first of all nice to see that someone tried to develop an own video codec, that makes use of new technologies (as covered by h.264/mpeg-4)!
Quote:
|
There are three main reasons I have my own codec. First, I needed to store many of the original MPEG2 features to restore the original. Second, I had this idea that a real transcoder would be better than a decoder/encoder approach.
|
hm as being someone who knows a bit about video compression, i find these two reasons (which i see as one actually) very strange:
there are two possibilities to get down the bitrate from an existing video stream:
1) bitrate peeling, which some codecs supports (eg ogg vorbis), but isnt supported in mpeg-2 (the video stream from the dvd)
2) decoding and reencoding
for me his reasons sound like he does 1) which he names as "transcoding" in contrary as "dec/encoding", but this sounds very strange and impossible in this case
i cant imagine that there is a way to simply modify an existing mpeg-2 stream partially (eg change the quantisation) and than get the same stream at a lower bitrate
especially this isnt done in the case of ratdvd as it includes things like deblocking and changed GOPs, which clearly tells that the mpeg-2 stream from the dvd gets decoded and than encoded with the ratdvd video encoder
Quote:
|
Third, H.264 and XviD (and other standard codecs) are covered by a wide range of patents and patent pools and I wanted to keep the program completely free.
|
this is strange for two reasons:
1) his tool already uses ac3 technology, so it violates patent rights already (in most western countries) and is therefore illegal (unless he pays licenses to the patent right holders of ac3 of course)
2) a patent-pending situation is not only caused when you are fully compliant to mpeg-4, but you are also using patented technology when you simply use one algo, eg covered in mpeg-4, that is patented of course
as m$' wmv9/vc-1 showed its practically hardly possible to develop a video codec that does not infringe existing patents. of course you can hide the useage of such tech by staying closed source/format (as m$ did for a long time)
so if the goal would have been to not have to pay licenses because of using patented technology (and therefore be a totally lega tool), i only see two codecs which could have been used:
ogg vorbis for audio encoding
ogg theora for video encoding