How to Record Your Family's Oral History
There is a real difference between recording a conversation and recording an oral history, and it isn't the questions. It is that an oral history is made to be used by someone who wasn't in the room — possibly in eighty years — so it needs permission, sound you can actually hear, a file format that will still open, an index of what's on it, and a copy somewhere that survives your hard drive. A phone voice memo captures the words. This page is about the other half: the parts that decide whether your grandmother's recording is still listenable and findable in 2110. To place the relatives she names as you go, use the cousin calculator.
Where are you in the process? Tap the closest match.
Tap an option above. The recording you can find and hear in fifty years is the only one that counts.
What oral history actually is
Worth being clear about, because the word gets used for two different things.
A family recording is a keepsake. An oral history is a source. The keepsake is for people who knew the person. The source is for people who didn't — a great-grandchild, a local historian, a researcher looking at what factory work was like in a particular town in 1968. The difference in how you make them is small; the difference in how long they last is enormous.
It is a first-hand account, deliberately recorded, with the context attached. The context is the part people drop: who was speaking, who was asking, when, where, and under what permissions. A recording without those four things is very hard for anyone else to use, and archives generally won't accept it.
The person speaking is usually called the narrator, not the subject or the interviewee. That word choice is not fussiness — it reflects that the recording is their account, in their words, and that they retain rights in it.
It is not the same as an interview technique guide. How to prepare, what to ask, how to handle silence and tangents is a separate skill, and it's covered properly in how to interview a relative. Everything below assumes you can already hold the conversation.
And it works best as a series, not a single event. Two shorter sessions of 60 to 90 minutes beat one exhausting three-hour marathon. Older narrators tire, and the second session is almost always better because they have been thinking about it in between.
Permission, and why the paperwork exists
This is the part that separates an oral history from a recording, and it is the part most families skip.
Get spoken permission on the recording itself. In the first thirty seconds, say the date, the place, your name, the narrator's name, and then ask directly: "Are you happy for me to record this, and for it to be kept for the family and possibly given to an archive?" Their "yes" is now part of the file. If nothing else on this page happens, do this.
Then use a written release as well. A single page: the narrator's name, the date, what the recording may be used for, whether it can be published or put online, and a signature. Archives will ask for it, and it is the document that lets a grandchild use the recording confidently in thirty years without wondering whether they're allowed to.
Rights in a recording are genuinely split, and the details differ by country. Broadly, the speaker has an interest in their words and the recordist in the recording, which is why a release that transfers or licenses those rights is standard practice rather than bureaucracy. If the recording will only ever circulate within the family, this matters less — but you cannot know that in advance.
Restrictions are normal and you should offer them. A narrator can agree to the recording while closing part of it: sealed for 25 years, closed until after their death, or "the bit about my brother is not for anyone outside the family". Write the restriction down with timestamps. Offering this in advance makes people far more candid, not less.
Ask again at the end. "Is there anything in there you'd rather I took out or kept private?" People say yes surprisingly often, and it is much easier to honor then than to discover the problem years later.
Keep the release with the audio, not in a drawer. Scan it, store it in the same folder as the recording, and name it to match — the file-handling side of that is the same discipline as how to digitize family documents.
The room matters more than the microphone
If you only fix one technical thing, fix this one. It is free.
Bad sound is the most common way a family recording is lost. Not deletion — unlistenability. Nobody sits through ninety minutes of a voice buried under a fridge compressor, so the recording quietly stops being used.
Choose a small, soft room. Carpet, curtains, a sofa, bookshelves. Hard kitchens and conservatories are the worst rooms in most houses because the sound bounces and everything smears.
Kill the hum before you start. Fridge and freezer off at the plug for the hour, central heating and air conditioning off, ticking clock into another room, television off, windows shut. Put both phones into flight mode; a phone hunting for signal buzzes through everything nearby.
Sit closer than feels natural. Distance is what lets the room in. Around an arm's length, at an angle rather than face-on, works well.
Listen to the room for ten seconds first. Stand still and just notice what you can hear. You will find at least two noises you had completely tuned out.
Record thirty seconds of room tone at the end — silence in that room, with nobody speaking. It sounds pointless and it is what lets anyone later patch or clean the audio without an audible jump.
