The game company wants us to localize the web-based news and update logs from English to Traditional Chinese on the game domain.
As the title mentioned, this pilot project is divided into two projects, to determine the effectiveness of using a machine translation (MT) engine, and to suggest a TMS for Valorant, web-based news and update logs from English to Traditional Chinese on the game domain
We created a small pilot company called GTZL to help our client to finish this project. Group members are Jihong Zhang, Ying Liu, Jason Tang, and Vivian Gao, myself.
Let’s talk about them one by one.
In the first project, the Riot Games required our company to translate their game–Valorant’s news from English to Chinese (Taiwan). The goal of this pilot project is to determine the effectiveness of using a custom-trained machine translation (MT) engine. We have several post-edited machine translation (PEMT) criteria, if the engine meets the teams, it will be considered viable. For more details, please refer to the Project Proposal below.
Although our ideal is that the customer-trained machine translation engine meets the needs of game localization since the game contains lots of characters, weapons, and place names, our trained engine cannot distinguish them, so some of the translation parts look wired. For more details, please refer to the Lessons Learned below.
Our second part of the project aims to find a “perfect” TMS for Valorant. We listed 7 business requirements and assigned different weightings to each: continuous Localization, TM and TB support, concordance searches, easy-to-use, build-in translation environment tools, cost-efficiency, and build-in invoicing. After about one week’s discussion and testing, We finally decided to recommend MS as the preferred TMS for our client company. For more details, please refer to the TMS Evaluation Project below.
Hope you enjoy our project. If you have any questions, feel free to leave your comments.