The Coolness of B-Frames [Archive] - Page 2

skal

19th January 2005, 11:51

Originally posted by DeathTheSheep

Tell me why,
Pray tell,
The B-frames are blocky
As hell

For I
Don't see
Why oh why
That could be

" Thou shall not munch P-frames" shall be thy answer ;)

More seriously: i'm certainly not able to say for sure why your b-frames are a block soup, but one thing is for sure: the surrounding p-frames are already a mess, so one can hardly expect the b-frame to look real better.

Not, a piece of hint that might help you get better encode at that bitrate/framerate. Let's look at 3 frames at ~8fps (the upper ones in the following -ugly- sketch), and watch the coding in progress:

http://skal.planet-d.net/munch_me.jpg

Arghh! When the encoder is coding frame P1, look at what he has as reference! It's P0, coded at QP=45. A mere block feast!
Question: How do you expect to ME to find a good match between frame P0 and frame P1 ??? The first one is a block soup, but P1 still has lot of details. One could pick any block from the first frame and match it with the third, with a SAD of ~51340986542 (roughly) ;)
The motion vector field will be erratic, and the middle b-frame will find it difficult using the mighty BDirect mode (not to mention that it'll be hard to match the frame P0 / P1, too).

Now the hint: you should make an intermediate encode at less brutal bitrate/framerate (say: 200kbps @ 15 Hz), and use this as a source for the final encode. Instead of trying to jump directly at the extreme settings. This is shown in the last row of pics above. Matching midly decimated pics (QP=32) might be easier for ME.

just my 0.02euros, as usual.

-Skal

yaz

19th January 2005, 14:35

hmm ... what about making an intermediate 'hi-def' reference stream by downsizing and downsampling to the desired scale but using one of the lossless codecs or using sg like xvid w/quant=1or2 and w/I-frame only. (it'd be sg like 'capturing' the source) this way u can also check whether this stream is watchable at all. w/that fps it seems like rather a cartoon not a film to me.

the bests
y

[/EDIT]@SMD
that's all i have about the 'coolness of b-frames at the moment :-))[EDIT]

DeathTheSheep

19th January 2005, 22:33

Wow. :goodpost: :goodpost: :-) Good advice, I s'pose.

DeathTheSheep - thank you for the pictures. now we can easily see what you mean. BOTH look terrible and unwatchable.
I revised my tests and now use much lower quantizers and full framerate.

w/that fps it seems like rather a cartoon not a film to me

Yeah, it is! lol
you should make an intermediate encode at less brutal bitrate/framerate (say: 200kbps @ 15 Hz), and use this as a source for the final encode.

But as you shall soon see....
However, now I AM trying a less brutal framerate (24fps vs. 7-8fps) and a less brutal quantizer (31 instead of 45). (Like proposed intermediate, kinda).

The results? The B-frame encoded clips are terrible. TERRIBLE. The motion is so horrendous, and the quality is EXTEREMELY inconsistent.

Screen of EXACT SAME FRAME (same thing happens for just about every frame, too, not just this one. I'd say 1/3 or at max 1/6 frames, and even worse, some of them.

Screen:
NO DANG B-FRAME MADNESS!! Smooth pic (AND SUPER smooth motion.)
http://img129.exs.cx/img129/1251/smoothnobframes5qe.png

B-Frame madness.
http://img129.exs.cx/img129/3933/blockybframes0pp.png
If you think the pic is bad, just wait 'till you see the motion. It seriously gives you a headache. The characters go through trees! Through them. And the inconsistent quality, wow. First, there's a block-free I frame. Then, 2 block-free P frames. Then WAAM, your senses are bombarded with hideous blocking. And then, the blocks smooth out, then WAM!! double the blocks....
it's pretty clear that b-frames are not good under these (very rare) conditions.
Well, this is 320*240. Quant=31. Deblock on 2 (on 6 it almost gives me seizures to see the quality decrease-increase-decrease-increase...) 24fps. Max motion prediction on everything. 5 subpxl refinement (max). 30 refs. MAX b-frame=1 or 2 ... I forget. Rare, eh? I guess it was... Well, now it ain't! ;-)

[EDIT]
I will now address other matters.

many people have been drawn into this "Coolness of B-Frames" discussion under false pretences!

