Skip to content

Settings

The Preferences tab controls how ForgePlayer turns a scene into a signal — which form of haptics it prefers, how it synthesizes from a funscript, and timing offset.

Content preference

  • Funscripts (live synth) — default. Synthesizes from the funscript in real time. Lets you pick the synthesis algorithm (continuous / pulse-based) and adjust haptic offset live. The funscript is the authored intent, and live synthesis now plays just as cleanly as a pre-rendered file — the DirectSound-default fix (v0.0.6) ended the USB-DAC pops/clicks that used to make pre-rendered sound the safer pick.
  • Sound files (.wav / .mp3) — pre-rendered stim audio, used instead of synthesizing live. No algorithm choice — whatever the file was rendered with. Pick this if you have older 312/2B-era hardware, or a scene that only ships a stim audio file and no funscript.

Either preference falls back across forms at Haptic 1 (silent stim is worse than wrong-form). Haptic 2 only ever plays prostate-specific sources for the matching preference, otherwise mirrors H1.

Generation algorithm (when preference is funscript)

  • Pulse-based (default) — power-efficient waveform with shaped discrete pulses. Slower numbing. For modern audio-based stereostim hardware (Tingler, EstimHero, and similar). ForgePlayer defaults to this because that's what the FunscriptForge content pipeline targets.
  • Continuous — classic restim waveform. Best for 312 / 2B-style hardware. Lower power efficiency. Works well at ~100 Hz. Flip to this once if that's your box.

(The default applies to a fresh install; an existing preferences.json keeps whatever you last chose.)

Haptic offset (s)

Shift the stim signal relative to the video. Positive = stim leads video; negative = stim lags. Compensates for USB / driver latency. Captured at launch — changing mid-playback applies on the next launch.

Debug mode

Top bar Debug toggle (off by default).

When on:

  • Every event streams to ~/.forgeplayer/debug-stream-<timestamp>.jsonl in real time (one event per line — survives a hard crash).
  • The ⚑ Mark button captures a timestamped marker. Press it whenever you hear / feel / see something weird during dogfood. The count next to the icon shows how many events have been recorded.
  • Export… writes the in-memory event buffer to a single ~/.forgeplayer/debug-<timestamp>.json.
  • Clear wipes the in-memory buffer (the on-disk stream survives).

What's logged:

  • UI clicks (transport, seek, library activate, setup changes)
  • Library scan / variant resolution / pin save / pin replay
  • Engine launch / teardown / fullscreen toggles
  • Stim source dispatch (stim.dispatch), prostate resolver (stim.aux_resolved), source-probe peak / RMS at t=0 / t=30s
  • Audio-thread events (auto-resync, seek envelope stages)
  • mpv device list

When off, every recording call returns immediately. No measurable overhead.

Debug environment variables

For pop investigation only — set before launching ForgePlayer:

  • FORGEPLAYER_SYNTH_ISOLATION — off (default), constant, alpha, alpha_beta, alpha_beta_volume. Strips funscript axes one at a time so you can bisect which axis (if any) drives audible artifacts. constant puts a steady carrier with no modulation at all; each next mode adds one axis on top.
  • FORGEPLAYER_RECORD_STIM_DIR — path to a directory. When set, every StimAudioStream callback writes its output block to a per-stream int16 WAV in that directory. Open in Audacity to visualize / listen to exactly what we sent the audio device.

Both are off by default. Production paths pay nothing.