In the year 2013, after YEARS of waiting, I have finally decided to localize AviUtl(exedit) so that I can prepare video for NicoNico and Youtube easier. Even though I can read Japanese, those technical terms are simply confusing and difficult for the eye (actually still confusing for some terms, but looking them up from Wikipedia became a much easier process.)

The first early release dated back to August. By now, major bug have been fixed and a significant amount of plugins have been localized.

While I have done the heavy work unpaid, the community will NEED to involve for a number of reason:

1> If I go broke or died, no one will follow up…

2> Computer works consume electricity, backup need HDD and I need to feed (while unemployed)

3> I can’t tell what kind of tutorial, in what format, and in what level, best suite the major public. Honestly, I don’t feel like writing tutorials for primary school kids or the mentally retarded. In addition, making real-time video tutorial is really troublesome. I would expect someone who is going to use an NLE to do some research and exploring on their own. Therefore, tutorial writers, particularly producers for tutorial videos, is urgently needed.

For one thing, the translated README for the core AviUtl seems never be read by anyone…. [This is an annoying phenomenon nowadays that people ignore README file and refuse to read manual.... and they call for help when they have problem... ]

4> Need promotion. Many people asking questions on AV-related forums set the question specifically on certain program, like– “How to perform X in program Y”. While Y is rarely “AviUtl” but mostly “Virtualdub(mod)”, “Avisynth”, “Vegas” and “AfterEffect”. Hard selling AviUtl on such thread is somewhat off-topic…. so, if you have some experience with AviUtl and find the EN-version useful, please post tutorial/announcement threads on forums, and recommended it to schools, studios and amateur producers.

For one thing, there is NO RECENT THREAD concerning AviUtl on Doom9, Doom10, or the AnimeSuki forum. (I am not going to troll around all these forums.)

5> Need feedback on translation improvement and help on code hacking.

So far, no one tell me if some technical terms was translated wrongly… but probably somewhere is wrong…. Also, I can’t locate where the heck in this program determine the dialog and its control’s size… and this force the use of obscure abbreviations….

Hackers will be needed to reverse engineer and better analysis the (assembly)code in order to get around this limitation.

6> Script and plugin development

SO, As a user of AviUtl(EN), please do not just grab this program and run away. Contribute, if you can.

[an exemplary early contributer, is "racer-x", on the videohelp forum. Thx!]

[DOWNLOAD]

7.96KB
MD5: E4A3E4E56EE25F1ADAAB298A878B5A0B

———————————————————–

The previously released “for GPU” version requires Win7 and a decent GPU.

Surprisingly, many people are still using WinXP, and of course many people do not have an independent graphics card.

So, here is a stripped down version…. using CPU and working on Luma noise only.

———————————————————–

NL-Means Light

■概要
nodchip氏のNL-Means filterをGPUの恩恵を受けられない環境向けに、一部の機能を削ることで高速化してみました。
A Cut-down version of NL-Means filter for systems that don’t have a powerful GPU.

・輝度成分のみを扱います Works on Luma component only.
・時間軸方向のフィルタリングは出来ません。No temporal filtering.
・参照範囲は4まで. Reference range is up to 4.

■オリジナルの配布元 Original code base.
AviUtl NL-Means filter by nodchip

http://kishibe.dyndns.tv/

■動作環境
Microsoft Visual C++ 2010 SP1 再頒布可能パッケージ (x86)が必要です。
Microsoft Visual C++ 2010 SP1 redistributable package (x86).

http://www.microsoft.com/downloads/details.aspx?familyid=C32F406A-F8FC-4164-B6EB-5328B8578F03&displaylang=ja

[DOWNLOAD]

50.7KB
MD5:3CF7F6472B1D9A1BD3796931B270F9B7

——————————————————————————————–

Wavelet Noise reduction and setting helper by GNB
==============================================

The package consist of two plugins.
waveletNR_G.auf is the actual plugin that do the heavy duty;
TypeG_Helper.auf is the one for automating settings.

The wavelet denoiser attempts to remove noise in the spatial component with minimum blur.

Installation
================
Copy the two auf files into .\Plugins folder.

Usage
========
1> Open up the setting dialogs—
[Settings> Type-G Helper]
[Settings> Wavelet_NR Type-G]

2> Load your video by drag-and-drop your file on to AviUtl’s main window

3> Using the slider at the bottom, navigate to somewhere the noise is obvious and overlaying on a uniform background.

4> Tick the checkbox on the top-right of Type-G Helper dialog

5> A square should appear.
Adjust the “Adj.X” and “Adj.Y” sliders to move the square to the uniform area with noise.

