Codec independent lossy audio compression detection

In this paper, we propose a method for detecting marks of lossy compression encoding, such as MP3 or AAC, from PCM audio. The method is based on a convolutional neural network (CNN) applied to audio spectrograms and trained with the output of various lossy audio codecs and bitrates. Our method shows good performances on a large database and robustness to codec type and resampling.

[1]  Louis Dunn Fielder,et al.  ISO/IEC MPEG-2 Advanced Audio Coding , 1997 .

[2]  Erkam Uzun,et al.  Methods for identifying traces of compression in audio , 2013, 2013 1st International Conference on Communications, Signal Processing, and their Applications (ICCSPA).

[3]  Patrick Aichroth,et al.  AAC encoding detection and bitrate estimation using a convolutional neural network , 2016, 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[4]  Gerhard Stoll,et al.  ISO-MPEG-1 Audio: A Generic Standard for Coding of High-: Quality Digital Audio , 1994 .

[5]  Marco Fontani,et al.  Detection and localization of double compression in MP3 audio tracks , 2014, EURASIP Journal on Information Security.

[6]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[7]  Jürgen Herre,et al.  Analysing Decompressed Audio with the "Inverse Decoder" - Towards an Operative Algorithm , 2002 .

[8]  Gerald Schuller,et al.  Efficient Cross-Codec Framing Grid Analysis for Audio Tampering Detection , 2014 .

[9]  Yun Q. Shi,et al.  Mp3 bit rate quality detection through frequency spectrum analysis , 2009, MM&Sec '09.

[10]  Gerald Schuller,et al.  Estimating MP3PRO Encoder Parameters From Decoded Audio , 2013, GI-Jahrestagung.

[11]  Jürgen Herre,et al.  Analysis of Decompressed Audio-The -Inverse Decoder- , 2000 .

[12]  Jiwu Huang,et al.  Detecting digital audio forgeries by checking frame offsets , 2008, MM&Sec '08.

[13]  Rui Yang,et al.  Defeating fake-quality MP3 , 2009, MM&Sec '09.