MP3 is the most popular compressed audio format in our daily life but it can be doctored very easily by pervasive audio editing software. Thus it is necessary to develop authentication methods for MP3. Different from JPEG compression for image, MP3 compression has its own characteristics. Thus existing forensics methods for JPEG compression is unable to be applied to MP3 compression directly. In this paper, we propose a passive approach to detect doctored MP3 audio by checking frame offsets. As the audio samples are divided into frames to encode, each frame has its own frame offset after encoding. Forgeries lead to the broken of frame grids. So the frame offsets are good indication for locating forgeries, and the frame offsets can be detected by the identification of quantization characteristic. In this way, the doctored positions can be automatically located. Experimental results demonstrate the validity of the proposed approach on detecting some common forgeries, such as deletion, insertion, substitution and splicing. Under different bitrates, the detection ratios are above 94%. To the best of our knowledge, this piece of work is the first one to investigate digital forensics on MP3 format.