A quiet room, which is free and matters most
Room noise is the one recording problem that cannot be fixed afterwards. Hum from air conditioning, traffic through a single-glazed window, a fridge two doors away, and the reverb of a hard-surfaced office all end up permanently in the file, and no amount of editing removes them cleanly.
The best recording space in most museums is a small carpeted room with soft furnishings and no windows onto a road, recorded outside opening hours. A store room with a coat rail of textiles in it will beat a boardroom with a glass wall every time, and it costs nothing to try both and listen.
A microphone, and where the spending stops
A USB microphone in the range most organisations would sign off without a second thought is enough. The audible difference between that and professional studio equipment disappears entirely once the file is compressed for the web and played through a phone in a busy gallery.
What does make an audible difference is a pop filter and getting the narrator to sit a consistent distance from the microphone, both of which are close to free. Headphones for the person recording matter too, so problems get caught during the session rather than a week later.
A USB microphone, not a laptop's built-in one
A pop filter, or a sock over the microphone in a genuine pinch
Closed-back headphones for whoever is monitoring
Free editing software, which is more than good enough
A consistent seat position so levels do not wander between stops
Software you actually need
Recording and editing can be done entirely with free software, and the features you need are trimming silence, evening out volume between stops, and exporting to a compressed format. That is a short afternoon of learning, not a skill you need to hire for.
Export every stop at the same loudness. A guide where stop three is noticeably quieter than stop two makes visitors reach for the volume control mid-visit, which is one of the few things that reliably makes people give up on a guide.
The people, which is the real shopping list
Equipment is the cheap part. The list that determines whether the guide happens is a writer who knows the collection, a narrator who can read aloud without sounding like they are reading, a reviewer for every language you publish, and someone with authority to sign the script off so it stops changing.
That last role is the one most often left vacant, and it is why projects stall. A script with no owner gets revised indefinitely, and every revision after translation has begun costs money in every language.
What you do not need
You do not need a recording studio, handheld devices for visitors, an app, or a professional voice actor for every language. Devices and apps are delivery choices rather than requirements, and a browser-based guide reached from a QR code removes both.
You also do not need to cover the whole collection to start. Ten stops that are genuinely good will be listened to; forty average ones will not.
Frequently asked questions
What equipment do you need to make an audio guide?
A quiet room, a USB microphone, headphones, and free editing software. The room matters more than the microphone, because room noise is the one problem that cannot be fixed after recording.
What software do you need to record an audio guide?
Free recording and editing software covers it. You only need to trim silence, keep the volume consistent between stops, and export compressed files for the web.
Do you need a professional voice actor?
Not for every language. A human narrator suits your main language and anything emotionally weighted. Synthetic narration is what makes additional languages affordable, as long as someone who speaks the language checks the names.
Do visitors need any equipment?
With a browser-based guide, only the phone they arrived with. Some visitors bring their own headphones; a guide should still be listenable on a phone speaker, because most will not.
How long does recording take?
Around three to four times the finished audio length for a first attempt, so an eighteen minute guide is roughly a day including retakes and editing. It gets much faster once the narrator settles.