Eheh, you see, its ALL about the B-frames. There's a lot of stuff I'm concerned about regarding them. A quick recap:

1. Their coolness is off the charts (if you turn the chart upside down, heh)

2. I was merely asking why B-frames make the picture so terrible...

3. Other stuff (dynamic B-frame insertion, B-frame deblocking, B-frame renderer problems--no postfiltering of B-frames, etc)

Manao

19th January 2005, 23:00

Why are you complaining about blocking when you know the decoder you're using does'nt support deblocking on B-frames ? You're not happy ? Use another decoder or wait till deblocking on B-frames is implemented.

Appart from that, you should definitely try Nero's H264. B-frames are dynamic, and the decoder deblocks b-frames.

DeathTheSheep

19th January 2005, 23:09

Quick question... Can the Nero decoder somehow be used with AVI? Will it deblock x264 B-frames?

And yeah, I'm pointing out that B-frame deblocking is nonexistent, and that poses a huge problem without lowereing the filesize all that much.

So why is there all this hype about B-frames? They most certainly aren't nearly as "cool" as people make them out to be. Why the huge fuss over something that damages the quality?

At quantizers near 30, this is quite a problem. Many folks here didn't beleive me when I said The blocking and quality of B-frames is terrible...

Nero... The Best of The Best. Yep. But it's in MP4 format, which nothing I own supports... and it can't be edited (which I do a ton of, beleive it or not). The reason for x264 is the AVI container, man!!

That's a rockin' container. GO AVI. Wow, now I'm off topic... Baa

DeathTheSheep

19th January 2005, 23:16

My final word:

at lower bitrates B-frames are much better at increasing quality
False!
The results are a lot better with lower quant than those obtained with a higher B-framed quant at the same filesize...(In terms of blocking and overall quality)
True as rain, baby.

Manao

19th January 2005, 23:19

But will you quit falling always to the same (wrong) conclusion ? You're judging something with the wrong tools, so the results are expected.

DeathTheSheep

19th January 2005, 23:21

Wrong tools? ...

Oh, that explains it... :cool:

I knew B-frames were too crappy to be true in my tests...*sigh*

So how can I go about finding the right tools? I'm curious now...

bond

19th January 2005, 23:50

to get correct deblocking on b-frames you have to do the following:

1) avi -> raw with mpeg4ip's avi2raw tool
2) raw -> mp4 with mpeg4ip's mp4creator tool (make sure to use the latest version (~1 day old))
3) decode with nero's dshow .mp4 parser and decoder

Originally posted by SeeMoreDigital
All the samples are Mpeg4/AVC (VideoSoft H.264 Codec 2 to be more precise).nope they have been encoded with x264 (only the fourcc has been changed to VSSH)

dragongodz

20th January 2005, 01:29

quote:"at lower bitrates B-frames are much better at increasing quality"

False!
no not false at all. try an easy experiment. make 2 encodes with xvid with full framerate and a reasonable frame size(lets say atleast 352x288/240) and set a smallish target size(something that should average around 400kbs would be good). do 1 with Bframes and 1 without, THEN you may see Bframes do actually work.

I knew B-frames were too crappy to be true in my tests...*sigh*
your test ,as has been already said by many, is the problem. at the ridiculously low bitrate nothing looks good and there is not enough detail left for motion vectors to make any sane decisions, Bframes OR Pframes, Bframes just get screwed up worse in this example.

