DotCheck/helmholtz-audio-v2

Apache-2.0 audio AI-likeness head for DotCheck. This repo includes the live .npz head, model card, license, and notices.

Field	Value
Hub id	DotCheck/helmholtz-audio-v2
Wire id	inhouse-audio@2
Label	Helmholtz
Artifact	dasheng_base_linear_head_audio_v2.npz
Backbone	mispeech/dasheng-base (Apache-2.0)
Protocol	4 s / center crop / up to 2 windows / max aggregate (serve may single-pass prewindowed clips)
Output	p ∈ [0,1]
Serve	CPU FastAPI /v1/analyze-audio; Express Dashboard + video soundtrack fuse

Model description

Frozen Dasheng-Base masked audio encoder + logistic head. Scores how AI-like speech, music, and general audio sound. Product Covenant seats soundtrack windows with Muybridge frame bags (AU6 fuse; default max) when a video has usable audio. Covenant seating is DotCheck-private; this repo is the open audio person only.

Files in this repo: README.md, LICENSE, NOTICE, CITATION.cff, and the .npz head file(s) listed above.

Architecture

audio bytes
  → mono 16 kHz
  → protocol window(s) (4 s)
  → frozen Dasheng-Base emb → logistic head → p
  → multi-window: max

Missing audio head at serve → /v1/analyze-audio returns 503 (fail closed).

Inference

Open weights: the live .npz head(s) in this repo (Apache-2.0), for use with the frozen upstream backbone named above. This is not a transformers AutoModel.from_pretrained("DotCheck/…") package.

Product scoring: Check/Dashboard or Pro API (below). Leviathan (shared memory and related product path) is not in these files.

curl -sS -X POST "https://dotcheck-server-c221c1f32c68.herokuapp.com/analyze-audio" \
  -H "Authorization: Bearer dc_YOUR_KEY" \
  -F "audio=@clip.wav"

UI: https://dotcheck.ai/check · Dashboard audio: https://dotcheck.ai/dashboard · API: https://dotcheck.ai/api · docs: https://dotcheck.ai/docs

Training data

Split	Content
Fit reals	LibriTTS + FSD50K (+ music-from-FSD)
Fit AI	parler-mini + dasheng-audiogen (+ contingency codecfake-fit / dfadd-fit)
Holdout speech	CodecFake / DFADD (sha-disjoint)
Holdout music/SFX	AudioGen odd-seed partition

Evidence: audio_gates_audio_v2.json · AUDIO_GATES_OK.

Evaluation

Holdout (public claim SSOT)

Metric	Target	Measured
mean P(AI) | real	≤ 0.12	0.004
mean P(AI) | AI	≥ 0.85	0.984
bal_acc @ thr	≥ 0.90	0.986

Intended use

• Standalone audio AI-likeness in DotCheck inference.
• Soundtrack windows fused with video frames in product serve.
• Citation of the holdout table above.

Out of scope

• Product scoring SLA / Leviathan / FUP via Hub download
• Courtroom deepfake forensics / speaker identification
• Legal determinations

Limitations

• Protocol windows are short; long-form structure is not modeled.
• Codec / compression / unseen generators can shift scores.
• Public floats are standalone audio gates; fused video bags are not a separate claim table yet.

License

LICENSE — Apache License 2.0 for DotCheck heads in this repo. Upstream backbones: see NOTICE.

Citation

CITATION.cff · wire inhouse-audio@2 / Helmholtz@2 · https://dotcheck.ai/docs