On this page
If you want to develop your own audio guide, you start with the text, not the technology. How many stops will your route have? How long can one recording last? And how do you write something that sounds good in an earbud while your visitor is looking at a painting, a façade or the edge of a wood? This guide goes through those choices in the order you make them.
How many stops does an audio tour need?
The number of stops follows from your route, not from a target. Walk the route and note every place where a visitor naturally pauses: in front of an object, on a square, at a viewpoint, at the start of a gallery. Those are your candidates.
Then start crossing them off. A stop earns a recording when there’s something to point at and something to tell that the visitor can’t read for themselves on the label next to it. Left with a stop where all you have is a name and a date? Merge it with its neighbour, or drop it.
Look at the distance between stops too. Two recordings five metres apart push your visitor into a stop-and-start rhythm; a ten-minute gap on foot makes the tour fall flat. A stop every few minutes of walking keeps it moving.
How long should one recording be?
The visitor is standing still. That’s the main limit. As long as there’s something to see that the recording is about, attention holds; once the text has moved beyond what’s in front of them, people walk on with a voice in their ear that’s still talking.
So measure your recordings in spoken time, not in words on paper. Read your text aloud with a stopwatch, or let the text-to-speech in your platform read it out and listen along. What looked short on paper often turns out half a minute longer when spoken.
Vary the lengths. An opening recording that explains the route can be longer than a quick pointer halfway. A key object that everyone stands in front of for ten minutes anyway can take more text than a stop people pass on the way. If every recording is the same length, your visitor hears a list; when the length follows what there is to see, they hear a story.
Stick to one rule: one recording, one subject. As soon as the same recording jumps from the building’s history to its current use to an anecdote about the architect, it has become too long. Split it, or choose.
Writing for the ear
A museum label is read; an audio recording is heard once, with no way to flick back. That changes your sentences.
Keep sentences short, with one thought in each. Put the subject first and the verb soon after it, so the listener doesn’t have to wait until the end of the sentence to know who does what. Subordinate clauses that work fine on paper lose their thread when spoken.
Write the way you talk. Address the listener informally if that suits your organisation. Say “this building stood here before the city had its walls” rather than “the structure dates from the period preceding the construction of the city walls”. Avoid lists of more than three items; nobody keeps a run of seven names in their head.
Point things out. An audio tour has one advantage a text panel lacks: you know where the visitor is standing and what they’re looking at. Use it. “Look at the third window arch from the left; it’s narrower than the rest.” Pointers like these guide the eye and make it clear straight away that the listener is in the right place.
Write numbers and abbreviations the way they sound. Put “seventeen fifty” or “fourteen metres high” in your script if that’s what the voice should say. If you work with AI voices, you hear straight away how a date or an abbreviation comes out and can adjust the spelling in your text to match.
And read everything out loud. Every tongue-twister, every sentence where you gasp for air halfway through, is a sentence your visitor won’t follow either.
Spreading the story across the route
You have one story and ten to fifteen separate recordings. The trick is to keep the story flowing while each recording stands on its own, because a visitor who misses a stop or walks the route the other way round shouldn’t lose the thread straight away.
Start with the big picture. What is this tour really about? Not “this museum”, but one idea: the harbour that made the city rich and then left it behind, the factory that changed its product three times, the forest that was planted by people. That line decides which stops are in and which, however interesting, fall outside this tour.
Then spread that line across the route:
-
Opening. Get the visitor oriented and make the tour’s promise. How long does it take, where does it end, what will they hear? Keep it short, then get walking.
-
Early stops. Tie your theme to something concrete. No nuance yet, but something the visitor can look at.
-
Middle sections. This is where the depth goes: the conflict, the change, the details that make the picture more complex. Alternate between big history and small, personal stories.
-
High point. One stop is the answer to the promise made in the opening. Give it the most time and the strongest text.
-
Closing. Briefly tie back to the main line and let the visitor go, perhaps with a pointer to what else there is to do.
Make every recording work on its own. A reference like “as you just heard” goes wrong with the first visitor who walks a different way. Better to briefly repeat what’s needed: one clause is enough to bring someone up to speed.
If you use GPS triggers, where the visitor stands decides which recording starts, not your running order. All the more reason to let each recording stand on its own feet and keep the common thread going through short callbacks rather than a strict sequence.
Writing for several languages: what to keep in mind now
If you’re going to translate your tour, your script should take that into account from the start. Wordplay, dialect and references that only locals understand get lost in translation or stop making sense. Write them out, or give the explanation a visitor from abroad needs in the recording itself.
Keep length in mind as well. The same content takes more words in German or French than in English, so a recording that’s just right in your own language can end up slightly too long in another. Leave a little room in your timing and you can fine-tune by hand later instead of rewriting.
Finally, write names and numbers so they work in every language. You’ll usually leave a local street name or building name untranslated; add one descriptive word so the listener knows what they’re looking at.
From script to tour
Script finished? Then the rest is a matter of filling it in. In the Wandrz platform you add your stops to the tour, paste the text into each stop and let text-to-speech turn it into a natural voice, in your own language and in the languages you add. Want to shorten a recording? Change the text and listen again; no recording studio, no technical hassle.
The editing is the real work. The number of stops, the length of each recording, sentences that are easy on the ear and a line that runs through the whole route: when those are right, your tour is live within a few hours.
Would you like someone to read your script before you record it? Then request a demo straight away.