#texttutorial

Installing Avisynth+ and KNLMeansCL:

There is a newer version of Avisynth+, but I will be showing how to use this version of Avisynth+ because I’ve tested it heavily.

Step 1: Updating Avisynth-

When you get to the AVS+ download page, scroll to the assets dropdown and click on the latest release’s EXE file. Click on the one that includes “vc_redist”.

Once you download it, run the installer. When it asks if you want to keep your old version of Avisynth, make sure the option that preserves your old Avisynth is selected. This way, you will still be able to install plugins to the Avisynth 2.5 plugins folder.

Step 2: Installing KNLMeansCL -

Once you download KNLMeansCL, unzip the file and copy the KNLMeansCL.dll file to the plugins folder here: C:/Program Files (x86)/AviSynth 2.5/plugins.

Step 3: Fixing the AvsPmod Error-

For some reason AvsPmod will show an error once you install Avisynth+, but this can be fixed with the library.zip file you downloaded!

Place the library.zip file in the C:/video/AvsPmod/lib folder. DO NOT UNZIP IT. Now when you run AvsPmod no errors should pop up.

Step 4: Using KNLMeansCL-

Run your normal Avisynth process. After you paste your settings from the resizer into AvsPmod, paste the KNLMeansCL settings you would use in Vapoursynth after debilinear/resamplehq and after (or before) your fine sharpening.

If you are confused on what settings to use, look at KNLMeansCL's official wiki

Example:

debilinear(630,354) finesharp(sstr=0.5) KNLMeansCL(1, 2, 4, 1.2) resamplehq(srcmatrix="TV.709",dstcolorspace="RGB32")

crop(181,3,-181,-1)

Adjust the last number (1.2) to control the strength. Lower values weaken the denoising and higher values strengthen the denoising.

Place this code in your script whenever you want to use it : )

*NOTE: KNLMeansCL has a few requirements to run properly, but they should already be automatically installed on your computer. You can view these requirements here if you face any errors. You probably won’t get any, but message me if any errors pop up*

*Optional* Adding KNLMeansCL to the Resizer:

I understand that it may be easier to just copy code from the resizer, so I have altered the resizer to include this feature. 

*NOTE: I would save the old resizers somewhere safe before replacing them :)*

If you are using brandinator's resizer-

Download the edited resizer from here and place it in C:/video/temp. Now the resizer will look like this:

If you are using my Avisynth multiple resizer scripts-

Download the edited template from here and place it in C:/video/MultipleResize. Now the next time you run one of the scripts the resizer will look like this:

*NOTE: The new resizer is set up so that you control the strength of KNLMeansCL’s default settings which is knlmeanscl(1,2,4,1.2). If you want to change it, just open the resizer.html/HtmlTemp.txt file in notepad, search for knlmeanscl, and change 1,2,4 to your 3 values. Example: changing 1,2,4 to 0,4,6. Check the official wiki to understand what the first 3 values do.*

I really hope this helped you. I am terribly sorry for the long delay in response; I had to make sure I tested this thoroughly : )

*Note: removing the logos combined with good coloring makes the logo less noticeable. I always remove logos before I sharpen, denoise, color, etc*

First Method - Blurring the Logo:

This method is really easy and straightforward. Blurring the sharp logos help make them look less noticeable

I will be blurring the channel logos on this gif scene of Hanbin

Step 1: Convert To Video Timeline- 

For the first method you must be in Video Timeline mode. The picture above shows the steps to convert your frame animation to video timeline. Once your animation is a smart object you need to duplicate it; this will be your blurred layer. Because the bottom logo pops on later, I duplicated the smart object again, so I can blur this layer separately

Step 2: Select The Logo-

Select the logo by using the rectangular marquee tool or lasso tool then press the “Add layer mask” button (as shown in the gif above). After you make the layer mask, click on the layer thumbnail to select it so that you can blur the image and not the mask

Step 3: Blur-

Go to Filter>Blur>Gaussian Blur and change the radius setting until the logo is blurred to your liking

Step 4: Trim Duration of Blur-

If you have a logo that is temporarily on the screen you may want to trim the duration of the blur. Since the bottom logo showed up later, I repeated steps 1-3 and dragged the end of the video timeline layer to the time that the bottom logo appears 

This is the result 

Second Method - Using a Content-Aware Fill script:

Unlike blurring, with content-aware fill you never know exactly what the result will look like. With content-aware fill, Photoshop fills in a selection with random parts of the image in close proximity to the selection. Because the algorithm is random, each time you run this script it will produce a different result

