding on whether to use the NAVER CLOUD PLATFORM.
Although we’d tried a few speech-recognition APIs, there was no service like Clova, which is user-friendly and uses only simple code.
The greatest advantage of the NAVER CLOUD PLATFORM's Clova Speech Recognition feature, apart from its speed, was being able to choose from five levels of speech analysis. Another benefit was being able to use different Naver artificial intelligence features in addition to speech recognition. We are planning to continue upgrading our app using Clova Face Recognition, also part of the NAVER CLOUD PLATFORM. That's because we think we'll be able to correct pronunciation more effectively if we can also analyze the way users move their mouths. In the future, we're planning to use NAVER CLOUD PLATFORM technology to help more of our friends get the speech therapy they need. Way to go, PPD! And way to go, NAVER CLOUD PLATFORM!

  • Background

    • Speech recognition feature to convert user's pronunciation into text is required
    • Issues of slow speed and text conversion result exist in case of using speech recognition API of other companies
  • Use

    • Converts user's voice into text quickly and accurately using Clova Speech Recognition (CSR)
    • Able to select by choosing the appropriate result among five types of text conversion result values
  • Results

    • Able to recognize user's pronunciation accurately with high recognition rate of Korean language
    • Even easier maintenance with simplified codes
    • Much faster conversion speed