The main advantage of the b-frames is not the detail gain, but the better motion handling imo
and the better detail preservation gained from the better motion handling ? ;)

there are still problematic blocks, but i think it's just due to the immature b-frame ME in x264
yes i already said Bframes are the least mature part of x264 and in another thread akupenguin reccomended only using 1 Bframe for the time being because of it. so saying Bframe as a whole are bad or whatever is plain wrong of course, which i dont think DeathTheSheep is ever going to understand. so i wont bother again here, i have better things to do than waste time convincing 1 person they dont know what they are talking about. :D

DeathTheSheep

20th January 2005, 22:21

Bframes are the least mature part of x264
True. I confess. I'm starting to see things clearly (glaucoma clearin' up, heh, lol).

akupenguin reccomended only using 1 Bframe for the time being because of it
...Although I did only use one in my higher bitrate tests, still, yeah, I get the picture.

so saying Bframe as a whole are bad or whatever is plain wrong of course...
Well, I never actually said that they were bad as a whole. But yeah, I did complain about the blocksing and all that, so I suppose. NERO has a good deblocking postfilter, which bond explained how to use (and fairly well, too!).

is plain wrong of course...

Eheh, beleive it or not (likely not, but who beleives sheep these dayz?), I'm truly not the only person who thinks that. Many folks I know who are partially involved in video encoding push against B-frames.

which i dont think DeathTheSheep is ever going to understand

Hey, I understand now! Ease off, man! I'm second person here, too... <[ sweat bead ]> Eheh...

PS
Although I'd recommend dragongodz read the thread a bit more thoroughly... I suppose you'd see that I didn't just test at ridiculously low bitrates. I did one at "near Bearable" quality, too (to my eyes, heh, which means quantizer ~30, which is recommended for x264.)

Dhampir

20th January 2005, 22:50

Originally posted by DeathTheSheep
Although I'd recommend dragongodz read the thread a bit more thoroughly... I suppose you'd see that I didn't just test at ridiculously low bitrates. I did one at near perfect quality, too (to my eyes, heh, which means quantizer ~30, which is recommended for x264.)

If that's considered "perfect quality" you are again having a case of glaucoma or you are just plain blind. Even that quant 30 encode looks pretty blocky with or without b-frames. The difference being that the p-frame is being deblocked while the b-frame is not, but even with the smoothing the blocking is still quite apparent in the p-frame. If the b frame was being deblocked I'd be willing to bet that they would look almost the same. Besides comparing a P frame to a B frame is a pretty meaningless comparison considering the conditions (the p-frame is not only a lower quant but is being deblocked) which would be like comparing an I frame to a P frame and saying that P frames suck cause they don't look as good as the I frame (again a ridiculous conclusion considering the comparison).

DeathTheSheep

20th January 2005, 23:40

Heh, sure. To each his own, as becomes apparent in this thread... lol

Yeah, even I can admit it's far from perfect...I just tried to get something that you guys would consider "bareable" or even "watchable for short periods of time without headaches" heh. Yeah, I admit, it's pretty far from perfect. I'll edit that out! Ya gotta love computers... :D

I hereby place a request to moderate the "glaucoma" forum. :p

Shinigami-Sama

16th February 2005, 05:00

Originally posted by Dhampir
Here (http://www.funknmary.de/movie/) is where the samples are. They are quite hideous to watch whether they have b-frames or not.
I'm ging to agre with sheep ehre
the b frames do look like crap
but I also have very sensitves eyes
even though I had shitty glases 0_o
anyways
but hte problem is hte bitrate is to low for acurate b-frame vectors
so hte de/edncoder doesn't know what hte hell do with b-frames
and it kinda can't increase the quants anymore because you seem to have hith te limit
also
with b-frames I can only read about 1/2 hte subs
and your audio blows even for that small a file size
I wonder what happened htere
but a certain plant help glacoma sheep ;)
maybe you wer useing a bit to much medicine when encoding/watching these eh sheep? ;)

KaiserS

16th February 2005, 05:22

Originally posted by Shingami-Sama
I'm ging to agre with sheep ehre
the b frames do look like crap

But the P and I frames look like crap as well.

Shinigami-Sama

16th February 2005, 06:38

but with that being said
it means the b-frames will look worse
seeing as how they;re hte more or les byproduct of hte p and i frames
there isn;t enough data left to justify useing b-frames at that low of a bitrate

SeeMoreDigital

16th February 2005, 10:24

I think this thread has run its course and should be closed now... It's doing more harm than good!

