[FFmpeg-devel] [PATCH] documentation: Fix several documentation bugs in trac ticket #2413

Stefano Sabatini stefasab at gmail.com
Tue Apr 9 01:12:44 CEST 2013


On date Saturday 2013-04-06 10:01:05 +0800, littlebat encoded:
> On Sat, 6 Apr 2013 01:59:19 +0200
> Stefano Sabatini <stefasab at gmail.com> wrote:
> 
> > On date Friday 2013-04-05 21:58:09 +0800, littlebat encoded:
> > > On Fri, 5 Apr 2013 12:40:40 +0200
> > > Stefano Sabatini <stefasab at gmail.com> wrote:
> > [...] 
> > > /opt/ffmpeggit/bin/ffmpeg -i in.avi -vf \
> > >   "movie=logo.png [logo]; [out] [logo] overlay [in_0]" out.avi
> > > 
> > > It seems the input video stream linked [out] automatically, and the
> > > command line works well! I suppose these two special links "[in]"
> > > and "[out]" have been dropped in the current ffmpeg, or they don't
> > > function properly if they are still designed as special labels. 
> > 
> > Interesting:
> > ffmpeg -i INPUT -vf  "movie=ffmpeg-logo.png [logo]; [foo][logo]
> > overlay [bar]" -y OUTPUT
> > 
> > Looks like if there is an unconnected filter input it is connected to
> > the default filter output, and the unconnected filter output is
> > connected to the default filter input. This was the result of some
> > relatively recent work on -filter_complex, but seems more like an
> > (undocumented) glitch rather than a feature.
> > 
> > The equivalent command fails with ffplay:
> > ffplay INPUT -vf  "movie=ffmpeg-logo.png [logo]; [foo][logo] overlay
> > [bar]" [...]
> > Output pad "default" with type video of the filter instance
> > "ffplay_buffer" of buffer not connected to any destination
> 

> I see, "[in]" and "[out]" are still designed as the special labels in
> the entire ffmpeg system. Using them is not harmful, but, lacking 
> them, some times, will cause problem, as your ffplay example above. My
> understanding is ok? 

Not in the entire FFmpeg system, they still are in ffplay -af/-vf,
while -filter_complex has an entirely different syntax. ffmpeg -af/-vf
has glitches, which are possible unwanted effects of the recent
changes in the code (can't say since I'm not the author of the code,
which didn't updated the docs).

> 
> >diff --git a/doc/filters.texi b/doc/filters.texi
> >index 483d8a1..773657f 100644
> >--- a/doc/filters.texi
> >+++ b/doc/filters.texi
> >@@ -5,8 +5,8 @@ Filtering in FFmpeg is enabled through the libavfilter
> >library. 
> > In libavfilter, it is possible for filters to have multiple inputs and
> > multiple outputs.
> >-To illustrate the sorts of things that are possible, we can
> >-use a complex filtergraph. For example, the following one:
> >+To illustrate the sorts of things that are possible, we consider
> >+the following filtergraph.
> 
> I found you still reserve the sentense "
>  In libavfilter, it is possible for filters to have multiple inputs and
>  multiple outputs.", but, the example here is just one video stream in
> and out, so, is this sentense a harmful influence on your
> example? Is it better to delete this redundant sentense? Or, add some
> thing to indicate this? for example, like below:
>  In libavfilter, it is possible for filters to have multiple inputs and
>  multiple outputs.
>  To illustrate the sorts of things that are possible, we consider
>  the following filtergraph(this filtergraph has only one input and
> output stream).

You're confusing filtergraphs and filters, in the example we have a
*filtergraph* with one input and one output, but there is a filter
with two outputs (split) and a filter with two inputs (overlay).

Since yours is a possible interpretation of the text, I tried to tweak
that in order to remove the confusion. Patch updated.
-- 
FFmpeg = Fantastic and Fiendish Mournful Portable Embarassing God
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0005-doc-filters-review-introductory-example-and-explanat.patch
Type: text/x-diff
Size: 3679 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20130409/fc61b131/attachment.bin>


More information about the ffmpeg-devel mailing list