Subtitle Edit - Check split lines when character limitation not met

[whatcrisis]

Hello guys,

First of all cheers to all. I have been reading these forums for a while, but never registered until now. Tight ship you guys have here.

I am a professional subtitler, and usually work with Subtitle Edit, but I am not really tech-savvy so maybe you can help me with this question.

Netflix specs ask that the character limitation is 42 characters per line. If the entire subtitle is below that limitation (and it is not a dialogue), it should be condensed into one line. Sometimes, when translating, you can do something like:

I wanted to let you know
that I am back.

This is 39 characters long, and would become 40 when condensed in one line. I wonder if there is a way in Subtitle Edit to create a rule that would automatically find and mark all events where the character size is 41 or less, and there are two lines. That way I could check that more easily before delivering the translation and ensure that I have not split subs where there should be one single line.

Does it make sense?

Thanks!

[abolibibelot]

This may be what you're looking for : « Tools => Merge short lines ». No, scratch that : this is for merging distinct subtitles with distinct timecodes.
But I just found out that SubtitleEdit had a dedicated “Netflix quality check” module. Well, I justed tested it : although it seems very thorough, it doesn't seem to report instances where the total length of two (non dialog) lines is below 42. So this should be requested for a future revision.


[The below suggestion doesn't work well actually, as I haven't found a way to restrict the search to pairs of lines wherein the total number of characters is below a given amount ; still, I'll let it there, it may give you or anyone ideas for another task...]

That could perhaps be done with a text editor that supports regular expressions, like TEDNotepad, allowing to perform advanced search, or search and replace, operations.
Syntax : “\n” means line break, “.” means any character, “+” means one or more instances of the character before, “^” means beginning of a line, “$” means end of a line, “\d” means any digit, “\d{,10}” means at most 10 instances of any number, “.{10,}” means at least 10 instances of any character, “\w{10}” means exactly 10 letters, brackets allow to define groups of several possible characters to match (or to match anything but said character(s) by putting “^” before), and parentheses allow to define groups to match in the replace pattern (each group in parentheses can be matched with \1, \2, up to \9, based on its sequential order in the search pattern).
For example :
Search (remove quotation marks) : « ^(\d+)\n^(.{12} --> .{12})\n^(.{,25})$\n([^\n]) »
Replace (remove quotation marks) : « \1\n\2\n\3 \4 »
The search pattern matches any SRT subtitle with two lines and a first line of 25 chacters or less. This can be translated as : a line beginning, any number (group \1), a line break ; a line beginning, exactly 12 instances of any character followed by “ --> ” then again exactly 12 instances of any character (group \2), a line break ; a line beginning, 25 or less instances of any character (group \3), end of line, line break ; any character that is not a line break (group \4).
The replace pattern effectively replaces the line break by a space if the first line is 25 characters or less.
For instance this :

564
00:45:20,718 --> 00:45:23,012
It's plump, juicy,
three inches thick.

Becomes :

564
00:45:20,718 --> 00:45:23,012
It's plump, juicy, three inches thick.

(From the 1993 movie Falling down which I discovered a few months ago and highly recommend. Some directors with a lousy reputation — in this case Joel Schumacher — have done at least one movie that is more striking, more relevant and enduring than the entire filmography of respected but tame contemporaries. Uwe Boll, however despised as he is, made at least three : Rampage, Darfur, and to a lesser extent Assault on Wall Street.)

[But again, it's not satisfying for this particular task, since, as it is, it won't detect if the total of line 1 + line 2 is below a target amount.]

[Ennio]

Did you try this?

In settings --> General tab: Single line max, length: 42. Unbreak subtitles shorter than: 43.

In "Fix common errors" window check "Remove line breaks in short texts (all except dialogs).
Click next. It should show all lines that meet those settings.

[whatcrisis]

Did you try this?

In settings --> General tab: Single line max, length: 42. Unbreak subtitles shorter than: 43.

In "Fix common errors" window check "Remove line breaks in short texts (all except dialogs).
Click next. It should show all lines that meet those settings.

Thank you, this seems to play the trick!