_______ _                                   ______ _ _           
 |__   __| |                 /\              |  ____(_) |          
    | |  | |__   ___        /  \     ______  | |__   _| | ___  ___ 
    | |  | '_ \ / _ \      / /\ \   |______| |  __| | | |/ _ \/ __|
    | |  | | | |  __/     / ____ \           | |    | | |  __/\__ \
    |_|  |_| |_|\___|    /_/    \_\          |_|    |_|_|\___||___/
                                                                    

The A-Files is a powerful audio steganography software that allows users to embed secret data within an audio signal, measure speech quality metrics, and test the audio signal's robustness against different types of attacks.

With The A-Files, users can ensure that their sensitive information remains private and protected.

• About
• Steganography algorithms
• Metrics
  • Speech Reverberation
  • Speech Intelligibility
  • Speech Quality
• Attacks
• Diagram
• References
• Licence
• Authors

The A-Files project contains:

• Audio watermarking methods
• Speech quality metrics
• Audio attacks

The A-Files is an audio steganography software that allows users to hide secret data within an audio signal. To do this, users can load both the audio file and the data they want to hide into the software. The A-Files supports several types of audio files, such as WAV or FLAC. Once the audio file and the secret data are loaded, the software can embed the data within the audio signal.

The A-Files offers several techniques for embedding secret data within an audio signal, such as LSB (Least Significant Bit) insertion, phase coding, and echo hiding. Each technique has its own strengths and weaknesses, and users can choose the technique that best fits their application's requirements.

The A-Files includes tools for measuring speech quality metrics, such as PESQ (Perceptual Evaluation of Speech Quality), SNR (Signal-to-Noise Ratio) and STOI (Short-time objective intelligibility) metric. By measuring these metrics, users can ensure that the audio signal has not been degraded during the process of embedding secret data.

The A-Files provides tools for testing the audio signal's robustness against different types of attacks. Attackers may attempt to remove the hidden data or alter the audio signal to render the hidden data useless. The A-Files includes tools for testing the audio signal's robustness against various types of attacks, such as noising, frequency cutting, and filtering. By testing the audio signal's robustness, users can determine the effectiveness of the audio steganography technique and make any necessary adjustments to improve the technique's strength and resilience.

Powered by some great GitHub repositories

List of implemented methods:

Each method extend abstract class SteganographyMethod

from abc import abstractmethod, ABC
from typing import List
import numpy as np


class SteganographyMethod(ABC):

    @abstractmethod
    def encode(self, data: np.ndarray, message: List[int]) -> np.ndarray:
        ...

    @abstractmethod
    def decode(self, data_with_watermark: np.ndarray, watermark_length: int) -> List[int]:
        ...

    @abstractmethod
    def type(self) -> str:
        ...

• MOSNet: Deep Learning based Objective Assessment for Voice Conversion [16]

• Bark spectral distortion (BSD) [7]
• Speech-to-reverberation modulation energy ratio (SRMR) [10]

• Coherence and speech intelligibility index (CSII) [7]
• Normalized-covariance measure (NCM) [7]
• Short-time objective intelligibility (STOI) [9]

• Signal - to - Noise Ratio (SNR) [12]
• Mel-cepstral distance measure for objective speech quality assessment [11]
• Segmental Signal-to-Noise Ratio (SNRseg) [7]
• Frequency-weighted Segmental SNR (fwSNRseg) [7]
• Cepstrum Distance Objective Speech Quality Measure (CD) [7]
• Log - likelihood Ratio (LLR) [7]
• Weighted Spectral Slope (WSS) [7]
• Perceptual Evaluation of Speech Quality (PESQ) [8]
• Speech Enhancement Metrics (Csig, Covl, Cbak, Composite) [13]
• Weighted Spectro-Temporal Modulation Index (wSTMI) [14]
• Spectro-Temporal Glimpsing Index (STGI) [15]
• Scale-invariant SDR(SISDR) [17]

Each metric extend abstract class Metric

from abc import ABC, abstractmethod
from numbers import Number

import numpy as np


class Metric(ABC):

    @abstractmethod
    def calculate(self,
                  samples_original: np.ndarray,
                  samples_processed: np.ndarray,
                  fs: int,
                  frame_len: float = 0.03,
                  overlap: float = 0.75) -> Number | np.ndarray:
        ...

    @abstractmethod
    def name(self) -> str:
        ...

List of attack on audio samples:

• Low pass filter
• Additive noise
• Frequency filter
• Flip random samples
• Cut random samples
• Resample (downsampling, upsampling)
• Amplitude scaling
• Pitch shift
• Time stretch

