How to use the full-width and half-width converter
Full-width characters come from Chinese, Japanese and Korean input methods. They are ordinary letters, digits and punctuation drawn wide enough to line up with ideographs in a grid. To a computer they are completely different characters from their normal counterparts, so a spreadsheet of orders or contacts can hold the same value twice, fail a search, or refuse to deduplicate, purely because one row was typed while the IME was in full-width mode.
The conversion follows the mapping Unicode defines. Full-width letters, digits and symbols sit in one continuous block from U+FF01 to U+FF5E, and subtracting 0xFEE0 from any of them yields the matching ASCII character. Adding the same amount to an ASCII character produces the full-width form. The ideographic space at U+3000 lives outside that block, so it is paired with the normal space separately.
Half-width katakana does not follow this rule. Its voiced and semi-voiced marks are stored as separate characters, so a simple offset cannot map them correctly. This tool leaves half-width katakana untouched and reports how many such characters your input contained. Hangul, kana in full-width form and ideographs are outside the mapping as well and are passed through unchanged.
Frequently Asked Questions
Its voiced marks are stored as separate characters, so a simple offset cannot convert it correctly. Rather than corrupt the text, the tool passes it through and reports how many characters it found.
No. Only full-width letters, digits, punctuation and the ideographic space are converted. Everything else is passed through unchanged.
One of them was almost certainly typed in full-width mode. Convert both to half-width and compare again and the mismatch usually disappears.