export const meta = {
title: "How to Search Interview Footage by Transcript and Build Selects",
description: "Find interview quotes across a video library, check their context, and give editors selected passages with source files and accurate time references.",
tldr: "Search transcripts to locate possible quotes, then watch the surrounding footage before choosing a select. Record the source file, version, speaker, and clearly labeled time range for every choice so an editor can recover the original moment. Group the source recordings in a collection, then use the logged time ranges to build a selects sequence in the editing application.",
slug: "how-to-search-interview-footage-by-transcript-and-build-selects",
publishedAt: "2026-10-31",
readingTime: 10,
thumbnail: "https://cdn.aspectlabs.dev/blog/how-to-search-interview-footage-by-transcript-and-build-selects/cover-f1da4bde5ec6.png",
authors: ["bright"],
primaryTopic: "post-production",
topics: ["post-production"],
tags: ["editing", "media-management", "review-workflow"],
faq: [
{
question: "How do I search video footage for a spoken quote?",
answer: "Create a transcript linked to the video, search for a distinctive word or phrase, and play the matching passage. If you do not know which interview contains the quote, use a library search that indexes spoken content across files. In Aspect, transcript search works within the open asset, while AI Search can find spoken content across the project."
},
{
question: "What should an interview selects log include?",
answer: "Record a unique select identifier, the exact source file and version, the speaker, the quote, and a start and end reference. Label whether those references are elapsed playback times, source timecode, or sequence timecode. Add the surrounding question, any important qualification, and enough information to find the synchronized production audio when it is separate."
},
{
question: "How do I use Aspect transcripts when editing in Premiere Pro?",
answer: "Export the Aspect transcript for your notes and quote log, keeping the source filenames and time references with each selection. Open those source clips in Premiere to build the sequence. If you want to insert clips by selecting transcript text, Premiere can transcribe the source media and add passages through its Text-Based Editing tools."
},
{
question: "Is a transcript timestamp the same as camera timecode?",
answer: "It may be a different reference. An elapsed timestamp measures position from the beginning of a file, while source timecode labels frames in the source media and may start at an hour or a recording time. Confirm the reference, frame rate, and media version before converting. A transcript from a combined interview or a different audio recording also needs a verified relationship to the camera clips."
},
{
question: "Does adding a search result to an Aspect collection create an edited subclip?",
answer: "Aspect collections group existing files and folders without moving or copying the originals. Adding a recording keeps the whole source available to the editor. Keep the selected time ranges in a quote log or time-range comments, then create the required subclips or sequence in your editing application."
}
],
draft: false
}
You may remember the point someone made in an interview without remembering the wording, the recording day, or which camera file contains it. Transcript search can find the answer, but the editor also needs its source file, time range, and the question that prompted it. Keep that information with each selected quote so the editor can find the footage and hear the answer in context.
Label timestamps carefully when copying quotes into a document. A timestamp from a camera clip may point somewhere different in a synchronized interview or yesterday's rough cut. The workflow below covers searching the interviews, choosing passages, and recording enough source information to turn those choices into an editable selects sequence.
Suppose you're producing a documentary about changes to a bus route and need interviews that explain how the new timetable affects passengers getting home from work. We'll use that story as the example for the search and quote log below.
## Prepare interviews so search results lead back to the right media
Before transcription, decide which recording the producer will review. A camera file with usable dialogue can be a straightforward reference. When the camera has only scratch audio, a synchronized viewing copy may be easier to understand, but it needs a record of the camera files and production sound used to create it. A transcript generated from a separate audio recording inherits that recording's timing, which may differ from the picture.
Keep the originals and their filenames intact. Add an interview identifier, speaker name, shoot date, and recording part in your logging system so two files named `C0001.mov` can't be mistaken for each other. The same identifier should appear in the transcript notes and the selects log. If an interview spans several camera files, preserve the boundaries or document how the combined viewing copy maps to them.
In Aspect, [transcription starts automatically after upload](https://aspect.inc/docs/asset-intelligence/ai-transcription), unless the location is excluded from AI indexing. Open the asset's Transcript tab when processing is ready. The transcript follows playback, and selecting a word takes you to its moment in the recording. Transcript availability can precede completed AI indexing, so an interview that can be read in the viewer may still need more processing before it appears in broader search results.