6> Click on the four buttons on bottom-right, from top to bottom in order
[Set YEdge Lv>SetY-NR> Set CEdge Lv> Set C-NR]

7> Uncheck the check box on top-right.

8> Tick the checkbox on top-right in the Wavelet_NR Type-G dialog box.
Check result from the main window.

9> If you are not happy with the result, try(UNCHECK the top-right checkbox in Wavelet_NR dialog first!)….
— Set “Mode” to “3″ in Type-G Helper, then start from step 4 again.
— Set YEdgeLv(Luma noise), CEdgeLv(Chroma noise) to a higher value, then start from step 4 again.
— Move to another frame and start from step 4 again

[DOWNLOAD]

5.87KB
MD5: F9255AFEBC8A6DC68F07CC64BF58EB5D

——————————————————————

A companion for NL-Means light-CPU

——————————————————————

AviUtl標準のノイズ除去相当のフィルタを色差成分のみに適用するフィルタです。
輝度のチャンネルも参照しているので輝度のチャンネルと共通の輪郭部分は潰れにくくなっています。

A noise removal plugin working only on chroma component.However, luma component is referenced during noise analysis.

動作環境
Requirement

SSE2が使えるCPU(チェックしていません)
CPU supporting SSE2

Microsoft Visual C++ 2008 SP1 再頒布可能パッケージ (x86) が必要です。
Microsoft Visual C++ 2008 SP1 redistributable package (x86).

http://www.microsoft.com/downloads/details.aspx?familyid=A5C84275-3B97-4AB7-A40D-3802B2AF5FC2&displaylang=ja

Since many people seems to be still using WinXP and get error when using the existing extra packs. I am going to release two versions of EP from now on, one for XP and one for Win7 or later, until there are insignificant need for working on XP.

DOWNLOAD ExtraPack 1.4 (Win7)

DOWNLOAD ExtraPack 1.4 (XP)

(in 7-zip format, please get 7zip from sourceforge to decompress the archives)

Main difference in these files:

1> Automatic Field Shift: Win7 version use AVX instruction for better performance

2> Resize filter. Disabled the AVX options in the XP version

3> QSVEnc. XP does not support QuickSync and therefore the whole plugin and related files are deleted.

New additions in 1.4:

— a huge expansion in repair filters[wavelet denoise, NL-Means denoise (with non-GPU stripped down version), Chroma denoise], delogo and subtitle removal

— expanded D&D filetype support for Adv.Editing

— included flac(lossless audio) option in the x264Gui/x265Gui/VCEEnc encoding plugin

—included mplex, for muxing to mpeg2/DVD format (NOT TESTED YET!)

— changed default bitdepth for L-Smash importer from 8 to 16

— rearrangement in video filter order

Since my video card died on boxing day and therefore cannot fully check hardware compatibility(now running on Intel HD3000… no access to recent Nvidia/AMD hardware) and encoding slow, please help to check if there are any problems in these new versions.

PLEASE report if they work or not (report even if it work), and your machine spec (a brief spec will be enough, mainly for the OS, CPU/APU and the video chipset).

DO NOT INFORM ANY SOFTWARE HOSTING SITES YET, UNTIL 10th Jan 2014.

[Download for Win7] 33.6MB   MD5:7683C6DB97A31EEF77BA98832740171E

[Download for XP] 30.5MB   MD5:F310CCD28B17082EE9B34033BE7EF35C

Their differences can be found HERE.

* If the program ask for a LDP file on launch, select the “logodata.ldp” inside the “.\Plugins” folder.
New SFX highlights:

Year 2014 coincide with the release of Extra Pack 1.4…. Here are the details:

IMPORTANT CHANGE
====================

NeroAAC encoder binary has been removed from this distribution after discovering its license prohibit re-distribution.

Please choose “fdk-aac”(actually ffmpeg’s libvo_aacenc) as audio encoder, or download NeroAAC Encoder from official site yourself (http://www.nero.com/enu/company/about-nero/nero-aac-codec.php)

Additions
====================

>Denoiser: Wavelet(and setting helper), NL-Means Light (and for GPU), Ringing Canceller, Chroma denoiser

>Restoration: Delogo, ImasUncaption (hard-coded subtitle remover)

>LUA Scripts: AutoTarget100 (~SureTarget in AE), Skydome (~Horizon in AE), Multi-color gradient, Cylinder Mapping with cross-sectioning, Thunder, Rainbow, sepia noise and many others.
[For usage, please investigate the included Sample projects (.aup files in the "SamplePrj" folder)]

>Expanded file type support for Drag-and-Drop (Adv.Editing)

Minor Changes
===================

>In L-Smash import plugin, changed AVS bit-depth to 16

>Removed x265-10bit binaries. (If anyone can get AviUtl working correctly with x265-10bit, please inform me.)

