x264 Rev.512

[guada 2]

Hello :) ,

I have a problem with this version.
I would like to know why, this version doesn't permit to use a bitrate chosen by an user.

For example:
VDM + x264 rev.512 cli + avs limitedsharpen

Rate control: all normal
MBs&Frames: all normal (max cons=2)
More: Parttion decision= 6b / Method= Exhaustive search / Max ref=5 / Deblocking filter= -2/-2

first past: 1 kbits
second pass: 4854 kbits

After the first past, i have this:

Error: 2pass curve failed to converge
targe: 4854, expected: 310,53 kbis,avg
Qp: 10?004
Try reducing target bitrate or reducing Qp.min (currently 10).
.

Then, what is my mistake, and why, I cannot have the bitrate as I wish.
To note: Xvid, RV10, Wmv9, VP6, DivX "NO PROBLEMO".

[foxyshadis]

1. always post command line with reports, and state whether it's svn (bob0r) or patched (eg, sharktooth).
2. exhaustive? why? o.O umh is essentially equivalent and much faster.

[guada 2]

Look at x264-2-1-1.stats:


#options: 720x576 fps=25/1 cabac=1 ref=5 deblock=1:-2:-2 analyse=0x3:0x113 me=esa subme=6 brdo=1 mixed_ref=0 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 chroma_qp_offset=0 slices=1 nr=0 decimate=1 bframes=2 b_pyramid=0 b_adapt=1 b_bias=0 direct=2 wpredb=1 bime=0 keyint=250 keyint_min=25 scenecut=40 rc=abr bitrate=1 ratetol=4.0 rceq='blurCplx^(1-qComp)' qcomp=0.60 qpmin=10 qpmax=51 qpstep=4 ip_ratio=1.40 pb_ratio=1.30
in:0 out:0 type:I q:21.00 itex:135354 ptex:0 mv:15619 misc:4627 imb:1620 pmb:0 smb:0 d:-;
in:3 out:1 type:P q:24.00 itex:11 ptex:6925 mv:4024 misc:1352 imb:4 pmb:571 smb:1045 d:-;
in:1 out:2 type:b q:26.00 itex:0 ptex:1035 mv:1497 misc:1044 imb:0 pmb:535 smb:1084 d:-;
in:2 out:3 type:b q:26.00 itex:0 ptex:1201 mv:2543 misc:1056 imb:0 pmb:529 smb:1082 d:-;
in:6 out:4 type:P q:32.00 itex:0 ptex:1379 mv:2687 misc:1198 imb:0 pmb:413 smb:1207 d:-;
in:4 out:5 type:b q:29.00 itex:0 ptex:1669 mv:1970 misc:1089 imb:0 pmb:450 smb:1169 d:-;
in:5 out:6 type:b q:32.00 itex:0 ptex:788 mv:1418 misc:1010 imb:0 pmb:314 smb:1306 d:-;
in:9 out:7 type:P q:40.00 itex:0 ptex:71 mv:1818 misc:1127 imb:0 pmb:311 smb:1309 d:-;
in:7 out:8 type:b q:37.00 itex:0 ptex:32 mv:454 misc:594 imb:0 pmb:92 smb:1528 d:-;
in:8 out:9 type:b q:40.00 itex:0 ptex:6 mv:136 misc:322 imb:0 pmb:29 smb:1591 d:-;
in:12 out:10 type:P q:48.00 itex:0 ptex:37 mv:1063 misc:924 imb:0 pmb:207 smb:1413 d:-;
in:10 out:11 type:b q:45.00 itex:0 ptex:0 mv:0 misc:176 imb:0 pmb:0 smb:1620 d:-;
in:11 out:12 type:b q:48.00 itex:0 ptex:0 mv:5 misc:187 imb:0 pmb:1 smb:1619 d:-;
in:15 out:13 type:P q:51.00 itex:2 ptex:41 mv:923 misc:786 imb:1 pmb:171 smb:1448 d:-;
in:13 out:14 type:b q:51.00 itex:0 ptex:0 mv:0 misc:176 imb:0 pmb:0 smb:1620 d:-;
in:14 out:15 type:b q:51.00 itex:0 ptex:0 mv:0 misc:176 imb:0 pmb:0 smb:1620 d:-;
in:18 out:16 type:P q:51.00 itex:10 ptex:48 mv:1094 misc:864 imb:14 pmb:197 smb:1409 d:-;
in:16 out:17 type:b q:51.00 itex:0 ptex:0 mv:2 misc:182 imb:0 pmb:1 smb:1619 d:-;
in:17 out:18 type:b q:51.00 itex:0 ptex:0 mv:0 misc:176 imb:0 pmb:0 smb:1620 d:-;
in:21 out:19 type:P q:51.00 itex:11 ptex:37 mv:1168 misc:936 imb:10 pmb:214 smb:1396 d:-;
in:19 out:20 type:b q:51.00 itex:0 ptex:0 mv:0 misc:176 imb:0 pmb:0 smb:1620 d:-;
in:20 out:21 type:b q:51.00 itex:0 ptex:0 mv:0 misc:176 imb:0 pmb:0 smb:1620 d:-;

etc.........

