[FFmpeg-devel] [RFC][PATCH] Soft target presets

[Michael Niedermayer]

On Wed, Nov 17, 2010 at 04:49:25AM +0100, Michael Niedermayer wrote:
> On Sat, Nov 13, 2010 at 01:30:44AM +0100, Stefano Sabatini wrote:
> > On date Friday 2010-11-12 23:20:22 +0100, Michael Niedermayer encoded:
> > > On Fri, Nov 12, 2010 at 08:29:21PM +0100, Stefano Sabatini wrote:
> > > > On date Friday 2010-11-12 20:13:59 +0100, Michael Niedermayer encoded:
> > > > > On Fri, Nov 12, 2010 at 07:30:03PM +0100, Stefano Sabatini wrote:
> > > > > > Hi,
> > > > > > 
> > > > > > this should allow program presets of the kind:
> > > > > > ffmpeg -i IN.flv -pre dvd-pal OUT
> > > > > > 
> > > > > > which should deprecate the -target option and deliver more flexibility
> > > > > > to the ff* tools.
> > > > > 
> > > > > and break all other tools in the process by storing ffmpeg.c specific command
> > > > > line options in preset files that are supposed to work with all libav* tools
> > > > > The preset file is supposed to map things to AVOptions, and we even have
> > > > > AVOptions for samplerate and so on or could have but you implement a way to
> > > > > bypass all this and feed it back in the previous layer.
> > > > > This is really ugly and broken
> > > > 
> > > > Patch #8, ffmpeg specific presets should be prefixed by "ffmpeg-",
> > > > same can be easily generalized to the other tools, this just
> > > > implements a convenient way to provide complex presets through a
> > > > bundled configuration file, rather than hardcode them.
> > > > 
> > > 
> > > > This would be very useful for specific targets (e.g. ipod, ipad, PSP,
> > > > ...).
> > > 
> > > I agree its usefull and there exists
> > > ffmpeg -i myfile `cat dvd_opts` outfile
> > > 
> > > -pre dvd_opts
> > > is just 1 char less
> > 
> > So why are we not shipping target-like presets?
> 
> lack of patch
> 
> 
> > 
> > > what makes -*pre better than `cat X` is that its able to automatically adjust
> > > the options based on codec and that its not ffmpeg.c specific but the presets
> > > can easily be used by GUIs too.
> > 
> > No, see acodec/vcodec/scodec.
> 
> If you try to indicate that the average GUI developer is too stupid to read the
> codec out of AVStream/AVCodecContext and load the correct file then well i have
> no comment ATM otherwise i didnt understand what you meant
> 
> 
> > 
> > > None of this applies to what you implement in this patchset, thus i fail to
> > > see the point of it compared to just using posix tools
> > 
> > So what's the point of -target in the first place?
> 
> who knows ...
> 
> 
> > Also not all
> > shells/systems support the `cat X` feature,
> 
> true but practically noone is using the "shell" in these systems people use
> GUIs there or cygwin/mingw with some more powerfull shell
> 
> 
> > and IMO supporting that
> > even in those cases is worth it, especially considering that the cost
> > in complexity is low.
> 
> iam not objecting if people want it, i just object if only you want it ;)
> because i think this is rather useless bloat

Also if this is implemented it can be done simpler.
id guess 10 lines in parse_options() might be enough

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Awnsering whenever a program halts or runs forever is
On a turing machine, in general impossible (turings halting problem).
On any real computer, always possible as a real computer has a finite number
of states N, and will either halt in less than N cycles or never halt.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20101117/4e9867eb/attachment.pgp>