Why you should opt
Konan Listener
From real-time subtitle creation to transmission in
a one-stop solution!
If you need to create learning content through subtitle data,
Konan Listener will become your solution.
Konan Listener is a solution that automatically creates subtitles for various content,
including university and public education videos,
utilizing high-quality speaker adaptive end-to-end speech recognition engine,
video processing, and natural language processing (NLP) technologies.
BENEFITS
Benefit 01.
Reduced Time and Costs for Subtitle Production
Compared to obtaining subtitles by making requests to stenographers, using the solution cuts the subtitle service provision time and subtitle production costs in half.
Benefit 02.
Enhanced Work Efficiency
The automation of subtitle registration brings efficiency to your work.
The entire process of subtitle creation, management, and transmission is also automated for systematic subtitle file management.
Benefit 03.
Increased Learning Subtitle Management Convenience
You can expect to increase convenience in learning data management by searching content quickly through use of metadata and integration with relevant systems.
FEATURE
By applying the Conformer Neural Network, the latest encoder technique in the end-to-end speech recognition field, Konan Listener improved its recognition error rate by around 17% compared to its key competitors’ technologies that utilize Transformer Neural Network-based encoders.
Through the time code-based editing environment, which enables the semi-automatic modification of recognition results and spelling corrector, the solution can also perfectly correct even the most trivial recognition errors and immediately reflect it in the client’s service.
Post-Correction Management
- Independent operations of video replay and subtitle editing allow simultaneous video playing and subtitle editing
- Edited subtitles are reviewed in real-time on the video replay screen at the same time as the subtitle editing
- Provides various editing shortcut keys for rapid editing
- Subtitle display area adjusted in detail based on the time code (video frame unit)
- Rapidly configures subtitle display and division sectors based on voice wavelengths (silent sections)
- Provides various convenience features, including section repeat, current section replay, speed up and slowdown, and subtitle size adjustments.
Spell Check
- Includes paragraph separation, punctuation mark indication, and spelling correction features
- Spelling, errors, and spacing features
- Expression correction feature that overlays active voice over passive voice and deletion features
End-to-End Speech Recognition Model
- Improved recognition accuracy for specific speakers through speaker-adaptive learning
- Meta-information checking and management features for language and voice models
-
Utilizing the state-of-the-art deep learning-based speech recognition engine, Konan Listener provides a high recognition accuracy even in real-time environments.
-
Konan Listener provides spelling correction features based on standard language, tailored to the pronunciation, voicing, and dialect of the speaker.
-
Providing various time code-based subtitle editing features, Konan Listener provides efficient subtitle editing functions.
-
Natural language processing features and management functions for synonyms, stop words, and dictionaries allow you to enhance your subtitle accuracy.
-
The solution’s speaker-adaptive end-to-end speech recognition model allows you to enhance the recognition accuracy of a specific speaker.
-
The solution provides the standard RESTful API that you can connect with education platforms, such as LCMS and remote video lectures.
TECHNOLOGY
USE CASES
To keep up with the times of breaking out of traditional campus-based learning to pursuing a contactless learning system, Konan Listener utilizes the converged next-generation education platform, Edu-Tech, to meet the needs of the shifting learning paradigm and changes in global environment, leading the establishment of an online learning environment and qualitative growth of the educational environment.
The solution provides the “real-time subtitle service” for hearing-impaired students’ video classes and the “spell check and subtitle editor” features that allow ordinary individuals to conduct subtitle editing tasks faster than that of stenographers.
CASES01. Development of AI Technology-Based Speech Recognition Subtitle Support Program
- Plans to establish equal learning environments for students with special disabilities
- Automatic subtitle support service based on the AI speech recognition engine integrated with remote video lectures and the learning support platform
- Enhanced access through intuitive UI and UX development, tailored to various learning environments
- Provides subtitle editing and management features as well as monitoring and statistical functions
CASES02. Establishment and Development of University Online Learning and Broadcast Support System
Established ONU’s broadcast speech-recognition subtitle creation system
- Established broadcast and video lecture content management system
- Established an environment that provides automated subtitle information by applying speech recognition technology integrated with existing CMS
- Enhanced subtitle quality through learning each speaker’s speech, language model, and post-correction
- Utilized as course material for students with hearing disabilities and foreign students, as well as for general students
Development of Speech Recognition System With Automatic Subtitle Creation Abilities
- Automatic subtitle creation method using the latest speech recognition technology
- Applies the subtitle management system’s replay/upload/download, subtitle information updates, editing, and transmission workflow
- Secures at least 80% accuracy for automatic subtitle extraction from speech recognition. For result accuracy below 80%, configurations and applications are made to achieve at least 80%
- Designed to easily review and edit final results
Provides High-Quality Subtitle Service and User Convenience
- Consistently provides high-quality subtitle services for students with hearing disabilities for undergraduate media lectures
- Enhanced subtitle accuracy through learning model upgrades and post-processing feature improvements
- Provides user convenience features through web and app function improvements for video subtitle readability improvements
- Maximized video subtitle readability and quality completion level through improvements in subtitle player features
