A Colombian Spanish Voice Pair for OVOS

Spanish text-to-speech voices are not rare. Voices that sound like Colombia rather than Madrid are rare. A member of the Colombian Spanish community asked for one. Two are the answer: phoonnx_es-CO_miro_espeak (male) and phoonnx_es-CO_dii_espeak (female). They are the same donor-voice pair OVOS uses for its other languages.

What they are

Both are VITS models trained with phoonnx, exported to ONNX, and phonemized with espeak-ng, which is what the _espeak suffix means. They synthesize on a CPU without a GPU or a network connection. Each model repository carries the .onnx weights, the phoonnx JSON config, a Piper-format config and the token table.

The training data is public: TigreGotico/tts-train-synthetic-miro_es-CO and tts-train-synthetic-dii_es-CO. Miro and Dii are the recorded voices of two real people, and the voice identities belong to TigreGótico. The models are published under CC BY-NC-ND 4.0: non-commercial use, unmodified, with attribution. Commercial use or a derivative voice needs an agreement with TigreGótico.

Try it

Install phoonnx with pip install --pre phoonnx (the run below is 1.91.2a1). The espeak-ng binary must be on the machine, because the phonemizer calls it. Then:

import wave
from huggingface_hub import hf_hub_download
from phoonnx.config import SynthesisConfig
from phoonnx.voice import TTSVoice

repo = "OpenVoiceOS/phoonnx_es-CO_miro_espeak"
voice = TTSVoice.load(hf_hub_download(repo, "miro_es-CO.onnx"),
                      hf_hub_download(repo, "miro_es-CO.json"))
config = SynthesisConfig(noise_scale=0.667, length_scale=1.0, noise_w_scale=0.8)
with wave.open("es_co.wav", "wb") as f:
    voice.synthesize_wav("Buenos días, ¿cómo amaneció?", f, config)

es_co.wav: RIFF (little-endian) data, WAVE audio, Microsoft PCM, 16 bit, mono 22050 Hz

In OVOS, the ovos-tts-plugin-phoonnx plugin ships inside the phoonnx package. Point it at either voice in mycroft.conf:

{
  "tts": {
    "module": "ovos-tts-plugin-phoonnx",
    "ovos-tts-plugin-phoonnx": {
      "voice": "OpenVoiceOS/phoonnx_es-CO_dii_espeak"
    }
  }
}

Limits

The voices are trained on synthetic data built from a donor voice. They carry the donor's timbre with Colombian pronunciation, not a Colombian speaker's own voice. Pronunciation quality is bounded by espeak-ng's Spanish rules. A word espeak-ng gets wrong, the voice gets wrong. There is no hosted demo, and the public TTS server serves en-US only, so the snippet above is the way to hear them. Pronunciation feedback and word lists go to the phoonnx issue tracker.

This work is part of the OpenVoiceOS From Beta to Breakthrough milestone, funded through the NGI0 Commons Fund, a fund established by NLnet with financial support from the European Commission's Next Generation Internet programme, under the aegis of DG Communications Networks, Content and Technology under grant agreement No 101135429. Additional funding is made available by the Swiss State Secretariat for Education, Research and Innovation (SERI).

Help Us Build Voice for Everyone

OpenVoiceOS is more than software, it's a mission. If you believe voice assistants should be open, inclusive, and user-controlled, here's how you can help:

💸 Donate: Help us fund development, infrastructure, and legal protection.
📣 Contribute Open Data: Share voice samples and transcriptions under open licenses.
🌍 Translate: Help make OVOS accessible in every language.

We're not building this for profit. We're building it for people. With your support, we can keep voice tech transparent, private, and community-owned.

👉 Support the project here