Grustnify

Grustnify (from the Russian word "грусть" — sadness) is an audio-processing tool that transforms any audio file into a “sadified” version of itself. The application slows down audio, lowers the pitch, adds spacious reverb, and exports the processed result in the original format (typically WAV).

The project is implemented in C++, using:

• FFmpeg for decoding and encoding
• Qt for UI and application framework
• Custom DSP algorithms for reverb and time-stretching
• GoogleTest for unit testing
• spdlog for logging

---

Features

✔ Decode audio using FFmpeg

Supports common audio formats such as WAV, MP3, FLAC, AAC, etc.

✔ Apply DSP effects

• Reverb (Schroeder reverb architecture: comb filters + allpass filters)
• Slow-down with pitch drop (time stretching resampling)

✔ Encode the processed audio

Currently outputs to WAV (float PCM).

✔ Simple UI integration with Qt

Load → Process → Save workflow.

✔ Tested via GoogleTest

Audio decoding and processing are unit-tested using generated or stored WAV samples.

---

Processing Pipeline

The audio is transformed through the following stages:

1. Decode → AudioBuffer The input file is decoded into an interleaved floating-point sample buffer.

2. Reverb A room-like reverb tail is generated using parallel comb filters followed by allpass diffusion.

3. Slowdown (Pitch Down) A simple resampling-based time-stretch algorithm stretches the audio by a factor (e.g. ×1.5).

4. Encode & Save The processed audio is written back as a WAV file, using the same sample rate and channel layout.

---

Example Result

Input:

song.mp3

Output:

song_grustnified.wav

A slower, deeper, reverberated version designed to sound melancholic or atmospheric.

---

Build Instructions

Requirements

• C++20 compiler
• CMake
• FFmpeg development libraries
• Qt 6 (Widgets)
• GoogleTest (automatically fetched or system-installed)
• spdlog

Build

cmake -B build
cmake --build build

Run tests

./build/tests/run_tests

---

Project Structure

src/
  core/
    audio_decoder.cpp/hpp
    audio_encoder.cpp/hpp
    audio_buffer.cpp/hpp
    reverb / time-stretch algorithms
  app/
    app.cpp/hpp
    main.cpp
  ui/
    main_window.cpp/hpp
  log/
    log.cpp/hpp

tests/
  test_audio_decoder.cpp
  data/
    sine_440.wav

---

DSP Overview

Reverb

Based on the classic Schroeder reverberator:

• 4 parallel comb filters (≈30–45 ms delays)
• 2 serial allpass filters
• Mix, room-size, and damping parameters

Produces a smooth, diffuse reverb tail.

Slowdown + Pitch Drop

Simple resampling approach:

[ y(n) = x(n / \text{speed}) ]

Result:

• longer audio
• lower pitch
• more melancholic tone

---

Roadmap

• MP3/OGG encoding support
• Adjustable effect parameters in the UI

---

Author

Created as an experimental project to explore FFmpeg, DSP, Qt development, and emotional audio transformation.