A pickup is a replacement recording for part of an existing performance. The new words may be correct yet still announce themselves through a different tone, pace, or room sound. Make the pickup match before attempting to hide the join. This is especially useful when a name, instruction, or product detail changes after the main narration is recorded.
Recover the original conditions
Find the unprocessed recording and the working version. You need both: the original helps reproduce the source sound, while the edited version shows the final context. Check any notes about the room, device, recording controls, reading posture, and script position.
If the original setup no longer exists, identify the most audible difference first. A new room tail is a different problem from a softer performance. Compare a test at roughly the same playback level before deciding the microphone itself is responsible. The room-echo workflow can help separate location changes from performance changes.
Listen to the thought before the target line
Play enough preceding audio to hear the speaker's intention and momentum. Then read the preceding sentence, replacement sentence, and following sentence as one unit. Record this extra material even if you expect to use only the middle.
For example, replacing “Choose the blue folder” with “Choose the green folder” can fail if you record only “green.” The original adjective may lean into “folder,” and a detached word introduces a new rhythm. A full phrase gives you more natural choices for the edit. ACX refers to rerecorded sections as pickups in its editing and spacing discussion.
Match intention before matching loudness
Describe the original delivery in usable language: “a helpful reminder after a demonstration” is clearer than “sound exactly the same.” Note whether the sentence begins with momentum, hesitation, reassurance, or a change of subject. Those decisions affect its shape.
Record one close match and one slightly more relaxed alternative. Announce their labels outside the usable speech. Do not generate a long collection of indistinguishable takes. If the new version is more formal because you are concentrating on a correction, read the surrounding paragraph once for meaning before trying again.
For generated narration, save the original text and selected voice, then test a coherent replacement passage. Reusing a voice does not guarantee identical phrasing. Keep the approved surrounding audio until you know which section can be replaced cleanly.
Place the edit where the sound allows it
In an external editor, compare possible boundaries between phrases. Avoid slicing through a vowel or a consonant that continues into the next word. Retain enough material on either side to adjust the join.
Match level with a local adjustment before reaching for compression. If the room background changes abruptly, use compatible recorded room tone where appropriate. Apply a short dialogue crossfade only after choosing a sensible boundary; a fade cannot make mismatched words become the same performance.
Review without staring at the edit
Listen from a few sentences earlier while looking away from the waveform. Mark the moment that attracts attention, then ask whether it is the tone, timing, meaning, or background. Fix that specific issue and listen again.
A pickup is complete when the replacement conveys the revised meaning and flows with the surrounding narration. The new sentence does not have to be indistinguishable under forensic comparison. It does need to avoid distracting an ordinary listener or creating an unintended emphasis.
Keep the replacement take, its script version, and the edited master together so another revision can start from the same evidence. For text-based alternatives, create a short comparison in the VocalCopyCat studio, download the candidates, and make the final continuity decision in the context of your project.
