From bc98463b28abda8c7fd1c194ca751e50b0ca06aa Mon Sep 17 00:00:00 2001 From: Robbert van der Helm Date: Tue, 17 Jan 2023 01:53:38 +0100 Subject: [PATCH] Add a basic amp envelope to Buffr Glitch --- plugins/buffr_glitch/CHANGELOG.md | 2 + plugins/buffr_glitch/src/envelope.rs | 85 ++++++++++++++++++++++++++++ plugins/buffr_glitch/src/lib.rs | 63 +++++++++++++++++---- 3 files changed, 139 insertions(+), 11 deletions(-) create mode 100644 plugins/buffr_glitch/src/envelope.rs diff --git a/plugins/buffr_glitch/CHANGELOG.md b/plugins/buffr_glitch/CHANGELOG.md index 90493f66..a637a26a 100644 --- a/plugins/buffr_glitch/CHANGELOG.md +++ b/plugins/buffr_glitch/CHANGELOG.md @@ -12,6 +12,8 @@ Versioning](https://semver.org/spec/v2.0.0.html). - Added an option to crossfade the recorded buffer to avoid clicks when looping around. +- Added attack and release time controls to avoid clicks when pressing or + releasing a key. - Added an optional velocity sensitive mode. ### Removed diff --git a/plugins/buffr_glitch/src/envelope.rs b/plugins/buffr_glitch/src/envelope.rs new file mode 100644 index 00000000..b12a9e0d --- /dev/null +++ b/plugins/buffr_glitch/src/envelope.rs @@ -0,0 +1,85 @@ +// Buffr Glitch: a MIDI-controlled buffer repeater +// Copyright (C) 2022-2023 Robbert van der Helm +// +// This program is free software: you can redistribute it and/or modify +// it under the terms of the GNU General Public License as published by +// the Free Software Foundation, either version 3 of the License, or +// (at your option) any later version. +// +// This program is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +// GNU General Public License for more details. +// +// You should have received a copy of the GNU General Public License +// along with this program. If not, see . + +/// The most barebones envelope generator you can imagine using a bog standard first order IIR +/// filter. We don't need anything fancy right now. +#[derive(Debug, Default)] +pub struct AREnvelope { + /// The internal filter state. + state: f32, + + /// For each sample, the output becomes `(state * t) + (target * (1.0 - t))`. This is `t` during + /// the attack portion of the envelope generator. + attack_retain_t: f32, + /// `attack_retain_t`, but for the release portion. + release_retain_t: f32, + + /// The value the envelope follower should try to achieve when not in the release stage. + target_value: f32, + /// Whether the envelope follower is currently in its release stage. + releasing: bool, +} + +impl AREnvelope { + pub fn set_attack_time(&mut self, sample_rate: f32, time_ms: f32) { + self.attack_retain_t = (-1.0 / (time_ms / 1000.0 * sample_rate)).exp(); + } + + pub fn set_release_time(&mut self, sample_rate: f32, time_ms: f32) { + self.release_retain_t = (-1.0 / (time_ms / 1000.0 * sample_rate)).exp(); + } + + /// Completely reset the envelope follower. + pub fn reset(&mut self) { + self.state = 0.0; + self.releasing = false; + } + + /// Only reset the release state, but don't reset the internal filter state. + pub fn soft_reset(&mut self) { + self.releasing = false; + } + + /// Set the maximum value the envelope follower should achieve. + pub fn set_target(&mut self, target: f32) { + self.target_value = target; + } + + /// Get the next value from the envelope follower. + pub fn next(&mut self) -> f32 { + let (target, t) = if self.releasing { + (0.0, self.release_retain_t) + } else { + (self.target_value, self.attack_retain_t) + }; + + let new = (self.state * t) + (target * (1.0 - t)); + self.state = new; + + new + } + + /// Start the release segment of the envelope generator. + pub fn start_release(&mut self) { + self.releasing = true; + } + + /// Whether the envelope generator is still in its release stage and the value hasn't dropped + /// down to 0.0 yet. + pub fn is_releasing(&self) -> bool { + self.releasing && self.state >= 0.001 + } +} diff --git a/plugins/buffr_glitch/src/lib.rs b/plugins/buffr_glitch/src/lib.rs index a3620cbb..ec322008 100644 --- a/plugins/buffr_glitch/src/lib.rs +++ b/plugins/buffr_glitch/src/lib.rs @@ -18,6 +18,7 @@ use nih_plug::prelude::*; use std::sync::Arc; mod buffer; +mod