Subtitle text filtering
If you'd like to filter out specific instances subtitle text, one way to do so is by using a regular expression (regex). asbplayer can match any sequence following a specified regex pattern and remove the matches.
Configure the regex filter
Under the misc section in asbplayer settings, locate the "Subtitle regex filter" textbox. Enter an appropriate regex to filter desired content. You can replace filtered content similarly by entering a string into the "Subtitle regex filter text replacement" textbox. Leaving this blank will simply remove the content.
Remove speaker names
Remove names enclosed by parenthesis to indicate speakers (e.g. "（山田） 元気ですか？")
Use this filter
([\(（]([^\(\)（）]|(([\(（][^\(\)（）]+[\)）])))+[\)）])
Preserve Japanese readings
Remove parenthesized names or descriptions while keeping Japanese readings immediately after Japanese text. For example, （山田）漢字(かんじ)を読む。 becomes 漢字(かんじ)を読む。. Both (かんじ) and （かんじ） are preserved when immediately preceded by a character in the kanji, hiragana, or katakana ranges shown in the regex.
This is a heuristic: it also keeps descriptions immediately after Japanese text, such as こんにちは（笑）, and removes readings separated from their base text by a space. It does not handle nested parentheses. Leave "Subtitle regex filter text replacement" blank.
Use this filter
(?<![一-龯ぁ-ゟ゠-ヿ])[\(（][^\)）]*[\)）]
Merge subtitle lines
Some subtitles are split in several lines and this regex forces them into a single line. For this filter to work, you must also put $1 $2 in the "Subtitle regex filter text replacement" field.
When combining this pattern with others, keep it last so a more specific alternative can match first at the same position. Alternatives do not run as separate filtering passes. Every match uses the same replacement text, and adding capturing groups in other alternatives changes the numbers needed in $1 $2.
Use this filter
(.*)\n+(?!-)(.*)
$1 $2
Remove newlines
Remove all newline characters without inserting spaces. For example, two lines containing こんにちは and 世界 become こんにちは世界. Leave "Subtitle regex filter text replacement" blank. To insert a space between lines instead, use the merge subtitle lines example.
Use this filter
\n
Remove bracketed descriptions
Remove indications enclosed by square brackets that sound or music that is playing (e.g. "[PLAYFUL MUSIC]" or "-[GASPS]")
Use this filter
-?\[.*\]
Remove uppercase descriptions
As an alternative to the above, filter out descriptions written in capital letters, but without the square brackets (e.g. "PLAYFUL MUSIC"). If your language has additional letters with diacritics, you feel free to add them to this list.
Use this filter
^[\-\(\)\.\s\p{Lu}]+$
Remove music symbols
Any combination of symbols on their own that represent playing music (e.g. ♪♬♪)
Use this filter
[♪♬#～〜]+
Remove sound effect cues
Remove subtitle cues consisting entirely of a bracketed description or music symbols, such as [THUNDER RUMBLING], -[GASPS], or ♪♬♪. The ^ and $ anchors require the entire subtitle text to match, so dialogue such as Hello [GASPS] or ♪ Hello ♪ is kept. Leave "Subtitle regex filter text replacement" blank.
Use this filter
^-?\[.+\]$|^[♪♬#～〜]+$
Remove left-to-right marks
Remove U+200E, the invisible left-to-right mark (LRM). For example, text containing Hello, an LRM, and world becomes Helloworld. Enter the literal escape \u200E in the regex field and leave "Subtitle regex filter text replacement" blank.
Use this filter
\u200E
Remove bidirectional controls
Remove all Unicode bidirectional control characters, including the left-to-right mark (U+200E), right-to-left mark (U+200F), and embedding, override, and isolate controls. Leave "Subtitle regex filter text replacement" blank. These characters can affect the display order of right-to-left or mixed-direction text, so use this filter only when you want to remove them.
Use this filter
\p{Bidi_Control}
Combining regexes
Regular expressions can be combined with the character | (no spaces needed inbetween). E.g., if you want to use the two regexes from the list above, you can use -?\[.*\]|[♪♬#～〜]+. You can combine as many regexes as you wish this way.
Remove parenthesized text and newlines
To remove parenthesized text and all newlines together, use:
Leave "Subtitle regex filter text replacement" blank. For example, （山田）こんにちは followed by a newline and 世界 becomes こんにちは世界. This removes both ASCII and full-width parentheses and their contents, including Japanese readings, and joins lines without adding spaces. It does not handle nested parentheses; use the speaker names example for those.
Use this filter
[\(（][^\)）]*[\)）]|\n
Combine all removal filters
To also remove bidi controls, parenthesized speaker names (including nested parentheses), bracketed descriptions, and music symbols, use:
Leave "Subtitle regex filter text replacement" blank. This combines the removal examples above and joins lines without adding spaces. It removes Japanese readings too; use the reading-preserving example if you want to keep them.
Use this filter
\p{Bidi_Control}|([\(（]([^\(\)（）]|(([\(（][^\(\)（）]+[\)）])))+[\)）])|-?\[.*\]|[♪♬#～〜]+|\n
Invisible characters and blank subtitles
An apparently empty subtitle can contain invisible characters such as an LRM. asbplayer drops text cues containing only whitespace, control characters, or default-ignorable Unicode characters after parsing and filtering. Direction marks within meaningful text are preserved unless your regex explicitly removes them.
Invisible characters at the beginning or end of a cue can prevent an anchored pattern such as ^\[.+\]$ from matching. If the pattern does not match, asbplayer retries it without surrounding whitespace and invisible controls, but only drops the cue if that retry leaves no meaningful text. For example, [THUNDER] followed by an LRM is removed with this pattern and an empty replacement. The retry preserves the original cue when its replacement would contain meaningful text.
When combining an anchored pattern with a bidi-removal alternative, remember that replacements happen in one pass. For example, ^\[.+\]$|\u200E applied to [THUNDER] followed by an LRM removes the LRM, but does not retry the bracketed description afterward. To remove both in the same pass, use the unanchored \[.+\]|\p{Bidi_Control} with an empty replacement. This also removes bracketed descriptions within dialogue.
Learn
Learn how to write and test custom regular expressions at Regex Learn - Playground.