Listen to sections containing names, specialist terms, overlapping speech, or a quiet speaker before copying the text. Aspect detects speakers automatically; add their names to the quote log after listening so the editor knows who said each line.
Aspect connects transcript words to playback, letting a producer jump from a remembered phrase to the corresponding moment in an interview. That makes it practical to listen to the answer around a search result before copying a quote into the editor's selects log.
## Search across interviews, then narrow within the recording
The first search depends on what you remember. If you know the interview and recall a distinctive phrase, open that file and search its transcript. If you remember only the subject, search the library first so you don't have to open every recording individually.
Aspect's [AI Search](https://aspect.inc/docs/asset-intelligence/ai-search) can find spoken content across a project using an exact phrase or a description of the topic. It also searches other information, including visual content and metadata. Phrase the request around speech when that's what you need: `interviews where someone talks about missing the last bus after an evening shift` asks for an interview answer more clearly than `buses at night`.
For the bus-route story, start with the question the film needs to explore: how the new timetable affected people getting home from work. Search for that idea, then examine the filters Aspect created. A date filter or folder restriction may exclude an older interview that's still relevant, while a broad project search may include promotional videos unrelated to the reporting.
Your first matches can help you learn the words people actually used. One passenger might say “evening shift,” another “closing,” and another “the last service.” Search those phrases as well. Once a likely file is open, transcript search helps locate repeated mentions and compare the first explanation with a later, more specific answer.
If a search returns no matches, check that the interview has finished indexing and is included in the selected folders. Try a nearby word if a proper name was transcribed incorrectly, then listen to the likely section manually. A missed search result doesn't mean the subject was never discussed.
Use Aspect's [asset intelligence features](https://aspect.inc/features/asset-intelligence) to search across the interview library, and keep a note of which folders and topics you reviewed. When several producers divide the work, that note shows what each person covered and which recordings still need attention.
## Review the whole answer before accepting a quote
Play the interviewer's question, the answer, and the passage that follows. A speaker may qualify an apparent claim a few seconds later, or use “they” to refer to a different group from the one the producer expects.
In the route-change example, a search might surface “I can get home before midnight.” The next sentence might explain that this is true only when the speaker leaves work early. Dropping that qualification would change the meaning even though the selected words were transcribed correctly.
Evaluate the delivery as well as the statement. A hesitation can express doubt, a laugh can signal irony, and overlapping speech can make an otherwise clear sentence difficult to edit. Note these observations separately from the quotation so a paraphrase or your interpretation never becomes indistinguishable from what the person actually said.
Keep a useful alternative when several answers cover the same point. One may be shorter, while another includes the context needed for a different cut. Give each selected passage its own identifier and record the reason for keeping it. For example, note when a longer answer explains the cause of a problem that the shorter quote leaves out.
## Give every select a source reference and an explicit time basis
You can keep the selects log in a shared document or spreadsheet. Include enough detail for another editor to recover the same moment without repeating your search. Use one entry per selected passage, including when several passages come from the same asset.
A quote-log entry for the bus-route story could look like this:
| Field | Example entry |
| --- | --- |
| Select identifier | `ROUTE_014` |
| Speaker | Maya, interviewed about the evening service |
| Verbatim quote | “Before the timetable changed, I waited outside the depot after every late shift. Now I can get home before midnight, but only if I leave work ten minutes early.” |
| Source | `INT_MAYA_A_003.mov`, original uploaded version |
| Location | `RouteFilm / Interviews / Maya / CameraA` and the asset link |
| Transcript reference | Elapsed playback from `00:12:18.400` to approximately `00:12:33.800` |
| Source timing | Constant 25 fps; first source frame is `01:00:00:00` |
| Source start reference | Expected `01:12:18:10`; verify in the editing application |
| Context note | Answer concerns late shifts; keep the qualification about leaving work early |
| Selection status | Selected for review; editor still needs to set final In and Out points |
The example distinguishes elapsed time from source timecode. Elapsed time starts at the beginning of the viewed file. Source timecode labels frames in the recording and can begin at an hour value or a recording time. [Adobe's timecode guide](https://helpx.adobe.com/premiere/desktop/organize-media/apply-labeling/about-timecode.html) explains the hours, minutes, seconds, and frames convention.
