---
title: 'CompSpoof: Component-Level Audio Anti-Spoofing'
url: https://www.emergentmind.com/topics/compspoof
type: topic
---

# CompSpoof: Component-Level Audio Anti-Spoofing

CompSpoof is a component-level audio anti-spoofing problem setting, dataset, and learning framework in which authenticity is assigned not only to an utterance as a whole but separately to its constituent acoustic components, specifically **speech** and **environmental/background sound** [2509.15804]. It addresses a manipulation regime in which only one component may be forged while another remains genuine, a situation that departs from conventional anti-spoofing assumptions that an utterance or segment is entirely bona fide or entirely spoofed. In response, CompSpoof introduces a five-class benchmark and a separation-enhanced joint learning framework that first separates the mixture and then applies dedicated anti-spoofing models to each component, with joint optimization intended to preserve spoof-relevant information [2509.15804].

## 1. Conceptual scope

CompSpoof defines **Component-level audio Spoofing (Comp-Spoof)** as a setting in which only specific components of an audio signal are forged, replaced, or generated while other components remain authentic [2509.15804]. The paper concentrates on two components, **speech** and **environmental/background sound**, and treats their authenticity states independently. This yields cases such as spoofed speech with genuine environment, genuine speech with spoofed environment, or simultaneous manipulation of both components [2509.15804].

The defining claim is that conventional anti-spoofing datasets and methods are too coarse because they assign a single bona fide/spoofed label to an utterance or segment. Even prior **partial spoofing** work is framed primarily along the **time axis**, asking which temporal region is manipulated; CompSpoof instead asks which **acoustic component in the mixture** is manipulated [2509.15804]. In the paper’s formulation, this is the distinction between utterance-level authenticity and source-level authenticity within a composite scene.

A useful formalization given in the description is
\[
x = x^{\text{speech}} + x^{\text{env}},
\]
with separate component labels
\[
y^{\text{speech}} \in \{0,1\}, \qquad y^{\text{env}} \in \{0,1\},
\]
and a final utterance label derived from the pair together with whether the sample is an original recording or an artificial mixture [2509.15804]. This suggests a structured classification problem rather than a single binary decision.

## 2. Dataset design and label taxonomy

The CompSpoof dataset is presented as the **first dataset for component-level audio anti-spoofing** [2509.15804]. It contains **five classes**, each with **500 samples**, for a total of **2,500 audio samples** [2509.15804]. The class system distinguishes both authenticity combinations and the difference between natural co-recording and artificial mixing.

| ID | Label | Brief description |
|---|---|---|
| 0 | `original` | Original bona fide speech with its corresponding naturally captured environment |
| 1 | `bonafide_bonafide` | Bona fide speech mixed with another bona fide environment |
| 2 | `spoof_bonafide` | Spoofed speech mixed with bona fide environment |
| 3 | `bonafide_spoof` | Bona fide speech mixed with spoofed environment |
| 4 | `spoof_spoof` | Spoofed speech mixed with spoofed environment |

This taxonomy supports both **five-class utterance classification** and implicit **component-level authenticity labels** for speech and environment [2509.15804]. The framework later uses those component labels explicitly through separate speech and environment anti-spoofing branches.

The distinction between `original` and `bonafide_bonafide` is structurally important. The former corresponds to naturally co-recorded bona fide speech and environment, whereas the latter is an artificial mixture of genuine components that were not originally recorded together [2509.15804]. The paper treats this as part of the utterance-level label space, not merely as metadata.

The dataset is partitioned with **stratified sampling** into **70% train**, **10% development**, and **20% evaluation** [2509.15804]. Audio durations range from **5 to 21 seconds** [2509.15804].

## 3. Data sources and construction protocol

CompSpoof is assembled from multiple source corpora [2509.15804]. The paper specifies the following provenance:

- **Bona fide speech**: **ASVspoof5** and **CommonVoice**
- **Spoofed speech**: **ASVspoof5** and **SSTC**
- **Bona fide environmental sound**: **VGGSound**
- **Spoofed environmental sound**: **VCapAV**
- **Original co-recorded bona fide audio**: **VGGSound**

The generation pipeline proceeds by selecting speech clips with clear voice activity and environmental clips from diverse scenes such as **indoor**, **street**, and **natural settings**, then resampling all audio to **16 kHz** [2509.15804]. For artificially mixed classes, the shorter signal determines the final duration and the longer signal is truncated. Speech and environmental sound are then mixed, and the environmental amplitude is adjusted to achieve a predefined **SNR relative to speech** [2509.15804].