[1] Alsabhany, Ahmed A., Ahmed Hussain Ali, Farida Ridzuan, A. H. Azni, and Mohd Rosmadi Mokhtar. “Digital Audio Steganography: Systematic Review, Classification, and Analysis of the Current State of the Art.” Computer Science Review 38 (2020): 100316. https://doi.org/10.1016/j.cosrev.2020.100316
[2] Hu, Hwai Tsu, and Ling Yuan Hsu. “Robust, Transparent and High-Capacity Audio Watermarking in DCT Domain.” Signal Processing 109 (2015): 226–35. https://doi.org/10.1016/j.sigpro.2014.11.011
[3] Natgunanathan, Iynkaran, Yong Xiang, Guang Hua, Gleb Beliakov, and John Yearwood. “Patchwork-Based Multilayer Audio Watermarking.” IEEE/ACM Transactions on Audio Speech and Language Processing 25, no. 11 (2017): 2176–87. https://doi.org/10.1109/TASLP.2017.2749001
[4] Saadi, Slami, Ahmed Merrad, and Ali Benziane. “Novel Secured Scheme for Blind Audio/Speech Norm-Space Watermarking by Arnold Algorithm.” Signal Processing 154 (2019): 74–86. https://doi.org/10.1016/j.sigpro.2018.08.011
[5] Zhao, Juan, Tianrui Zong, Yong Xiang, Longxiang Gao, Wanlei Zhou, and Gleb Beliakov. “Desynchronization Attacks Resilient Watermarking Method Based on Frequency Singular Value Coefficient Modification.” IEEE/ACM Transactions on Audio Speech and Language Processing 29 (2021): 2282–95. https://doi.org/10.1109/TASLP.2021.3092555
[6] Nugraha, Rizky M. “Implementation of Direct Sequence Spread Spectrum Steganography on Audio Data.” Proceedings of the 2011 International Conference on Electrical Engineering and Informatics, ICEEI 2011, no. July ( 2011). https://doi.org/10.1109/ICEEI.2011.6021662
[7] Philipos C. Loizou. “Speech Enhancement. Theory and Practice, Second Edition” CRC Press ( 2013). https://doi.org/10.1201/b14529
[8] Miao Wang, Christoph Boeddeker, Rafael G. Dantas and ananda seelan. “PESQ (Perceptual Evaluation of Speech Quality) Wrapper for Python Users” Zenodo 2022. https://doi.org/10.5281/zenodo.6549559
[9] C.H.Taal, R.C.Hendriks, R.Heusdens, J.Jensen 'A Short-Time Objective Intelligibility Measure for Time-Frequency Weighted Noisy Speech', ICASSP 2010, Texas, Dallas. https://doi.org/10.1109/ICASSP.2010.5495701
[10] Tiago H. Falk, Chenxi Zheng, and Way-Yip Chan. A Non-Intrusive Quality and Intelligibility Measure of Reverberant and Dereverberated Speech, IEEE Trans Audio Speech Lang Process, Vol. 18, No. 7, pp. 1766-1774, Sept.2010. https://doi.org/10.1109/TASL.2010.2052247
[11] R. Kubichek, "Mel-cepstral distance measure for objective speech quality assessment", Proceedings of IEEE Pacific Rim Conference on Communications Computers and Signal Processing, Victoria, BC, Canada, 1993, pp. 125-128 vol.1, https://doi.org/10.1109/PACRIM.1993.407206.
[12] https://en.wikipedia.org/wiki/Signal-to-noise_ratio [13] Yi Hu and Philipos C. Loizou, "Evaluation of Objective Quality Measures for Speech Enhancement", IEEE TRANSACTIONS ON AUDIO, SPEECH, AND LANGUAGE PROCESSING, VOL. 16, NO. 1, JANUARY 2008 229 [14] A. Edraki, W.-Y. Chan, J. Jensen, & D. Fogerty, “Speech Intelligibility Prediction Using Spectro-Temporal Modulation Analysis,” IEEE/ACM Trans. Audio, Speech, & Language Processing, vol. 29, pp. 210-225, 2021. [15] A. Edraki, W.-Y. Chan, J. Jensen, & D. Fogerty, “A Spectro-Temporal Glimpsing Index (STGI) for Speech Intelligibility Prediction," Proc. Interspeech, 5 pages, Aug 2021. [16] Lo, Chen-Chou and Fu, Szu-Wei and Huang, Wen-Chin and Wang, Xin and Yamagishi, Junichi and Tsao, Yu and Wang, Hsin-Min, "MOSNet: Deep Learning based Objective Assessment for Voice Conversion", arXiv preprint arXiv:1904.08352, 2019, https://arxiv.org/abs/1904.08352 [17] Roux, Jonathan Le and Wisdom, Scott and Erdogan, Hakan and Hershey, John R, "SDR – Half-baked or Well Done?", ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2019, https://dx.doi.org/10.1109/ICASSP.2019.8683855

[1] https://github.com/kosta-pmf/audio-watermarking
[2] https://github.com/ktekeli/audio-steganography-algorithms
[3] https://github.com/schmiph2/pysepm
[4] https://github.com/ludlows/PESQ
[5] https://github.com/mpariente/pystoi
[6] https://github.com/jfsantos/SRMRpy
[7] https://github.com/jasminsternkopf/mel_cepstral_distance [8] https://github.com/nglehuy/semetrics [9] https://github.com/aminEdraki/py-intelligibility [10] https://github.com/aliutkus/speechmetrics

The A-Files is an open source software under GPLv3 license.

• Paweł Kaczmarek (@pawelkaczmarek12) - Military University of Technology, Faculty of Electronics
• Zbigniew Piotrowski - Military University of Technology, Faculty of Electronics