meguIV: The Official Akiba-Online DVD Encoder (v1.0.1.1)

[Rollyco]

Thanks for trying Neo63 but I was looking for a human translation. According to guy and isityours it's a machine translation which is bound to be confusing to the target audience.

 

[isityours]

Thanks for trying Neo63 but I was looking for a human translation.

there were posts to that effect on the first page of this thread but they came to nothing (that ive seen anyway). i just assumed it had been conceeded that most of the members that would benefit from a japanese version of the instructions were anonymous leechers that dont rip or post anyway, therefore defeating the purpose. that and the abysmal lack of foreign language support offered by this site. that said, if a translation (probably closer to an explanation about what meguIV does and how to use it, rather than how it works) is still considered desirable, i am happy to do what i can.

(guy might even pitch in too, if asked nicely...)

 

[Rollyco]

If you and/or guy want to attempt a translation, I'd be very interested. The idea is to also have someone write a short "meguIV visual guide - now in Japanese!" news post linking to the new translation for our Japanese visitors. I think that would be an effective way of spreading the word.

 

[Authorised-BMN]

60 v 30 fps

A question for those who know a lot more about encoding than I do: a lot of folks are doing 30 fps and 60 fps encodings of the same video. Is there an easy, reliable way of taking a meguIV 60 fps rip and dropping half the frames to create the 30 fps version? Based on Google, it seems there are two choices of avisynth commands and some potential audio sync issues. Is this something someone has already thought about and solved? Could it be incorporated into Vitreous's "rerip" protocol? This might be an opportunity to save some of our encoders a lot of CPU time, but it would take me ages to figure it out -- if it's even possible. Any thoughts?
erfectplan:lease:

 

[isityours]

i also wondered how make a 30 and 60 from the same hfyu file in meguIV(it) (if possible). i assume it can be done, im just not sure how. some help would be appreciated.

 

[Vitreous]

Is there an easy, reliable way of taking a meguIV 60 fps rip and dropping half the frames to create the 30 fps version?

Edit:I slightly missed the original question. This is how to efficiently make both a 60fps and 30fps from the same ISO

This can be done with a bit of manual tweaking in MeguIV(it):

- Run MeguIV(it) and go to "Options>>Settings"
- In the "Main" tab ensure "Delete Intermediate Files" is unchecked

- Make a 60fps rip as normal ("Add pre-rendering job" in the One-Click advanced Encoder Config must be checked - it will be unless you've been tinkering). Don't close MeguIV(it) after the rip finishes
- Rename your 60fps mkv at this point to distinguish it

- Go to the output video folder and in Notepad or similar, edit the file with a name like "hfyu_VTS XX X.avs"
- It will contain a single AviSource line, leave that untouched. Add lines so the result looks like this:

global MeGUI_darx = 16
global MeGUI_dary = 9
AviSource("Yourpath\hfyu_VTS XX X.avi")
SelectEven()

- The first two lines should be the correct aspect ratio, 16 & 9 or 4 & 3. The last line is the one that will drop half the frames for 30fps
- Save this file (over the original)

- Now in the main MeguIV(it) window we will use the top half of the main "Input" tab
- Drag the "hfyu_VTS XX X.avs" you just edited into the "AviSynth Script" box near the top
- Select a [Vit] 30fps choice from the "Encoder settings" below
- Select "RAWAVC" in the "File format" below
- Uncheck "Add pre-rendering job"
- Press "Enqueue" (the one in the top half of the window) - the 30fps raw video will be encoded

- We need to mux this into a 30fps mkv manually - go to "Tools>>Muxer>>MKV Muxer"
- Drag the file "hfyu_VTS XX X.264" into the muxer "Video Input" (ensure you choose the correct .264 file, the one that begins "hfyu_")
- Drag the file "VTS XXXX audio.m4a" into the muxer "Audio 1>>Input". The file *might* be a .wav in some cases
- Drag the file "VTS_XXXX Chapter Information - OGG.txt" into the muxer "Chapters File"
- Edit the name for your 30fps mkv in "Muxed Output"
- Press "Queue". The 30fps mkv will be created

- You'll need to clean up the intermediate files yourself. The huge "hfyu_xxx.avi"is the main one you want to delete to recover space.

 

[Authorised-BMN]

... from existing 60fps?

