On Wed, Jun 5, 2024 at 6:00 PM Mark Filipak <markfilipak.i...@gmail.com>
wrote:
> On 05/06/2024 09.44, Paul B Mahol wrote:
> >
> > What is OFFmpeg ? and how to use it?
>
> It is Object-FFmpeg. It is a fully video-object-oriented language, not
> functions pasted into python.
>
> Objects are video/audio/subtitle/chapter streams, frames, fields,
> pictures, halfpics, scans.
>
> There's very few global functions. Objects own methods, instead, of course.
>
> Methods are owner-specific, so they 'know' how to work on the object type
> that owns them. So,
> similar methods of two different objects can have the same name -- they
> don't need different names
> or calling options to differentiate them -- and there are _no_ calling
> parameters. Methods are
> simple, simple, simple.
>
> New use-case patches do not add function code. They add new properties and
> new methods to handle the
> things in the object type that are unique to the use-case but was
> previously unknown and not covered.
>
> Properties are also owner-specific. All the MPEG tags are properties for
> example.
>
> Properties and methods are inherited, so streams that are spawned
> (constructed) from parent streams
> are already defined and already know how to do almost everything. For
> example, child pictures know
> what their parent frame's MPEG tags were because it inherited the tags.
>
> Constructors are methods that spawn children that inherit properties and
> methods _plus_ are given
> those properties and methods that make them different from their parents.
> Constructors act like
> structs but they make objects.
>
> OFFmpeg execution -- the process pipe -- is state-machine driven. State
> machines are encapsulated
> and functionally complete by design. State-machines have 100% coverage and
> are 100% testable. When a
> state-machine works, it's done _forever_. When you are done writing code,
> you know you're done!
>
> There are two unique global methods: stride-match and ?-test. They are
> very powerful and complete.
> Stride-match was inspired by '-vf shuffleframes' but greatly expanded.
> ?-test was inspired by
> var=val?yes():no() but greatly expanded.
>
> There are global datatypes. For example,
>
> "pts" is the same as "x=pts" where 'x' is a temporary anonymous object
> upon which operations that
> follow that line work on until the end of code or until a different object
> is invoked or spawned.
>
> "pts'12345'" is the same as "pts=12345".
>
> "pts'+12345'" is the same as "pts+=12345".
>
> "?pts'12345'(abc:xyz)" is the same as "if (pts==12345) abc(); else xyz();"
> (though OFFmpeg methods don't pass parameters).
>
> "..n" means "from zero through n by indexing an owner's property" and it
> provokes automatic looping.
> So, for example,
> ?pts'..67890'(pts'+123':pts'-456')
> is the same as something like this:
> for (i=0, I=pts[0]; pts[i]; i++) {
> if (pts<=67890) pts[i]+=123;
> else pts-=456;
> }
> (I apologize for the 'C'-sh code above. I know the array indexes should
> really be pointers to a
> struct, but I don't know pointers to structs, I don't know 'C'. So, I
> tried to convey the 'flavor'
> of what I'm writing about.)
>
> There are drawbacks to OFFmpeg:
> It is very terse, very cryptic, hard to learn.
> OOP is difficult for folks used to functions.
> There will be a lot of pushback, especially against state machines. That
> will be true even though
> '-filter_complex' is already a partially implemented state machine.
>
> OFFmepg is easier than other OOP languages because there is only 1 level
> of ownership, and because
> the sole datatype is string.
>
> OFFmpeg is more accurate than other languages because It uses exact
> numerical expansions denoted
> like this: "23.[976023..]" and "3753.75[0..]" -- remember, all values are
> strings. (Yes, FFmpeg has
> 24000/1001 and 90090/24, but 23.[976023..] and 3735.75[0..] are easier to
> turn into decimals having
> the best native precision.)
>
> Finally, OFFmpeg has a ton of operators.
>
> Kindly forgive me if I've made any inadvertent mistakes above.
>
Where can I obtain OFFmpeg product/solution?
>
> --Mark.
>
> _______________________________________________
> ffmpeg-user mailing list
> ffmpeg-user@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-user
>
> To unsubscribe, visit link above, or email
> ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
>
_______________________________________________
ffmpeg-user mailing list
ffmpeg-user@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-user
To unsubscribe, visit link above, or email
ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
