Which Polaxis Talking Machine Is For You?

image_pdfimage_print

Polaxis was born from a simple obsession: making machines talk—inspired by Kraftwerk and sparked by discovering the
SPO256 AL2.

Back in the 80s, storing real human voice in small devices wasn’t an option. If you wanted a talking object, you had
to synthesize the voice in real time instead of just playing samples. That synthetic, slightly robotic voice has its
own charm – especially when you can bend pitch, speed, accents, and make it react to what’s happening in your
system.

Today, computers can speak and sing with realistic prosody, but embedding a flexible speech synthesizer in an
instrument, a Eurorack case, or an art installation is still not trivial. Polaxis focuses exactly on that: bringing
speech synthesis to embedded electronics and, in particular, to the Eurorack modular world.

This page will help you choose the right combination of host (Talko, Emy, Terminal, Kraftor) and voice modules (Vox
boards) for your project.

1. Start from what you want to do

Pick the sentence that sounds most like you. Each one points to one or more Polaxis options.

  • “I want a simple, characterful Speak & Spell–style voice in my rack or on my desk.”
    • Start with: Talko 2 (Eurorack or Bricky standalone).
  • “I want a deep, patchable Eurorack speech module with menus, SD card, and a big vocabulary.”
  • “I want a desktop talking box, not a Eurorack module.”
    • Start with: Emy or Talko 2 in Terminal desktop case (Emy Terminal
      sold out for now, Talko Terminal coming soon).
  • “I want to control speech mainly from MIDI (DAW, sequencer, keyboard).”
  • “I want to explore different synthetic voices (Dectalk, French formant, Japanese, Chinese, etc.)”
    • Start with: Emy or Kraftor as a host, then add Vox boards (Vax, Mea, Kaiwa, Sino, Nexus, SpeakJet, SPO, Robo…).
  • “I want a future-proof Eurorack module with FX and MIDI on board.”
    • Watch for: Emy 2 (coming soon), which aims to extend Emy with onboard FX and TRS MIDI.

You can always combine several: for example, Talko 2 in your rack for instant Speak & Spell style sounds, plus a Kraftor + Vox combo on your desk for text-to-speech in multiple languages.

2. A quick primer on speech synthesis (Polaxis style)

When most people think “talking machine,” they think of samples: short recordings of words played back from memory.
Polaxis focuses instead on synthetic voices generated in real time.

  • Vocoder: A vocoder takes an existing audio signal (like a synthesizer) and shapes it using the spectral characteristics of another signal (the modulator, usually your voice).
  • Speech synthesis: On the other hand, generates speech from scratch—either from text
    (TTS) or from phoneme sequences—without needing an external audio source.
    Polaxis products are mainly speech synthesizers: they create synthetic voices internally, though some (like Robo Vox) can also work in vocoder mode with an external carrier signal.
  • Text-to-speech (TTS)
    Convert plain text into speech using a built-in voice engine. Great for reading sentences, doing digital poetry,
    or generating phrases on the fly.
  • Phoneme / allophone–based speech
    Build words from small sound units (phonemes). This gives a more robotic character (classic chips like SPO256,
    SpeakJet, etc.) because of the lack of prosody—the natural rhythm, stress, and intonation of speech. Perfect for
    glitch, chopped speech, and experimental vocals.
  • Formant and glitch / drones
    Some voice engines can be pushed away from “clear speech” into singing, drones, or noise-like textures. This is
    where the synthesizer behaves like an instrument, not just a talking box.

Why it matters in modular / embedded world:

  • Everything is real time: you can modulate pitch, speed, accent, or FX with CV or MIDI while the module is speaking.
  • The voice can react to gates, triggers, sensors, or conditions in your system.
  • You can mix intelligible phrases with totally alien textures using the same hardware.

Polaxis products are built around this idea: small embedded machines hosting one or more synthetic voices that you can play like instruments.

3. The Polaxis ecosystem: hosts and voices

Think of the ecosystem in two layers:

  • Hosts = the “brains” and user interface
    These are the modules or boxes you actually patch or plug into:

    • Talko 2 (available as Eurorack, Bricky standalone,
      or Terminal desktop case)
    • Emy (available as Eurorack or
      Terminal desktop case)
    • Kraftor (MIDI host)
    • Emy 2 (upcoming evolution of Emy)
  • Vox boards = the “voices”
    Small hardware speech modules you plug into a compatible host (Emy, Kraftor). Each Vox has its own language, character, and synthesis engine.

A simple way to see it:

  • If you want “one self-contained talking module” with a fixed voice and a simple workflow → Talko 2.
  • If you want “a platform that can host many different voices” → Emy, or Kraftor plus Vox boards.

4. Product overview – which host is for you?

4.1 Talko 2 – the playful entry point

Talko 2 is the new generation of the original Talko. It focuses on the classic Speak & Spell–style experience, but powered by a modern MCU (RP2040), so you can also apply real-time effects. Talko 2 emulates the TMS5220 and TMS5100 speech synthesis chips.