[guada 2]

At x264-2-1-2.stats:

#options: 720x576 fps=25/1 cabac=1 ref=5 deblock=1:-2:-2 analyse=0x3:0x113 me=esa subme=6 brdo=1 mixed_ref=0 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 chroma_qp_offset=0 slices=1 nr=0 decimate=1 bframes=2 b_pyramid=0 b_adapt=1 b_bias=0 direct=2 wpredb=1 bime=0 keyint=250 keyint_min=25 scenecut=40 rc=2pass bitrate=4854 ratetol=1.0 rceq='blurCplx^(1-qComp)' qcomp=0.60 qpmin=10 qpmax=51 qpstep=4 cplxblur=20.0 qblur=0.5 ip_ratio=1.40 pb_ratio=1.30
in:0 out:0 type:I q:10.00 itex:406990 ptex:0 mv:24673 misc:4841 imb:1620 pmb:0 smb:0 d:-;
in:3 out:1 type:P q:10.00 itex:582 ptex:174135 mv:9144 misc:843 imb:6 pmb:1438 smb:176 d:-;
in:1 out:2 type:b q:12.00 itex:0 ptex:38959 mv:6082 misc:1055 imb:0 pmb:788 smb:754 d:-;
in:2 out:3 type:b q:12.00 itex:0 ptex:38309 mv:7707 misc:1088 imb:0 pmb:696 smb:774 d:-;
in:6 out:4 type:P q:10.00 itex:3852 ptex:185130 mv:10643 misc:735 imb:20 pmb:1415 smb:185 d:-;
in:4 out:5 type:b q:12.00 itex:0 ptex:34588 mv:8151 misc:933 imb:0 pmb:707 smb:796 d:-;
in:5 out:6 type:b q:12.00 itex:2 ptex:34791 mv:7803 misc:956 imb:1 pmb:678 smb:807 d:-;
in:9 out:7 type:P q:10.00 itex:5927 ptex:215955 mv:12520 misc:670 imb:34 pmb:1442 smb:144 d:-;
in:7 out:8 type:b q:12.00 itex:106 ptex:41729 mv:9624 misc:813 imb:1 pmb:802 smb:679 d:-;
in:8 out:9 type:b q:12.00 itex:3 ptex:48637 mv:10829 misc:947 imb:1 pmb:751 smb:696 d:-;
in:12 out:10 type:P q:11.00 itex:4876 ptex:182433 mv:11447 misc:684 imb:24 pmb:1442 smb:154 d:-;
in:10 out:11 type:b q:13.00 itex:3 ptex:33866 mv:9702 misc:869 imb:1 pmb:767 smb:702 d:-;
in:11 out:12 type:b q:13.00 itex:259 ptex:39620 mv:10391 misc:906 imb:3 pmb:754 smb:692 d:-;
in:15 out:13 type:P q:11.00 itex:4904 ptex:182316 mv:11764 misc:648 imb:24 pmb:1422 smb:174 d:-;
in:13 out:14 type:b q:13.00 itex:0 ptex:33123 mv:10103 misc:862 imb:0 pmb:788 smb:689 d:-;
in:14 out:15 type:b q:13.00 itex:0 ptex:34639 mv:8724 misc:965 imb:0 pmb:749 smb:744 d:-;
in:18 out:16 type:P q:11.00 itex:5262 ptex:190861 mv:12160 misc:645 imb:27 pmb:1425 smb:168 d:-;
in:16 out:17 type:b q:13.00 itex:0 ptex:35333 mv:10228 misc:879 imb:0 pmb:783 smb:694 d:-;
in:17 out:18 type:b q:13.00 itex:0 ptex:37409 mv:9779 misc:852 imb:0 pmb:745 smb:710 d:-;
in:21 out:19 type:P q:11.00 itex:7434 ptex:192190 mv:12078 misc:674 imb:35 pmb:1423 smb:162 d:-;
in:19 out:20 type:b q:13.00 itex:3 ptex:35363 mv:10295 misc:955 imb:1 pmb:772 smb:702 d:-;
in:20 out:21 type:b q:13.00 itex:278 ptex:36056 mv:9562 misc:888 imb:3 pmb:753 smb:715 d:-;

etc......

[Pomyk]

Bitrate should be the same for both passes. You have bitrate=1 in the first pass which makes no sense.

