Gridspace MIT IAP 2025

building sound:
sound & speech synthesis

Gridspace presents a lecture series on the fundamentals of constructing sound with machines.

January 2024
monday
tuesday
wednesday
thursday
friday
6
Bytes to Beats
Engineering a song from scratch.
7
8
First Vibrations
Acoustics and signals.
9
Harmonic Scaffolding
Additive Synthesis. Modular synths. Special guests: Noise Engineering
10
13
Wave Sculpting
Subtractive synthesis.
14
15
Atoms of Sound
Wavetable, granular, formants, sampling.
16
17
FM Synthesis
Special guest: John Chowning (Stanford).
20
Physical Modelling
Special guest: Ange Yaghi.
21
22
Atoms of Speech
Parametric and concatenative TTS.
23
Callbacks
History of speech synthesis feat. Gene Frantz (TI) and TBA.
24
27
Learning to Speak
Early neural TTS. Seq2seq models. WaveNet.
28
29
Neural Machines
Tacotron and spectrogram synthesis. Vocodors and spectrogram inversion.
30
31
Modern TTS
Variational methods. Few shot voice replication. VITS.

About Gridspace

Gridspace is a speech and language technolgy company that teaches machines to listen and speak. We process and automate billions of minutes of speech data per year.

Questions, queries, concerns and excitement can be directed to [email protected].

Last year's IAP lectures can be found here.

Wanna come hang out?

Links to register for in-person IAP lectures are down below.

Lecture 1: Register here

Lecture 2: Register here

Lecture 3: Register here

Lecture 4: Register here

Lecture 5: Register here

Lecture 6: Register here

Lecture 7: Register here

Lecture 8: Register here

Lecture 9: Register here

Lecture 10: Register here

Lecture 11: Register here

Lecture 12: Register here