Subtitle formatting

So there are a lot of “project threads” in the subtitling section, and I thought I’d like to add an actual discussion on subbing.

Many channels have their own rules about how to format translations/subs, and they’re usually pretty similar. This is my (most updated) personal list, if anyone wants to weigh in on it.

Note on HTML tags


add a line break (like when you press “enter”)


italicize text; if you are italicizing the entire subtitle, you only need to add the opening tag at the start of the text (the end tag </i> will be automatically appended - note the forward slash in the closing tag if you’re adding it manually)

When to italicize text

  • Off-screen speaker - speaker is not in the shot, or in some cases, not the focus of the shot
  • Thoughts and/or voiceovers
  • “Short” flashbacks - maybe six lines, give or take; for longer flashbacks, italicize first two or three and last two/three
  • Song lyrics - (I also like the use of musical note characters, like ♫, at the start of each line)

Multiple speakers

Format: - First speaker here.<br>- Second speaker here.

- First speaker here.
- Second speaker here.

  • The hyphens/dashes before the text indicate and differentiate
    multiple speakers
  • The line break is used for a similar reason
  • I like the space between the hyphen and text…because it looks nicer.

On screen text

[Here’s the text on the screen.]

  • I like brackets; they look nice - but using parenthesis, or equal signs or hyphens on both ends of the text also get the point across

Translation continuity

This is more of a grammar thing, but it relates a bit. Theoretically, one segment = one sentence. However, since segmenters aren’t always fluent in the languages they segment, we’ll get split sentences. (And of course, some sentences are just split because they’re really long.)

When translating, use punctuation according to what’s spoken - not according to segments.

In other words, it’s not necessarily true that you would capitalize the first word in each segment, nor is it true that you’ll always end a segment with a period.

Miscellaneous things

  • Italicize team credits (Subtitles brought to you by…)
  • Italicize titles and episode numbers for TV shows (Episode ##) - or use all capital letters or onscreen text formatting
  • Try to minimize subs to two lines?

Line breaks <br>

You’ll have noticed, I’m sure, that when subs get too long for one line, you end up with two lines. But depending on the length of the sub, this gets to look kind of ugly…

Some people don’t really like line breaks, because subtitles do appear differently depending on the device you’re using. But the way I like to use them, it doesn’t really matter… So since I like them, I’ll discuss it anyway.

Like mentioned earlier, <br> is essentially like pressing “enter” on the keyboard, but pressing “enter” doesn’t really add a break in the subtitle editor.

You don’t add line breaks when there are nearly two full lines of text. When you have that much text, they’ll end up on two lines and look nice either way.

You want to add line breaks when text is around 1 to 1.5 lines long.

Here is some sample text. Now imagine the line ends | there.

By default, you’ll end up with: (alignment aside)

Here is some sample text. Now imagine the line ends

But by using a line break somewhere in the middle of the whole line, you can end up with something like:

Here is some sample text.
Now imagine the line ends here.

It doesn’t have to break at the end of a sentence, by the way. The idea is to center the whole segment of text, as best as you can, keeping both lines around the same length.

Definitely, in part, it’s because it looks nicer. But consequently, it’s also easier to read - since the text is less spread across the screen, your eyes only need to focus on one spot.

The tricky thing with line breaks is getting the feel for when to add them, based on whether you’re working on a licensed or fan channel, as subtitles display differently. …I’ve gone on a bit about this now, so I’ll stop. But if you’d like to discuss it further, I’d love to chat ^^

(PS. It’d be great if Viki would add some feature that automatically cuts a line in the center once it hits the end of a line, or reaches a certain length. …But I guess this will do for now.)

Of course, as this is my two cents about it, if you’re planning on subbing on a channel (that’s not managed by me), be sure to read up on those channel guidelines.

I’m almost positive I’ve forgotten something I wanted to add… But we’ll leave it at this for now.


Thanks. I was wondering how to italicize ans break up subs into two lines.

A few questions:

  1. I read in a comment that I can no longer locate, that different viewer apps see subtitles differently. It was suggested that spaces be placed after tags to ensure that those tags are applied by all apps. e.g., italics space break space hyphen space text…same on other side of text.

I would like to see the use of spaces addressed in the NSSA Subtitling guide, but if you can comment, I would appreciate it. When I am editing, I see spaces used arbitrarily throughout episodes.