Both RadicalEd (http://forum.doom9.org/showthread.php?s=&postid=597725#post597725) and Skal (http://forum.doom9.org/showthread.php?s=&postid=597918#post597918) have pointed out why the original posters B-VOP comments are based on flawed assumptions!

In essence... DeathTheSheep's B-frames look poor because his I and P "reference" frames were poor to begin with!

Elias

16th February 2005, 14:10

Originally posted by DeathTheSheep
Yeah, I've heard all the hype. I see it written about in posts all the time. Never a day does pass without questions pertaining in some form or another to B-Frames ("When are B-Frames gunna be enabled for AVC?!!!!"). But now I must ask, what the heck is up with B-frames anyway?

I heard they are a form of bidirectional frame encoding/rendering. At this, I scratched my head long and hard and almost pulled a "pancake bunny." So what?!

B-Frames are complicated, they take long to encode, they are BLOCKY as HECK, and they increase filesize. Need I upload a sample with and without B-Frames fer y'all to see? Why don't people stick to the IPPPPPPPIPIPPPPPPPPPP... method? Complexity? If so, I don't see it.

So what is all this B-Frame craziness? What use to they serve? Or are they only useful at collosal bitrates (yeah, I'm livin the >56kbps road)? NeeheeharharharHAR.... No, seriously,
Inform Da Sheep Afore He A-Kills Hisself...BAAAa Wow, was that English? I'm impressed, you should become a literature writer. As for B-Frames, they're good. The image becomes somewhat softer, as (opposed to sharper) but you loose lots of macroblocks. So, conclusion is to not use B-Frames with higher bitrates. Only with low bitrates is it to be used. Besides, it degrades the compatibility for the MPEG-4 file. I've completely stopped using B-Frames.

Tommy Carrot

16th February 2005, 14:26

Originally posted by Elias
As for B-Frames, they're good. The image becomes somewhat softer, as (opposed to sharper) but you loose lots of macroblocks. So, conclusion is to not use B-Frames with higher bitrates. Only with low bitrates is it to be used. In my opinion b-frames are always beneficial, except when the codec reaches the maximum quality (but this is only problem with mpeg ASP, as AVC becomes saturated at MUCH higher bitrate), and - as DeathTheSheep's example showed - at ridiculously low bitrates, where the reference frames are already useless. The softening effect is not necessary, it's quite the opposite, the bitrate saving of b-frames helps to preserve even more details.

Elias

16th February 2005, 14:31

Originally posted by Tommy Carrot
In my opinion b-frames are always beneficial, except when the codec reaches the maximum quality (but this is only problem with mpeg ASP, as AVC becomes saturated at MUCH higher bitrate), and - as DeathTheSheep's example showed - at ridiculously low bitrates, where the reference frames are already useless. When I said low bitrates, I meant like 1024 kbit/s etc. using very very very low bitrates... it doesn't matter what you use, it will look shitty either way. Honestly, I don't see the point with B-Frames at bitrates such as 3 mbit and up. It'll only make the video not playable on some DVD standalones and... well, Quicktime :) Who cares about Quicktime anyway. Hehe, seriously though, there are lots of software players that can't handle B-Frames/GMC etc.

yaz

16th February 2005, 14:44

Originally posted by Elias
... seriously though, there are lots of software players that can't handle B-Frames/GMC etc. wouv ... just point me only one :) anyway, handling b-frames is not a player issue when sw environment is considerded but a codec one. hw players are another part of this prob, of course.

however (just my simpleton approach) why do u think so many ppl of brilliant minds & skills would put so much effort on implementing b-(s-, n-)frames in their codecs if it(they) were such useless a/o harmful as u state ? ever thought of that ?

the bests
y

Elias

16th February 2005, 14:53

Originally posted by yaz
wouv ... just point me only one :)
Philips MPEG-4 Player.
Originally posted by yaz
however (just my simpleton approach) why do u think so many ppl of brilliant minds & skills would put so much effort on implementing b-(s-, n-)frames in their codecs if it(they) were such useless a/o harmful as u state ? ever thought of that ?