>Updated FFmpeg

>Rearranged video filter order due to expansion of plugins. Please revise the order yourself if you find something not working as expected.

>All components, except x265Gui, should now be displaying GUI in English

Using AviUtl as a function plotter!

Functions/Parametric Plotter script(Custom Objects)
[DOWNLOAD]

There are a number of known methods to create a flame in AviUtl. Each have their characteristic and cause different CPU loads.

The key idea is like this:

1. - Add a simple circle
2. - deform the circle, or make it part of a “Particle output” object
3. - Apply gradient (red-orange for normal fire)
4. - Add Emission
5. - Apply non-symmetric radial blur
6. - Add dynamics to flame using “ripple” or “raster”

More sophisticated method involve the use of “Noise”, but that will also add to CPU load.

[Sample Project Download]

Video filters often cast their effects on the whole video area. However, there is something called “partial filter” (which I translated it incorrectly as “sub-filter”…) that can apply a filter effect on a defined area only.

In this tutorial, I am going to show you how to cast mosaic on certain “sensitive area” and let the mosaic area “follow” the sensitive area’s movement.

To get start, you will need to have a sample video ready.

So, let’s begin…

1> Load your video file on to the timeline and navigate to where you want the mosaic starts showing up.

2> Add a “Sub-filter” below the video’s track

3>Move the start pt. of the sub filer to a few frames in front of current frame (just to be safe)

4>Associate a “monochromatic” filter with the sub-filter object and set it to some easy-to-spot color. This step is simply making it easier to track the filter’s location.

5>Move the shape to cover the sensitive area

6>Click once on Adv.Edit’s title bar to gain its focus, then use the arrow keys on the keyboard to move a few frames ahead, where the sensitive area nearly gets outside the colored shape. Press “P” key once when you get the point right. An inverted triangle on the sub-filter object indicates a mid-point(or control point, key frame, whatever…) is added.

7> Left-click on “X” and choose “rectilinear movement”, so that we can animate the mosaic.

8>Adjust the LEFT side of the coordinate parameters(i.e. the starting value of this segment) to cover up the sensitive area completely. (Hint: Click and drag on the coordinate values may be easier than dragging on the main window for fine tuning.)

9>Repeat the “move a few frames forward, add mid-point, and adjust segment’s starting coordinate” operations until you have cover the duration when the mosaic is needed.

10> Then associate the “Mosaic filter” with the sub-filter object.(Hint: You may also use the blur effect to replace mosaic)

11> Uncheck “monochromatic” to view the actual mosaic effect. Tune the mosaic/sub-filter parameters if necessary (e.g. Size)

12>Move back a few frames, check that the mosaic is in the correct places over the required duration.

13> Reduce the duration of the sub-filter object to cover only the required duration.

14> Proceed with further editing and encode.

Say, you have a few layers of objects and composed something like this—

2x “Square” + particle output

And now you want to work on this whole image and give the following—

OR look like this —

Before you go forth and create a few duplications of the Square and Particle output, or even planning to perform several rounds of encoding, consider the use of Frame Buffer:

A Frame Buffer object on Layer N represents the flattened image from Layer 1 to Layer N-1

 

Hence, with this feature in mind, you can apply effects on the whole composite image without costing too much on CPU(but cost memory) and ensure the duplicated image is a faithful copy of the original. Particularly useful when Particle output is involved. (Particle output is CPU intensive and two different PO objects will likely have different patterns at any moment).

The arrangement of objects for the above image will look like this:

And you can find this “Frame buffer” object in the usual place:

So, what happen if two FB objects are placed on consecutive layers?

The previous definition still holds— Anything above this Frame buffer layer (including previous frame buffer objects) are combined and represented by this frame buffer object.

Recently someone posted here claiming he/she cannot reply to a comment. After investigating in this morning, I found that the “Crayon syntax highlighter” is giving a lot of PHP errors and deactivating it restore commenting functions.

As a result, I have removed crayon syntax highlighter but some older post that utilized its feature will give a code block without coloring.

I’m putting up this post to serve as a central place for discussing AviUtl’s translation.

Due to the complexity of this program and my lack of vocab concerning CG and design, it is certain that typos will appear and some features got a strange translation.

When starting a discussion/suggestion, please observe the following:

> State clearly where the text can be found, not just the text itself.

> Limit the translation to a max of 10 alphanumeric characters, if possible.

Thank you for your feedbacks!

