How to Put Live Translated Subtitles on a Screen

Someone is speaking in one language and part of the room does not follow. The usual answers are an interpreter in a booth or nothing at all. There is a third option: put the translation on a screen, live, in letters big enough to read from the back. This guide covers how to set that up, and where the limits are.

The short version

Start any live mode on the device that hears the speaker, open the settings, and choose Show on a second screen. You get an address. Type that address on the beamer laptop, the TV browser or a second computer, and the translation appears there as large white text on black. Nothing to install on either device, and the screen does not need an account.

The screen can be any device. It does not have to be the one doing the translating, and it does not have to be in the same room.

Sizing it for the room

Two settings decide whether people at the back can actually read along. Text size scales with the width of the screen, so the same address works on a 1280x720 projector and a 4K panel. Sentences on screen decides how many lines stay visible at once, from one up to eight.

Fewer lines are easier to read at a distance and feel calmer. More lines help when people are taking notes or when the speaker moves fast and someone looks away for a moment. Start at two and adjust while someone stands at the back of the room. You can change both while the session is running and the screen updates immediately, which matters because the beamer is usually the one device nobody can reach once the event has started.

When the room speaks more than one language

A screen shows one language. If your audience is mixed, the screen is the wrong place to make everyone agree on a single one.

Use live broadcast instead. You speak once, and every listener opens a link on their own phone and picks their own language. The screen in the room then becomes simply one more listener, showing whichever language suits the room, while everyone else reads along in theirs. This is the part most video conferencing tools cannot do: they carry a single subtitle track, so everybody gets the same language or nobody does.

If you are streaming as well as presenting

For a hybrid event the subtitles can go into the video itself rather than onto a screen in the room. The projector view has a transparent mode meant for OBS and vMix: add it as a browser source and only the text composites over your video, with no black box behind it.

It looks broken in a normal browser tab, which is expected. Transparency only means anything inside software that composites layers.

Recorded material is a different job

If the talk already exists as a recording, do not translate it live. Upload the file instead and you get a subtitle file back: .srt or .vtt for a video player, or plain text to read and edit. The timings come from the recording itself, so the subtitles line up with the video without anyone retiming them by hand.

That route is cheaper, and the quality is better because nothing has to be decided in real time. For a course recorded in advance it is almost always the right choice. See audio file translation.

Practical things worth knowing beforehand

Sound quality decides everything. A microphone in front of the speaker beats a laptop microphone at the back of the room by a wide margin. If there is a mixing desk, feed it in directly. Every problem downstream starts with audio that was hard to hear in the first place.

Test with the actual screen. Beamers wash out contrast and a TV browser can be awkward to type an address into. Ten minutes in the room the day before saves a scramble at the start.

Stop the laptop from sleeping. The projector view asks the browser to keep the screen awake, which Chrome and Edge honour. Firefox does not, so leave it on a machine that is not set to sleep.

Long events run in parts. Uploads are capped at six hours per file, so a multi-day course is several uploads rather than one.

What it does not do

This is machine translation, not a certified interpreter. It is very good at making a talk followable for someone who would otherwise understand none of it, and it is not the right tool where a mistranslated sentence has legal or medical consequences. For a lecture, a workshop, a service or a conference session, following along imperfectly is a great deal better than not following at all.