If you choose to remove soft returns, they are converted to spaces or ignored where an adjacent character is a space. If you do not remove them, they are converted to the <\n> tags.

The placement of discretionary hyphens is dependent on the structure and pronunciation of words in the source language, so usually you should choose to remove rather than extract them.

If this option is selected, the character entities that have a Unicode character are replaced by that character, and other character entities are converted to standalone tags.

<\[> thin space
<\{> hair space
<\z> zero width space
<\8> figure space
<\p> punctuation space
<\_> em dash
<\!-> non-breaking hyphen
<\!s> non-breaking space
<\m> em space
<\e> en space
<\j> word joiner
<\5> three-per-em space
<\$> four-per-em space
<\^> six-per-em space

Kerning and tracking information applied to source text is usually too closely aligned to the source language text layout to be useful for the translated text, so usually you should not process these tags.

Note: Even if you choose to skip them, where those tags are combined with other tags, the information is represented in tags in the target document.

If you choose this, hard returns (forced line breaks) are converted to tags that appear inside the segments. If you do not choose this, they are converted to structure tags and so cause a segment break.