Turn Your Video Notes Into a Useful Shot Log
Written by the AZET team, the team at AZET, operator of azet.io. This illustrative guide is for reference and creative exploration, not a demonstrated AZET feature; all footage notes and examples below are hypothetical.
As of 2026-10-12
“Good hand shot” is a frustrating note when three files show the same hand. In our hypothetical example, a maker has filmed a short sequence of carving a rubber stamp. They need to find the take where the carving tool stays visible, not a paragraph praising the footage.
We would turn the maker's playback notes into a small shot log. The source is text that the creator supplies after watching their own recordings. Nothing in this exercise assumes that AZET sees footage, recognizes actions, chooses clips, or edits a video.
Give each observation an address
Before organizing anything, keep the original file name and a take identifier beside every note. A description without an address cannot reliably send you back to the footage.
For this hypothetical project, the creator supplies the following observations. Each time range is elapsed time within its own clip, written as minutes and seconds. These are rough listening and viewing notes, not frame-accurate edit points.
STAMP_021.mov / take A / 00:03–00:09
Left hand holds block. Tool makes first groove. Tool tip leaves frame at end.
STAMP_022.mov / take B / 00:02–00:11
Same first groove. Tool stays visible. Sleeve crosses the upper corner.
STAMP_023.mov / take C / 00:04–00:08
Ink roller crosses finished stamp. No note about sound.
STAMP_024.mov / take D / 00:01–00:07
Paper lifts to reveal print. Right edge of print outside frame.Separate observation from editorial preference
Use one field for what the creator saw and another for what they might do with it. “Tool stays visible” is an observation supplied in the notes. “Candidate for showing the carving step” is an editorial suggestion.
That separation also prevents a missing note from becoming a positive claim. For take C, sound is “not described.” It is not “clean,” “silent,” or “usable.” The creator can fill that field after listening.
A compact hypothetical log could look like this:
| Take | Observed action | Limitation | Possible use |
|---|---|---|---|
| A | First groove cut | Tool tip leaves frame | Early part of carving action |
| B | First groove cut | Sleeve enters corner | Candidate carving explanation |
| C | Roller crosses finished stamp | Sound not described | Candidate transition to printing |
| D | Paper lifts from stamp | Print edge outside frame | Partial reveal, pending review |
In the working copy, include the full filename and range from the source notes beside each row. The shortened display here lets us compare the editorial decisions without repeating those addresses.
Ask for a log, not an imagined edit
A hypothetical prompt for this task:
Organize these creator-written playback notes into a shot log. Keep each filename, take ID, and supplied range unchanged. Separate observed action, observed limitation, missing information, and optional editorial use. Mark suggestions as candidates. Do not infer image quality, camera movement, audio quality, or continuity that the notes do not describe. Do not combine different files into a single take.
Ask for unresolved questions after the table. In this example, useful questions would include whether take C has usable sound and whether take D has enough of the print visible for the intended reveal.
The failure to watch for: invented coverage
Suppose a proposed log says, “Take D: wide shot of the complete print.” That sounds convenient, but it contradicts the note that the right edge is outside the frame. Correct the row before using the log to plan a sequence.
A subtler mistake is treating takes A and B as consecutive actions. Both describe the first groove. Placing them in sequence might suggest two different grooves unless the edit makes the repetition intentional. Mark them as alternatives for the same action until playback establishes otherwise.
We would also resist a neat story that requires a shot the notes never mention. If there is no observation of setting the ink tray down, list that coverage as absent from the notes. Do not invent a file for it.
Review with the source clips open
Pick one row at a time and open its exact file. Jump to the supplied range. Confirm that the action, limitation, and take identity match. If the clip differs, fix the source observation before rewriting the suggested use.
Then compare alternatives side by side in your notes. For this hypothetical sequence, decide whether the visible tool in B matters more than the sleeve intrusion. Write that choice as a reason tied to the intended scene.
FAQ
Do I need precise timecodes to begin?
No. Rough elapsed ranges can help you find material. Label them as rough, retain the original files, and verify exact edit points in your editing environment when needed.
Can one clip contain several log entries?
Yes. Give each observed action its own range and keep the shared filename. Do not create new take identities merely because you split a long clip into rows.
What should I do with a take I dislike?
Keep its observation and the reason for rejecting it. A cropped reveal might still supply a detail later. The log is an index; deciding to omit a take does not require deleting the recording.
The maker introduction we published on October 3 describes an AI assistant email waitlist. Visit azet.io for that waitlist and current notices if you want to follow the project. We have not verified present availability in this guide.