[NUT-devel] [nut]: r615 - docs/nut4cc.txt

[Rich Felker]

On Tue, Feb 19, 2008 at 03:26:02AM +0100, Michael Niedermayer wrote:
> On Mon, Feb 18, 2008 at 03:43:17PM -0500, Rich Felker wrote:
> > On Mon, Feb 18, 2008 at 09:25:47PM +0100, Michael Niedermayer wrote:
> > > On Mon, Feb 18, 2008 at 02:12:57PM -0500, Rich Felker wrote:
> > > > On Mon, Feb 18, 2008 at 08:02:00PM +0100, Michael Niedermayer wrote:
> > > > > On Mon, Feb 18, 2008 at 09:39:39AM -0700, Loren Merritt wrote:
> > > > > > On Tue, 12 Feb 2008, cladisch wrote:
> > > > > > 
> > > > > > > Log:
> > > > > > > clarify which of the MPEG-4 video codecs is meant
> > > > > > >
> > > > > > > -MP4V    ISO MPEG4 Video
> > > > > > > +MP4V    ISO MPEG4 Part 2 Video
> > > > > > 
> > > > > > Does this mean ISO MPEG4 Part 2 Advanced Simple Profile? Or are all 
> > > > > > profiles allowed and the rest just probably not supported by any given 
> > > > > > implementation?
> > > > > 
> > > > > I would say all are allowed, simply because the alternative is more work
> > > > > than any volunteer here is likely going to do. We already lacked one just
> > > > > writing this list ...
> > > > > Also this problem would apply to mpeg1/2 h264, h263, ... as well.
> > > > 
> > > > Does part 2 include all the ridiculous stuff like sprites?
> > > 
> > > ----
> > > Information technology  Coding of
> > > audio-visual objects 
> > > Part 2:
> > > Visual
> > > ...
> > > 7.9    Generalized scalable decoding 241
> > > 7.9.1  Temporal scalability ..241
> > > 7.9.2  Spatial scalability...246
> > > 7.10   Still texture object decoding.251
> > > 7.10.1 Decoding of the DC subband ...251
> > > 7.10.2 ZeroTree Decoding of the Higher Bands.252
> > > 7.10.3 Inverse Quantisation .257
> > > 7.10.4 Still Texture Error Resilience265
> > > 7.10.5 Wavelet Tiling.268
> > > 7.10.6 Scalable binary shape object decoding ..270
> > > 7.11    Mesh object decoding ... 276
> > > 7.11.1 Mesh geometry decoding . 276
> > > 7.11.2 Decoding of mesh motion vectors... 279
> > > 7.12    FBA object decoding. 281
> > > 7.12.1 Frame based face object decoding.. 281
> > > 7.12.2 DCT based face object decoding . 282
> > > 7.12.3 Decoding of the viseme parameter fap 1  284
> > > 7.12.4 Decoding of the viseme parameter fap 2  284
> > > 7.12.5 Fap masking ... 285
> > > 7.12.6 Frame Based Body Decoding... 285
> > > 7.12.7 DCT based body object decoding 286
> > > 7.13    3D Mesh Object Decoding  287
> > > 7.13.1 Start codes and bit stuffing .. 288
> > > 7.13.2 The Topological Surgery decoding process... 288
> > > 7.13.3 The Forest Split decoding process.. 291
> > > 7.13.4 Header decoder.. 292
> > > 7.13.5 partition type .. 293
> > > 7.13.6 Vertex Graph Decoder... 294
> > > 7.13.7 Triangle Tree Decoder... 298
> > > 7.13.8 Triangle Data Decoder... 299
> > > 7.13.9 Forest Split decoder .. 303
> > > 7.13.10 Arithmetic decoder  309
> > > 7.14    NEWPRED mode decoding... 314
> > > 7.14.1 Decoder Definition. 314
> > > 7.14.2 Upstream message 314
> > > 7.15    Output of the decoding process .. 314
> > > 7.15.1 Video data... 315
> > > 7.15.2 2D Mesh data.. 315
> > > 7.15.3 Face animation parameter data 315
> > > 8       Visual-Systems Composition Issues... 315
> > > 8.1     Temporal Scalability Composition... 315
> > > 8.2     Sprite Composition 316
> > > 8.3     Mesh Object Composition  317
> > > 8.4     Spatial Scalability composition 318
> > > -----
> > 
> > Am I the only one who thinks this face animation, mesh, sprite, etc.
> > crap that's never been implemented and probably never will be
> > implemented in the real world belongs tucked away in its own obscure
> > fourcc?
> 
> Well we would need a proper definition of what belongs in
> mp4v and what belongs in the other one. Also i fear if we want to do this
> half professional, we would have to provide sound arguments and tests why
> these things are useless and should be seperated. The general feeling that
> they are, which i think we all share is a little weak ...
> 
> We also could sidestep above and just give each profile its own fourcc but
> that is going to be alot of work, and higher complexity on the muxer side,
> which has to figure out which profile a input mpeg4 is then ...
> 
> Besides that i vote for the new fourcc to be "SHIT" if we end up doing
> a 4cc split :)
> 
> Anyway, iam slightly leaning toward just leaving all in one 4cc. Because i
> just dont see any practical gain from the split, its just work, and work
> for which i dont vounteer ...

Can't we just say MP4V means MPEG-4 ASP? Or are there real features in
use that are outside the domain of ASP?

Rich