Key ideas:

  • Speaks numbers, alphabet, phonemes, and classic phrase banks.
  • Real-time control of pitch, speed, and FX (via knobs and an “Error” / glitch button).
  • A 7-segment display shows the current bank or effect.
  • Available as:
    • Eurorack module (for your rack).
    • “Bricky” standalone version (inspired by Error Instruments), for use on a desk or with pedals.
    • Terminal desktop case (coming soon), a robust desktop box with jacks, MIDI, and power.

Choose Talko 2 if you want:

  • Instant fun, robotic speech with hands-on controls.
  • A simple way to add a talking voice to your modular or desktop setup.
  • An entry-level Polaxis product with a strong personality.

4.2 Emy – the modular speech workstation

Emy is a multi-voice Eurorack speech synthesizer. It
combines:

  • A software speech engine (TMS5220-style LPC synthesis) reading compressed data from an SD card.
  • A rear mikroBUS connector to host hardware Vox
    boards    .
  • Real-time control of speech parameters via knobs and CV for each main control.
  • An OLED display and menu system to browse speech files and settings.
  • USB connectivity for firmware updates, serial communication, and even USB keyboard text entry.

Emy is available as a Eurorack module, and also in the Terminal desktop case (sold out for now), which provides the same functionality in standalone desktop box with CV and MIDI options.

Typical use cases:

  • Complex spoken phrases, texts, and vocabularies from SD card.
  • Live performance where you modulate pitch, speed, stretch, and energy via CV.
  • Experimental setups where you patch speech like any other audio source in your rack.

Choose Emy if you:

  • Work in Eurorack and want one main “brain” to handle multiple voices.
  • Need SD-based vocabularies and menu navigation.
  • Want a front-panel interface, not just MIDI.

4.3 Kraftor – MIDI host for all Vox boards

Kraftor is a compact MIDI instrument designed to host the entire
Polaxis voice board catalog without any need for
Eurorack.

Key features:

  • Hosts all Polaxis Vox boards via a mikroBUS
    port.
  • USB connection for:
    • Power
    • Firmware updates
    • MIDI control
    • Serial communication
  • TRS MIDI (type B) input for external sequencers, keyboards, or DAWs.
  • Built-in amplifier with volume control, plus speaker out and line out jacks.
  • Grove connector with I2C isolator for extras (OLED, sensors, NFC text reader).

Choose Kraftor if you:

  • Want to drive voices mainly from MIDI (notes, CCs, sequences).
  • Work in a DAW or with hardware sequencers rather than CV/gate.
  • Need a small standalone box that can talk, sing, and glitch without a modular system.

I can also build custom Kraftor-based instruments with a keyboard, display, speaker, or any I2C interface on demand.

4.4 Emy 2 – the next step

Emy 2 is the upcoming evolution of Emy. The goal is
simple:

  • Do everything Emy 1 does (Eurorack host with
    SD card, menu, Vox compatibility),
  • Add onboard FX (for more sound design possibilities),
  • Add TRS MIDI directly on the module.

If you’re starting from scratch and want a “do everything” Eurorack speech hub with MIDI and FX built in, Emy 2 is the one to watch.

5. Voice modules (Vox) – choosing your synthetic
voice

Vox boards are small voice modules you plug into Emy or Kraftor. Each one is a different personality:

  • Some focus on text-to-speech in English, Spanish, Chinese…
  • Some are phoneme-based, with 8-bit metallic voices.
  • Some can do singing, chant-like tones, drones, and glitch.

 

  • Vax Vox – DECTalk text-to-speech with multiple voices, capable of singing lines with expressive prosody.
  • Mea Vox – French formant generator for phoneme-based speech, drones, bass notes, and rhythm-locked vowels.
  • Kaiwa Vox – Japanese romaji speech from SD card files, with chant-like elongation and different character chips.
  • Sino Vox – Chinese and English TTS with five unique voice types.
  • SpeakJet Vox – English allophones plus sound effects for complex, synthetic phrases.
  • SPO Vox – the vintage SP0256-AL2 chip for classic 8-bit metallic phonemes.
  • Nexus Vox – RC8660-based English TTS with control over speed, volume, pitch, expression, and reverb.
  • Robo Vox – The famous Kraftwerk Speech engine with the option to use an external audio carrier for extra effects

You can start with one Vox that matches your project (for example, Vax for text-to-speech or Mea for French phonemes), and later add others to expand your palette.

6. Still unsure? Ask for a recommendation

If you are not sure which combination fits your project:

  • Tell me if you work mainly in Eurorack, desktop synths, DAW, or installations.
  • Tell me if you prefer text-to-speech, phoneme-based voices, or more glitch/drone textures.
  • Mention if you need specific languages.