FFmpeg vs FFvfw vs Mencoder ? - Page 23

audioslave · #**441** 04-05-2004, 05:01 PM

Thank you for your help! Now - FINALLY - I'm able to use Mencoder!

On the downside I'm not really impressed with the quality I got from the intro scene from the Abyss

. I'll look into the tweaking parameters to see if I'll get better results...

audioslave · #**442** 04-05-2004, 05:06 PM

Hmmm, I'm getting a buffer underflow message the line before the encoding start. What can be causing this

bilu · #**443** 04-05-2004, 05:18 PM

Quote:

Originally Posted by audioslave

Hmmm, I'm getting a buffer underflow message the line before the encoding start. What can be causing this

We all get that message at the beginning, it means that the buffer isn't full yet

Bilu

audioslave · #**444** 04-05-2004, 05:34 PM

Alrighty then!

Any idea on why the encoded video look so bad? The frames look almost as if I have encoded an interlaced source without a deinterlacing filter - which I haven't

.

EDIT:
Maybe you wanna see some screenshots, to better understand what I'm talking about

bilu · #**445** 04-05-2004, 05:37 PM

What was your command-line?

Try lmin=1 and no scplx_mask for the greatest quality.

Bilu

incredible · #**446** 04-05-2004, 05:45 PM

@Digi.Doc

Did you CHECK your stream in Bitrateviewer???
When setting lmin the Q curve sometimes plays crazy! Means sometimes it rises with bitrate (as wanted) and still keeps that high Q like constant! for about a minute (but no needing to keep that for a minute) even if already no compelx scenes/very low bitrate do happen!.

That risky bitrate rise (now 12000kbit) scene in "K19" is a scene where a man jumps over from the submarine into the ocean = complex water rumble scene! A hell!

audioslave · #**447** 04-05-2004, 05:51 PM

@bilu

WOW

Amazing quality!!! But (here we go again

) now I noticed the colors seems washed out. Once again the intro scene from the Abyss: The first cut to the interior of the submarine where the room is lit up by red lights - they aren't so red anymore. Look much more orange to me

Any tweaks for that?

bilu · #**448** 04-05-2004, 06:06 PM

@audioslave

Try removing the intra_matrix and inter_matrix settings.

Bilu

audioslave · #**449** 04-05-2004, 06:07 PM

@bilu
That would be the Notch matrix, right?

audioslave · #**450** 04-05-2004, 06:12 PM

Damn, this encoder is faaaast

I'm getting between 25-30 fps encoding speed on my AMD 2000+

@bilu
Removing the matrix didn't help with the color issues

.

bilu · #**451** 04-05-2004, 06:29 PM

try trell:mbd=2

audioslave · #**452** 04-05-2004, 06:38 PM

No difference. Wierd!

bilu · #**453** 04-05-2004, 06:43 PM

Remove vf=yuvcsp.

audioslave · #**454** 04-05-2004, 06:46 PM

I just ran my encoded video stream through ReStream and couldn't find any B-frames

This surely can't be right, can it?

@bilu
Testing your latest tip right now...

bilu · #**455** 04-05-2004, 06:50 PM

add vmax_b_frames=2.

audioslave · #**456** 04-05-2004, 06:52 PM

You poor thing. Helping an idiot like me all night long

. Don't get me wrong, I'm sooo grateful

EDIT:
I really think this color problem is an important thing to solve as fast as possible so we don't get these washed out looking movies. Don't you agree?

bilu · #**457** 04-05-2004, 07:08 PM

@Incredible

Quote:

of=rawvideo=1
ovc=lavc=1
nosound=1
noskip=1
vf=yuvcsp
lavcopts=vcodec=mpeg2video:vrc_buf_size=1835

reme =2

recmp=2:ildct=1
:ilme=1:vstrict=-1:autoaspect=1:vqblur=0:vrc_minrate=300:vbitrate=3 000
:vrc_maxrate=9800:vrc_eq=avgTex:vqmin=1:mbqmin=1:l min=2.49

