Can the text show both speakers?

Not readably in a single line during genuine crosstalk. Where the two take turns quickly rather than truly overlapping, alternating lines work fine.

Who stays in the picture?

The frame follows whoever is speaking, and on paid plans split screen keeps both in shot, which suits an exchange better than cutting between them.

Should I fix the text by hand here?

Overlaps are the first place worth checking, because two voices at once is the hardest thing for any system to separate. You can adjust the words before the clip is produced.

Is crosstalk worth clipping at all?

Often yes. Disagreement holds attention. It just needs more setup than a clean single speaker moment does.

And when a third voice joins in?

The same rule, more strictly. Pick the one thread the clip is about and let the others be sound rather than text.

Does separate microphones help?

Considerably. One microphone per speaker is the single biggest improvement available for any multi person recording.

Should an overlap be cut out?

Rarely. Disagreement holds attention. What it needs is a few seconds of setup before it so the viewer knows what is being argued about.

Subtitles When Two People Talk at Once

Whose words should the text follow?

Whoever is carrying the point. In an interruption that is usually the person being interrupted, because their unfinished sentence is what makes the interruption land. Trying to show both at once produces a line nobody can read in the second it is on screen.

Deciding whose words to follow

Genuine overlap is usually shorter than it feels, often two or three seconds inside a much longer exchange.

Situation	Follow	Because
One interrupts the other	The person interrupted	Their unfinished sentence is what makes it land
Both answering the same question	Whoever the clip is about	Two threads in one line cannot be read
Quick turns, not real overlap	Both, alternating	There is time for each line to be read
Someone agreeing underneath	The main speaker	The agreement is sound rather than content
Three or more at once	The single thread the clip is about	Anything else becomes noise

Keeping the exchange understandable

When voices overlap, show the words that carry the point and drop the rest. Viewers need to follow the exchange, not read every syllable. A short interruption can often be left out of the text entirely if it does not change the meaning.

Check that the result still reads naturally.

Cutting around crosstalk

Sometimes the best solution is to trim the clip so the overlap is shorter, starting just after the interruption or ending before it. The moment keeps its energy while the text stays clear.

Preserve any reaction that makes the exchange funny or tense.

Showing who is speaking

If both people matter, a layout showing both faces helps viewers match words to speakers. Text position or colour can also separate the two voices.

Keep the approach consistent within a clip.

Recording to reduce crosstalk

Separate microphones for each person make overlapping speech easier to handle later. Asking guests to finish their points before responding also helps during key moments.

Lively conversation is good; constant overlap is hard to clip.

One thread through a noisy moment

Find where the overlap starts and ends. It is usually shorter than it feels, often two or three seconds.
Decide which voice the moment belongs to. Ask what the clip is about. The text should follow that thread, not the loudest voice.
Cut so the exchange has a beginning. An overlap dropped in cold reads as noise. A few seconds of setup makes it read as an argument.

Take an argument from your own recording and see what comes back.

Questions people ask

Can the text show both speakers?: Not readably in a single line during genuine crosstalk. Where the two take turns quickly rather than truly overlapping, alternating lines work fine.
Who stays in the picture?: The frame follows whoever is speaking, and on paid plans split screen keeps both in shot, which suits an exchange better than cutting between them.
Should I fix the text by hand here?: Overlaps are the first place worth checking, because two voices at once is the hardest thing for any system to separate. You can adjust the words before the clip is produced.
Is crosstalk worth clipping at all?: Often yes. Disagreement holds attention. It just needs more setup than a clean single speaker moment does.
And when a third voice joins in?: The same rule, more strictly. Pick the one thread the clip is about and let the others be sound rather than text.
Does separate microphones help?: Considerably. One microphone per speaker is the single biggest improvement available for any multi person recording.
Should an overlap be cut out?: Rarely. Disagreement holds attention. What it needs is a few seconds of setup before it so the viewer knows what is being argued about.