Seikon no Qwaser BDMV

[romeyo007]

Hello
I have the BDMV for seikon no qwaser Anime and I'm trying to encode it to 1080p MKV the problem with that is the BDMV's are interlace here the Media info for the m2ts

any advice in encoding them? I'm using MeGUI for my encode
thank you

[mp3dom]

You can try to see if it can be inverse telecined.

[romeyo007]

is that option available in MeGUI
when i tried to analyse the dga file it gave me this setting:
LoadPlugin("C:\Program Files\MeGUI\tools\dgavcindex\DGAVCDecode.dll")
AVCSource("D:\Anime\RAWS\SnQVol.1\QWASER OF STIGMATA VOL1\BDMV\STREAM\00004.dga")
LoadPlugin("C:\Program Files\MeGUI\tools\avisynth_plugin\TIVTC.dll")
LoadPlugin("C:\Program Files\MeGUI\tools\avisynth_plugin\EEDI2.dll")
LoadPlugin("C:\Program Files\MeGUI\tools\avisynth_plugin\TDeint.dll")
edeintted = AssumeTFF().SeparateFields().SelectEven().EEDI2(field=-1)
tdeintted = TDeint(edeint=edeintted,order=1)
tfm(order=1,clip2=tdeintted).tdecimate(mode=1,hybrid=1)
#crop
#resize
#denoise

but the problem still there

[Blue_MiSfit]

Post a stream sample. You're probably not handling ivtc/deinterlace correctly.

You can definitely do a better job than TDeint+EEDI2 for the deinterlacing.

Derek

[romeyo007]

do u want an encoded it sample or sample for the m2ts?

[RunningSkittle]

post sample from m2ts on mediafire or somewhere similar

[romeyo007]

post sample from m2ts on mediafire or somewhere similar

how can I cut a part from m2ts and save it as m2ts?

thanks

[Hiritsuki]

What!! BDMV is interlaced, I first heard.

[romeyo007]

What!! BDMV is interlaced, I first heard.

I know i was surprise to but it's the truth
check the txt file

[Hiritsuki]

I cant download txt to view, it's means "Attachments Pending Approval".

[romeyo007]

here:
General
ID : 0
Complete name : D:\Anime\RAWS\SnQVol.1\QWASER OF STIGMATA VOL1\BDMV\STREAM\00002.m2ts
Format : BDAV
Format/Info : Blu-ray Video
File size : 4.76 GiB
Duration : 23mn 36s
Overall bit rate : 28.9 Mbps
Maximum Overall bit rate : 48.0 Mbps

Video
ID : 4113 (0x1011)
Menu ID : 1 (0x1)
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L4.1
Format settings, CABAC : Yes
Format settings, ReFrames : 2 frames
Codec ID : 27
Duration : 23mn 36s
Bit rate mode : Variable
Bit rate : 25.4 Mbps
Maximum bit rate : 38.0 Mbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : MBAFF
Bits/(Pixel*Frame) : 0.409
Stream size : 4.19 GiB (88%)
Color primaries : BT.709-5, BT.1361, IEC 61966-2-4, SMPTE RP177
Transfer characteristics : BT.709-5, BT.1361
Matrix coefficients : BT.709-5, BT.1361, IEC 61966-2-4 709, SMPTE RP177

[Hiritsuki]

Wow!, that so small of normal BDMV size.
It's not AVCHD of DVD9 size?

[romeyo007]

Wow!, that so small of normal BDMV size.
It's not AVCHD of DVD9 size?

this is one episode from the BDMV

[Blue_MiSfit]

Interlaced BluRay is perfectly normal. I've seen tons of them, though they're not nearly as common as 1080p24.

Don't attach things to the forum, upload a piece (cut using MKVMerge or something) to mediafire or another free hosting. No rapidshare please, unless you have a paid account.

Derek

[x110x]

hi, try ripbot x264

chose your m2ts file

click to properties
http://pepperpics.biz/i/r2.png

search with right arrow "deinterleace"
http://pepperpics.biz/i/r1.png

?

[romeyo007]

Interlaced BluRay is perfectly normal. I've seen tons of them, though they're not nearly as common as 1080p24.