This construction deliberately exposes the benchmark to both speech spoofing and environmental spoofing. The paper does not treat background sound as a nuisance variable; it is itself a manipulable target. A plausible implication is that anti-spoofing systems that rely on global utterance statistics can confuse authentic and manipulated evidence when they coexist in the same mixture.

## 4. Separation-enhanced joint learning framework

The proposed method is a **separation-enhanced joint learning framework** built from four modules: a **binary mixture detection model**, a **UNet-based separation network**, a **speech anti-spoofing model**, and an **environment anti-spoofing model** [2509.15804]. The anti-spoofing branches are implemented with **XLSR-AASIST**, while the separator operates in the **STFT domain** [2509.15804].

The first stage is a binary **mixture detector** that distinguishes \(c_0\) from \(c_{1,2,3,4}\), that is, `original` versus all other classes [2509.15804]. The second stage separates the mixture into estimated speech and environmental waveforms. Given mixed waveform \(x\), its complex spectrogram is
\[
X(f,t) = \mathrm{STFT}(x).
\]
The separator predicts a complex speech mask \(M_{\text{speech}}(f,t)\) and computes
\[
\hat{S}(f,t) = M_{\text{speech}}(f,t)\odot X(f,t),
\]
followed by
\[
\hat{x}^{\text{speech}} = \mathrm{ISTFT}(\hat{S}).
\]
The environmental component is formed from the residual
\[
r = x - \hat{x}^{\text{speech}}.
\]
With \(S(f,t)\) the magnitude spectrogram of the separated speech and \(R(f,t)\) the magnitude spectrogram of the residual, the dynamic scaling factor is
\[
\alpha = \frac{\mathrm{mean}(|R(f,t)|)}{\mathrm{mean}(|S(f,t)|)+\epsilon},
\]
and the environment mask is
\[
M_{\mathrm{env}}(f,t) = 1 - \tanh\Big( \frac{|S(f,t)|}{|R(f,t)| + \epsilon} \cdot \alpha \Big).
\]
The estimated environment is then reconstructed from the masked residual-domain representation [2509.15804].

The two component detectors are trained as binary classifiers with class groupings specified in the paper: the **speech detector** uses **c13 vs c24**, and the **environment detector** uses **c12 vs c34** [2509.15804]. At inference, the binary outputs are combined with the mixture decision to recover the five-way label. The implied rule is: if the mixture detector predicts \(c_0\), the output is `original`; otherwise the speech and environment decisions map to `bonafide_bonafide`, `spoof_bonafide`, `bonafide_spoof`, or `spoof_spoof` [2509.15804].

The paper’s main methodological claim is that **separation alone is insufficient** because reconstruction-oriented separation can suppress forensic cues [2509.15804]. To counter this, the separator and the component detectors are trained jointly. A consistency term compares anti-spoofing outputs from reference and separated components:
\[
\begin{aligned}
L_\mathrm{cons} &= L_\mathrm{cons}^{env}+L_\mathrm{cons}^{speech} \\
&= \mathrm{KL}(p_\mathrm{ref}^\mathrm{env} \,\|\, p_\mathrm{sepa}^\mathrm{env}) + \mathrm{KL}(p_\mathrm{ref}^\mathrm{speech} \,\|\, p_\mathrm{sepa}^\mathrm{speech}).
\end{aligned}
\]
The overall objective is
\[
L_\mathrm{joint} = \kappa * L_\mathrm{sepa} + L_\mathrm{cls}^\mathrm{mixed} + L_\mathrm{cls}^\mathrm{speech} + L_\mathrm{cls}^\mathrm{env} + L_\mathrm{cons},
\]
with \(\kappa = 10\) [2509.15804].