This method works best for scenes that do not have a lot of detail near the logo (e.g., gradients, solid colors, blurriness, etc.) and/or small logos

*Note: This method was tested on PS CC 2014 & 2017 previous versions might not work. Also, I noticed that 2017 did a better job at removing the logo. I recommend using a newer version of PS for this method* 

I will be removing the “Seventeen” text from this gif scene of Wonwoo. Sorry for the quality... inkigayo ain’t shit

*You will need to download the .jsx Photoshop script file from this folder*

Step 1: Install Script-

You don’t have to install the script, but if you want it to show up in the file menu for easy access instead of browsing for it, copy the file to C:\Program Files\Adobe\[your version of ps]\Presets\Scripts

Step 2: Convert To Frame Animation-

For the first method you must be in Frame Animation mode. The picture above shows the steps to convert video timeline to frame animation

Step 3: Make Selection and Save Current State-

As I mentioned before, Content-Aware Fill is unpredictable, so you may have to run the script more than once to get the result you desire. The only problem is you can only press Ctrl+Z so many times, and by default PS only stores 50 history states. If you have more than 49 frames being processed, you will not be able to return to the state your animation was in before you ran the script. The solution is to make a snapshot of the current state before you run the script

Delete the frames you do not need, make your selection with a marquee tool or lasso tool, and then open the history window and click the camera icon to save the current history state. No matter how many frames the script processes, you can always click on “Snapshot 1” to restore the state before the script was ran

Step 4: Run the Script-

Start the script by clicking the menu item or browse for where you have saved it. The script will pop up 3 prompts: 

The first asks you to verify that you are in frame animation mode and you have selected the logo

The second asks what frames you want to process. You enter this in a,b format. For example, If you only want to remove the logo on frames 1 through 10, you would enter 1,10. If you want to process all frames just press ENTER on the keyboard

The third asks if you want to use color adaptation (color adaptation is not available for PS versions older than PS CC 14). If you enter ‘y’, Photoshop will attempt to blend the fill with the surrounding colors more. Sometimes it looks better sometimes it doesn’t. Test it to see if you like using it

...or you could just press enter a bunch of times until the prompts go away

This is optional, but sometimes after the script runs, you may want to use the spot healing brush tool to fix some errors content-aware made

This is the result. Here, content-aware fill does a pretty decent job because the background is a gradient and his hair does not have much detail due to the trash source quality.

In this example you can clearly see... the Content-Aware script is not flawless. If you look closely, you can see that Bobby’s mouth ended up on his neck sjkjskjsk. To fix this problem, You can make a different selection and run the script again hoping that you get good results, use the healing brush tool, or give up like me and just blur it instead.

I have put a .zip folder with psds of the un-altered gif frames of Hanbin and Wonwoo in my Tutorial Resource Folder. You may use them for practice or whatever you please

Denoising with a program called Neat Video > NIK Collection - Color Efex Pro 4: Detail Extractor> Topaz Detail > Topaz Denoise

I DO NOT recommend using the NIK Collection on gifs because although it makes your gifs super sharp it takes forever to process. I am only mentioning it because I want to be transparent that I did not only use topaz so you may need to adjust the settings.

Here were my Topaz Detail Settings:

Everything else is set to default. Adjust where necessary. 

For Topaz DeNoise:

I changed these settings alot (especially the overall strength slider) depending on how much noise was in the gif.

So those were my Topaz Settings in this ridiculously long answer.

I might make a video on how existing Avisynth users can use Neat Video because it basically does everything Topaz does in less than 1 minute.

This may seem like a lot, but when you practice, the process goes by much faster than what it may seem now.

Step One: Open your video with the .Bat file-

After every file is placed correctly and Avisynth is installed, you want to drag your video file into one of the .bat files in the video folder.

You will probably not use the .bat files that are highlighted in red, but if you want me to explain their uses, message me.

.Bat Files Highlighted in Green: 

Normal.bat

Drag .ts/.tp video files into it

Creates 17 second video with original dimensions 

You will have to Crop and resize in photoshop

Normalwebm.bat

Drag .ts/.tp video files into it

Creates 17 second video

You can crop video and resize using Html Resizer

Normalwebmrange.bat

Drag .ts/.tp video files into it

Creates a custom trimmed video

You can crop video and resize using Html Resizer

Reencode.bat

If you make a mistake when saving your avs.avs script, edit and save it, and then double click on this bat to convert it to an .avi file

x264Lossless.bat

Drag any video file into it (.mp4, .mkv, .avi, etc)

Creates 17 second video with original dimensions 

You have to crop and resize in photoshop

Use this if you got an error using the other .bats

x264Losslesswebm.bat

Drag any video file into it (.mp4, .mkv, .avi, etc) 

Creates 17 second video

You can crop video and resize using Html Resizer

Use this if you got an error using the other .bats

*NOTE: You will usually get an error with the normal .bats if your .ts/.tp file is not mpeg-2 format so use x264Lossless .bats for those videos.*

I use Normalwebmrange.bat most of the time.

