Familiar Player vs Multimode Ladder Filter

Familiar is a self-hosted music player that understands the sound of your music, not just its metadata. Ask for "something that sounds like rain on a window" and it actually works. Point Familiar at your own home server (macOS Docker, Linux, Synology, OpenMediaVault — whatever you run). Nothing leaves your network. No cloud accounts, no subscriptions, no tracking. • AI chat and playlist creation — Powered by the Claude API using your own key. Describe a mood, a setting, a vibe; Familiar pulls from your library. • Semantic audio search — CLAP embeddings let you match by how music sounds, not just artist or genre. • Deep feature analysis — BPM, key, energy, brightness, danceability, valence, instrumentalness, and more, extracted on your server. • Ambient mode — continuous drone plus ghostly snippets from your library, tuned to the current key. • Community analysis cache — opt in to share anonymized audio fingerprints so others skip hours of analysis, and benefit from theirs. • Smart playlists — rules-based playlists that update automatically as your library changes. • Offline playback, lock-screen controls, synced lyrics (LRCLIB), Last.fm scrobbling. • Works with FLAC, ALAC, MP3, AAC, Ogg, Opus. Familiar is open-source (MIT) and requires you to run your own backend. Setup guides for Docker, macOS, and common NAS platforms are on GitHub.

A 4-pole (24 dB/oct) ladder filter modeled after the topology designed by Robert Moog in 1965. Four cascaded one-pole stages with tanh saturation at input and feedback produce warm analog-style distortion that scales with drive and resonance. Five filter modes (lowpass, highpass, bandpass, notch, and resonant) cover everything from classic subtractive sweeps to sharp self-oscillating leads. Resonance drives the filter to self-oscillation at high values, generating a sine tone at the cutoff frequency. A dry/wet control enables parallel filtering.