[MEncoder-users] How does one interlace progressive content?

[Grozdan Nikolov]

On Wednesday 15 August 2007 14:47, Reimar Döffinger wrote:
> Hello,
>
> On Wed, Aug 15, 2007 at 12:38:59PM +0200, Pierre Catello wrote:
> > 2007/8/14, Reimar Döffinger <Reimar.Doeffinger at stud.uni-karlsruhe.de>:
> > > What means "encoded as 576 interlaced"? They could have simply used the
> > > interlaced motion estimation and interlaced dct, that works perfectly
> > > fine for progressive data, it just wastes lots of bits. That kind of
> >
> > Of course encoding as interlaced means processing the source as separated
> > fields and hence using interlaced-aware dct etc.
> > You're playing on words... If "encoded as interlaced" is not clear, then
> > lavc ilme, idct options are not clear, xvid interlaced encoding is not
> > clear, x264 interlaced encoding is not clear etc.
>
> No, I am not only playing with words, the original poster asked about
> how to "make it interlaced", 

Yes you are right. Let me explain why I started this topic... Because of my 
limited knowledge on how interlacing is made, I first though that they shoot 
the content in progressive mode at say 25fps and then use some filters or a 
process which makes interlaced content out of the progressive one, but still 
preserves the 25fps framerate... Because of this, I also thought that you can 
use some MEncoder filter to do the same procedure, but I was wrong.

Now, as Loren Merrit has explained to me in a previous reply, they don't make 
interlaced content out of progressive, but use a interlaced camera when they 
shoot the footage. I very well know that the ilme and ildct options don't 
make interlaced content out of progressive one. They are just there so that 
if one wants to encode interlaced content, the motion estimation and DCT can 
switch to a "special" mode to better treat the interlacing. The same is true 
for the software scaler. When you enable interlaced mode in the scaler, does 
it make interlaced content out of the progressive one? Nope... it just 
switches to this "special" mode so it can deal correctly and do its magic 
when it scales down/up interlaced content during encoding

PS: I'm not encoding my progressive stuff here to interlaced, I'm just trying 
to understand how interlacing is made... just to be clear :)


> and I really don't think he meant what ilme 
> and ildct. Those do not make anything interlaced, they just use an
> encoding method optimized for interlaced material. If you don't care
> about bitrate/quality, you can use them for progressive material or you
> can leave them out for interlaced material.
> That is not even considering things like pullup that some people do and
> some don't consider as interlacing.
>
> > thing can happen easily if you have the usual person without a clue
> >
> > > using an encoder that is crap (and looking at some of my DVDs, there
> > > are a lot of both in that field - but with most using dual layer DVDs
> > > that's not really much of a problem).
> >
> > You're wrong. Interlaced encoding can arise from other reasons, and with
> > people who do have clues about video processing.
> > But, you, do you have any clues about industrial processes (in particular
> > DVD production), or broadcast processes. Do you really think that someone
> > is "manually encoding" each content that you find on a DVD or a DVB
> > stream ?? You of course know that in many cases hardware encoder are
> > used, or single pass real time encoder in the case of DVB, without
> > information about the incoming or source stream.
>
> DVB is something completely different, it can actually have realtime
> constraints, though honestly if it was really wanted you could probably
> change the process to use two-pass encoded stuff for at least 50 % of
> content.
> And in either case, there is no need for manual encoding, it only has to
> be known if the source is progressive or interlaced - especially for
> DVDs the effort of that compared to the effort the sometimes put into
> menus etc. should be quite small (excluding DVDs where interlaced and
> progressive is mixed in one title, which I haven't had so far).
> But of course I do know about "industrial processes", but even if the
> decisions made make perfect economic sense and the results are far in
> the "good enough" range (not always the case either), that alone doesn't
> make it not crap (from a technical standpoint) - especially using
> hardware encoders for DVDs is not really such a great idea at least
> nowadays.
> Btw. there is one reason why wasting bits on DVDs actually _does_
> matter - the reason is that your data rate is limited to 1x DVD speed,
> which can end up hurting video quality especially with lots of
> high-quality audio streams.
>
> Greetings,
> Reimar Döffinger
> _______________________________________________
> MEncoder-users mailing list
> MEncoder-users at mplayerhq.hu
> http://lists.mplayerhq.hu/mailman/listinfo/mencoder-users