## 5. Training protocol and implementation details

The baseline is a direct five-class extension of **XLSR-AASIST** [2509.15804]. The proposed framework keeps **XLSR-AASIST** for the mixture, speech, and environment branches, while adding the UNet separator [2509.15804].

For separation, the system uses **STFT-domain** complex spectrograms with audio sampled at **16 kHz**, **window size 64 ms**, and **hop length 16 ms** [2509.15804]. Separation-based methods process audio in **4-second windows** with **2-second hop** [2509.15804]. Segment predictions are aggregated to the file level by **majority voting** [2509.15804].

Optimization uses **Adam** with learning rate
\[
1\times10^{-3}
\]
for the separator and
\[
1\times10^{-5}
\]
for the anti-spoofing models [2509.15804]. In the joint framework, models are trained **independently for the first 4 epochs** and then **jointly from epoch 5 onward** [2509.15804]. The paper does not describe additional data augmentation beyond resampling, mixing, SNR control, and chunking.

The evaluation metrics differ from common ASV anti-spoofing practice. Whereas many speech spoofing studies emphasize **EER** and **t-DCF** [2007.05979], CompSpoof reports **Precision**, **Recall**, and **F1** at file level for the five-class task, together with segment-level Precision, Recall, and F1 for speech and environment anti-spoofing [2509.15804].

## 6. Empirical results, significance, and relation to adjacent work

The experimental comparison includes three systems: the utterance-level **Baseline**, the **Separation-Enhanced Framework (SEF)** without joint learning, and **SEF+JL** with joint learning [2509.15804]. On the **eval** set, overall F1 is **0.827** for the Baseline, **0.668** for SEF, and **0.908** for SEF+JL [2509.15804]. On the **dev** set, the corresponding F1 scores are **0.840**, **0.705**, and **0.912** [2509.15804]. The central empirical conclusion is therefore twofold: naive separation degrades performance, but separation with joint learning outperforms the direct utterance-level classifier.

Per-class eval F1 further localizes the gains [2509.15804]. For the Baseline, the paper reports **0.980** for `original`, **0.843** for `bonafide_bonafide`, **0.745** for `spoof_bonafide`, **0.829** for `bonafide_spoof`, and **0.738** for `spoof_spoof`. For **SEF+JL**, the corresponding numbers are **0.990**, **0.899**, **0.871**, **0.905**, and **0.874** [2509.15804]. The most important improvements occur in the mixed-content classes where only one component is spoofed. That pattern supports the paper’s core premise that component-wise modeling is necessary when authentic and manipulated evidence coexist in the same recording.

Segment-level results reinforce the same interpretation. On the CompSpoof eval set, **speech anti-spoofing** improves from **F1 = 0.720** without joint learning to **F1 = 0.863** with joint learning, while **environment anti-spoofing** improves from **F1 = 0.718** to **F1 = 0.849** [2509.15804]. The paper also notes that environment anti-spoofing remains weaker than speech anti-spoofing, suggesting that an **XLSR-AASIST-based environment detector may not be ideal** [2509.15804].

CompSpoof occupies a specific position within a broader shift from utterance-level to more structured spoof analysis. In speech anti-spoofing, datasets such as **LlamaPartialSpoof** stress **fully and partially fake speech** and show that current fake speech detection systems struggle to generalize to unseen scenarios, with best performance reported as **24.49% equal error rate** [2409.14743]. That line of work localizes manipulations in time and semantics; CompSpoof instead localizes authenticity across **simultaneous acoustic components** [2509.15804]. A plausible implication is that future anti-spoofing benchmarks may need to consider both axes jointly: temporal partial spoofing and component-level spoofing.

The paper also exposes several limitations. The dataset scale is **2,500 samples**, the component taxonomy is restricted to **speech** and **environment**, and many classes are formed by artificial mixing rather than by naturally edited recordings [2509.15804]. The method itself depends on separation quality, and the poor performance of SEF shows that component extraction can damage spoof evidence if it is optimized only for reconstruction [2509.15804]. Even so, CompSpoof establishes a distinct research program: anti-spoofing in composite audio scenes, where authenticity is a property of individual sources rather than only of the utterance.

Source: https://www.emergentmind.com/topics/compspoof