Techniques for media content tamper protection and detection are described. In some embodiments, the protection method is performed at an encoder. The encoder obtains a first sequence of signatures within a group of pictures and a second sequence of signatures for audio frames corresponding to the group of pictures. The encoder then combines the first sequence of signatures and the second sequence of signatures to generate a combined signature, and the combined signature is then embedded in a transport stream for transmission to a client device. In some embodiments, the detection method is performed at a client device, where the client device, in response to detecting a first message in the transport stream, starts calculation of signatures for a group of pictures and corresponding audio frames. The client device then uses the calculated signatures to validate an expected signature for the group of pictures and the corresponding audio frames.