envelope; /// The maximum number of octaves the sample can be pitched down. This is used in calculating the /// recording buffer's size. @@ -37,8 +38,9 @@ struct BuffrGlitch { /// The gain scaling from the velocity. If velocity sensitive mode is enabled, then this is the `[0, 1]` velocity /// devided by `100/127` such that MIDI velocity 100 corresponds to 1.0 gain. velocity_gain: f32, - /// The gain from the gain note expression. - gain_expression_gain: f32, + /// The envelope genrator used during playback. This handles both gain smoothing as well as fade + /// ins and outs to prevent clicks. + amp_envelope: envelope::AREnvelope, } #[derive(Params)] @@ -55,6 +57,14 @@ struct BuffrGlitchParams { #[id = "octave_shift"] octave_shift: IntParam, + /// The attack time in milliseconds. Useful to avoid clicks. Or to introduce them if that's + /// aesthetically pleasing. + #[id = "attack_ms"] + attack_ms: FloatParam, + /// The attack time in milliseconds. Useful to avoid clicks. Or to introduce them if that's + /// aesthetically pleasing. + #[id = "release_ms"] + release_ms: FloatParam, /// The length of the loop crossfade to use, in milliseconds. This will cause the start of the /// loop to be faded into the last `(crossfade_ms/2)` ms of the loop region, and the part after /// the end to be faded into the first `(crossfade_ms/2)` ms of the loop after the first @@ -73,7 +83,7 @@ impl Default for BuffrGlitch { midi_note_id: None, velocity_gain: 1.0, - gain_expression_gain: 1.0, + amp_envelope: envelope::AREnvelope::default(), } } } @@ -103,9 +113,32 @@ impl Default for BuffrGlitchParams { max: MAX_OCTAVE_SHIFT as i32, }, ), + + attack_ms: FloatParam::new( + "Attack", + 2.0, + FloatRange::Skewed { + min: 0.0, + max: 50.0, + factor: FloatRange::skew_factor(-2.5), + }, + ) + .with_unit(" ms") + .with_step_size(0.001), + release_ms: FloatParam::new( + "Release", + 2.0, + FloatRange::Skewed { + min: 0.0, + max: 50.0, + factor: FloatRange::skew_factor(-2.5), + }, + ) + .with_unit(" ms") + .with_step_size(0.001), crossfade_ms: FloatParam::new( "Crossfade", - 0.0, + 2.0, FloatRange::Skewed { min: 0.0, max: 50.0, @@ -160,6 +193,7 @@ impl Plugin for BuffrGlitch { fn reset(&mut self) { self.buffer.reset(); self.midi_note_id = None; + self.amp_envelope.reset(); } fn process( @@ -190,7 +224,9 @@ impl Plugin for BuffrGlitch { } else { 1.0 }; - self.gain_expression_gain = 1.0; + + self.amp_envelope.soft_reset(); + self.amp_envelope.set_target(self.velocity_gain); // We'll copy audio to the playback buffer to match the pitch of the note // that was just played. The octave shift parameter makes it possible to get @@ -201,11 +237,12 @@ impl Plugin for BuffrGlitch { .prepare_playback(note_frequency, self.params.crossfade_ms.value()); } NoteEvent::NoteOff { note, .. } if self.midi_note_id == Some(note) => { - // A NoteOff for the currently playing note immediately ends playback + // Playback still continues until the release is done. + self.amp_envelope.start_release(); self.midi_note_id = None; } NoteEvent::PolyVolume { note, gain, .. } if self.midi_note_id == Some(note) => { - self.gain_expression_gain = gain; + self.amp_envelope.set_target(self.velocity_gain * gain); } _ => (), } @@ -216,10 +253,14 @@ impl Plugin for BuffrGlitch { // When a note is being held, we'll replace the input audio with the looping contents of // the playback buffer // TODO: At some point also handle polyphony here - if self.midi_note_id.is_some() { - // TOOD: This needs to be smoothed, but we this should be part of a proper gain - // envelope - let gain = self.velocity_gain * self.gain_expression_gain; + if self.midi_note_id.is_some() || self.amp_envelope.is_releasing() { + self.amp_envelope + .set_attack_time(self.sample_rate, self.params.attack_ms.value()); + self.amp_envelope + .set_release_time(self.sample_rate, self.params.release_ms.value()); + + // FIXME: This should fade in and out from the dry buffer + let gain = self.amp_envelope.next(); for (channel_idx, sample) in channel_samples.into_iter().enumerate() { // This will start recording on the first iteration, and then loop the recorded // buffer afterwards