Separate score from scene sound
Audience-only underscore belongs to the score layer. A radio, ringtone or live performance belongs to the scene and follows location, distance and story action.
Drama Skills normally leaves score realization to the sequence timeline, so individual files do not each begin a new piece. Per-shot musical intent still matters. Honor an accepted baked-music or single-container approach when that is the project’s choice. Density, strength and whether to use music follow the scene’s experience.
Map the music across cuts
Decide whether neighboring shots need continuity, an intentional stop or space for a line. This manual author table is a plan, not an executed mix report.
| State | Specify | Listen for |
|---|---|---|
| Enter | Trigger and position in the chosen track | A line masked or emotion revealed too early |
| Continue | Same timeline position across the cut | Restart, sudden new music or level jump |
| Yield | Dialogue span and recovery point | Relative voice, room and music levels |
| Exit | Trigger and ending shape | An accidentally chopped tail |
A hard music cut can serve the story, such as a sudden loss of safety. State that intent and judge the actual listening result. Smoothing should not erase the desired break.
Original example: a three-shot dome test
This is an edit design, not generated or auditioned footage. A technician adjusts the star map before opening. The projector fan is an established scene sound; M1 is the example project’s audience-only score.
| Shot | Content | Score |
|---|---|---|
| A01 | First stars appear on a dark dome; fan continues | Enter M1 with the stars |
| A02 | Focus-ring detail; “Jupiter’s edge is still blurry” | Continue M1 and yield for this line |
| A03 | Stable stars, wide dome view | Continue then exit; fan remains |
Yielding score does not mute all sound. Keep the line intelligible; preserve or reduce the fan according to the accepted scene plan.
If A02 contains music, inspect whether that layer can be treated separately. If so, adjust it and join timeline M1. If music, speech and fan share one baked file, assess the actual usable material. Request replacement footage when the complete line cannot be preserved; do not assume lossless separation.
From plan to an actual mix
In this pinned version, built-in render retains source audio. Its sound line records intent; it does not execute score entry/exit, crossfades or dialogue ducking. Use an external edit or mix tool when needed, record source material, track position, treatment spans and actual output, then review the finished sequence.
Listen across A01→A02 and A02→A03 for track position, complete speech and fan space. An acceptable overall level does not prove each line is clear. Recheck a join when a shot is regenerated rather than carrying an old listening result onto a new file.
See room-tone continuity for ambience, and scene sound design for why music or silence belongs here.
FAQ
Must every shot be music-free?
No. Timeline score is a default; accepted baked music or unified delivery can remain. Separate scene sound from score and check actual joins.
Do it with the skill
Use $short-drama-edit with existing footage, script and edit list. Separate scene sound and score; map entry, continuation, dialogue space and exit. Identify external mixing work and review the actual output afterward. Without media or output, deliver a plan rather than an audition pass.
Read the method: Cross-shot music and ambience · Audio seams, captions and enacted mixing · Choose cuts by information and performance