> 2. Proposal (A) > > Modify libav interface so, that there were series of hooks on most > significant stages of encoding
Then you're looking in the wrong place, as libx264 is the encoder that most people use for video. It's not part of libavcodec. >and decoding. How does this even make sense? What's the point of a "watermark" that relies on the decoder, so a user can simply switch decoders and the watermark goes away? > Why? As everybody know > encoding is very expensive operation, so DW currently are developed in > compressed domain. And watermarking engine embedded in decoder, stops > at some stage, say quantized DCT coefs and reconstruct bitstream with > some modifications. If you use a compressed-domain watermark, e.g. parity of quantized coefficients or some other steganography, you can simply re-encode the video and it goes away. That makes it rather useless, as everyone re-encodes videos when pirating them anyways. Jason _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
