[MEncoder-users] interlaced/non-interlaced

[Andrew Berg]

 On 10/5/2010 7:01 PM, Toerless Eckert wrote:
> On Sat, Oct 02, 2010 at 12:53:40AM -0500, Andrew Berg wrote:
> > There are two big reasons. First, progressive video is more compressible
> > than interlaced video with any codec. Second, the video would only need
> > to be deinterlaced once instead of every time you want to watch it.
>
> There are billions of TV sets out there dong deinterlacing every
> time they receive something so the second part of the argument isn't
> all that persuasive ;-)
It's inefficient because takes power and CPU cycles to deinterlace. With
hardware players, this isn't a huge issue, but it's unnecessary overhead
for software players.
> > I am not sure how 25p -> 50i is handled. It's possible that you are
> > receiving 25p video flagged as interlaced if the original is recorded at
> > 25p. You can play video back frame-by-frame with MPlayer by pausing it
> > and pressing the period key to advance a frame. If you see full frames,
> > it's progressive. If you see combing, it's interlaced.
>
> Unfortunately i always see full frames. Have you tried it ? mplayer
> seems to alwas be showing a full frame somehow interlaced from two field
> *sigh*.
If you're not seeing any combing, then it's very likely the video is
progressive. Broadcast MPEG-TS flags can never be trusted to indicate
whether a video stream is interlaced or progressive. I use this method
to help determine whether video is true 30i or telecined 24p if I'm
unsure. I tried this with pure 30i content and it's not as obvious as I
thought to see interlacing. You'll have to look very closely for
combing, especially if your input is the same resolution as your display.
> > If you use a deinterlacer on progressive content, you will likely have
> > very noticeable distortions. If the 25p video was interlaced before
> > broadcast, you've already lost (a very small amount of) quality.
>
> "if the 25p video was interlaced" ? What do you mean with that ?
I mean if the original progressive was interlaced (the verb, not the
adjective) before broadcast. 25p -> 50i.
> Why not ? If i receive an MPEG2 stream that indicates it's 50 fields
> per second interlaced, but every 2 fields actually belong to a single
> frame, then my deinterlacing operation is simple to combine the two fields
> into a field without any interpolation.
>
> Right ?
No. If you have frames, you leave the video alone. Honestly, I don't
know what's allowed to be in a broadcast stream in Europe, so I wouldn't
know if a station could broadcast pure 25p alongside pure 50i or not.
> > You will get better quality if you deinterlace to 25p (Yadif can output
> > one frame for every pair of fields) than if you stay with 50i. Yes,
> > technically, quality is lost when you deinterlace, but it's not much of
> > a loss.
>
> Somehow this seems to be a mantra everybody repeats without knowing
> real evidence.  How can you maintain 50 field/sec smoothness of motion
> after deinterlacing into 25p ?
You can't, but you are free to not deinterlace. If it is such an issue,
either don't re-encode or live with the fact that interlaced video is
less compressible than progressive video and have either larger files or
reduced quality. Try encoding with a constant quality (higher quality is
better since you're testing) and compare bitrates and quality among
25p/50i/50p. If you still think the higher bitrate is less of a problem
than the temporal resolution (I think that's the right term) loss, then
by all means, encode interlaced.
> But everybody else here seems to
> want me to deinterlace in all cases ;-)))
That's because we consider the tradeoff to be very worthwhile.
> Well, TV recordings usually willhave some pre and post roll which can
> have separate encodings from the main program recorded. In addition,
> if the recording is a made-4-tv news report or the like then it will
> be cut together from all type of input clips, some of them 4:3 AR,
> some 16:9, some interlaced, some 25p.
If the stream truly switches like that, then you have a real problem for
which I have no solution.
> So far i' ve not been too unhappy with mencoder and interlaced, except
> for what i think are bugs, and maybe those bugs are more common in
> less often used combination of options, and if everybody here subscribes
> to the church of deinterlacing, then i can see why i have more bugs in my
> encodings ;-))
It's far from dogma. It's a matter of opinion and most people would find
the efficiency gain worth the quality loss from deinterlacing after
doing their own tests.
> What rates do you think is sufficient for deinterlaced PAL (720 at 576x25p) ?
> I am using 1500kbps for interlaced. I alredy know it's not ideal
> for noisy (older) content, but its pretty good for newer (less noisy)
> content.
That depends on what you consider to be sufficient quality and on what
kind of content you're dealing with (noise is far from the only factor).
I've never dealt much with MPEG-4 part 2 and I don't ever plan to, so I
can't help you here.