Make a 60fps rip as normal

Thanks so much, Vitreous! This should save time for those who post both 60fps and 30fps (as I have done a few times). What about an existing 60fps encoded by someone else? For example, IMOJ-011 was encoded by Kaleb2727: here. If I want a 30fps version (without loss of quality), do I have to encode it from the DVDISO (which I am in the process of doing)?
uzzled:

 

[Vitreous]

If I want a 30fps version (without loss of quality), do I have to encode it from the DVDISO (which I am in the process of doing)?

It is possible to create a 30fps mkv from a 60fps mkv. But it isn't just a simple muxing thing, a reencode would be needed, which will lose quality. If you really want to know the method I can advise, but you will get better quality encoding from the ISO.

Edit: Actually, the process for 60fps mkv to 30fps mkv in MeguIV is more complex than I'd expected because MeguIV doesn't include the appropriate plugins for decoding MKVs. You'd need to start adding new tools to the Sandbox - I don't have the time to experiment with the exact requirements. This kind of thing is best done in MeGUI proper with a full encoding configuration (avisynth + plugins + ffdshow + Haali + etc.).

 

[astrayred]

Interestingly, akiva has been posting a few 60fps rips recently. I'm not sure whether he did those rips or not, but I'm guessing two things here:

1) Akiva is Japanese
2) The Japanese ripping community is taking interest in 60fps encodes

Unfortunately, it seems like those 60fps encodes aren't done using TGMC/QTGMC, but with a lesser quality deinterlacer. Perhaps Yadif as that is very common. If a Japanese version of MeGUIV is done I think we can really expect to see some nice encodes come straight out of the nice people who have been rippping videos for us

 

[isityours]

Rollyco, attached is a slightly simplified version of the original english instructions that i propose to use as a basis for a translation. could you take a look please and tell me if there is anything i shouldnt omit etc? i have also included a quick mention of meguIVit. the diagrams have also been left out for now, for simplicities sake.

 

[Rollyco]

Thanks isityours. I think that the full text (minus the changelog) would be a suitable object for translation, so I edited posts #2 and #3 to clarify that.

 

[guy]

[meguIV guide Japanese translation (12/23/2010 update)]




【meguIV】 めぐ・アイ・ヴィ （名詞）

meguIVとは: 専らアイドルイメージビデオ（IV）とアダルトビデオ（AV）だけの為に作られた、最新特級デインタレース・アルゴリズムや特製のプリセットを利用特定なワン・クリック・エンコーダー。 ベースとして人気エンコーダーMeGUIから作られたポータブルアプリ。 やっと好きなAV女優のケツのニキビまで見れるようになった！

――――――――――――――――――――――

meguIVを使うと特級ビデオクオリティが証されるから、Akiba-Onlineでスレタイトルで[HQ]の接頭を使えることとなる。 今までmeguIVでエンコードされたビデオはこちら。

ポイント

• ポータブルアプリだからインストールや設定不要
• いつもシャープ、キレイな特級ビデオクオリティ
• MP4とMKVファイル形式オプション
• MP3とAAC音声フォーマットオプション
• DVDチャプターを自動に作られる
• 縦横比を自動に選定される
• スクイーズビデオ出力　（720でエンコード）
• マルチコアCPU利用の場合はマルチスレッド・デインタレース（TempGaussMC）可

履歴（英語のみ）
[hide]v1.0.1.1:

• Fix: messed up package.

v1.0.1.0:

• Fix: x264 (video encoder) not starting on 64-bit systems.
• Fix: The first upload of 1.0.0.0 was the wrong build. To be safe, please upgrade.

v1.0.0.0 (Fixed version uploaded April 20th, 2010):

• Upgraded MeGUI, TempGaussMC, and x264.
• Video output is sharper and cleaner now (see attached screenshots).
• 30-50% speedup! Medium-quality presets are not needed anymore.
• About 25% smaller output video filesizes.
• New pre-rendering step requires 15-25GB free space for temp files.

