diff --git a/src/reels/mod.rs b/src/reels/mod.rs
index 17def48..c6bfd68 100644
--- a/src/reels/mod.rs
+++ b/src/reels/mod.rs
@@ -207,7 +207,7 @@ fn finish_job(
 
 /// Render version: bump to invalidate every cached reel after a rendering /
 /// scripting change that should produce a fresh result.
-const RENDER_VERSION: u32 = 5;
+const RENDER_VERSION: u32 = 6;
 
 /// Narration expressiveness — Chatterbox's `exaggeration` knob. A slight bump
 /// over the ~0.5 default warms up otherwise-flat narration without over-acting;
diff --git a/src/reels/render.rs b/src/reels/render.rs
index 4e9431b..c8ddc04 100644
--- a/src/reels/render.rs
+++ b/src/reels/render.rs
@@ -39,7 +39,7 @@ const NARRATION_TAIL_SECONDS: f64 = 0.6;
 /// gentle dip; burst photos get a much snappier fade so the difference between
 /// a held shot and a quick burst is obvious.
 const SINGLE_FADE_SECONDS: f64 = 0.35;
-const BURST_FADE_SECONDS: f64 = 0.08;
+const BURST_FADE_SECONDS: f64 = 0.12;
 
 /// Video-clip framing. A clip plays at most this long, with its live audio
 /// ducked to `CLIP_DUCK_VOLUME` under the narration.
@@ -560,7 +560,7 @@ mod tests {
         assert!(g.contains("[v0][v1][v2]concat=n=3:v=1:a=0[v]"));
         assert!(g.contains("[3:a]apad[a]"));
         // Burst uses the much snappier fade (vs 0.35 for a held shot).
-        assert!(g.contains("d=0.08"));
+        assert!(g.contains("d=0.12"));
         assert!(!g.contains("d=0.35"));
     }