Originally posted by Elias
As for B-Frames, they're good. The image becomes somewhat softer, as (opposed to sharper) but you loose lots of macroblocks. I said it was good, not harmful. I rest my case.

KaiserS

16th February 2005, 15:11

Originally posted by Elias
[B]Philips MPEG-4 Player.

I have the DVP-642 and it handles my XviD encodes with 2 consecutive b-frames just fine.

Elias

16th February 2005, 15:13

Originally posted by KaiserS
I have the DVP-642 and it handles my XviD encodes with 2 consecutive b-frames just fine. I was talking about the software player.

KaiserS

16th February 2005, 15:16

Originally posted by Elias
I was talking about the software player.

Oh, my bad.

Elias

16th February 2005, 15:17

Originally posted by KaiserS
Oh, my bad. No problem :)

rawr

16th February 2005, 22:45

I second the 'use snow' call a while ago. Sure, it may not look much like the source at ultralow bitrates, but it's certainly pretty. :D

rawr.

DeathTheSheep

17th February 2005, 06:30

Page 3, 3rd post from the bottom. It's a much better quality example of the point I was making. No blocks at all, very visible picture, crystal clear on all P-frames. But if you look at the B-frame version, it is visibly worse. And this isn't a glaucoma picture like the horrible quality thing I showed at the top of page 3. Disregard the bad quality ones for download, and look at the good quality one. (I had it up for download in super high quality, but I was forced to remove it from angelfire (not enough room).)

Speaking of glaucoma, anyone know what kind of plant cures it?

Mindless prattle ensues; spare yourselves boredom and confusion by deigning not to heed it...

BaaAAAa, I'd LOVE to be a novelist. Problem is, who would buy my sheeply satires? (HAH, look, what a joke, I misspelled it...and that TOO!! NEHEHE, ok I'll shut up now... I wonder how long the longest post in history was... What would happen if France sunk to the bottom of the sea...FrAtlantis...What if ffdshow cures my glaucoma...wait i dont have glaucoma....wt*..wouldn't the world be cool if anime was real...no, not RealVideo...altho that would be nice too...)

Shinigami-Sama

17th February 2005, 07:58

Originally posted by DeathTheSheep

Speaking of glaucoma, anyone know what kind of plant cures it?
locoweed is cure mate ;)
best is avable in none other than my home province of B.C..:cool:

Originally posted by DeathTheSheep
Page 3, 3rd post from the bottom. It's a much better quality example of the point I was making. No blocks at all, very visible picture, crystal clear on all P-frames. But if you look at the B-frame version, it is visibly worse. And this isn't a glaucoma picture like the horrible quality thing I showed at the top of page 3. Disregard the bad quality ones for download, and look at the good quality one. (I had it up for download in super high quality, but I was forced to remove it from angelfire (not enough room).)

yeah anglefire sucks..
I can copy/paste hte link ot hte homepage
and keep hitting enter and every few times I get a pornsite load up -_-
and a bad one at that
and by 'good'qual you mean hte non-2meg ones?

Sharktooth

17th February 2005, 13:28

Ok, wait...
B frames were never meant to look like P or I frames. They are there to spare bits so the codec can redistribute them for a better overall quality.
They were intended to look worse than P and I... but as long as you keep the max cons. b-frames value to 1 or 2, during playback, your eyes won't be able to spot the difference (that's the theory behind b-frames...).
A frame by frame comparison will show the difference but that's a nonsense coz you dont watch movies frame by frame...

Mug Funky

17th February 2005, 13:40

well, actually, b-frames at the same quantizer is the rest of the video i imagine would look very much the same (this wouldn't apply to very high quants though). the trick is that b-frames are able to deliver comparable quality at a higher quantizer than p and i-frames.

the reason we have no more than around 2 consecutive b-frames is because they don't predict from other b-frames, but rather the nearest p or i frame. this means the reference frames are further apart, and thus the motion-compensation becomes less useful. this is where DeathTheSheep's problems come in - he's decimated the frame-rate, which does practically the same thing as increasing consecutive b-frames (except the overall motion is jerkier). you have the combination of slow playback giving the eyes more time to catch badness, and poor motion-compensation because the interframe differences are much higher.

k... hope that makes sense. i'm asleep at the keyboard here :)

dragongodz

17th February 2005, 13:56

is there an echo ? or is that a scratched record that keeps repeating ? funny i can hear the same things being repeated over and over and over and over......

leaves just one thing to say really.