  1. Are there shortcuts for a single musical note or an Em dash or do we have to make our own shortcuts?

  2. I would like to see a comment under subtitling rules in Team Notes as to whether American English or British English speech and grammar rules are preferred by the Language Moderator or if the Moderator wants to leave it up to the General Editors. I usually come to an agreement with other General Editors on a project so that the language and grammar used throughout the project is consistent (e.g., mum, mom).

  3. I have come across subtitles with this structure: Spkr A (Do you want to go?) Spkr B (Okay.) Spkr A (Right!). These are displayed on one line and sometimes with a break after Speaker B. Should these be hyphenated on three lines? Cutting in an additional segment is hardly worth the effort. How would you do this if you were editing an existing subtitle?

  4. Would like someone to be more specific regarding
    and how to use/format breaks so that text will appear pretty in all viewing apps–or at least in the predominant apps. I like your comment about not using breaks when there are nearly 2 full lines of text.

  5. Would like to see a section in the NSSA Subtitling PDF listing subtitling shortcuts, Keyboard shortcuts not available on the Timer, spacing between HTML tags that result in the desired visual effect for the dominant viewing apps.

  6. I would like to see a section in the NSSA Subtitliing PDF that addresses music lyrics and when to include them or not. Some Segmenters cut segments that include dominant dialog and lyrics together (this can be annoying when too much is on the screen). Other Segmenters do not include lyrics with dominant dialog and pick up the lyrics when a segment does not include dominant dialog. Since music lyrics are repeated over and over in a drama, I am inclined for the most part to exclude them from segments with dominant dialog, especially when the dialog is critical to the moment or plot.

I really appreciate your effort to consolidate formatting concepts/rules. I hope someday to see a PDF on the NSSA site that augments the current very helpful document,providing more details and guidelines.

1 Like

I’ll answer what I know about… c:

  1. No, these have to be made by yourself - meaning to copy e.g. a single musical note from somewhere. The Em dash can however be easily made: just press down “alt” on your keyboard while writing “0151” with the numbers on the right side of your keyboard. There shold also be a code for the music note, but I don’t remember it…

  2. In general, these things shold be written as following. When they are not, it is often due to a typing error (e.g. “<br” instead of “

  • Do you want to go?
  • Okay.
  • Right!
    Also, if we are not able to see one of the people speaking (e.g. talking in the phone or that the person simply just isn’t captured by the camera) this person’s lines would be written like this.
  1. I actually have no idea how to make it look pretty in apps and via website. But yes, do not use it if there is a lot of text in a line. You should definitely use it if a person is talking while some song is playing and the lines in the song are translated. Also if two people are talking together in the same segment. I prefer to make a break if there are two lines of medium length, but I don’t do it much since I’m not quite sure how it will turn out for the viewers.

  2. You can find shourtcuts in every subtitle editor. Just a part of a screenshot for reference:

    I don’t know if this may help since you want NSSA to show it somewhere, but hey, might work…?

  3. The general rules for including music lyrics is more or less: “if the lyrics fits the scene (e.g. what happens, the feeling and such) then you should include them”. Don’t worry too much about this if you are subbing in any other language than english (since everything is translated to english before other languages). I can’t say much more than this ^~^’

I definately agree with this, the sentences get ridiculously long and it’s hard to read them and watch at the same time when your eye wanders all the way from left to right on a larger screen; at times I have to pause the video in order to read the sentence(s). Anyone struggles with this too?


There was a whole thread about breaks. I had started it, because viki issued a recommendation exactly one year ago (I think it was April 16th, 2016) saying to please never use breaks unless it’s a dialogue or a footnote.
The reason being that what looks nice on my or your PC screen might look completely messed up in a smartphone, especially with some applications with strange names I don’t remember.
They make their own automatic break according to the width of the smartphone/tablet screen. So their break adds up to your forced break, with the result that most of the screen is covered with text!!!

On all the dramas where I moderate I make it a point to tell people never to put decorative breaks. And so do most veteran moderators I’ve worked with.
Here are the Translation Guidelines written by senior viki editor cgwm808 and used by most people here.
If don’t have the time to go and check it out, I’m copy pasting the relevant sentence.

20) Breaks. Put a break in a subtitle only if more than one person is speaking in the dialogue or the subtitle segment contains both transcription of text on the screen and dialogue. Don’t put breaks into the subtitle for appearance’s sake only. [/QUOTE]

Of course, although very widely used, these are not official viki rules, nor are they written in stone; every moderator has the liberty of choosing other guidelines, and make their own. In my own projects, I use them but with some modifications.