[/hide]

必要動作環境

• Windows XP/2003以上（32bit/64bit）
• SSE3付マルチコアCPU
• 15-25GB空き容量（エンコード中に一時ファイルのため）
• Microsoft .NET Framework 2.0以上

――――――――――――――――――――――

使い方
1: 「One-Click」ボタンをクリック。

2: 「Input file」列で入力VOBファイルを選択。

• DVDのVOBファイルの中から一番目のVOBファイルを選びなさい。 （例: VTS_○○_1.VOB）　その後のVOBファイルは同じフォルダの中から順番で自動に見付けられる。
• ISOファイルの場合中のVOBファイルをWinRARや7-ZipなどでHDDに解凍必要。 解凍しないとエラーが表されて、DVDチャプターは自動に作られなくなる。
• 出力フォルダを好きなフォルダに選択可。

3: 「Advanced Options」にチェックを付く。

4: 「Advanced Config」タブに移動。

5: 「Working Directory」列でエンコード中の一時ファイルを保存するフォルダを選択。

• エンコード完了に一時ファイルは自動に削除される。
• 15-25GB空き容量必要

6: 「Project Name」にエンコードの名前を付ける。

• この名前は出力ファイル名となる。

7: 「Encoder Config」タブに移動。

8: 音声フォーマットを選択（MP3、AAC）。

• 不安の場合にはAACを選びなさい。 AACの音声クオリティの方が高い。

9: 出力ファイル形式を選択（MKV、MP4）。

• 不安の場合にはMP4を選びなさい。 MP4に対するPS3などの互換機の方が多い。
• 「Device Type」列を「Standard」のままで。

10: 「Go!」をクリック。 エンコードはジョブキューに追加されて、自動開始。

――――――――――――――――――――――

備考

• 別のエンコーダーアプリ（元MeGUI, AviSynthなど）が同時にインストールされている可能。 meguIVはポータブルアプリとしてシステムファイルや設定、又は別のアプリに障らない。
• meguIVが動かない状態になる場合、又は設定などを元に戻したい場合には、meguIVの「Sandbox」フォルダを削除するとmeguIVは元の状態に戻る。
• meguIVの「MeGUI Update」アップデート機能不利用。
• 特級クオリティのために、meguIVは大変遅い。 60分再生時間のイメージビデオを3GhzのC2D（2コア）でエンコードするのは普通に7時間かかる。 3GhzのC2Q（4コア）を利用するとそのエンコード時間は半分になる。 最新Core i7を利用している方は超ラッキー！

 

[isityours]

so it seems guy went ahead and did a translation (why im not sure as i already said i would do one) making mine redundant to a certain extent. i did however include hotlinks to the japanese versions of relevant pages and a part explaining use of meguIVit as i feel that it has more than proven itself to be not only an improvement on, but, more accurately, surpasses the original meguIV 'Sandbox'. i hope you will consider including it (even if only in the japanese translation). otherwise there are no major differences between the translations. i can only suggest having a japanese member look at them to determine which parts are most readily understandable to a native speaker.

 

[guy]

Somehow I missed where you said you would do a translation.
:bow-pray:

 

[isityours]

.....if a translation (probably closer to an explanation about what meguIV does and how to use it, rather than how it works) is still considered desirable, i am happy to do what i can.

(guy might even pitch in too, if asked nicely...)

sorry, i didnt mean it like that. and i wasnt just assuming that you read my posts but i thought you would have seen it as it was related directly to the previous post you commented on too. its no biggie, it just made sense that collaborating would have saved us both some time.

 

[astrayred]

Hi Vitreous! I read on your recent Yuri Takase topic that you have a new version of QTGMC coming out. A few questions, if I may...

1) Any estimate of the release date? I'm really looking forward to giving it a go as you can probably guess.

2) How does the quality improvements affect the speed? It sounds like there's a lot more additional processing going on!

3) Will the improvements affect progressive mode as well?

Thanks for the great work, as always.

 

[Vitreous]

Hi Vitreous! I read on your recent Yuri Takase topic that you have a new version of QTGMC coming out. A few questions, if I may...

1) Any estimate of the release date? I'm really looking forward to giving it a go as you can probably guess.

2) How does the quality improvements affect the speed? It sounds like there's a lot more additional processing going on!

3) Will the improvements affect progressive mode as well?

Thanks for the great work, as always.

1) No idea really, depends how much time I get to work on it. Need time to create a more usable interface. I've already added the new settings to QTGMC, but they are very delicate - they require a great deal of TGMC algorithm knowledge to use effectively. I need to add a new kind of preset to simplify their use. Also will take time to do point 2 below. Two weeks maybe?