DIE THREAD. JUST BLOODY DIE ALREADY. WHY THE HELL WONT YOU UP AND CARK IT ALREADY ? :devil:

DeathTheSheep

17th February 2005, 22:03

DIE THREAD. JUST BLOODY DIE ALREADY.

We hear ya, mate.

If you hear the same things, something's gone wrong... People aren't looking at page 3, 3rd post from the bottom.

No, framerate isn't decimated. 23.976fps. That post (3rd page, 3rd from bottom) will cure all. Even better than locoweed.

That's right, 3rd page, 3rd post from the bottom is a different test--not the one that's 2mb.

So load up on your locoweed guys, and head on over to take a peek at the post everybody missed.

One last time, the post that everybody missed is on the 3rd page, and it't the 3rd post from the bottom.

Then this thread will die. I'll die it myself. With locoweed. *cough*.

ChronoCross

18th February 2005, 01:00

Originally posted by DeathTheSheep
Eheh, I think I'll pass. Let me work one up for you, man! And at a lower quant. Baa! :p

Ah, here goes.

Bframes: (not same frame according to VDub, because B-frames through off frame count, another example of B-frame coolness hehe)

As you can see, b-frames give terrible blocking while movement is taking place, and I can't tell what the heck is going on. How many characters do you see? Maybe no the best example, but you get the point. Sample 1, frame 58 (B) or 55 (P):
http://img49.exs.cx/img49/1268/bf0jl.jpg

NO BF
While still pretty blurry (as i said, not the best example), it's a heck of a lot worse. I'll try higher br later. How many characters are there? Fundemental estimation difficulties with low framerate.
http://img49.exs.cx/img49/6277/nobf6bb.jpg

Convinced?

You should be shot for ruining a scene from naruto. Did you try reencoding a already horrible codded fansub rip from Dattebayo? If so you are #1 working with a bad source...trying to recode a poorly done xvid into a smaller xvid will only make it worse. either that or it was a divx raw provided by saiyaman.

B-frames in 95% of cases, especially with anime will improve the quality of encodes. the common rule to follow when you are encoding anime is 175MB per 23.5 minutes of video, even xvid at defaults should handle this quite nicely. also filtering can have a big effect on the quality. Most people are in too big of a hurry to apply proper filtering of TV rips. TV rips most consistently when captured from a video capture card are oversharpened which takes up valuable bitrate so you should smooth it down slightly. Additionally you should always return a 120fps encode to 29.997 or 24 fps in order to save space because there will be almost no lack of smoothness if it is decimated correctly, another thing that a majority of people lack. If you feel like bframes are still totally screwing the quality then you should lower the sensetivity to reduce the number used. But saying that bframes lower the quality considerably is totally bogus. in most cases it will even increase the quality tremendously.

btw that clip is of chouji, shikamaru, Naruto and Neji Chasing the sound 4 through the forest to rescue sasuke.....

Shinigami-Sama

18th February 2005, 02:17

umm
those are from a very small source
and naruto isn't all that great now anyways it aws but...
anyways
bt;s qual isn'tthat bad better tha nsome of hte groups
and saiyaman blows to start with
that was db because it;s subed hte same way
those are like a quant of 45 I think
and untill you said it
I did;nt see anything in that pic
Sheep is using bitrates so low
that you can't get any motion vectors out of hte p&i frames to make bframes useable
like said in pretty much every 3rd post...

DeathTheSheep

18th February 2005, 02:45

Originally posted by DeathTheSheep

I revised my tests and now use much lower quantizers and full framerate (24fps vs. 7-8fps) and a less brutal quantizer (31 instead of 45). (Like proposed intermediate, kinda).

