Convolution Reverb on a Daisy Seed

by

for Proxima

a new work

Here ideas riffing on convolution reverb and Daisy Seed.

Prompt “generate code for a daisy seed for a convolution reverb”

See also https://electro-smith.com/products/daisy-seed

Haven’t tried below but looks plausible.

Imagine microphone input devices hanging in a gallery. Upon entering a gallery, a visitor scans a QRCode on their smartphone, selects some signal generation, hold the output of the phone to the device in the gallery that is connected to specialized loudspeakers.

Convolution Reverb Code Generated using Google AI

Implementing a convolution reverb on the Daisy Seed requires utilizing SDRAM for the long impulse response (IR) samples and efficient DSP for processing, typically found in the library’s .

Below is a foundational C++ code structure using and to create a mono-to-stereo convolution reverb.

Prerequisites

  • DaisyToolchain installed.
  • An Impulse Response WAV file (16/24-bit PCM mono) placed in the project folder.

Convolution Reverb Code ()

#include "daisy_seed.h"
#include "daisysp.h"
#include <string>

using namespace daisy;
using namespace daisysp;

// -- Configuration --
#define IR_SIZE 48000 // 1 second at 48k
// Allocate IR in SDRAM
float DSY_SDRAM_BSS ir_buffer[IR_SIZE]; 
float DSY_SDRAM_BSS conv_buffer[IR_SIZE * 2]; // For processing

// -- Objects --
DaisySeed hw;
FastConvolver l_conv, r_conv;
Parameter drywet_param;
float drywet_val;

// -- Audio Callback --
void AudioCallback(AudioHandle::InputBuffer in, AudioHandle::OutputBuffer out, size_t size)
{
    drywet_val = drywet_param.Process();

    for (size_t i = 0; i < size; i++)
    {
        // Get input, apply drywet
        float dry = in[0][i];
        
        // Process Convolution
        float wet = l_conv.Process(dry); 

        // Output Wet/Dry mix
        out[0][i] = (dry * (1.0f - drywet_val)) + (wet * drywet_val);
        out[1][i] = (dry * (1.0f - drywet_val)) + (wet * drywet_val);
    }
}

int main(void)
{
    // -- Initialize Hardware --
    hw.Init();
    hw.SetAudioSampleRate(SaiHandle::Config::SampleRate::SAI_48KHZ);
    float samplerate = hw.AudioSampleRate();

    // -- Load IR (Placeholder for actual WAV loading logic) --
    // In a real project, use FatFs to load WAV into ir_buffer
    for(int i=0; i<IR_SIZE; i++) ir_buffer[i] = 0; 
    ir_buffer[0] = 1.0f; // Impulse (dummy)

    // -- Initialize Convolvers --
    l_conv.Init(ir_buffer, IR_SIZE, conv_buffer);
    r_conv.Init(ir_buffer, IR_SIZE, conv_buffer + IR_SIZE);

    // -- Initialize Parameters --
    drywet_param.Init(hw.adc.GetPtr(0), 0.0f, 1.0f, Parameter::LINEAR);

    // -- Start Audio --
    hw.StartAudio(AudioCallback);

    while(1) {}
}

Key Implementation Details

  1. SDRAM Usage: Convolution requires long buffers. The .
  2. : provides a .
  3. Impulse Response Loading: The example uses a dummy impulse ( at index 0). To make this functional, you must use the library to read a WAV file from an SD card into the during .
  4. Wet/Dry Mix: The output is a mix between the original signal () and the convolved signal (.

To Compile and Flash

  1. Create a new project folder in .
  2. Add the (copied from another example).
  3. Add the code above.
  4. Run and .

AI responses may include mistakes.

Other Links (not from AI)

Tags