v1.34.0

kevinraymond/fosforav1.34.0Aug 20, 2026by github-actions[bot]

AI Summary

Improves audio analysis reliability by fixing drop detection timing and implementing actual section boundary detection using self-similarity.

Key Highlights

Drop event now fires correctly on drops, with new 'Arm hold' and 'Jump window' sliders for better control.
Section boundaries are now actually detected using self-similarity novelty thresholds.
Trail and feedback effects now measure trail length in seconds for consistency.
Audio panel's Drop sliders trade sensitivity against false cues.

New Features

Arm hold slider
Jump window slider

Full Release Notes


### Fixed
- **Trail and feedback effects measure trail length in seconds now, not frames**, so the same
  preset looks the same on a 30 Hz laptop and a 240 Hz desktop — at a decay of 0.82 the image
  surviving one second ran from 2.6e-3 at 30 fps to 2.1e-21 at 240. Nothing changes at 60 fps;
  above it, feedback backgrounds that accumulate particles settle brighter than they used to,
  and below 30 fps trails run a little long instead of flashing dark on a stutter.
- **The drop event fires on drops now** — it had gone off once in 141 reference drops across
  374 tracks, because it demanded 4 seconds of unbroken build-up before it would even watch
  for one. On nine hand-labelled EDM tracks it now catches 7 of 23 drops at three wrong cues
  for every right one; it still fires on build-ups and breaks. The audio panel's Drop sliders
  trade sensitivity against false cues.
- **The drop jump window now lasts the 1.5 seconds it always claimed.** It was sized against
  the analysis rate but filled at a tenth of it, so the setting was off by nearly ten times —
  and by different amounts at different sample rates (12.9 s at 44.1 kHz, 14.1 s at 48 kHz).

### Added
- **Two new Drop sliders in the audio panel**: "Arm hold" keeps the drop armed through a cut
  before the drop, for material where the build-up drops out entirely first (off by default);
  "Jump window" sets how far back the loudness-jump baseline looks.

## Downloads

| Platform | File | Notes |
|----------|------|-------|
| **macOS** | `.dmg` | Universal binary (Apple Silicon + Intel) |
| **Linux x86_64** | `.tar.gz` | glibc 2.35+ (Ubuntu 22.04+, Debian 12+) |
| **Windows x64** | `.zip` | Windows 10+ |

## Getting started

**macOS**: Open the DMG, drag Fosfora.app to Applications (or run directly). Signed and notarized.

**Linux / Windows**: Extract, run `fosfora` from the extracted directory. The `assets/` folder must be next to the binary.

## Requirements

- **GPU**: Vulkan (Linux/Windows) or Metal (macOS)
- **Audio**: Built-in mic or line-in for audio-reactive visuals
- **NDI®** (optional): [NDI® SDK runtime](https://ndi.video) for network video output
- **Video playback** (optional): `ffmpeg` on PATH for media layers

---
NDI® is a registered trademark of Vizrt NDI AB.