Hey Doubleblass!

Sorry for the delay, quite concentrated on this channel and got more than 1 post, pms to answer, good lesson for me to learn before posting a lot, I feel I have to answer for politeness or answering one of your questions for ex :wink:

I will only answer to questions that other contributors didn’t answer to make it less repetitive:

I asked a Super Editor on Viki whom I totally have faith in because got to work more than once and her edition explanations are really detailed and clear. So she told me that for avoiding mobile crash:
breaks only: put a space before you type break shortcut so in the green subtitle box, it appears like that
- Hello! <br>- Hello! and not this one:
- Hello!<br>- Hello!

Other than this case, putting a space after a <br> or before/after <i> is not necessary.

Wonder if people accept both? I’ve never met this case! For me, people from America and from England will have the same subtitles, so one way or another, if we favor one audience, the other one is not favored xd
When I studied English at school, it was British English that was learnt (audio, school books), sometimes we have the equivalent in American English (not always and not for every equivalent) but that’s all, same for my other friend translator here and I just checked on the internet, the official translation language for European Commission is British English: Source here:
So for 28 countries of EU (27 with Brexit) and some past colonies from GB where they still have some traces left: I think that they were taught British English in general (and so for Europeans viewers who watch in English on Viki XD)

But yes, if people wonder about, good point to inform on TN.

I almost asked the same question but was more about segmentation first to one nssa staff member a few months ago (and her explanations are one of the best in segmentation + subtitling!), questions&answers so you can understand fully and her explanations are so complete that I don’t want to cut them and it can help others with segmentation & subtitle because by understanding segmentation, I learnt a lot about how it could be subtitled and why the segmenters cut it like that and how they saw the subtitles in their head (credits to her, she’s awesome and feeling really super grateful):

[details=12) Multiple talking:]Q1) Is it allowed to make only 1 segment with 3 people talking in it if they talk at the same time and if it’s not possible to cut 3 segments per character? And in the case where we want to respect 2 people max by segment but to do that we have to break the “full sentence” rule:
¤ 1st seg: A/B talking
2nd seg: end of B sentence/C talking at the same time

¤ Or 1 seg: A/B/C talking.
Which one is better?

Is it allowed to have 3 lines of subtitles in some special cases?

A1) Right, so if it’s unavoidable, you can put 3 people together. I had a scenario similar to this yesterday while I was subbing. The segger created a segment with 3 voices. It was okay because all the voices had short lines (1-2 words each.) I created a two line sub-title no problem.

But yes, if they are talking all at the same time, you have no other choice but to create the one segment.

However, if Speaker C comes in after A and B start talking, it might be possible to start Seg #2 when Speaker C starts speaking. It sort of depends on what is being said.

So, use this rule. Whoever is the main speaker, go with that. If B is main speaker the whole time and A and C are just chiming in with “Yes”, “No”, etc. type things, but go with one segment to keep Speaker B in the forefront.

Seg #1: A/B/C

I might sub it like this:

`- Speaker B…

  • Speaker A. - Speaker C`

Q2) We can put “- Speaker A. - Speaker C” in the same line? I’ve never seen this kind of subtitle when I subtitle. Is it a general rule that everybody uses?

A2) Yes, it happens rarely, so that’s why you probably have never seen it, but as long as the subtitle is short, this is doable.

If the speech happens in order and they are not overlapping each other, then you could just make a 3-line subtitle:

`- Speaker A.

  • Speaker B.
  • Speaker C.`

Q3) But if Speaker A is saying “Mom!” and Speaker C is saying “Aunt!” or if A “Yes, I agree.” and C “No, I don’t agree” or A “I like blue” and C 'I like red".
How do we know who is saying what if we don’t understand the language: Like “Yes, I agree” > who is saying that? It might impact the understanding of viewers.
If we don’t know the language, should I risk doing putting A and C in the same line?

A3) Right, it’s harder if you don’t know the language. You’ll just have to pick a format and hope for the best. But again, that’s why we have editors. When I edit for channels, this is the type of stuff I change. But what I appreciate is when segmenters were thoughtful to give me something that is easily editable. So, as long as I can tell what the segmenter was trying to do, and it’s clear, then I’m grateful and it’s easy to edit. So, pick something that is simple and logical. You’ve given this a lot of thought, so I can tell you will come up with something that will work. Also, remember, there are usually multiple solutions and most of them will work.[/details]

I think the best way to have an answer is actually to contact nssa, maybe it’s even in their agenda :slight_smile: they’re all very nice and very helpful!

