XviD-27072002-1

[Swede]

@uManiac
If your reading this, your SMP build link seems to be invalid. Maybe you set this purposely until the next build? He's still on vacation but I've just checked his dir's here and there's no SMP build at all in there... Prolly the SMP-builds has been broken somewhere. Perhaps Koepi can make one for you.

[Emp3r0r]

I just like to let everyone know about a test I did today using the newest XviD build and Divx5.02 Pro.

I created a video of Planet of the Apes

Divx5 Pro all default
w/ Bframes, qpel, gmc (no physco)

with XviD 27072002-1
UltraHigh
modulated
300 max kf spacing
1 min kf spacing
luma enabled
min I 1, max I 21
min P 1, max P 29
Iframe boost% 22
Below i-frame dist 8
iframe bitrate reduction %40
all else default (no special curve)

both videos were almost spot on the 744650 kbytes requested size.

Now for the intersting part, I sat my brother down in front of the two files and showed him the battle scene about 1.5 hours into it. I told him to look closely and see if one looked better than the other. Keep in mind he doesn't know how to type, let alone know the difference between XviD and Divx.

His reply, I can't tell a difference. Look again, closely I tell him. I can't much tell a difference he replies.

So there you have it, to the average eye, both codecs produce simular results.

For me, I also had no opinion which looked better, though I think XviD is more enjoyble to use :D Both had problems in the same areas and when there was noticable blocks, both had them. They both produced a wavey effect on some of the smoke scenes, but maybe that is how the DVD looks. Anyway, great job XviD team! It is amazing how you guys produce a codec that rival's a corporations money fed codec. Don't stop what yall are doing! ;)

avs file below ----------------

#
# Created with Gordian Knot
#
# http://thewef.nav.to
#
# PLUGINS
# get them from http://users.win.be/dividee
LoadPlugin("C:\PROGRA~1\GORDIA~1\mpeg2dec.dll")

# SOURCE
mpeg2source("C:\DVD's\Planet of the Apes\ape.d2v")
#
# TRIM
trim(0,162423)
#
# CROPPING
crop(5,55,709,365)
#
# RESIZING
BicubicResize(640,272,0,0.5)
#

[spyder]

Credits are very unstable here too. Lots of I-frames distributed throughout.

[athos]

Koepi> Does you latest build (ie. the name of this thread) include the latest changes regarding "consecutive iframes correction" by Foxer?
27.7.2002 21:40:
U vfw/src/2pass.c (rev.1.6) Foxer:
- consecutive iframes correction
U vfw/src/config.c (rev.1.15) Foxer:
- consecutive iframes correction
U vfw/src/config.h (rev.1.12) Foxer:
- consecutive iframes correction
U vfw/src/config.rc (rev.1.15) Foxer:
- consecutive iframes correction
U vfw/src/resource.h (rev.1.11) Foxer:
- consecutive iframes correction

[Koepi]

Athos,

I developped that stuff together with foxer and that build was testing the code. So you had it before it was in CVS. Since there where no bug-reports concerning that code i told foxer to finally commit it.

tses.
I start wondering what you guys think or if you are incapable to read the changelogs.

Regards,
Koepi

[athos]

i thought these were additional changes by foxer after your latest build, but now i know better

[rui]

Ok, it was about time that i did some tests with this new build :)

I used a small trailer (so it’s a not so significant test as you guys tests), and tried a number of diferent settings of the new i-frame parameters.

I left all default except motion search precision 6, modulated quantization at 2nd pass and used hinted ME.

Then i used :

Test 1 – below i-frame distance... 10
i-frame bitrate reduction % - 40

Test 2 – below i-frame distance... 20
i-frame bitrate reduction % - 40

Test 3 – below i-frame distance... 10
i-frame bitrate reduction % - 30

Test 4 – below i-frame distance... 10
i-frame bitrate reduction % - 50

Like all of you , i too noticed a great improvement in quality, especially since this was a small test. In a full movie i expect the quality diference to be even greater.

But the most important thing to me is that i couldn’t spot significant diferences between the diferent settings. This is good, because one can’t ruin a encode because of a not so good parameter choice.
But on the other side is bad because those experts out there can’t “play” so much with the settings to improve that final bit to get the near perfect encode. But since the quality is already very good i think that this isn’t very significant.

As always, it only my opinion (based on small tests, remember), others might think diferently.

[rui]

Hi Koepi:

I just saw at uManiac's site that you introduced "Support for XVID_GREYSCALE " in xvid.

Sorry for my complete ignorance :(, but what does it means? Judging for what is in uManiac's site, it seems some kind of method to save data from being encoded, alowing to achieve even more compression (?) :confused:

I tried to get some info both here and in xvid.org (i can't get to it, it seems to be down), but couldn't find any, except this post:
http://forum.doom9.org/showthread.php?s=&threadid=28250&highlight=Support+for+greyscale
A small quote from poptones: Layers are most powerful when used with masks. So, there's a mask() command to go along with layer(). Mask will take an RGB clip and apply the greyscale version of that clip as a mask to the first clip.

[Koepi]

Greyscale means:

throw away chroma information.

If you have unclean (noisy) b/w materia there is still some colour information in it which xvid encodes. In such cases, you can save up to 5-10% if you use that "enable greyscale" switch.

Btw., it was introduced by gruel but he doesn't use windows, so I stepped in and added the VFW support for it.

[rui]

Many thanks for the explanation. :)

So, i assume this means good news for the fellows out there encoding old black and white movies (that most probably are noisy material), but not much for the others that are encoding regular colour movies.

[Koepi]

You're welcome rui :)

Hm, of course it's useful for the others as well, just imagine you'd like your movie like in the old times.
Just leave away the sound and make it b/w!
Maybe you have some weird piano-tunes which you can loop and use it to give a proper musical experience ;)

Regards,
Koepi

P.S.: some people requested this feature, so here it is. Some features are useful for only a few people, you have to live with that or make up new using-ideas (like erm... forgot whom it was... abuses oggmux to split OGM files afterwards...) ;)

[rui]

Originally posted by Koepi
P.S.: some people requested this feature, so here it is. Some features are useful for only a few people, you have to live with that...

In fact, this is one of the wonderful things about xvid. Since it’s developed by lots of people, it always takes care of the wishes of the little guy.

Imagine divx doing something like this (i mean, listening to the wishes of the users) :rolleyes:

[sillKotscha]

Originally posted by Koepi
or make up new using-ideas (like erm... forgot whom it was... abuses oggmux to split OGM files afterwards...) ;)

it was me :p ... from now on I'll always "abuse" OggMux to split my OGM files

regards Sill

[Koepi]

Hehe, Sillkotscha, this wasn't meant bad!

I sure like if some people are so creative to find "new" ways to use a software. This "abuse" was meant as in: the software wasn't meant to be used _that_ way ( I coded it as Muxer, not as splitter ;) ), but I like it that it is capable of doing that.

If only some more people were that successfully creative! :)

Regards,
Koepi

[cult]

great news.I encode lot of b&w...hope u will ad it to ur binary soon koepi

[Koepi]

Hm, I could build a new binary...

Let's see...