The same authors in proposed to detect double MP3 compression through the use of support vector machine classifiers with feature vectors formed by the distributions of the first digits of quantized MDCT coefficients in particular, a global method was proposed, where the statistics on the first digits of all quantized MDCT coefficients are taken, and the computed probability distributions of nine digits are used as features (nine dimensions) for training a support vector machine (SVM). According to this, a detector was proposed that just measures the number of MDCT coefficients assuming ☑ values and compares this value to a given threshold T: if it is lower than T, the file is a fake quality one, otherwise it is a singly compressed one. In, to defeat fake quality MP3, authors observed that there are many more quantized modified discrete cosine transform (MDCT) coefficients with small values in a singly compressed MP3 file than in a fake quality MP3 file, no matter which bit-rate the fake quality MP3 was transcoded from. The works dealing with MP3 audio files and double audio compression proposed in the current literature are briefly reported in the following. In particular, our aim is to propose a forensic scheme able to detect fake quality audio tracks and to provide information on the first compression bit-rate, as well as to localize possible tampered portions in audio files. Basing on such digital footprints, in this paper, we will assume the role of the forensic analyst that wants to discover two of the main forgeries that a MP3 audio track can undergo: (i) fake quality, i.e., the audio file is recompressed at higher bit-rate to pass it off as a high-quality track and (ii) tampering, i.e., a portion of the audio track has been edited or deleted. In particular, similarly to the image forensic field, the analysis of the artifacts due to double compression has received a lot of attention. In the last years, the research in multimedia forensics started to consider audio contents for investigating their origin and authenticity.