Drag your video into the .bat you want to use. After opening your video with whichever .bat you want, a command prompt window will pop up. Here you enter the start timestamp. This is where you want your video to start. For normalwebmrange.bat, it lets you choose what time you want the video to end too. The rest of the .bats make a 17 second clip from your start time. 

*NOTE: Make sure you type the timestamp in HH:MM:SS format*

After the the command prompt is done (it takes about 40-60 seconds ), Google Chrome will open the resizer.html and here you will resize and crop your video

Step 2: Resizing and Cropping-

In the Resizer you want to choose your gif size. By default it is set to 268px width by 350px height. You can set the width to any of the the Tumblr sizes and have the height be whatever you want. Then drag the lower right corner to scale the video and drag the video to fix the position. The opacity slider helps you to better visualize what your video will look like cropped.

After cropping and resizing the way you want, you need to choose a Preprocessor:

None

Use this if you are using files that are not .ts/.tp files

QTGMC 30

Keeps original frame rate

QTGMC 60

Doubles frame rate which produces a smoother gif

*NOTE: I would use this option for 178px and 268px width gifs because more frames = higher file size. You might go over the Tumblr 3MB limit with bigger gifs*

Fast vs. Slow

Fast of course takes less time but is lower quality

Slow takes a lot longer but it is higher quality

Sometimes there is not a lot of difference in quality between the two. Test and see which one you like better. I use qtgmc 30 slow most of the time because i love placebos ): 

Then choose one of the three resizers: resamplehq, debilinear, and dithertools. I normally choose “debilinear” because it makes the video sharper than the other options.

Choose your desired options, add extra sharpening if you wish, and copy the text in the white box. If you want to crop in Photoshop instead, Do not copy the last line of text

Close the Html Resizer and the program Avspmod should open if it has not already done so

Step 3: Avspmod Trim Video and Save Script-

Paste the text that you copied from the resizer in google chrome (on line 17)

*SUPER IMPORTANT NOTE: PLACE A “#” BEFORE QTGMC ON LINE 17. IF YOU DO NOT, WHEN YOU TRIM, AVSPMOD WILL FREEZE AND STOP RESPONDING*

And then if you selected QTGMC 60 in the resizer, remove the “#” from line 7. This will make sure that your video moves smoothly.

Now trim your video. Click on the up arrow in the bottom left corner to show your video, and then scrub through the video to find where you want your gif to start and end. Press the home key on your keyboard (or go to Video > Trim Selection Editor > Set Selection Startpoint) to set the start, and then find where you want to end and press the end key on your keyboard (or go to Video > Trim Selection Editor > Set Selection Endpoint).

In the trim editor make sure you have my settings shown in the above, then click on line 8, and lastly press “Apply”

remove the “#” from qtgmc and save the script by pressing CTRL + S and close out of Avspmod

*NOTE: ONCE YOU REMOVE THE “#” DO NOT CLICK ON YOUR VIDEO AGAIN OTHERWISE AVSPMOD WILL FREEZE*

Step 4: Wait for VirtualDub to Process and Open the Output in Photoshop-

After you close Avspmod, Virtual dub will start to process your video. If you want to see the progress, click on the virtualdub icon in the taskbar go to View > Show Status Window and you should see the current frame being processed, the time that has elapsed, and the estimated time it will take to finish.

*NOTE: IT USUALLY WILL TAKE 1-3 MINUTES TO CONVERT. Mine took longer because I had multiple programs running and I was screen recording*

After VirtualDub is done processing, you can find the output video in the folder titled “temp”

Here is your video.avi file. Now you can open it in Photoshop, color it, use topaz, etc, and then you can export your gif. 

*NOTE: If you want to run this process again, rename the video.avi file to something else so that it doesn’t get deleted*

Final Gifs:

These two gifs will be in my Edit Gallery. You may use/edit them as you please. I have put psds of the un-altered gif frames in my Tutorial Resource Folder