[guada 2]

@Pomyk

It is not an obligation, because the first past makes only an analysis.
Whereas in the second pass, the bitrate is operational.

[akupenguin]

Bitrate is required on the 1st pass... or use equivalent quality in some other mode, i.e. cqp or crf.

You're thinking of XviD, which ignores bitrate on the 1st pass and always uses cq2 to matter what you ask for.

x264 encodes exactly as you specify. And analysing a 1kbps video (as if it could even go that low) gives no information at all about how to encode a 5mbps video. Hence why ratecontrol is confused.

[Pomyk]

In your example it encodes every frame at q=51. What analysis is that? Useless
Do you have any reason not to use the same value for both passes?

[guada 2]

Thanks to both.

But, just by curiosity,

What did you says about it?

When i make 2pass with the same bitrate (4854 kbits).
VDM in job contol (1st pass) write WARNINGS.

and the second pass 2pass WAITING, but for a long time (I don't know exactly because i follow again.

An idea...

[Sirber]

VDM and x264 CLI? :confused:

[guada 2]

Sirber


What is the problem?
if you use VDM with it, then...

look at this link:
http://mirror01.x264.nl/

[ChronoCross]

you can't use the cli with VDM. your using the vfw version which isn't as good as the cli. SO it's harder to diagnose your problem. however pengvado is probably right.

[Sagittaire]

Sirber


What is the problem?
if you use VDM with it, then...

look at this link:
http://mirror01.x264.nl/

it's not problem ... it's only warning

VDM or VD make that with slow codec ...

[tomos]

Bitrate is required on the 1st pass... or use equivalent quality in some other mode, i.e. cqp or crf.

You're thinking of XviD, which ignores bitrate on the 1st pass and always uses cq2 to matter what you ask for.

x264 encodes exactly as you specify. And analysing a 1kbps video (as if it could even go that low) gives no information at all about how to encode a 5mbps video. Hence why ratecontrol is confused.

sorry about taking this OT for a sec but i wanted to check this post.

say i do a first pass with say 5meg bitrate. There would be more info in the stats file than if i do the first pass and specified 4meg bitrate?

If thats right, and i wanted to encode and get the max compression on a high quality source - could you do 3 passes with each pass lowering the bitrate?

e..g say your target bitrate is 4mb/sec, could you do the 1st pass @ 8mb, 2nd @ 6mb and finally the 3rd @ 4mb?

is that possible? and could it lead to improvements in the image over just a standard 3-pass encode @ 4mb?

[bob0r]

Sirber


What is the problem?
if you use VDM with it, then...

look at this link:
http://mirror01.x264.nl/

WTF is with mirror01? LOL
damn wikipedia:
http://en.wikipedia.org/wiki/X264
External links
The latest x264 builds daily (x264.nl) > http://x264.nl/
x264 binary with graphical user interface > http://mirror01.x264.nl/x264/ :rolleyes:

What the hell is this? They don't test the links? Oh well, weirdness will always happen i guess.

I know this has nothing to do with the current topic, but nor does this quote make any sense.

[ChronoCross]

I think the point is the first pass tries to aim for a specific bitrate. if you choose a different one in the second pass it'll always give the 2nd pass curve failed to converge message. the point of running passes is to greater refine the curve to control for a specific bitrate.

[foxyshadis]

It won't always fail to converge. As long as there's enough breathing space in the quant distribution of the first pass you'll be okay with any normal encode, perhaps slightly suboptimal if it's too far off. But a first or second pass that clings too close to the extremes won't mesh with one that doesn't very well, too much quantization noise in the stats. :p

Anyway, the guis all set it up to encode both at the same bitrate all the time, or this reason, so unless you're testing second pass bitrates at 50, 500, and 5000, with one first pass, it should never be a problem.

[guada 2]

Hello to all. :)

I'm very sorry.......
I found my problem. All came of me.

In fact, I wanted to force a bitrate on a small file.
After having to use Divx 6.2.2, CCE, and Nero digital (max avc).
I compared the real bitrate and those displayed in playing with Nero Show time 3. it is nearly similar. but below for what I hoped.

And there, I remembered GK, which recalculated automatically the ideal bitrate.
And it is exactly what do any codec, when the bitrate is very high.

But a question challenges me.
When one speaks bitrate in high definition ( 8mbps, 12 mbps, etc...); is it the maximal bitrate used, a middle bitrate or a value of reference?

[guada 2]

When one speaks bitrate in high definition ( 8mbps, 12 mbps, etc...); is it the maximal bitrate used, a middle bitrate or a value of reference?
No idea

Other questions on the high bitrate:
What are the limits (in term of bitrate) of the codec x264?

Bye.

[Sagittaire]

No limit I think ... just quantizer limit q1 for lossy quality. 1080p q0 lossless mode can certainely produce stream with more than 100 Mbps ...