A (sometimes) manually controlled speech-like sound generator
Based on ESP32 with external UDA1334A DAC. Very much active work in progress. It is basically working, but changing day-to-day.
These docs in need of an update!
Slightly older prototype demo/intro (YouTube video) :
Not long ago I rediscovered an electronics magazine article I first saw whilst still at school, around 1979 : Chatterbox. It was a design for a simple speech synthesizer, controlled by means of a joystick, switches etc. I wanted to build one then; now ever more so.
The (mighty elegant) implementation then was with a handful of op amps & CMOS logic ICs, now I'd like to try something similar based around microcontroller(s). Also ties in nicely with other projects I have on the go.
The human interface is a very appealing part of the original Chatterbox. So joystick, bunch of switches, a few pots. Use an inexpensive microcontroller to do the work. But given such a setup, it should be straightforward to add other interfaces, notably Web (via WebSockets) and MIDI.
Work-in-progress
Changing on a daily basis. Check Quasi-blog Notes.
I built up the hardware & software to proof-of-concept, it can work. The hardware is pretty close to being finalised. The software kind-of works, but is rather a mess, now refactoring to allow later experimentation to figure out a good final setup.
2020-05-30 version chatterbox_1.1.0
Spent a lot of time transferring everything into a permanent pretty box (pictured above). Have made some progress in refactoring. Soon should be able to get back to the fun side of the coding.
2020-04-15 version chatterbox_1.1.0
I'd got part way through implementing the Web interface and the code was getting way too messy to work with. So I'm currently refactoring everything into more manageable C++ classes. To help with this I've moved from the Android IDE to PlatformIO (on VSCode). I should have done this weeks ago!
Also, although this stuff is all here on GitHub, I haven't really thought in terms of releases yet. So far the code versioning has been ad hoc, really just for backup purposes. But this PlatformIO environment is so much better, once I've done the refactoring and cleaned up minor bugs I can easily shift to doing proper releases (with Semantic Versioning).
2020-04-03 version chatterbox_1.0.18 many hours of work later. Added basic WebSocket comms, started a browser-based UI. Added a bunch more switches to the hardware, now 4 toggle, 8 push, maxing out the ESP32 I/O except for a UART I'm reserving for MIDI and also there's a free I2C I/O which may go to a display, memory card, other physical I/O, who knows..?
2020-03-17 version chatterbox_1.0.10 work on
2020-03-11 : mini-milestone! Added a minimal Web server.
2020-03-10 : Milestone! All core functionality basically working, see video above. version chatterbox_1.0.7
2020-03-05 : All available pots set to filter controls, using a state variable. Between stages have added soft clipping (tanh), works a treat!
2020-03-03 : Milestone! first formant filter essentially working.
2020-02-02 : I've got pot 5 controlling the pitch of a wavetable-generated larynx-like tone. Pot 4 controls the waveform. (Switch 0 is hooked up to change between test waveforms). First attempts at a filter, cutoff freq on pot 3. Something very wrong - it's industrial noises in a box. Well, pulsed farty noises...
2020-02-29 : I now have the main controls hooked into the code : joystick, 4x potentiometers, 5x push-button switches. Of these only 2 pots and 1 switch are actually doing anything.
2020-02-26 : First prototype case with controls put together (Front panel v1.0). Digging into software.
2020-02-21 : First sounds from core hard/software (stereo 7kHz sine from wavetable @ 44.1kHz, 16 bit). This repository created.