2) Hard to say at this point. The rip I posted was *very* slow to make, effectively equivalent of a "Medium" pass followed by a "Placebo" pass. But that was me experimenting, I believe there is considerable scope to make the new process more efficient and get it all into one pass. Standard QTGMC processing speed is unaffected.

3) Progressive isn't supported with the new modes at the moment. I have pondered it though. For interlaced, the improvements come from doing a lossless deinterlace in the first pass: result is progressive, full of detail, but shimmering. Then use a second progressive pass to stabilize that output. That workflow gives a marked improvement over standard QTGMC processing, but the first pass doesn't make sense for progressive. Progressive might be able to gain with more source-accurate output, but that won't seem like much.

Notes:
There are a couple of other additions in the new QTGMC. There's a new setting for progressive input when using InputType=2 or 3, which improves quality markedly in stable areas. There's also a minor request I had regarding chroma repairs.

I've also been looking at the AviSynth source code to find the multi-threading problems. The problem is deep-seated: the method used to interface between AviSynth and the plugins is inappropriate. It needs a different architecture, but one that would be incompatible with all existing plugins (i.e. the interface code for all plugins would need to be updated). I guess that's why no-one seems to want to fix it properly.

I've come up with a simple, but hideous hack to prevent the most frequent crashes. Don't ask about it, it's very wrong. I've recompiled Avisynth and all the plugins used by QTGMC to use this new code. I can't get it to do random crashes now (ignoring out of memory issues with complex scripts). I also removed all plugin dependencies (so they don't require any redistributables) and compiled with Intel C++ 2011 with maximum optimizations for some minor speed-up. These new dlls will be in the next MeguIVit and I hope that will make it more usable/stable for everyone.

 

[astrayred]

Thanks for the reply!

Regarding the progressive mode, I asked because recently I have been noticing an increasing amount of DVDISOs that's either fake interlaced (DGindex says interlaced but it's actually progressive), or a mix of interlaced and progressive (Main movie is progressive, interview is interlaced >.>) It's a pain to work with. I recently stumbled upon this nice piece of software called Splash Pro. It's a video player that has a GPU motion interpolation engine. Works pretty good for the low amount of resources that it consumes, so I have been using progressive mode for those pesky progressive DVDISOs instead of interpolating using MVTools.

About the Avisynth problems, perhaps you could post about it on D9? Perhaps it would get development moving again. I personally get random crashes very rarely though.

 

[no__One]

Mr.Astrayred's commentary motivate me to do a little back on my experience with the last meguIVit_0.2.1.

Like Mr.Astrayred, I also face random crachs. This happened with this title:
http://www.akiba-online.com/forum/showpost.php?p=1070261&postcount=2452

The only solution I have for this situation it seems to reactivate the pre-rendering. (Currently disabled by default, and aside from the aforementioned DVD, I haven't no more or less crash than before).

By cons, my main problem is the encoding of sources other than DVD (DivX, XviD,...): if the pre-rendering not activated = Crach and even with it... it's so-souzzled:
I know this is not its primary purpose. But for a novice, profiles + meguIVit are disconcerting ease of use in terms of quality and obtained without requiring expertise in fluid engineering studies and ellipsoidal curves ....

I hope my little report will be of any use.

Regards.
:study:

 

[Vitreous]

Disabling pre-rendering on Placebo settings runs the risk of out-of-memory issues and crashes. Really there's no good reason for memory issues on SD material, but unfortunately another problem with AviSynth is that it's caching algorithm is not smart enough. Caches too much, runs out of memory, crashes. Very much affects multi-threaded ripping.

Best solution would be to rewrite AviSynth's caching (and threading) code. No-one wants that job.
Second best solution is to make a 64-bit MeguIVit. It's on my to do list.

Edit1 @Astrayred: I'll up the threading hack-fix code & dlls on D9, but I don't think it will kickstart any development work. The authors already know that Avisynth needs to be re-architected, and have had some discussions about it. They failed to agree and didn't appear eager to do the work in any case.

Edit2 @Astrayred:

I have been using progressive mode for those pesky progressive DVDISOs instead of interpolating using MVTools.

Using progressive mode on interlaced material will give you near duplicate frames - something much like 30fps. Is that your intent?