Game Localization · Glossary
Translation Quality Assurance — TQA in Game Localization
Translation Quality Assurance (TQA) is the systematic process of reviewing translated content against defined quality standards before delivery. In game localization, TQA combines automated file-level checks — terminology consistency, number accuracy, formatting, placeholder validation — with human linguistic review to verify that translations meet quality specifications before import into the game engine.
TQA vs. LQA — Two Distinct Quality Layers
TQA and LQA are complementary but distinct. TQA (Translation Quality Assurance) operates at the string file level — it checks translated strings before they are imported into the engine. Automated TQA tools (Xbench, Verifika, ErrorSpy) run regex-based checks for missing translations, number format mismatches, terminology violations, and tag errors. Human TQA reviewers check for fluency, consistency, and terminology adherence. LQA (Localization Quality Assurance) operates inside the working game build — it catches layout breaks from text expansion, font rendering failures for CJK characters, strings that appear grammatically correct in isolation but are contextually wrong in the game UI, and functionality issues in RTL builds. Both layers are required for a shipped-quality localization. File-level TQA catches what LQA would find slowly; in-engine LQA catches what no file-level check can detect.
TQA Error Categories
Game localization TQA uses standardized error taxonomies to categorize and severity-weight issues. Common error categories: Accuracy errors — mistranslation, omission, addition; Fluency errors — grammatically incorrect or unnatural text; Terminology errors — deviation from the approved game glossary (wrong character name, incorrect item name); Formatting errors — missing or extra tags, incorrect placeholders, truncation, encoding issues; Style errors — tone inconsistency, incorrect register. DQF-MQM (Multidimensional Quality Metrics) is the current industry-standard TQA framework, providing a consistent rubric for scoring translation quality across projects.
Automated TQA Tools
Automated TQA tools run deterministic checks on translated string files that would take hours to find manually. Xbench is the most widely used game localization QA tool — it checks terminology consistency against a glossary, runs regex-based quality rules, and compares source and target segments for number and tag accuracy. Verifika provides similar checks with a different rule set and is common in Eastern European localization workflows. ErrorSpy covers formatting and tag validation. These tools catch mechanical errors; they cannot assess translation quality, cultural appropriateness, or contextual accuracy — that requires human review.
TQA in the Game Localization Workflow
In a professional game localization workflow, TQA occurs in three stages. Pre-translation TQA runs automated consistency checks on source strings before translation begins — catching inconsistencies that would propagate into all target languages. Post-translation TQA runs automated checks on all translated files and flags issues for human review. Pre-delivery TQA confirms that all flagged issues have been resolved before files are returned to the client. These stages together ensure that LocQA in the game engine focuses on context-dependent issues — rendering, layout, functionality — rather than discovering basic terminology and formatting errors that file-level TQA should have caught earlier.
SandVox and Translation Quality Assurance
SandVox integrates automated TQA into every translation project — Xbench runs against a per-project terminology glossary before delivery, catching formatting errors, placeholder mismatches, and terminology deviations at the file level. In-engine LocQA runs in your actual game build after import, catching rendering and layout issues that no file-level tool can detect. Both layers are included — not sold separately.
Related terms: Lqa Testing · Localization Qa · Cat Tool · Translation Memory · Localization Style Guide
Frequently Asked Questions
What is the difference between TQA and proofreading?
Proofreading is a general linguistic review — checking grammar, spelling, and readability. TQA is a structured evaluation against defined quality criteria, using an error taxonomy, severity ratings, and pass/fail thresholds. TQA produces a measurable quality score; proofreading produces a revised document. For game localization, TQA with an error taxonomy is the professional standard — proofreading alone does not provide the structured feedback needed to manage quality across large, multi-language projects.
Can automated TQA tools replace human review?
No. Automated TQA tools catch mechanical errors — missing translations, tag mismatches, number format issues, terminology violations — with high reliability. They cannot assess meaning, tone, cultural appropriateness, or contextual accuracy. A string that passes all automated checks can still be a poor translation if the meaning is subtly wrong or the tone is inconsistent with the game’s style. Automated TQA and human review are complementary layers, not substitutes.
How does TQA work for CJK languages?
CJK (Chinese, Japanese, Korean) TQA adds language-specific checks beyond the standard set. For Japanese: honorific consistency (formal/informal register), rendering of katakana loan words, half-width vs. full-width character usage. For Chinese: Simplified vs. Traditional character set consistency, character encoding (must be UTF-8 throughout). For Korean: spacing rules (Korean applies different spacing rules than European languages), hangul rendering with font coverage verification. These checks are in addition to the standard terminology, formatting, and accuracy checks applied to all languages.
Need Expert Game Localization?
SandVox provides end-to-end game localization including translation quality assurance — for narrative games, mobile titles, webtoons, and interactive fiction.