IAudioNode

#include "crosspad/audio/IAudioNode.hpp"

Generated from crosspad/audio/IAudioNode.hpp in crosspad-core. Edit the header’s doxygen comments to change this page.

Types

crosspad::IAudioNode

A DSP unit in the audio module’s node chain.

Nodes operate on interleaved stereo float32 buffers in [-1.0, 1.0]. Generators (synths, samplers) override mix() to add into the bus buffer. Effects override process() and may be in-place (in == out is allowed).

── Lifecycle ─────────────────────────────────────────────────────────────

• Constructed on any thread.
• Registered into a module/mixer (addNode / addChannel).
• Module/mixer calls onPrepare(maxFrames, sampleRate) once during setup and again on any restart-time change. This is where SR-dependent state (filter coefficients, delay lines, scratch buffers sized to maxFrames) must be allocated.
• Audio thread calls mix() / process() repeatedly, never beyond maxFrames.

── RT-safety contract for mix() / process() ───────────────────────────── These run on the audio thread and MUST be real-time safe:

• No heap allocation (no new, no vector::resize, no string formatting).
• No blocking primitives (no mutex, no condvar, no ring-buffer wait).
• No syscalls / logging / file I/O.
• No printf-family or any unbounded code path. Communicate with non-audio threads via std::atomic on POD scalars or lock-free SPSC queues. Any setup-time work belongs in onPrepare().

setEnabled() / enabled() and any setter exposed to UI code may be called from any thread; back them with std::atomic.

Members

Member	Description
process	Effect-style processing.
mix	Generator-style mixing.
isGenerator	Marks this node as a generator (mix() called) vs effect (process() called).
name
enabled
setEnabled
onPrepare	Called by the host (IAudioModule or AudioMixerEngine) once both the audio buffer size and sample rate are known, and again on any restart-time change.
onSampleRateChange	Legacy SR-only notification.

Methods

process

virtual void process(const float *in, float *out, uint32_t frames)

Effect-style processing.

Interleaved stereo [L,R,L,R,…] in [-1.0, 1.0]. Default impl is pass-through; override for actual DSP. RT-safe contract applies — see class doc.

mix

virtual void mix(float *, uint32_t)

Generator-style mixing.

Adds into existing bus buffer; default no-op. RT-safe contract applies — see class doc.

isGenerator

virtual bool isGenerator() const

Marks this node as a generator (mix() called) vs effect (process() called).

name

virtual const char * name() const =0

enabled

virtual bool enabled() const

setEnabled

virtual void setEnabled(bool)

onPrepare

virtual void onPrepare(uint32_t, uint32_t sampleRate)

Called by the host (IAudioModule or AudioMixerEngine) once both the audio buffer size and sample rate are known, and again on any restart-time change.

Allocate scratch buffers sized to maxFrames * 2 (stereo) here so subsequent mix()/process() calls stay alloc-free. Default impl forwards to legacy onSampleRateChange() for nodes that only care about sample rate.

onSampleRateChange

virtual void onSampleRateChange(uint32_t)

Legacy SR-only notification.

Prefer overriding onPrepare() for new nodes that need the buffer size for pre-allocation.

Declared in crosspad/audio/IAudioNode.hpp:36.