The results? The B-frame encoded clips are terrible. TERRIBLE. The motion is so horrendous, and the quality is EXTEREMELY inconsistent.

Screen of EXACT SAME FRAME (same thing happens for just about every frame, too, not just this one. I'd say 1/3 or at max 1/6 frames, and even worse, some of them.

Screen:
NO DANG B-FRAME MADNESS!! Smooth pic (AND SUPER smooth motion.)
http://img129.exs.cx/img129/1251/smoothnobframes5qe.png

B-Frame madness.
http://img129.exs.cx/img129/3933/blockybframes0pp.png
If you think the pic is bad, just wait 'till you see the motion. It seriously gives you a headache. The characters go through trees! Through them. And the inconsistent quality, wow. First, there's a block-free I frame. Then, 2 block-free P frames. Then WAAM, your senses are bombarded with hideous blocking. And then, the blocks smooth out, then WAM!! double the blocks....

Well, this is 320*240. Quant=31. Deblock on 2 (on 6 it almost gives me seizures to see the quality decrease-increase-decrease-increase...) 24fps. Max motion prediction on everything. 5 subpxl refinement (max). 30 refs. MAX b-frame=1 or 2 ... I forget.

NOBODY CAN SEE THIS?!! :scared: :scared: Wow, everyone goes back to the crappy test, and everyone overlooks this one. Quant about 30, which is the recommended quant for anime. How on Earth can nobody not see this? Seriously, I'm just pointing this fact out, but it's kinda wierd. Maybe it only shows up on my computer. Aw, heck, I wish I could just delete the old one with the horrid 45 quant. NONONO!! Now everyone thinks I'm a jerk (I'm a SHEEP. A locoweed-ed SHEEEEP) because everyone thinks I'm going crazy with the low bitrates. And good folks like dragongodz are compelled to adopt expressions of confusion and frusteration and eventually anger at the sheep who JUST WONT GET IT... but it's not really ME who won't get it, because everyone skipped the most crucial post in the entire thread, assuming instead that good ol' Sheep's tests are WACK or borked (whatever "borked" means, jeez, it's not even in the dictionary)..
Ah, whatever. Someone autta slaughter the sheep--I mean, realize the sheep's message and leave without shearing the sheep's dignity... aw it's 11:30... Gotta do homework...
And what's wrong with Naruto? It's the best anime in the universe... It just keeps getting better!! Remember the old stuff with Haku and Zabuza? And then Gaara? And finally, Sasuke as evil... Man, it can't get much better.
And with x264 (WITHOUT B-FRAMES) I can compress an episode to 20MB (AAC-HE at 40khz, 31quant)...,

Shinigami-Sama

18th February 2005, 02:51

if you want us to use this as a test
post damed clips of it
unles that now wrong seeing as how it;s licensed

naruto got raly bad the last 40eps
and it jst stated to dig itself out of hte hole it dug
and I endoce anime at like err what quant I forget at like 1GB per 1.5hours so w/e mate you and your 20megs and ep

DeathTheSheep

18th February 2005, 03:17

Originally posted by Shingami-Sama
if you want us to use this as a test
post damed clips of it

Ah, I would, BAaa. The problem is, I have nowhere to host them. BTW, I believe short clips are permissible for demonstrational purposes, as long as the content matter as a whole isn't revealed...

So anywhere?

Have a good life. Death is around the corner. DeathTheSheep, that is. ;-)

ChronoCross

18th February 2005, 07:13

Ok you need to provide info about your source. because you might actually be one of the crappiest encoders alive if your bframes are that crappy, sorry for the language but this guy is totally off base with his encoding methods.

I need the following source info
Framerate: 120(NTSC/FILM lowest common Blend), 30(NTSC), or 24(FILM)
DeInterlaced by the capturer or by the encoder(you)
Capture codec: xvid or divx
Complete AVS
Complete Xvid settings.

It seems to me you are just doing stuff to purposely make bframes look really bad. I have done hundreds of tests and never run into the problems you seem to be having. Also are you using the xvid decoder or something else.