1 Like

For 7) I also asked the same question about lyrics and segments because I had a hard time with them:

[details=Lyrics and dialogue]
Q1) Okay, I’m lost here compared to previous answers about segmenting or not segmenting lyrics.
If I hear an interlaced lyric/dialogue: I should forget about lyrics and just concentrate on dialogue (Answer 7)
For example we have:
1st segment: A character line
and at the same time I hear “shining brighter these days” but the full lyric sentence is normally “The sun and the moon are shining brighter these days”.

Should I cut like this:
1st segment: A character line + “shining brighter these days”// (right after the end of the lyric)
Or like this:
1st segment: A character line + 1s extension (forget about lyric)

A1) Okay, sorry, right. If all this time it’s been interlaced and you decided not to segment lyrics, then choose option #2 (forget about lyric.) If you’ve been alternating lyrics with dialogue, then this is what I do.

1st segment: A character line + 1s extension

Now, let’s say I’m sitting in the middle of the word “brighter” after I do my right arrow 1s extension. To make the lyric segment I’ll first try to go to the beginning for the word “brighter” as long as it doesn’t go back further than 1 sec (my extension.) If it does (meaning, “brighter” starts when the voice is speaking, then I look for where “these” starts, and I make a segment there.

So, then the subber might do this:

Seg #1:
Speaker’s subtitle here.
shining brighter these days.

Seg #2:
shining brighter these days.

This is where the subber’s job is make the subtitle cohesive. Then the viewer just knows that the lyric went on and the speaker stopped speaking…

Not all subbers are this sophisticated and you end up with this:

Seg #1:
Speaker’s subtitle here.
shining brighter…

Seg #2:
these days.

So, when you segment voice and lyrics together, the lyric is like another voice but you get to pick which word to start at. If you want to give the voice segment more read time, then advance a 1 sec and look for the nearest lyric word there. If you think the voice segment has enough read time and you want to say start a lyric segment at the beginning of a phrase and it happens to start right at the end of a voice segment, then start it there…

Q2) What we really hear in Seg #1:
Speaker’s subtitle here.

What we really hear in Seg #2:
brighter these days.

I didn’t understand this part.

A2) Yeah… about the lyric segments. One of the reasons we don’t really teach this method is because it’s complicated and most subbers don’t know how to handle it, but here is what is happening.

So, first we segment to the dialogue. I think we got that part, right?

Secondly, we add in a layer for the lyrics. Let’s look at the segments in context:

[----- Speaker’s voice + 1s ext ----]
[----------------- Lyric: “Shining brighter these days” ---------]
[- Seg #1------------------------------------]

If you ignore the Lyric, you see Seg #1 for Speaker. That’s what we’ve been focusing on.

For illustration sake, let’s ignore the Speaker and look at how to compose the Lyric segment with two connected segments (i.e. no gap), so that there will be no flash to the viewer if the subtitle is exactly the same. For example:

If we subtitled it this way:

[----------------- Lyric: “Shining brighter these days” ---------]
[- Seg #1----------------------------------------][Seg #2-----------------]

Seg #1:
Shining brighter these days.

Seg #2:
Shining brighter these days.

To the viewer, you think there is only one segment, not two, because the subtitle didn’t change and the segments are connected. Without a gap being there, there’s no flash, and no indication that the subtitle actually changed.

So, now let’s add back the Speaker segment. I’m going to extend the original Speaker segment a tad so that Seg #2 can start at “these days.” I do this so Seg #2 doesn’t start in the middle of “brighter” which is confusing to subbers and also if the subber doesn’t know how to handle this interlaced Lyric, then they will sub “these days.” But if they know how to handle interlaced lyrics, they will subtitle it as if Seg #1 and Seg #2 belong to the Lyric and Seg #1 is the Speaker and the Lyric.

[----- Speaker’s voice + 1s ext ----]–]
[----------------- Lyric: “Shining brighter these days” ---------]
[- Seg #1----------------------------------------][Seg #2------------------]

Seg #1:
Speaker’s subtitle here.
Shining brighter these days.