#Inc
#:intra_matrix=8,13,18,23,28,33,38,43,13,18,23,28, 33,38,43,48,18,23,28,33,38,
43,48,53,23,28,33,38,43,48,53,58,28,33,38,43,48,53 ,58,63,33,38,43,48,53,58,63,
68,38,43,48,53,58,63,68,73,43,48,53,58,63,68,73,79 :inter_matrix=16,18,20,22,24,
26,28,30,18,20,22,24,26,28,30,32,20,22,24,26,28,30 ,32,34,22,24,26,28,30,32,34,
36,24,26,28,30,32,34,36,38,26,28,30,32,34,36,38,40 ,28,30,32,34,36,38,40,42,30,
32,34,36,38,40,42,44

#Notch
#:intra_matrix=8,9,12,22,26,27,29,34,9,10,14,26,27 ,29,34,37,12,14,18,27,29,34,37,
38,22,26,27,31,36,37,38,40,26,27,29,36,39,38,40,48 ,27,29,34,37,38,40,48,58,29,34,
37,38,40,48,58,69,34,37,38,40,48,58,69,79:inter_ma trix=16,18,20,22,24,26,28,30,18,
20,22,24,26,28,30,32,20,22,24,26,28,30,32,34,22,24 ,26,30,32,32,34,36,24,26,28,32,
34,34,36,38,26,28,30,32,34,36,38,40,28,30,32,34,36 ,38,42,42,30,32,34,36,38,40,42,44

Tried your matrix in comparison with Notch.

Inc
===
3815 KB
avg 875/2.30
peak 2639/3.68

Notch
====
4005 KB
avg 919/2.32
peak 2682/3.68

In BV they looked almost equal.
Looking at it I like Notch a bit more, more natual it seems.
But very difficult to distinguish between the two.

Bilu

bilu · #**458** 04-05-2004, 07:09 PM

Quote:

Originally Posted by audioslave

I really think this color problem is an important thing to solve as fast as possible so we don't get these washed out looking movies. Don't you agree?

Not right now, I still have other priorities. Others can walk this path

Bilu

bilu · #**459** 04-06-2004, 04:11 AM

My own compile, this one solves the B-frames issue.

http://clientes.netvisao.pt/bilu/bru...er20040405.zip

The last mplayerhq.hu compile was March 26th after all

Bilu

digitall.doc · #**460** 04-06-2004, 04:22 AM

Quote:

Originally Posted by incredible

@Digi.Doc

Did you CHECK your stream in Bitrateviewer???
When setting lmin the Q curve sometimes plays crazy! Means sometimes it rises with bitrate (as wanted) and still keeps that high Q like constant! for about a minute (but no needing to keep that for a minute) even if already no compelx scenes/very low bitrate do happen!.

That risky bitrate rise (now 12000kbit) scene in "K19" is a scene where a man jumps over from the submarine into the ocean = complex water rumble scene! A hell!

Inc, I did check my streams in BitrateViewer, this is the way I take my bitrate/Q values. And didn't notice the behaviour you refer, in my tests Q raises and falls fast, and if keeps constant (it really doesn't keep much constant, but when I used vqmin=2:lmin=1) it's at a low value.
And I tested mencoder with several bitrate-demanding scenes, and never get a too high bitrate (max about 10000 when used vqmin=1:lmin=0.5, but never again). Ahain, don't know if it's a matter of different mencoder compilation/version, or that I use vbitrate=9800 (differnet settings, I mean).
bilu, in your last test, when testing inc matrix (I still didn't

, sorry inc), I see you get nice quantizers (avg about 2.3 is not that bad) but really low avg bitrate: is it related to the source you used, or to your settings?. Did you compare this same slice with vrc_eq=tex vs avgTex?. I don't know if this same sample couldn't be encoded at lower quantizers and higher bitrate (that would mean of course bigger file size)... I can't forget we are encoding KDVD and we can use higher bitrates... Because we're encoding/talking about KDVD, don't we?
I find your results with inc matrix interesting, maybe he can tweak it more. And as I posted in the other thread, I'll also test vrc_eq=avgTex and see how it goes.
Cheers