Here, the first source frame is `01:00:00:00`, and the media runs continuously at exactly 25 frames per second. Twelve minutes, eighteen seconds, and 400 milliseconds after that first frame corresponds to `01:12:18:10`, because 400 milliseconds contains ten frames at that rate. This arithmetic applies to the stated example, not automatically to every interview.
A transcript generated from a trimmed export, a joined interview, or an independently recorded audio file needs its own verified mapping. Frame-rate conversions, variable frame rates, and timecode discontinuities also require care. Use the transcript to reach the passage, then set the final boundaries while listening in the editing application. Leave enough material around the words to preserve breaths, the question where needed, and room for an edit.
## Gather the source assets while keeping the quote ranges separate
After choosing passages, group the relevant files so the editor can browse the source material together. Aspect [collections](https://aspect.inc/docs/share-and-present/collections) reference existing assets and folders without moving or duplicating them. A collection named `Route film — interview selects` can therefore contain recordings from several shoot folders while their original organization stays intact.
Keep the selected time ranges in the quote log alongside the collection. If Maya has three useful answers in one recording, add the recording to the collection once and create three log entries. The editor can then set the In and Out points for each passage in the editing application.
Use [time-range comments](https://aspect.inc/docs/review-and-approve/commenting) to attach a select identifier and context note to a passage when that helps the team review it. Check the asset version as well: a replacement upload may change the content and trigger new transcription. The quote log must continue to identify the version on which the selection was based.
Aspect collections gather existing assets without copying the files or changing their original folders, and the same asset can appear in several collections. A producer can therefore organize interview sources by story topic while keeping the camera media in the structure the editor already uses.
## Build a selects sequence in the editing application
When you hand over the selects, include access to the media and notes on how to use the selected passages. Send the collection or source links with the selects log, and make sure the editor can open the referenced files. A transcript export alone doesn't supply the camera media, synchronized sound, or decisions about where each passage should begin and end.
On macOS or Windows, [mount the Aspect project](https://aspect.inc/docs/instant-access/mount-a-project) and use **Open in mount** on a search result to reveal its original file. Open that file in the editing application, then use the quote log to reach the passage you selected.
Compare the spoken words with the log entry before adding the passage to the sequence. If the production uses synchronized clips or a multicamera setup, edit with those prepared sources to preserve the correct audio relationship. Aspect's [Premiere workflow guide](https://aspect.inc/docs/workflows/premiere-pro) covers accessing shared media, while Premiere handles the edit itself.
Export the Aspect transcript when you want the full text beside the quote log or need to share it with someone preparing the story. Include the source filename and timing reference with the export so the text remains useful if it's separated from the handoff folder.
To assemble the edit by selecting words in Premiere, [transcribe selected source clips, highlight transcript text, and use Insert](https://helpx.adobe.com/premiere/desktop/edit-projects/edit-video-using-text-based-editing/add-clips-to-the-timeline-using-text-based-editing.html) to add the corresponding content to a sequence. Adobe also documents [Paper Edit](https://helpx.adobe.com/premiere/desktop/edit-projects/edit-video-using-text-based-editing/create-a-sequence-with-paper-edit.html), which previews selected transcript passages and creates a new sequence from them. Both tools use Premiere's transcript and the source clips open in the project.
Before handing over all the selects, have the receiving editor recover one entry from each source type, such as a camera clip and a synchronized interview. Confirm the picture, speaker, words, timing, and audio, then resolve any mismatch before repeating the process across dozens of selects.
The editor can then build [documentary string-outs and selects sequences](https://aspect.inc/blog/post-production/how-to-build-string-outs-and-selects-for-documentary-editing). The source references and context notes remain useful as the story changes, because the team can revisit an answer or find an alternative without repeating the original search.