Seg #2:
Shining brighter these days.

We don’t advocate this method to our students because it’s complicated and most subbers are not trained in this manner, but in some places it works. You just have to decide… It can get super complicated and not worth the effort. That’s why we advocate alternating between voice and lyric, or just voice because that is a lot easier to execute and a lot easier to sub.

Q3) I see I think I understand: Interlaced lyrics/dialogue + end of the lyric : we don’t subtitle what we hear if the lyrics are cut, we write the same lyric sentence.
But in this case, for the space that I let before Seg #1 for example because it was the beginning of the lyrics (like I didn’t do a segment for this but I let a space): “The moon and the sun are”. After reading your message, if I decided to make a #3 segment for the beginning of the lyrics.
Could I do the same like here:
[Seg #3 Subtitle: “The moon and the sun are shining brighter these days”][Seg #1 subtitle: Speaker’s voice + the same lyric sentence][Seg #2 subtitle: the same lyric sentence]

[----- Speaker’s voice + 1s ext ----]–]
[----------------- Lyric: “Shining brighter these days” ---------]
[- Seg #1----------------------------------------][Seg #2------------------]

The last extension for [----- Speaker’s voice + 1s ext ----]–] How long can this extension be?

I thought that we subtitled only what we hear, like if the song is cut, the subtitle must show that this is cut.
And same with the character voice, if a part of A’s voice (end of his sentence) is in the segment of B (interruption of B), the subtitle should show this.
But I understand that it’s not that :open_mouth:

A3) [Seg #3 Subtitle: “The moon and the sun are shining brighter these days”][Seg #1 subtitle: Speaker’s voice + the same lyric sentence][Seg #2 subtitle: the same lyric sentence]

[----- Speaker’s voice + 1s ext ----]–]
[----------------- Lyric: “Shining brighter these days” ---------]
[- Seg #1----------------------------------------][Seg #2------------------]

To keep things simple, you only want to do the lyric combined segments with the beginning of a speaker’s segment plus whatever lyric follows. So technically, Seg #3 should go with a previous Speaker’s segment. They work in pairs and this keeps things more orderly for the subber. But this is why we don’t encourage this interlaced method, because first off, it’s hard, and many seggers don’t execute it properly and it’s a mess for the subbers to figure out what to do with it. Secondly, the majority of subbers don’t even know this method. So, only do this if you know each lyric segment could also be its own standalone segment (i.e. the alternating method.)

Okay, so always remember, speakers get priority still in this interlaced format. So, first think, segment to the voice as if lyrics are not there, and if possible, then layer in the lyrics underneath. Personally, I almost never execute this, just so you know. But it’s a method that is good to know about.[/details]

Of course, I think the best is always asking directly to people who have experience for these types of things.

I forgot shortcuts (for tips for typing faster, and this one you asked for some so I put them here)

The pic kuromidesu sent is Viki shortcuts for subtitling and some more Windows/Linux/OS X shortcuts that work on Viki subtitling editor:

CTRL + LEFT ARROW: Go to the beginning of the previous word
CTRL + RIGHT ARROW: Go to the end of the next word
FN + LEFT ARROW: Jump at the end of the line
FN + RIGHT ARROW: Jump at the beginning of the line
FN + UP ARROW: Jump at the beginning of the first line (for multiple lines in the same subtitle box)
FN + DOWN ARROW: Jump at the end of the last tine (for multiple lines in the same subtitle box)
CTRL + BACKSPACE: Delete the last complete word that you typed in the subtitle box

For OS X (FN becomes OPTION and CTLR becomes CMD)
OPTION + LEFT ARROW: Go to the beginning of the previous word
OPTION + RIGHT ARROW: Go to the end of the next word
CMD + LEFT ARROW: Jump at the end of the line
CMD + RIGHT ARROW: Jump at the beginning of the line
CMD + UP ARROW: Jump at the beginning of the first line (for multiple lines in the same subtitle box)
CMD + DOWN ARROW: Jump at the end of the last tine (for multiple lines in the same subtitle box)
CMD + BACKSPACE: Delete the last complete word that you typed in the subtitle box

I don’t usually jump from words to words to correct them, I prefer to use a mouse for that.
But for lines, when I transcribe and that I hear later 2 voices and that I forgot a dash for dialogue - or italics <i>, I use jumping at the beginning of the first line to put them after I finished subtitling the subtitle box. It’s faster than going back and forth in the same sentence.
Same for deleting previous word, I use the shortcut.
Select all, cut, copy, paste are shortcuts I use a lot too.

