The words people use, explained.

The interpreter speaks at the same time as the speaker, lagging a few seconds behind. Listeners hear the translated version continuously, so the event runs at its normal length.

It is demanding work. Human interpreters usually work in pairs and swap every 20 to 30 minutes, which is why booking two people per language is standard rather than an upsell.

The speaker says a few sentences, pauses, and the interpreter renders them. Accurate and cheap in equipment terms, since no booth or receivers are needed.

The cost is time: it roughly doubles the length of everything that is said. It suits bilateral meetings and short speeches, not a conference programme.

Simultaneous interpretation where the interpreters are not in the room. They work from a hub or from home over a platform, and listeners in the venue receive the audio on a device.

It removes travel and booth costs but adds dependence on connection quality and on a clean feed reaching the interpreter. The usual failure is not the platform but the audio sent to it.

The soundproofed cabin at the back of a conference room where simultaneous interpreters work. Standards exist for size, ventilation and sightlines, because interpreters need to see the speaker and the screen.

Booths are either built in or rented and assembled for the event, which is a significant part of why interpretation at conferences is priced per language and per day.

Translating through an intermediate language. If nobody works directly from Finnish into Portuguese, the Finnish is first rendered into English, and the Portuguese interpreter works from that.

It makes rare language pairs possible, at the price of a second delay and a second chance to drift from the original.

The interpreter sits beside one or two listeners and interprets quietly, with no equipment at all. Also called chuchotage.

It works for a single guest at a meeting and falls apart beyond that, because a whispering interpreter is audible to everyone nearby and exhausting to listen to for long.

The umbrella term for anything that gets sound to a hard-of-hearing person more directly than the room's loudspeakers do. Hearing loops, FM and infrared systems and phone-based listening all fall under it.

The point of all of them is the same: take the signal at the microphone and deliver it past the room's acoustics, reverberation and background noise.

A wire loop installed around a room that creates a magnetic field carrying the audio. Hearing aids with a telecoil, often labelled T on the device, pick it up directly with nothing in the user's ear.

It remains the best answer for people whose hearing aid should receive the signal itself. Its limits are that coverage stops at the edge of the loop, it carries one channel and therefore one language, and installing it is building work.

A radio transmitter at the sound desk and a set of receivers handed out to listeners, who plug in headphones or a neck loop. Common in churches, theatres and lecture halls.

It works well and its cost is operational rather than technical: the receivers have to be charged, cleaned, handed out, collected and eventually replaced, and a listener has to ask for one at the door.

A portable transmitter worn by a guide and receivers with earpieces for the group, used on walking tours, factory visits and in museums. Sometimes called a whisper system.

It lets a group spread out and keeps the guide from raising their voice. The same handling burden as an FM system applies, multiplied by how often groups turn over.

A spoken track describing what is visible for blind and low-vision audiences: what is on stage, what an exhibit looks like, what is happening on screen.

The technology to deliver it is straightforward. The hard part is the writing, because it has to describe before it interprets, and it has to fit in the gaps between what is already being said.

Text of what is being said, produced as it is said. A human doing this professionally is a captioner or speech-to-text interpreter; CART stands for Communication Access Real-time Translation.

Automatic captioning covers a great deal at a fraction of the cost and is not equivalent. Where accuracy is contractual or legal, in hearings or examinations, a human is still the answer.

Transcription turns speech into text in the same language. Translation turns it into another one. Automatic systems do the first and then the second, so an error in transcription is inherited by every language downstream.

This is why audio quality matters more than any setting. A clean microphone feed improves every language at once, and a muddy one degrades all of them.

The delay between something being said and being heard. For raw audio over the internet this can be well under a tenth of a second, low enough that you can watch a speaker's mouth without noticing.

Translation adds to it, because a segment has to be transcribed and rendered first. A second or so behind the original is normal and does not feel like a delay while following along.

Line level is the standard signal strength between audio equipment, as opposed to the much weaker signal coming straight out of a microphone. When you ask a venue for a feed, this is what you are asking for.

Post-fader means the signal is taken after the mixing desk's fader, so it follows what the operator does during the event. A pre-fader send ignores those moves and is almost never what you want for a broadcast.

EN 301 549 is the European standard for accessibility requirements in ICT products and services. It is the technical reference that European accessibility legislation points at.

For anything web-facing it in turn leans on WCAG, the Web Content Accessibility Guidelines, usually at level AA. When a tender asks for WCAG 2.1 AA, this is the chain it comes from.