Don't attach things to the forum, upload a piece (cut using MKVMerge or something) to mediafire or another free hosting. No rapidshare please, unless you have a paid account.

Derek

here a sample (http://hotfile.com/dl/88415286/60d9dff/00002a.rar.html)

[jpsdr]

I've made for personnal use of case like it (anime) an IVTC filter under VDub, with purpose of not deinterlacing after. I must check on your sample if it works, as i've for now only used it on DVD (even if i've made update to handle blu-ray).
Otherwise, are you sure it's still the "pain" anime format used to see on DVD ? Telecine sequence change on almost each scene ?
Otherwise, it telecine sequence is stable almost all along, with a simple chain of PullDown() you can get it.
Can't download your sample for nor...

[romeyo007]

this what I'm trying to fix
http://img262.imageshack.us/img262/1668/96424682.png

[mp3dom]

The problem comes probably because the source is an upscale (most probably from an half-hd source like 960x540). The source suffers from huge aliasing and this confuses TFM to detect combed frames even if the frames are clean. In the screenshot that you've posted, the book's page suffer from aliasing and TFM think is combed and so it deinterlace it with the tdeint+eedi combo. If you disable the deinterlacing you'll get only 'clean' frames (the aliasing still remains, so probably you should use an antialias script).

If you try a plain
xxxsource("xxx.dga")
tfm(PP=1,display=true,mode=1,slow=2,micmatching=3,order=1)

you'll see that in your screenshot TFM flag the frame as combed on the book's page.
PP=1 tells TFM to detect combed frames but not deinterlace it. display=true shows the metrics used by TFM. You can remove the display=true if you want.

[romeyo007]

thanks I'll try that
but would fix everything or it fix some part and then another problem to fix?

[mp3dom]

On your sample it fix everything (in the sense that it will no longer deinterlace clean frames wrongly detected as combed). The aliasing still remains as it's in the source. Obviously I don't know if on the entire show it will fix everything. If the source is encoded at 60i probably there are some parts at 30p or some cross-dissolvences/edits at a real 60i (with no pulldown) so you should watch and treat those parts in a different way.

[jpsdr]

My VDub filter seems to work fine on this sample, are you interested by trying it ?
You can get them here (http://rapidshare.com/files/436131928/Filtres_JPSDR.rar).
Filter is made to work with YV12 video, so you can input your YV12 29.97 interlaced, and output YV12 23.976fps progressive, without any deinterlace filter/process needed !
I've tested it with the following avs input file :
DGSource("00002a.dgi",i420=false,deinterlace=0,fieldop=0,use_D3D=false)
(DGI file created with DGIndexNV).
Note : Because of the cut, the 3 first frame are interlaced, so when applied, the output of the 1rst frame is interlaced, because pattern of the begining is not standard telecine. Output of all the others frames is good.
You have to put in VDub the 2 following filters (in this order) :
IVTC v4.4.2
Remove Frames v1.0.0
Leave all the parameters by default.
You can first test on sample, and on one episode after if test on sample satisfy you. If you find it interesting, i'll explain the "Manual Mode".

[romeyo007]

My VDub filter seems to work fine on this sample, are you interested by trying it ?
You can get them here (http://rapidshare.com/files/436131928/Filtres_JPSDR.rar).
Filter is made to work with YV12 video, so you can input your YV12 29.97 interlaced, and output YV12 23.976fps progressive, without any deinterlace filter/process needed !
I've tested it with the following avs input file :
DGSource("00002a.dgi",i420=false,deinterlace=0,fieldop=0,use_D3D=false)
(DGI file created with DGIndexNV).
Note : Because of the cut, the 3 first frame are interlaced, so when applied, the output of the 1rst frame is interlaced, because pattern of the begining is not standard telecine. Output of all the others frames is good.
You have to put in VDub the 2 following filters (in this order) :
IVTC v4.4.2
Remove Frames v1.0.0
Leave all the parameters by default.
You can first test on sample, and on one episode after if test on sample satisfy you. If you find it interesting, i'll explain the "Manual Mode".

I don't have the DGIndexNV where can i get it?