Hope it can be useful to you :slight_smile: I think I didn’t forget other questions.

1 Like

Re: American versus English.
I had my subtitles corrected because I wrote “colour”, “neighbour” and so on instead of Americanese “color”, “neighbor”. But there wasn’t anything in the Team Notes.
So I believe that it’s better if the English Editor gives a specific instruction about this too, so that everybody knows beforehand.
Maybe s/he can ask the various translators and editors which one they are most comfortable with and then go for the majority, to make things easier for them.

1 Like

I brought up this matter for the following reason:

When an episode or all episodes in a series are G-edited in different styles, it appears to the viewer that there are many errors in the subtitles and this makes the subtitling look amateurish. Also grammar elements may be significantly different between American vs English grammar styles (not to mention classic vs modern styles). This not only applies to the expected agreement of tense and number, but punctuation marks that suggest meaning, such as sarcasm.

ONE GRAMMAR STYLE IS NOT BETTER THAN ANOTHER. My aim is simply “consistency to the extent possible.” (Perhaps British Style would use single quotes with the final quotation mark inside of the period.) :slight_smile:

I would offer the suggestion that when there are several G-editors on a project, they could decide among themselves which style they want to utilize. I am happy to use whatever style a project’s G-editors prefer to use because I don’t enjoy watching dramas and films that use mixed styles. A G-editor might even enjoy the challenge of learning both systems.

I hope this clarifies what I was trying to communicate in my prior message.

Of course the Chief editor should give guidelines to everybody, so it is consistent. That makes perfect sense.

Are there rules regarding punctuation & capitalization when you have to split two subtitles at a place where ordinarly a dash or semicolon would be used?

For instance:

1st subtitle
Speaker 1: - It was so real
Speaker 2 - Here’s the thing -

(extended pause)

following subtitle
Speaker 2: whenever you’re having a nightmare like that and you can’t get out of it, all you have to do is say, “One, two, three, wake up!”

If they’re independent clauses like you have, just treat them like two sentences and capitalize the second subtitle.

You cannot end subtitle 1 with a semicolon (or colon, which would be more appropriate if it were a continuous sentence). If it’s not interrupted speech, the em dash also isn’t an option. Only the ellipsis would do the trick. I would put an ellipsis (three dots) which marks a hesitation, thinking how you will phrase your next sentence. But if there is an extended pause, I agree that you can also treat it as a separate sentence.

Ok thanks. I suppose either solution would work. But if I were to go with the ellipsis, should the second subtitle then begin with a capital letter or lowercase?

Either would work, here too. If there’s a bit pause, I would treat it as a trailing sentence, and then picking up again afresh, and put a capital letter. But lowercase would be fine too, as the continuation of the thought started in the first subtitle. So it’s your choice, really.


Along the same lines, if I had a pair of subtitles that read:

I was on a campout

(extended pause)

…far away.


I was on a campout…

(extended pause)

…far away.

Would the first or second one be more appropriate, or does it matter that much?

I would put ellipsis after the first subtitle, as if the person’s thoughts were trailing…
and then start the next one without ellipsis. With either lowercase or even uppercase, if you like. The uppercase meaning that after his thoughts were trailing, he decided to give some more details about it. Especially if the pause is more than 2-3 seconds, I think it qualifies as almost a separate sentence.

I was on a campout…
Far away.

Thanks, I’m finding this quite helpful. If you could bear with me, I just have one more subtitle pair that continues along this vain but I think might be a little different:

“I don’t think one second of this is funny -” (tersely and with definitiveness)

(extended pause)

- sir." (as an afterthought with a touch of sarcasm).

Might an ellipsis after the first subtitle be too contradictory with the speaker’s tone, while, conversely, an ellipsis added before “sir” clarify that it’s an afterthought? Or would there be a better way altogether to punctuate this subtitle pair?

That’s a difficult one. As you say, the ellipsis takes away from the assertiveness. On the other hand, you don’t have many choices. Maybe the fact that the viewers listen to the voice and they can understand the intent from there is a help: the subtitles don’t have to do the whole job- except for those hearing impaired.
Usually we don’t put ellipsis before a sub, here at viki. Of course some people may do it, there are no hard and fast rules. But so far I haven’t seen it.