What to record with, honestly
You can spend nothing at all, or a fair amount. Here is what each level actually buys you.
| Setup | Rough cost | What it gives you | Best for |
|---|---|---|---|
| Phone voice memo, phone on the table | Nothing | Clear words, thin sound, no control over levels | Any conversation you'd otherwise not record at all |
| Phone plus a clip-on lapel mic | Low | A large jump in clarity — the biggest value-for-money step on this list | Most family oral histories |
| Handheld recorder with built-in mics | Moderate | Proper levels, headphone monitoring, WAV recording | A series of interviews, or a project you want archived |
| Recorder plus two lapel mics | Moderate to high | Both voices captured evenly, usable in a slightly imperfect room | Two narrators, or a couple interviewed together |
| Video camera or phone on a tripod | Varies | Gestures, face, hands, the room — families value this enormously | Willing narrators; some people freeze on camera |
The single best upgrade is a clip-on lapel microphone. Clipped to the shirt about a hand's width below the chin, it sits close to the voice and stops the room from getting in. It costs less than a takeout and improves the recording more than any technique.
Use headphones while you record. Not to be professional — to catch the disaster. Clothing rustle on the lapel mic, a hum you can't hear in the room, a mic that came loose. You cannot fix what you never heard.
Watch the level meter. Aim for the loudest moments to peak somewhere around three-quarters of the way up, well short of the top. Audio that hits the ceiling is distorted permanently; audio recorded a bit quietly can be brought up later.
Always run a second device. Your phone, recording the same conversation from the table, costs nothing and has saved a great many interviews from a flat battery or a full memory card.
Test and play back before you begin. Ten seconds, listened to properly. This one habit catches almost every failure that ruins an interview.
Formats, settings and naming
The technical decisions that determine whether the file opens in fifty years.
Record WAV if your device offers it, at 24-bit and 48 kHz. That is the ordinary archival standard for spoken word. If your device only does 16-bit and 44.1 kHz, that is fine too — it is what CDs used and it is perfectly good for speech.
Keep the WAV as the master and never touch it. Make MP3 copies for emailing round the family and for listening in the car. MP3 throws information away permanently, so a family that only has MP3s has already lost something they can't get back. Storage is cheap; an hour of WAV is roughly half a gigabyte.
Do not edit the master. Trim, tidy and cut a listening version if you want, but the unedited original stays as it was recorded. That is what makes it a source rather than a production.
Name files so a stranger can read them. Something like `2026-03-14_Doreen-Hall_b1938_interview-01.wav`. Date first in year-month-day order so files sort chronologically, then who, then which session. A file called `REC0043.WAV` is functionally lost.
Write a short deposit sheet for each recording: narrator's full name and date of birth, interviewer, date, place, duration, equipment used, language, permissions and any restrictions. One side of paper, stored alongside the audio. Archives ask for exactly this, and it is also what lets you cite the recording properly later — see how to cite genealogy sources.
Back it up the same day, three copies, two kinds of media, one off-site. Computer, external drive, cloud. Do this before you do anything else, including listening to it. Irreplaceable recordings are lost to broken phones every year and the fix takes two minutes.
Logging: the twenty minutes that make it usable
An unlogged two-hour recording is a locked box. This step is what most distinguishes a real oral history from a family voice memo.
A log is a list of timestamps and topics. Every two to five minutes, or whenever the subject changes: `00:14:20 — leaving school at fifteen, apprenticeship at Hartley's, names foreman Mr Doyle`. That's it.
Do it within a week, while you still remember the conversation. You will catch names and places that a stranger listening cold would never make out.
Record every proper noun with its spelling. People, streets, firms, ships, villages, regiments. Names are what future researchers search for, and they are also what automatic transcription gets most reliably wrong.
Flag the good bits. Mark the two or three minutes you would actually play to someone. Every archivist does this, and it is what gets a recording used rather than shelved.
Note the restricted sections by timestamp, matching whatever the narrator asked for. Vague restrictions get ignored by accident; timestamped ones don't.
One page per hour of audio is enough. Nobody needs more, and the temptation to write a full transcript at this stage is what stops people finishing the log at all.
Transcribing, and how much is enough
Full transcription is transformative and slow. Be realistic about which version you'll actually complete.
Manual transcription runs roughly four to six hours per hour of audio for someone who hasn't done much of it, more with a strong accent or two people talking over each other. That figure is why so many family projects stall here.
