TD-PSOLA

Time-domain
pitch shifting | time stretching | formant shifting
Built on Time-Domain Pitch-Synchronous Overlap and Add (TD-PSOLA)

---

✨ What is this?

A small experimental TD-PSOLA engine that can:

• Pitch shift
• Time stretch
• Preserve/adjust formants

---

🧪 Results & Experiments

input | 12 semitones up | 12 semitones down

drive distortion | fry distortion

_{for more experiments, please check audio folder}

---

⚙️ Installation

Dependencies:

pip install numpy soundfile praat-parselmouth

Note: praat-parselmouth is used for pitch estimation and epoch marking

---

🎈 Usage

Example: 12 semitones shift

import soundfile as sf
from td_psola import td_psola # or paste the function into your script

audio, sr = sf.read("input_audio.wav")

result = td_psola(
    audio, sr,
    pitch_semitones=12,
    stretch_factor=1.0,
)

sf.write("output_audio.wav", result, sr)

---

_{made with questionable sleep schedules please I am suffering}