Speaking of Witch

What it does

A sampler that builds its sounds from the speech synthesis already on the device. Type a word or a whole sentence into one of sixteen pads, pick a voice from the system language list, and it renders a sample to play from the pads or a keyboard. Each pad keeps its own text, voice, speed, pan, gain, colour and MIDI channel, with trim handles, one shot, reverse, loop and ping pong playback, and drawable pitch and amplitude curves over the phrase.

In the demo

Jade works in standalone and then in AUM, resynthesising every pad to a single voice, drawing pitch bends across a sentence, and reversing and looping words. She splits the pads across multi outs so each one takes its own effects, then builds a board live from phrases the chat types in.

Worth knowing

The voices come from the operating system, so the set depends on what is installed. Jade hit a bug where some British voices did not load in the AUv3, which she had reported to the developer.