[jpsdr]

You can get it here (http://neuron2.net/dgdecnv/dgdecnv.html), but it's not free (even if price is very small), it's by the same authot of DGAVCDecode, but you need an nVidia card wich support some minimal features (if you have an nVidia card, but realy too old, it may not work).
The only problem with DGAVCDecode, it's that it begins to be a little old and not supported/updated anymore, but otherwise, i don't see why you couldn't use it. Nothing prevent you to test my filters with DGAVCDecode.

[romeyo007]

You can get it here (http://neuron2.net/dgdecnv/dgdecnv.html), but it's not free (even if price is very small), it's by the same authot of DGAVCDecode, but you need an nVidia card wich support some minimal features (if you have an nVidia card, but realy too old, it may not work).
The only problem with DGAVCDecode, it's that it begins to be a little old and not supported/updated anymore, but otherwise, i don't see why you couldn't use it. Nothing prevent you to test my filters with DGAVCDecode.

My PC Spec:
Core i5 750@ 2.67Ghz
4 GB DDR3
Win7 Ultimate 32bit
Geforce GT 220
23" HDMI monitor

[mp3dom]

The major problem of DGAVCDec is that it's single-threaded so your decoding speed will be slow (compared by your cpu possibilities). If you can't/don't want to spend for DGIndexNV you can use FFMS2 which is multi-threaded and have frame accurate indexing (as DGAVCDec) for m2ts and mkv containers.

[romeyo007]

after using this function
tfm(PP=1,display=true,mode=1,slow=2,micmatching=3,order=1)

General
Complete name : D:\Anime\RAWS\SnQVol.1\QWASER OF STIGMATA VOL1\BDMV\STREAM\00002[test1][1080].mkv
Format : Matroska
File size : 909 MiB
Duration : 23mn 36s
Overall bit rate : 5 380 Kbps
Writing application : x264 r1649 c54c47d
Writing library : Haali Matroska Writer b0

Video
ID : 1
Format : AVC
Format/Info : Advanced Video Codec
Format profile : High@L5.0
Format settings, CABAC : Yes
Format settings, ReFrames : 8 frames
Codec ID : V_MPEG4/ISO/AVC
Duration : 23mn 36s
Bit rate : 5 273 Kbps
Width : 1 920 pixels
Height : 1 080 pixels
Display aspect ratio : 16:9
Frame rate : 29.970 fps
Color space : YUV
Chroma subsampling : 4:2:0
Bit depth : 8 bits
Scan type : Progressive
Bits/(Pixel*Frame) : 0.085
Stream size : 891 MiB (98%)
Writing library : x264 core 98 r1649 c54c47d
Encoding settings : cabac=1 / ref=8 / deblock=1:0:0 / analyse=0x3:0x113 / me=umh / subme=9 / psy=1 / psy_rd=1.00:0.00 / mixed_ref=1 / me_range=24 / chroma_me=1 / trellis=1 / 8x8dct=1 / cqm=0 / deadzone=21,11 / fast_pskip=0 / chroma_qp_offset=-2 / threads=6 / sliced_threads=0 / nr=0 / decimate=1 / interlaced=0 / constrained_intra=0 / bframes=3 / b_pyramid=2 / b_adapt=1 / b_bias=0 / direct=1 / weightb=1 / weightp=2 / keyint=250 / keyint_min=25 / scenecut=40 / intra_refresh=0 / rc_lookahead=40 / rc=crf / mbtree=1 / crf=18.0 / qcomp=0.60 / qpmin=10 / qpmax=31 / qpstep=4 / ip_ratio=1.40 / aq=1:1.00
Language : English



it kinda fixed the problem but now there is weird squares are blurry in the start and in dark areas

[Blue_MiSfit]

Damn. That's a seriously screwed up source you've got there, mate :(

I don't know how to fix it. There's nasty aliasing in the source fields. Yuck.

Derek

[jpsdr]

Core i5 750@ 2.67Ghz
Geforce GT 220

It's Ok.
So, have you try my VDub filters, even with DGAVCDec ?

[mp3dom]

The image is under pending approval so we cannot see it. Can you upload it elsewhere? Thanks.

