Unicode assigns a number to each character. 国 and 國 are different code points. A font is a drawing of those code points. If the file has the right character but the font has no glyph, you see a box. If the file has the wrong bytes, you see mojibake.
This is why conversion and fonts are related but not the same job. OpenCC can change 国 into 國. It cannot install 微软雅黑, Songti TC, or Noto Serif CJK on someone else's computer. People send a converted file, the colleague sees tofu squares, and both sides blame the converter. The squares are usually a font coverage problem.
Code points are not pictures
A Word file stores characters as numbers. The font file stores drawings keyed to those numbers. When the drawing is missing, the operating system may:
- show a box
- fall back to another font for that one character
- show a different region's glyph for a code point that has localized variants
Fallback can make a sentence look like a ransom note: one character in Song, the next in a sans-serif system font. That is ugly and still readable. A hard box is worse, because you cannot tell whether the conversion happened.
Some code points have region-specific glyph variants. The number is the same; the preferred drawing differs for Taiwan, Hong Kong or mainland sources. A font designed for one region may draw 骨 or 起 slightly differently. Readers notice. This is not OpenCC changing the wrong character. It is the font's locale tables. See Taiwan vs Hong Kong Traditional Chinese for character-level differences that are different code points.
One writing system, many fonts
- Simplified-focused fonts may lack uncommon Traditional forms.
- Traditional-focused fonts may lack some Simplified forms.
- A large CJK font, or a pair of SC/TC fonts, is the usual fix.
Operating systems already ship fallback fonts such as Microsoft YaHei, PingFang, Source Han / Noto CJK, and Songti. This site uses those system fonts in CSS rather than bundling font files. Your document should do something similar: pick a font family that actually contains the target standard.
When you embed fonts in PowerPoint or PDF, you may embed only the glyphs you used. After conversion, new glyphs are required. A file that was "self-contained" in Simplified can stop being self-contained in Traditional. Re-embed a full CJK font, or tell recipients which font to install.
Documents after conversion
When you convert a Word, Excel or PowerPoint file, the characters in the file change. Recipients still need a font that can draw them. If a colleague opens your file and sees boxes, send them this page and the garbled Chinese checklist before you convert again.
A practical review machine has:
- a Simplified-capable font
- a Traditional-capable font
- the same Office version you will use to present
If you only review on the laptop that created the Simplified original, you may not see the Traditional fallback failures your client will see.
Unicode is why converters can exist
Because 门 and 門 are different numbers, a dictionary can replace one with the other without caring how a brush would write them. Simplified vs Traditional Chinese is a social standard. Unicode is the technical one. OpenCC sits between them.
Older encodings — Big5, GB2312, GBK — stored Chinese in smaller maps. Files from the 1990s still appear in attachments. If those bytes are read as UTF-8, you get garbage, not boxes. Garbage is encoding. Boxes are fonts. Do not run OpenCC on garbage.
Frequently asked questions
Is there one font that covers all Chinese? Large CJK families such as Noto CJK / Source Han cover a great deal. "All of Unicode" is still a lot. For office work, a current SC font plus a current TC font is enough.
Should I convert to images so fonts do not matter? That defeats editing, searching and later conversion. Fix fonts instead.
Why does my browser look fine and Word look wrong? The browser and Word do not share the same fallback list. Test in the application the recipient will use.
Does this site embed Chinese fonts in downloads? The converted office files keep the font information they already had, plus the new characters. They do not install a font on your computer.
Are 国 and 國 the same character in Unicode? No. They are a Simplified/Traditional pair with different code points. That is what makes conversion a real operation.