Real-Time Language Translation Device for Video Calls

The patent introduces a server as an intermediary component that handles the translation processing between users of different languages. The server receives voice recognition results from the first terminal, performs translation to the second language, and returns the translated text. This mediator approach allows the translation functionality to be added without significantly increasing the complexity of the terminal devices themselves, as the complex processing is offloaded to the server.

The translation system is segmented into distinct functional modules: voice recognition module (converting speech to text in the first language), translation module (translating from first language to second language), and text output module (displaying translated text). This segmentation allows each component to be optimized independently and facilitates parallel processing, reducing overall system complexity while maintaining real-time translation capability.

The system performs preliminary actions by continuously receiving and buffering voice recognition results before final translation is needed. The terminal continuously obtains voice recognition text in the first language and prepares it for translation, so that when translation is required, the processing can be completed more quickly. This preliminary preparation reduces the perceived processing time during actual communication.

The translation process operates continuously rather than intermittently. The terminal continuously receives voice recognition results, continuously translates them to the second language, and continuously displays the translated text. This continuous operation eliminates gaps in translation processing, maintaining communication fluency without requiring users to wait for batch processing to complete.

The patent replaces manual translation operations with an automated electronic translation system. Instead of requiring users to manually switch languages or wait for delayed translation, the system automatically substitutes the translated text in real-time through electronic processing. This substitution of mechanical/manual operations with automated electronic systems dramatically improves response time and user experience reactivity.

The system creates a copy of the voice recognition result text and translates this copy to the second language, then displays both the original and translated versions. This copying approach allows the translation process to run in parallel with the original voice recognition process, without delaying the display of either version. The translated copy is generated and substituted immediately, providing reactive user experience while maintaining system efficiency.