Edit: Have you seen if the source contains 30p or true 60i zones? Because if it can be full ivtc'ed you should add TDecimate (or you can use hybrid parameter to convert 30p to 24p with blend)
Also, try to use latest version of x264

[romeyo007]

The image is under pending approval so we cannot see it. Can you upload it elsewhere? Thanks.

Edit: Have you seen if the source contains 30p or true 60i zones? Because if it can be full ivtc'ed you should add TDecimate (or you can use hybrid parameter to convert 30p to 24p with blend)
Also, try to use latest version of x264

http://img8.imageshack.us/img8/4576/00002test11080mkvsnapsh.png

where can i know if the source is 30 or 60?

i know the frame rate is 29.970 fps

thanks

[Blue_MiSfit]

I think hybrid or not is the least of this footage's problems. The fields are horribly aliased.

[romeyo007]

I think hybrid or not is the least of this footage's problems. The fields are horribly aliased.

what is the best setting and res for it you think?

[mp3dom]

If you accept to loose a bit of sharpness you can think of keep it at 1080p (the source is anyway an upscale too):
http://img709.imageshack.us/img709/4758/1301j.png
Not the best result but at least it's more "watchable".

I think the source was made at 960x540. Some recent anime was made at that resolution and then upscaled to 1080p

[romeyo007]

If you accept to loose a bit of sharpness you can think of keep it at 1080p (the source is anyway an upscale too):
http://img709.imageshack.us/img709/4758/1301j.png
Not the best result but at least it's more "watchable".

I think the source was made at 960x540. Some recent anime was made at that resolution and then upscaled to 1080p

what setting did u use for that image?
look much better than mine

[VFR maniac]

I think the source was made at 960x540. Some recent anime was made at that resolution and then upscaled to 1080p
No, Qwaser was made at 1280x720 exactly since Cross-Conversion Correction (CCC) (http://www.geocities.jp/flash3kyuu/), one of AviUtl plugins, restores almost all its original resolution.
[before]
http://vfrmaniac.fushizen.eu/OtherStuff/QUASER_OP1.PNG
[after]
http://vfrmaniac.fushizen.eu/OtherStuff/QUASER_OP1_ccc.PNG

This plugin is the best option for interlaced Cross-Conversion (720i->1080i).
It's a pity that this plugin cannot be used in Avisynth. (Memory Access Violation)
If you hope to Avisynth to restore progressive resolution, here is one option. (This functions' is inferior to CCC.)
BCSInterlacedResize_mod 1.3 (http://www.avisynth.info/?%A5%A2%A1%BC%A5%AB%A5%A4%A5%D6)

[Blue_MiSfit]

Wow. That's seriously impressive!

[Fadeout]

Does it help if I can see the encoding options in a file that was encoded solving that problem?

I know it can be done, just don't know how.

[romeyo007]

Does it help if I can see the encoding options in a file that was encoded solving that problem?

I know it can be done, just don't know how.

i can post the avs file if that what u asking?

[Fadeout]

This is how it should look. No idea how it was made, though.

http://img819.imageshack.us/img819/2415/qws.png

[romeyo007]

This is how it should look. No idea how it was made, though.

http://img819.imageshack.us/img819/2415/qws.png

where did u got that?

[romeyo007]

If same like to see how the main source looks I upload it the m2ts for the first episode.(I hope that okay to do here)
(warning some mature scene)

http://hotfile.com/dl/88856612/34655ae/00002.part01.rar.html
http://hotfile.com/dl/88859663/316a728/00002.part02.rar.html
http://hotfile.com/dl/88887123/6333869/00002.part03.rar.html
http://hotfile.com/dl/88862122/d4b7010/00002.part04.rar.html
http://hotfile.com/dl/88862123/42554a2/00002.part05.rar.html
http://hotfile.com/dl/88865810/556cb50/00002.part06.rar.html
http://hotfile.com/dl/88866325/da4d74e/00002.part07.rar.html
http://hotfile.com/dl/88873520/7a616cb/00002.part08.rar.html
http://hotfile.com/dl/88880318/b247df5/00002.part09.rar.html
http://hotfile.com/dl/88880816/c4f817b/00002.part10.rar.html

if someone can encode the episode and tell me what the best setting I should use for it. since all the other group failed with this project.

thanks

[jpsdr]

Have-you try my VDub pluggin ?

