[x264-devel] Re: x264 roadmap

[Andrea Barbieri]

thank you for the URL pointer, I was just hoping that for consolidation 
matters the todo (and its wereabout knowledge) was part of the source code 
repository rather than in some other places like Doom9 forum (with quite a 
high level of noise).

I'm currently looking at the Win32 runtime (needed to modify the .sln 
and .vcproj and very few files for .NET 2003) and using the latest VTune 
analyzer... I need to perform more tests... the reason being the current 
timings with brdo=0 and trellis=0 show x264_quant_8x8x_trellis being the 
second most expensive function (~7%) which doesn't make sense.

normally with the monochrome format tool the 10bit tool goes hand to hand
(even though the former is part of the High profile and the latter of the 
High-10 one)

GP:=Guillaume POIRIER

GP> Hi,
GP> 
GP> On 3/5/06, Andrea Barbieri <andrea.barbieri at movingimageresearch.com> wrote:
GP> > LM:=Loren Merritt
GP> >
GP> > LM> On Sat, 4 Mar 2006, Andrea Barbieri wrote:
GP> > LM>
GP> > LM> > is a roadmap of x264 new features available for general consumption (no
GP> > LM> > great details nor exact timing, just an idea of what to expect next) or is
GP> > LM> > it just bug-fixes/performance enhancements?
GP> > LM>
GP> > LM> I do not plan features in advance.
GP> >
GP> > interesting observation
GP> >
GP> > I hope that my question didn't influence the removal of the TODO file from
GP> > the repository.
GP> >
GP> > the knowledge of what needs to be improved might also help developers
GP> > mediated contributions from a larger pool of expertise... and to know what
GP> > currently unsupported ISO/IEC 14496-10 features (mainly all of the FRExt
GP> > ones, open GOP) would score high in the developers agenda might improve as
GP> > well the chances of targetted efforts and contributions to be made.
GP> 
GP> Loren has a todo list online:
GP> http://students.washington.edu/lorenm/src/x264/todo.txt
GP> 
GP> In addition to that, there are 2 other functionality that several
GP> users have requested on this mailing list: grayscale encoding and
GP> interlaced encoding. Those two will eventually be implemented "when
GP> someone with enough knowledge and time will do it".
GP> 
GP> Guillaume
GP> 
GP> --
GP> Reinventing the wheel certainly is annoying, but as long as all other
GP> wheels are square...
GP> 
GP> Reimar Döffinger
GP> 
GP> 

-- 
Andrea Barbieri
KeyID=0x034DFD5A
KeyFingerprint=C1 68 EA 9A 71 89 53 8D  21 4F 12 81 A7 52 9F 32  03 4D FD 5A
Moving Image Research, The Workshop, Hampton Lane, Bristol, BS6 6LE, UK
Tel +44 117 9732200, FAX +44 117 9732210
http://www.movingimageresearch.com/