aufklarer / Sidon-ONNX

Sidon — ONNX (speech restoration / dereverberation)

Cross-platform speech restoration (denoise + dereverberation + bandwidth restoration) exported to ONNX, for ONNX Runtime on Linux / Windows / Android / macOS. Turns a noisy/reverberant clip into studio-quality 48 kHz speech — ideal for cleaning a voice-cloning reference before TTS (it preserves speaker identity).

Two-stage pipeline:

16 kHz audio → [w2v-BERT log-mel front-end] → predictor (w2v-BERT 2.0, 8 layers)
            → cleansed features [1, T, 1024] → DAC decoder → 48 kHz audio

Variants

Total 246 M params (predictor 193.6 M + DAC vocoder 52.4 M). Output 48 kHz. FP16 was produced with ONNX Runtime's transformer-aware converter (attention stays FP32). No int8 build: dynamic INT8 corrupts this conformer on the ORT CPU EP ( ConvInteger unsupported; MatMul-only int8 drops SI-SDR to −2.8 dB). Use fp16 for a compact model; on Apple, the CoreML build has a working int8.

Files

Quality (no-reference MOS)

FP32 is bit-exact to the PyTorch reference; FP16 is near-lossless (waveform cos 0.999).

Restoration lifts OVRL 2.90 → 3.29, driven by BAK 3.40 → 4.11 (reverb removed).

Front-end

The graphs take input_features [1, T, 160] from the w2v-BERT 2.0 SeamlessM4T feature extractor (16 kHz). The ONNX graphs use a dynamic time axis (chunk in the runtime). Front-end + chunking are handled by the consuming engine ( speech-core ).

Usage

import onnxruntime as ort, numpy as np
pred = ort.InferenceSession("fp16/sidon-predictor.onnx")
voc  = ort.InferenceSession("fp16/sidon-vocoder.onnx")
# input_features [1, T, 160] from the w2v-BERT 2.0 feature extractor (16 kHz)
feats = pred.run(None, {"input_features": input_features})[0]
audio = voc.run(None, {"features": feats})[0]   # [1, M] @ 48 kHz

Source

Exported from Sidon (sarulab-speech), checkpoint sidon-v0.1 ; paper arXiv:2509.17052 . Base SSL encoder: facebook/w2v-bert-2.0 ; vocoder: DAC (descript-audio-codec). All components are MIT-licensed.

Links

• speech-core — C++ engine (ONNX backend)
• speech-android — Android SDK
• soniqo.audio — website
• blog