Kokoro Grounded Documentary Voiceover Model

A broadcast-grade text-to-speech engine powered by Kokoro-ONNX, featuring the Grounded Documentary voice profile (55% bm_george + 45% bm_daniel, British RP).

Tuned specifically for long-form documentary, philosophical, and educational narration with:

Deliberate Pacing: Unhurried speed (0.92), natural sentence pauses (0.42s), and extended pauses (1.25s–1.35s) across major thematic shifts.
Grounded Tone: Deep, comforting lower-mid chest resonance preventing listening fatigue over long sessions.
Subtle Modulation: Narrative emphasis carried via pitch inflection rather than jarring volume spikes.
Targeted De-Essing & Mastering: Built-in broadcast mastering chain cutting harsh sibilance on high frequencies ('s', 'sh') with EBU R128 (-16 LUFS) loudness normalization.

Quickstart (Python)

pip install kokoro-onnx soundfile huggingface_hub

import numpy as np
import soundfile as sf
from kokoro_onnx import Kokoro
from huggingface_hub import hf_hub_download

REPO_ID = "gitmodelmujtaba/kokoro-grounded-documentary"

# 1. Download model weights and tuned voice profile
model_path = hf_hub_download(repo_id=REPO_ID, filename="kokoro-v1.0.onnx")
voices_path = hf_hub_download(repo_id=REPO_ID, filename="voices-v1.0.bin")
profile_path = hf_hub_download(repo_id=REPO_ID, filename="grounded_documentary.npy")

# 2. Initialize Kokoro
kokoro = Kokoro(model_path, voices_path)
voice_style = np.load(profile_path)

# 3. Generate Speech
text = "Look up at the night sky... To you, it might look like an endless ocean of cold vacuum."
audio, sr = kokoro.create(
    text,
    voice=voice_style,
    speed=0.92,
    lang="en-gb",
    sentence_pause=0.42,
    clause_pause=0.20
)

sf.write("output.wav", audio, sr)
print(f"Generated {len(audio)/sr:.2f}s of broadcast narration!")

Broadcast Mastering Chain (FFmpeg)

For broadcast-grade warmth and smooth de-essing matching YouTube documentary standards:

ffmpeg -y -i output.wav -af \
  "highpass=f=80,\
   equalizer=f=160:t=q:w=1.0:g=2.8,\
   equalizer=f=260:t=q:w=1.2:g=2.0,\
   equalizer=f=450:t=q:w=2.0:g=-2.0,\
   equalizer=f=3200:t=q:w=2.0:g=-1.6,\
   equalizer=f=5800:t=q:w=2.5:g=-3.5,\
   equalizer=f=7200:t=q:w=2.5:g=-3.0,\
   lowpass=f=11000,\
   deesser=i=0.45:m=0.5:f=0.55:s=o,\
   compand=attacks=0.03:decays=0.3:points=-80/-80|-35/-25|-16/-13|0/-3,\
   loudnorm=I=-16:TP=-1.5:LRA=8.5" \
  -c:a libmp3lame -b:a 192k output_mastered.mp3

Files in this Repository

kokoro-v1.0.onnx: Kokoro ONNX model weights (310 MB).
voices-v1.0.bin: Kokoro voice library (27 MB).
grounded_documentary.npy: Pre-blended float32 voice embedding vector (510 KB).
narrator.py: Self-contained Python helper script for single-command generation.

Downloads last month: -; Downloads are not tracked for this model. How to track