هـ
هَضْرَةHADRA AI 🇲🇦
Official Technical Specification

HADRA-TTS (v1.0 Specification)

A Dual-Tier Moroccan Darija Speech Architecture combining Flow-Matching DiT for studio-grade acoustic fidelity, VITS ONNX for sub-50ms conversational turns, and 0ms pre-synthesized Opus caching.

1. The Three-Tier Routing System

Rather than forcing a single model across contradictory latency-quality requirements, HADRA implements a coordinated 3-tier routing engine:

Tier 0: Instant Phrase Cache0.01 ms

24 pre-synthesized Opus clips at NFE 32. Fuzzy-matched (RapidFuzz ratio ≥ 90%) for instant conversational greetings, confirmations, and fillers.

Tier 1: F5-TTS Flow-Matching DiT (Quality Engine)Studio Grade

336M Diffusion Transformer, 22 layers, 1024-dim, Vocos 24kHz. Inpainting mel canvas conditioned on curated 4-6s golden reference clips.

Tier 2: Multi-Speaker Piper-TTS / VITS (Latency Engine)RTF 0.05 (30ms)

Single feedforward pass, ONNX-native, zero ODE numerical integration. Dedicated to real-time voice-to-voice conversation loops.

2. Flow-Matching ODE & The NFE-16 Sweet Spot

In Flow-Matching Conditional Vector Fields, the audio trajectory is integrated via ODE. While NFE=8 causes severe under-integration of fine spectral detail (slurring pharyngeals and uvulars like ح, خ, ق, ع), NFE=16 with Sway Sampling achieves ~95% of NFE-32 fidelity:

# Optimal Inference Recipe
nfe_step = 16
sway_sampling_coef = -1.0 # Front-loads early trajectory curvature
cfg_strength = 2.0
truncated_cfg_ratio = 0.55 # Disables CFG after 55% steps to cut compute 28%

3. DODa Dataset & Canvas Infill Hygiene

Unlike cross-attention models, F5-TTS prepends reference audio directly into the mel-canvas and text into character embeddings. Therefore, prompt hygiene directly determines synthesis stability:

  • Golden Length: 4.0 to 6.0 seconds (below 2.5s timbre destabilizes; above 8s wastes compute).
  • Silence Trimming: Hard trimming (< 50ms leading/trailing) to eliminate hallucinated pauses.
  • SNR & Reverb: Strict studio SNR > 20 dB with zero room reflections.
  • Intonation: Natural falling sentence cadence at medium tempo.

4. Real CPU Inference Benchmarks

Engine / PrecisionNFE StepsRTFQuality Grade
PyTorch FP32 BaselineNFE = 3210.60Studio (Pristine)
⚡ AMX-BF16 + Sway (-1.0)NFE = 163.40 (~14s)Studio (95% of NFE-32)
🚀 Piper VITS ONNX1 Pass< 0.05 (30ms)Realtime Conversational
⚡ Instant Opus Cache00.00 (0.01ms)Lossless Master

5. Citation & BibTeX

@misc{hadra_darija_tts_2026,
  title={Habibi-TTS-DODa-Darija: Open-Source Moroccan Darija Flow-Matching Voice Model},
  author={Hadra Open Initiative and Contributing Medical Student},
  year={2026},
  publisher={Hugging Face},
  howpublished={\url{https://huggingface.co/Jip7e/habibi-tts-doda-darija}},
  note={Fine-tuned on DODa Moroccan Arabic Dataset (12,105 Studio Utterances)}
}