Translated Japanese, Chinese, and Korean (CJK) text breaks the layout for two reasons: font substitution and line-breaking rules. When a translation tool outputs CJK characters that the document's original font does not contain, the system silently swaps in a fallback font that has different widths and spacing — so text shifts, overflows, or renders as empty boxes. On top of that, CJK scripts follow strict line-breaking conventions that Latin-based tools ignore, producing awkward wraps, punctuation stranded at the wrong edge of a line, and broken vertical text. A document-first translator applies the right CJK fonts and line-break rules so the output reads correctly. Bluente preserves CJK layout across 120+ languages.
Bluente is an AI-powered document translation platform used by 30,000+ professionals to translate files in 120+ languages while preserving original formatting. This article explains why CJK layouts break during translation and how to keep them clean.
What Is Font Substitution and Why Does It Break Translated CJK Text?
Font substitution is what happens when a document is set in a font that has no CJK characters, and translated Chinese, Japanese, or Korean text is poured into it anyway. The rendering system cannot show characters the font does not contain, so it either substitutes a fallback font — often a generic system font with different proportions — or displays the infamous empty rectangles ("tofu") where glyphs should be. Either way, the carefully spaced source layout falls apart.
The damage is subtle but real. A fallback font typically has different character widths, stroke weights, and metrics, so lines that fit before now overflow their boxes, columns misalign, and table cells expand. The document is still readable in a literal sense, but it no longer looks like the original — which is the entire point professionals hire a format-preserving translator to avoid. This is the "character set broke" problem teams describe when a translated file comes back looking nothing like the source.
Why Does Line Breaking Go Wrong in Japanese and Chinese?
Because CJK scripts do not put spaces between words and follow rules that Latin-centric tools were never built to respect. In Chinese and Japanese, text can wrap almost anywhere between characters, but a set of conventions called kinsoku shori governs where it may not: opening brackets and quotation marks must not end a line, and closing brackets, periods, and small kana must not begin one. A tool that applies English word-wrapping logic ignores these rules and strands punctuation in the wrong place, which reads as clumsy or unprofessional to a native reader.
Mixed scripts make it worse. A long Latin word — a product name, a URL, an untranslated technical term — dropped into a Japanese paragraph cannot be hyphenated the way the surrounding CJK can be broken, so it can force a large gap or push a line out of alignment. Teams working in Japanese often want a whole phrase kept together on one line for readability, and a tool that does not understand this will split it in a way that changes how the text is meant to be read.
Does This Affect Vertical Text and Ruby Annotations Too?
Yes, and these are where naive tools fail most visibly. Japanese and Traditional Chinese documents are sometimes set vertically, top-to-bottom and right-to-left, and Japanese frequently uses ruby (furigana) — small pronunciation glosses set above or beside characters. A translation tool that flattens everything to horizontal Latin-style text will destroy vertical orientation entirely and drop or misplace ruby annotations. Rebuilding either by hand is slow, specialist work.
A document-first platform recognizes these features as structure rather than decoration, preserving reading direction and annotation placement so the translated file behaves the way a CJK reader expects — not like a Western document with CJK characters pasted in.
Which Documents Are Most Affected?
Any high-stakes document translated into or out of CJK where appearance carries weight. Common cases include financial statements and annual reports translated between Japanese, Korean, Chinese, and English; regulatory and pharmaceutical filings where a submission approved in one language must match exactly in another; contracts and corporate governance documents where numbering and clause layout must survive; training decks and manuals for regional teams; and marketing collateral where typographic quality is the brand. In each, a substituted font or a badly wrapped line is not a minor blemish — it is the difference between a file that can be used as-is and one that needs a designer to rebuild it.
How Does a Format-Preserving Translator Keep CJK Layouts Intact?
By handling fonts and line-breaking as part of the translation, not as an afterthought. A document-first platform outputs CJK text in fonts that actually contain the required glyphs, applies the correct line-breaking rules for each script so punctuation and phrasing land where a native reader expects, and manages the text expansion or contraction that comes with switching scripts so boxes, tables, and columns stay aligned.
Bluente is built to do this. It preserves the document's structure — fonts, text boxes, tables, reading direction — while translating, so Japanese, Chinese, and Korean output renders cleanly instead of collapsing into substituted fonts and broken lines. The result comes back looking like the original, across PDF, DOCX, PPTX, and XLSX.
What Should I Check After Translating a CJK Document?
Review four things next to the source. First, no empty boxes or obviously mismatched fonts anywhere in the text. Second, lines wrap sensibly — no opening bracket stranded at a line end, no lone closing punctuation at a line start. Third, tables, columns, and text boxes still align and nothing overflows its container. Fourth, if the source used vertical text or ruby annotations, those are preserved. A quick side-by-side pass catches the issues that would otherwise reach a reader.
How Much Rework Does This Save?
The savings are large because CJK reformatting is specialist work. When a translated file comes back with substituted fonts and broken wrapping, fixing it means a designer who understands CJK typography reflowing text, reapplying fonts, and re-checking every line — often hours per document, and a bottleneck when a team is rolling out across Asian markets. Because Bluente returns CJK files with fonts, line-breaking, and structure already intact, a document that used to sit in a design queue comes back ready to use in minutes, typically in under two minutes per file.
Frequently Asked Questions
Q: Why does my translated Japanese text show empty boxes or squares? Those boxes ("tofu") appear when the document's font has no glyphs for the CJK characters, so the system cannot render them. A document-first translator outputs the text in a font that contains the required characters, avoiding the problem.
Q: Why does translated Chinese text overflow its table cells or boxes? Usually because a fallback font with different character widths was substituted, or because text expansion was not accounted for. A format-preserving platform manages fonts and reflow so cells and boxes stay aligned.
Q: Do translation tools respect Japanese line-breaking rules? Most Latin-centric tools do not — they strand punctuation at the wrong edge of a line. Bluente applies the correct CJK line-breaking conventions so text wraps the way a native reader expects.
Q: Can Bluente handle vertical Japanese text and furigana? Yes. Bluente treats reading direction and ruby annotations as structure and preserves them, rather than flattening everything into horizontal text.
Q: Which CJK languages does Bluente support? Japanese, Simplified and Traditional Chinese, and Korean are all supported, alongside 120+ languages in total, with layout and font handling appropriate to each script.
Q: Does this work for scanned CJK documents? Yes. Bluente applies OCR to rebuild the text layer of scanned files, then translates while preserving structure, so scanned CJK documents keep a clean layout instead of becoming flat, broken text.
Start translating documents for free. Bluente preserves your formatting across 120+ languages in under 2 minutes. Try BluTranslate free — no credit card required.