Automatic transcription has genuinely changed this. Speech-to-text now produces a serviceable first draft of clear speech in minutes. It still mangles names, dialect, place names and anyone with a heavy regional accent — so treat it as a draft that must be corrected against the audio, not as a finished transcript.
Correct the names first, then everything else. If you only have an hour, spend it fixing proper nouns. That's what makes the transcript searchable and useful.
Transcribe what was said, not what was meant. Keep the false starts, the "I don't know", the repetitions and the regional grammar. Smoothing a narrator's speech into standard English quietly removes the person from their own account.
Mark anything you can't make out as `[unclear 00:42:10]` with the timestamp rather than guessing. A guess that gets copied into a family history becomes a fact nobody can unpick.
The transcript never replaces the audio. Tone, hesitation, laughter and the moment someone's voice changes are the parts a family listens for. Keep both, together, and use the direct quotations when you come to write a family history.
Where to put a copy so it outlives you
Everything above protects the recording for your lifetime. This section is about after that.
Assume the family copies will disappear. Drives fail, cloud accounts close when nobody pays the bill, and formats drift. A single institutional copy is worth more than ten family ones.
Try local first. County record offices, city archives and local studies libraries in public libraries collect community recordings, and a narrator who worked in a local industry or lived through a local event is exactly what they want. University libraries and regional oral history projects are the next place to ask.
Ask before you assume. Almost every archive collects selectively, and they will want to know the subject, the length, the permissions and whether you have the signed release. That conversation takes one email and tells you what they need.
National options exist too. In the United States, StoryCorps interviews are archived at the American Folklife Center at the Library of Congress. In Britain, national and regional sound archives hold large oral history collections and can point you to the right local scheme. Availability and criteria change, so check current terms rather than trusting a list.
Give them the whole package, not just audio: the master file, the log, any transcript, the deposit sheet and the signed release. A recording with no paperwork is usually declined however good it is.
And leave instructions in the family. One line in your own notes saying where the masters are, what the passwords are, and which archive holds a copy. That sentence is what stops the whole project ending in a house clearance.
FAQ
What is oral history?
Oral history is a first-hand spoken account, deliberately recorded and documented so someone who was not present can use it later. What separates it from a family recording is the surrounding material: consent, an identification of who is speaking and when, an index of the contents, and a copy stored somewhere durable.
What equipment do I need to record an oral history interview?
A phone will do, but a clip-on lapel microphone is the single best upgrade and costs very little. Clip it a hand's width below the chin, monitor through headphones, and run a second device as backup. A quiet, soft-furnished room improves the result more than any microphone.
Do I need a consent form to record a relative?
For family-only use, spoken permission recorded at the start of the file is the minimum. If the recording might ever be published, put online or deposited with an archive, use a short signed release covering permitted uses and any restrictions. Archives will normally require that release before accepting a deposit.
What file format should I use for oral history recordings?
Record WAV at 24-bit and 48 kHz where possible, and keep that file unedited as the master. Make MP3 copies for sharing and everyday listening. MP3 discards audio data permanently, so a family holding only compressed copies has already lost quality it cannot recover.
Do I have to transcribe the whole recording?
No. A timestamped log — topics and names every few minutes, one page per hour of audio — delivers most of the practical benefit in about twenty minutes. Transcribe fully only if you have time; automatic transcription gives a usable first draft, but names and accents must be corrected by hand.
Where can I donate a family oral history recording?
Start with county record offices, city archives and local studies collections in public libraries, then universities and regional oral history projects. Ask before you send anything, since most collect selectively. Supply the master audio, the log, any transcript, a deposit sheet and the signed release together.
Record it once, properly
None of this needs to be perfect. A recording made with a lapel mic in a quiet living room, with a spoken permission at the head of the file, a signed release scanned beside it, a one-page log and a copy in the cloud is already better than the overwhelming majority of family recordings ever made — and it is maybe an extra hour of work on top of the conversation you were going to have anyway. Start with consent and the room, because those two cannot be fixed afterwards; everything else can. If you are not sure who to record first, the questions in family history interview questions and the parent-specific list in questions to ask your parents will tell you where the biggest gap is. Then use the cousin calculator to place everyone they name.
