The Tools We Use to Lock Dialogue to Picture
Rhys Pritchard describes the timeline markers, reference tones and checking habits he relies on when matching edited voiceover to video, and why the method matters more than the software.
Rhys Pritchard · · 6 minutes to read
Start with a reference you can trust
Before any editing tool is opened, I ask for one thing: a picture reference that will not change. That means a locked cut, or at least a version the producer has agreed will not move while the audio is being finished. Without it, every correction is aimed at a target that shifts under your hands.
In the studio we work from a flattened guide video with burnt-in timecode in the corner. It looks plain, but it gives a learner and an instructor the same point of reference when they discuss a particular line. Saying that the sentence at a given timecode lands late is far more useful than saying it feels a little off.
The guide track on that video is also kept. If the original recording used a scratch voice, we leave it in a separate channel so the finished voice can be compared against it by ear.
- Ask for a locked picture, or agree in writing when it will change
- Use a guide video with visible timecode
- Keep the scratch voice on its own channel
Markers, not guesses
The most reliable method I teach is also the simplest: place markers on the timeline at every point where sound and picture must agree. A hard consonant against a lip closing, a door slam, the first word after a cut, the last word before a title card. Each marker is named, so a revision note can point to it.
Learners often want to slide a whole clip until it looks right. I ask them to mark first and slide afterwards. Once the markers are down, you can see whether the error is a constant offset or something that grows, and those two problems need different remedies.
A constant offset is moved in one action. Growing error needs another look at how the file was made.
- Name each marker so feedback can refer to it
- Place them at plosives, impacts and cut points
- Decide offset or drift before moving anything
Checking for drift at both ends
Drift is the quiet problem in explainers and longer programme segments. A recording can match perfectly in the first minute and be a fraction of a second out by the tenth. It usually comes from a mismatch in sample rate or frame rate somewhere between the recorder and the timeline.
Our check is to look at the opening and the closing of the same piece, never only the start. If the first marker agrees and the last one does not, we measure the gap, note it, and compare it with the length of the audio file against the length of the picture.
When the two lengths differ slightly, the cause is usually a conversion, and it is better to fix the conversion than to stretch the voice by ear.
- Compare the first and last markers of every piece
- Measure the gap before correcting it
- Compare audio length with picture length
A clap, a tone and a short test
Where we have any say in how material is recorded, we ask for a clap or a short tone at the head of each take. It takes a second and makes later alignment almost mechanical, because the spike in the waveform lines up with a known moment.
For material that arrives without one, we create our own reference by finding a clear sound in both the recording and the guide track, such as a breath before the opening line, and treat that as the anchor. It is less exact than a clap, but it is consistent from one session to the next.
Learners then export a short test section, play it against the picture in a different player, and listen on headphones and on small speakers. A difference that appears in only one of those is worth chasing.
- Request a clap or tone at the head of each take
- Use a breath or hard consonant as a fallback anchor
- Test a short export on more than one player
Nudging by frames, then by ear
Fine movement is done in small, deliberate steps. We nudge by a single frame first, since that is the unit the picture works in, and only go finer when the sound itself demands it. Dialogue tolerates a little more lead than lag, so if a decision is close, we lean towards the voice arriving very slightly early.
Every adjustment is made on the clip, not on the whole track, so that the rest of the work stays untouched. We keep a short note of what moved and by how much, which saves a great deal of time if the producer later returns with a changed cut.
None of this needs unusual equipment. It needs patience and a habit of listening again after every small change.
- Move by whole frames before anything finer
- Adjust single clips rather than whole tracks
- Log each change and its size
What we hand back
When the work is checked, the review version carries the markers with it, so the producer can see the points we aligned and confirm them against their own copy. Anything we could not resolve is written down plainly, with the timecode, rather than hidden under a fix that might not hold.
Learners practise the same discipline. They finish a timeline, then write a short sync note that someone else could follow without asking questions: which reference was used, what offset was found, whether drift was present and how it was handled.
That note is as much a part of the work as the audio. It turns a private judgement into something another editor can check, repeat and trust.
- Return markers with the review version
- State unresolved points with timecodes
- Write a sync note another editor could follow
Synchronisation is a matter of method more than software. A fixed picture reference, named markers, a check at both ends of the piece and small, logged adjustments will serve you in any editing package. Work in that order, listen again after each change, and keep a note of what you did. The result is dialogue that sits with the picture and a process you can explain to the next person who opens the project.
Rhys Pritchard
Lead Instructor, Dialogue EditingRhys has edited interview films and recorded reports for independent producers across Wales for over a decade. He teaches cleanup and voiceover editing.