A sudden power failure turned off my server yesterday(while I was sleeping) and led to ~10 hr of down time.

Sorry for inconvenient caused.

If I were to release Extra Pack 1.5 now, there will be merely a few translation fix, and perhaps including one or two more plugins(a loupe, and a SmoothD plugin, optionally a color waveform scope[for really advanced color correction]). I wish to accumulate more changes before releasing the next extra pack.

SO, as a user, what do you need?

What do you think is essential for common tasks that is not yet included in EP1.4?

剛才去百度嘗試搜尋一下AviUtl和本網站, 發覺完全沒有本網站的條目(videohelp上的則有顯示)…難怪不少中國大陸民眾都不知道已經有全面英化的版本, 還有人重做AUL的漢化…

[DOWNLOAD]

96KB; MD5: 212304DDD7B0FB255AE234C34DC6E400

===============================================

Chapter Editing Plus
(with Silent search function and sync with frame removal)
codebase by Homu,
extended by Rutice(2012, [Src: https://github.com/rutice/chapter.auf/]),
translated into English by Maverick Tse.(2014)
—————————————————-
On top of the conventional “Chapter Editing” that allows you to
mark chapters and automatically add it to mp4/mkv, this plugin
provides some useful feature for cutting out commercials.

After loading the video into AviUtl, click on [Silence]. Then the plugin will mark(add a chapter point)
all the starting points of silent segments.
Note that it will not mark the end of silent segements however.

Tuning:
Frame# — For how many silent frames in a row should it count it as a true silent segment.
For filtering out very short silent segments.

Threshold — Below which volume level should the plugin treat the frame as “silent”.
Useful to skip BGM to some degree.
Set it to <10 if there are no BGM

SC options:
Take scene change into account so that chapter will less likely
to be created in the middle of a transiting scene.

Others:
This plugin will also update the chapter timing if you cut away part of
the video after creating the chapter list.

Remember to check “AutoSave” if you want chapter info to be included
in your MP4/MKV when using the x264GuiEx/QSVEnc/VCCEnc encoder interfaces.

#Installation
Put the .auf into the .\Plugins folder, rename/move the old one if necessary.

#Access

Display> Chapter Editing Visible

Recently, I am interested in the possibility of writing a plugin for AviUtl to achieve “Motion Tracking” like the one in After Effect.

Writing from scratch is impossible, so I am using the OpenCV2 library.

Since it is the first time I tried to write a full-fledged plugin, a lot of problem was encountered. So far, I managed to get the RGB image of any frame from AviUtl and feed it into OpenCV’s Mat object.

Both AviUtl and OpenCV have their unique ways handling coordinates and image, so it is quite hard to get them working together(esp. for first time).

If I get its core function working, I may release its source so that other people can improve on it. After all, many OpenCV examples are either too old(OpenCV v1.x) or too complicated and lacking sufficient comments.

(I am planning to use CamShift in the simplest sense, so do not expect the tracking work as well as commercial software.)

[DOWNLOAD]

1.18MB, MD5:010103E205388129CCF144AD1D4ED54D

Motion Tracking Plugin for AviUtl
by Maverick Tse
(alpha 0.01, 2014-02-01)
======================================
Intro
———
This plugin allows you to track any object in a video using OpenCV’s CamShift algorithm.
(Recognition by color)
Previously, similar function is only available in expensive commercial software.

System Requirement
———————–
-CPU with SSE2 support
-Win7 or later
-VC2012 Redistributable
-AviUtl v1.0 with Advanced Editing(拡張編集)　0.92

Installation
—————-
Extract all files in the archive and put into AviUtl’s folder

Usage
——-
1> Desginate the object to track using X,Y and width parameters
2> If you are using the English version of Advanced Editing, tick “English EXO”
3> If you want to generate a Sub-filter(部分フィルタ) instead of typical object, tick “As Sub-filter”
4> Click on [Analyze!]
5> Load tracking.exo (inside AviUtl’s folder)
6> Adjust as needed.

Tuning
——-
> Try to set an area with distinct color
> If tracking does not works well, adjust(increase) Smin, Vmin.

Please report bug/suggestion to:

http://mavericktse.is-a-geek.com/wordpress/

History
———
1 Feb, 2014: Initial release, alpha 0.01

2 Feb, 2014: Bugfix for interface and last frame, beta 0.1

Last year’s December, my EVGA GTX570SC died in action. Luckily it is still covered by warranty and fixed by by EVGA (Taiwan)  after 2 months of waiting.

As a result, works on MMD/3DCG can be resumed very soon.

The MMD Halloween series shall be finished in this season.