[romeyo007]

Have-you try my VDub pluggin ?

I wasn't able to do or I didn't know how to do it?

if u can explain the steps in more details then i can try it

thanks

[jpsdr]

As i said :
In the first time, you just have to use the 2 following filters (in this order) :
IVTC v4.4.2
Remove Frames v1.0.0
Leave all the parameters by default.

[romeyo007]

As i said :
In the first time, you just have to use the 2 following filters (in this order) :
IVTC v4.4.2
Remove Frames v1.0.0
Leave all the parameters by default.

what the function code for them? and is it okay to use it on dga or it have to be .dgi?

[Daemon404]

No, Qwaser was made at 1280x720 exactly since Cross-Conversion Correction (CCC) (http://www.geocities.jp/flash3kyuu/), one of AviUtl plugins, restores almost all its original resolution.
[before]
http://vfrmaniac.fushizen.eu/OtherStuff/QUASER_OP1.PNG
[after]
http://vfrmaniac.fushizen.eu/OtherStuff/QUASER_OP1_ccc.PNG

This plugin is the best option for interlaced Cross-Conversion (720i->1080i).
It's a pity that this plugin cannot be used in Avisynth. (Memory Access Violation)

I've put together a clean room AviSynth port here:

http://japland.org/ccc/ccc_v0.4_avs.zip

Just make sure call it before you crop/resize.

Enjoy!

[jpsdr]

I have no idea what you call function code...
You just have to select them in VDub.
A few minutes ago i would have said it's Ok to use on dga, but i've just downloaded the 1rst episode you've put, open it in VDub with DirectShowSource and... result was "awfull" !!! Nothing to do with the sample you've provided, but opened with DGIndexNV. So, now i'm transfering the 1rst episode on my PC dedicated for video (where DGIndexNV is installed), and see if there is a difference between DirectShowSource and DGIndex. Nevertheless, i strongly suggest you to try with .dga for now.

Edit : Same result with DirectShowSource and DGIndex, so, no problem to use on dga.
Problem i thought i see is what said : upscale was made in interlace mode : Each field have been upscaled independantly, and put back together. Shouldn't have impact on my filters, so, no problem to try them.

[romeyo007]

I will try the two advice as soon I'm done from my exams.
is there freeware for the DGIndexNV?
thanks

[jpsdr]

There's no freeware for DGIndexNV, so, if you don't want to buy it, you can continue with DGAVCDecode.

I've just take a quick look to the result with my IVTC filter, result is very good, but, as nothing is perfect, and i know both limits of my filters and where there is always problem, i'll have to explain to you :
- How to use "manual mode" and detect where there is problem.
- How to use my Deinterlace filter combined with "file mode".
But, it will be a little long, so, tells when you've done with your exams.
But basiquely, you can have very good result, but my filters have to be used in a 2 pass process.
After the 1rst pass, you have to watch the result to find out the small parts where filter fails, to manualy correct them. It's a little long, but you avoid the deinterlace of the whole video process.
As expected, you can first see problems on fades on the opening, so, some frames (but only a few), need to be deinterlaced, this is where my Deinterlace filter combined with "file mode" is used.
Not seen yet parts where IVTC failed, but i've made a VERY quick look.

When you'll check, or if anyone other want to try. After doing the whole process (in YV12), open the result file with a simple script like :

AVISource("File.avi")
PointResize(720,480)

This will make the part still interlaced more visible.
On the result file of my filter, problem are visible on 74 first frame fading, and on frames 2617 to 2620 (for exemple).
There is problem on fading almost because :
- Fading are made after pulldown and field based, and unfortunately, not frame based before pulldown. So, even if your IVTC is correct, result is still interlaced because fading is between fields.
- Fading between 2 scenes, often the 2 scenes have different telecine patterns, in that case IVTC is impossible. Even if pattern is the same, because fading is field based, result is still interlaced, even if your IVTC pattern match the telecine pattern.

@Daemon404
Thanks for the ccc filter, i've take a look at the code, strange behavior to work only on Y...
Nevertheless, code is very simple, and i'll look later to create a VDub filter and add it to my filters.
Maybe some SSE optimisation can be made, but i'll see this later.