Split Long Subtitle Lines
Or paste subtitle text
Options
One box, five choices, and each choice has a page of its own. Wrap re-flows every cue to the column count, which can shorten a line that was too long as easily as it can pull a short line back up. Split long lines breaks only the lines that are over the count and leaves the rest exactly as they were. Balance keeps the number of lines each cue already has and moves the words between them until the longest line is as short as it can be. Join puts each cue onto a single line. Limit leaves alone every cue that is already within the line count. This page fills the box in with the choice its title names and lists no other.
Counted in columns rather than characters, so a Chinese or Japanese character takes two of them and an accented Latin letter takes one. 42 is the usual figure for Latin-script subtitles. Lower it and the lines come out shorter but more numerous; nothing is ever cut off.
Which frame rate the timings are counted in. On the format pages this is what MicroDVD needs, because it stores timing as frame numbers. On the frame rate pages it is the rate the file was timed against — the version the subtitles were made for — and the page's second box takes the rate of the video you are going to use them with. 23.976, 24, 25, 29.97, 30, 50, 59.94 and 60 are the common ones. Getting it wrong shifts the whole timeline, and on the frame rate pages it shifts it by a ratio rather than by a fixed amount.
This page does one thing: it breaks the lines that are longer than the number of columns you set. Every line that is already within the count is left exactly as it is, down to the byte, including the lines sitting right beside the ones that get broken. If your file has a few runaway lines in it and the rest of the breaks are fine, this is the page that fixes those few and touches nothing else.
The number is counted in columns. A Latin letter takes one column and a Chinese or Japanese character takes two, so a limit of 42 holds forty-two Latin ones or twenty-one Chinese ones and both come out the same width. The default is 42, the figure Latin-script subtitles have used for decades.
Breaking happens at the boundaries between words and never inside one. A word that is longer than the whole limit, such as a URL or a long German compound, is given a line to itself and left over the limit, and the run says how many lines are still too wide. Cutting a word in half would be editing your text, and the title of this page promises to move line breaks rather than change words.
One thing to be clear about before you start: this page does not merge. A cue whose two lines are 12 columns each stays two lines even though it would fit on one, because pulling them together is re-flowing rather than splitting, and there is a separate page for that. Everything runs in your browser with no upload and no account, and the eleven formats that carry more than one line per cue are accepted: SRT, WebVTT, ASS, SSA, SBV, SAMI, LRC, MicroDVD, TTML, JSON and CSV.
Examples
One runaway line beside a line that is fine
Input
1
00:00:01,000 --> 00:00:04,000
I told you we would miss the last train if we stayed any longer
Short line
Output
1
00:00:01,000 --> 00:00:04,000
I told you we would miss the last train if
we stayed any longer
Short line
The first line is 52 columns against a limit of 42, so it breaks after the last word that fits, which ends it on exactly 42. The second line is 10 columns and comes back untouched, byte for byte, in the same position — it is neither re-measured nor considered for joining with anything. That second half is the whole difference between this page and the one that wraps: wrapping would have re-flowed both lines and might well have put this short one somewhere else.
A word longer than the entire line length
Input
1
00:00:01,000 --> 00:00:04,000
Say it: pneumonoultramicroscopicsilicovolcanoconiosis
Output
1
00:00:01,000 --> 00:00:04,000
Say it:
pneumonoultramicroscopicsilicovolcanoconiosis
The 45-column word cannot be brought under a limit of 42 by any break that keeps it a word, so it is given a line of its own and left over the limit, and the run reports one line still over. The two words before it move up to the first line. Cutting the word in two would produce something nobody can read and would be a change to your dialogue rather than to its layout, which is not what this page was asked to do.
A short line and a Chinese line over the same limit
Input
WEBVTT
00:00:01.000 --> 00:00:04.000
Good.
你好世界今天天气很好我们出去玩吧顺便买点东西回来
Output
WEBVTT
00:00:01.000 --> 00:00:04.000
Good.
你好世界今天天气很好我们出去玩吧顺便买点东
西回来
Two lines in one cue and only one of them is over. `Good.` is 5 columns and does not move; the Chinese line is twenty-four characters, which is 48 columns, and breaks after the twenty-first. Chinese has no spaces, so each character is a break point of its own and the break lands inside the sentence — which is normal for the script and is why the count is measured in screen width rather than in characters. The WEBVTT header and the timing line come back exactly as they went in.
Frequently asked questions
How is breaking long lines different from wrapping subtitles?
Wrapping re-flows every cue from scratch, so a short line beside a long one may be pulled up into it and the line breaks you already had can all move. This page only touches lines that are over the count. The two use the same column number and can produce the same file for a cue that is over the limit on its own, but they behave differently on every cue that is already fine: wrapping may change it, splitting leaves it alone byte for byte.
Why did the lines that were already short not change at all?
Because there was nothing to do to them. The page measures each line, breaks the ones that are over the count, and passes the rest through untouched rather than writing them out again from a re-measured copy. That is not just a claim about the output looking the same: the lines that are within the limit come through as the very same data the file held, which is what makes this page safe to run on a file whose breaks you have already fixed by hand.
Does breaking a long subtitle line change its start or end time?
No. Both lines of a broken cue are on screen at the same time, for exactly the span the cue already had. Nothing about the timeline changes, and the cue numbers are the file's own. If what you need is for the two halves to appear one after the other, that is a different job: the page that splits a long subtitle into two cues divides the time as well, which is a visible change to the film's timing rather than to its layout.
Where does a broken subtitle line actually get broken?
At a space or at a character boundary in a script that does not use spaces. A break inside a word is never made: the word would have to be cut in half, which changes what the viewer reads. Chinese and Japanese text has no spaces to break at, so every character is a boundary and the break can land anywhere in the sentence. That is how those scripts are set, and it is the reason the limit is counted in screen columns rather than in characters.
What does the report show after splitting long subtitle lines?
How many cues had a line broken, and how many lines are still over the limit when the run finishes. The second number is the one to look at, because it is the count of lines that could not be brought inside the count — a word longer than the limit is the usual reason. Cues that were already within the limit are not counted at all, so a report of one cue means one cue was actually changed.
Which subtitle formats can have their long lines split?
The eleven whose cues can hold more than one line: SRT, WebVTT, ASS, SSA, SBV, SAMI, LRC, MicroDVD, TTML, JSON and CSV. Plain text is left out because one line of text is one cue in that format, so there is no line inside a cue to break and a run on a text file would return the file unchanged. Line length is a property of the visible text, so a styling range that covers a break is kept on both halves rather than dropped.