How SYSTRAN helped MGI to speed-up Game Localization with tailored Machine Translation
![](https://www-staging.systransoft.com/wp-content/uploads/2023/10/SYSTRAN-Case-Study-MGI.png)
![](https://www-staging.systransoft.com/wp-content/uploads/2023/10/logo-mgi-group.png)
주요 수치
100M
registered gamers
+10
Top Massive Multiplayer Online Games
15%
Decrease in expsenses
2000
The year MGI was founded
Case study
MGI group 당면 과제
As a publisher of free-to-play Massive Multiplayer Online games, MGI Group was facing regular content updates that result in a constant high work load for localization teams. Some games receive content updates that equals the word count of the “Lord of the Rings” trilogy over the course of one year (more than 1000 pages).
The difficulty MGI Group was facing with human translations was time constraints. Games must be translated from Chinese into English and then into French, German and Spanish, in a very short time frame, making on-time delivery a real challenge.
![](https://www-staging.systransoft.com/wp-content/uploads/2023/11/Systran-illustration-Apprentissage-multilinge-education-768x800.png)
solution
시스트란 번역 서버
SYSTRAN has used MGI Group’s well structured data stored in their CAT tool (MemoQ) to train a specialized translation engine dedicated to their own game localization needs.
MGI Group uses SYSTRAN translate Server with a tailored MT engine with a seamless integration in their CAT Tool, thanks to SYSTRAN connector with advanced features, including Translation Memories, Tag Management, User Dictionaries, Glossaries, etc…
![](https://www-staging.systransoft.com/wp-content/uploads/2023/11/Systran-illustration-Translate-server-768x768.png)
Benefit
Improving the quality of Machine Translation for in-game content
Video games rely on nuanced dialogue for immersion. SYSTRAN’s Neural Machine Translation (NMT) technology, combined with MGI Group’s structured translation data, enabled the creation of a customized Machine Translation system for swift in-game dialogue localization.
The engine, specialized for in-game content, utilized MGI Group’s Translation Memories & Glossaries to train machine learning algorithms. SYSTRAN NMT engines recognized and enhanced text tags while preserving their integrity for a seamless localized gaming experience.
Ongoing translation quality was ensured through SYSTRAN’s User dictionary for real-time adaptation, with planned periodic engine re-training cycles for long-term performance enhancement.
![](https://www-staging.systransoft.com/wp-content/uploads/2023/11/Systran-illustration-Association-Machine-translation-768x800.png)
정보 MGI group
Media and Games Invest SE is a digital integrated games and media company with its main operational presence in Europe and North America. The company combines organic growth with value-generating synergetic acquisitions, demonstrating continuous strong, profitable growth with a revenue CAGR of 45% over the last 6 years. Next to strong organic growth, the MGI Group has successfully acquired more than 30 companies and assets in the past 6 years. The acquired assets and companies are integrated and cloud technology is actively used to achieve efficiency gains and competitive advantages.
Using Machine Translation for gaming
In an attempt to optimize cost/performance ratio, MGI Group aimed to increase efficiency by adding Machine Translation (MT) to their localization workflow. First, generic MT was used to translate from Chinese to English and the results of ten lead to a higher workload instead of a reduction.
One of the major challenges with MT were the machine-readable instructions in the text that need to stay exactly as they are in structure and placement, and should not be translated.
However, with MT, some instructions like e.g. [CURRENCY] have been translated. Additionally, the machine-readable instructions had been placed randomly and needed manual adaptation.
![](https://www-staging.systransoft.com/wp-content/uploads/2023/11/images.jpg)
Henrik Schröder
Head of Localization & Product Support gamigo MGI Group
SYSTRAN Machine Translation enabled us to gain more independence in our localization process, with a successful shift from human translation to post-edition. We’re now achieving delivery twice as fast, with a 15% decrease in expenses.