/u/ - Yuri

Photoshop Performance

Before doing anything in Photoshop it's best to set the preferences. Most of this is obvious enough and will be down to personal taste.

File Handling

I recommend disabling autosave. Just remember to save what you're doing frequently or get good and stop fucking up.

Disabling psd and psb file compression will take less time to save documents but will take up more HDD space.

Performance

This is the most important tab to set up correctly.

>Memory usage

This slider allows you to set the maximum allowed RAM that Photoshop can use. I allow Photoshop to use about 80% of 32GB. Remember to leave enough RAM for other programs.

>History and Cache

Document history in Photoshop are previous states of the document saved to scratch. The more history states you allow the further you will be able to step back (crtl+alt+z), undoing more changes. However allowing more history states will take up more scratch, so if you're working with a limited amount of RAM, don't set this too high.

Photoshop uses a cache stored in RAM to display the document in the window view. Using a higher level of cache will speed up document navigation using the hand and zoom tools. This is especially true of large documents. The exact setting will depend on how much RAM you have, as more cache levels takes up more RAM.

>Scratch Disks

A scratch disk is a temp file placed onto a physical drive where Photoshop will write history states to when it runs out of RAM. Running out of RAM and writing document changes to a HDD can cause severe slow down. I recommend using the fastest drive that isn't your boot drive. Please be aware that this can absolutely hammer hard drives. If you have an older drive with important data stored on it, don't use it for scratch. If you've got a free SATA port on your motherboard, get a 60GB SSD (Kingston A400 drives sell for around $30 on amazon) and let that get raped instead.

>Graphics Processor Settings

This is entirely dependent on what GPU you have. Photoshop uses OpenCL rather than CUDA, so there is no benefit to having Nvidia over AMD. Having a powerful GPU will enable Photoshop to use OpenCL acceleration to enhance certain filters and functions. For scanlation the most important tools that benefit from this are the transform tools and image resizing.

Cursors

While there's something nostalgic about the standard cursors as they have changed little since version 4.0, use the precise cursors for a better view of what you're doing. This is important for redrawing.

Processing Raws

Stage 01 - Step 01

Open up the raw scanned image files. I usually open up 10 at a time.

Notice that the raws have been scanned in RGB and are therefore using RGB colour space. In this example the images have green tint from the cathode in the scanner.

On the first image, rotate the image to the correct orientation.

Image > Rotation

Use the rectangular marquee tool to select around the page. Allow enough space for other pages to be at different angles.

Copy the selection (crtl+c), then open a new document. The document will automatically set the dimensions to the selection currently in the clipboard. Remember to set the document to 8-bit greyscale and paste the selection (crtl+p) into the new document.

Repeat this for every greyscale page in the doujin, pasting them into the document that you have just created.

If you having issues with copying large image files to the clipboard, right click on the layer and duplicate it to the new document instead.

Doing this will convert the raw images to greyscale and copy them into a single document in one step.

If a pasted image is larger than the document canvas, use the move tool (hotkey v) to position the page at the centre of the document.

Stage 01 - Step 02

You should have all the pages of the doujin on their own seperate layer in a single document. While this takes up more resources, it is more time efficient than working on each page it's own document.

Now is a good time to save Save it as a psb (photoshop big) file. psd (photoshop document) has a file size limit of 2GB which will be exceeded when saving high resolution, mutlilayer files.

The next thing to do is level each layer.

A quick way of making a layer visible when there are a lots of layers, is by holding the alt key and clicking the eye icon of the layer you want to make visible.

I level destructively by using the levels panel in the image menu (crtl+l) because it's faster and I've been doing this long enough that I rarely make mistakes.

If you're less confident, you may prefer to do this non-destructively by using an adjustment layer. You can merge the adjustment layer with the image layer by using merge layers (crtl+e) in the layers menu. Make sure you have the adjustment layer selected as this works by merging the selected layer with the layer below.

I under level raws because I believe that the final "correct" leveling should be done by the cleaner, not the raw provider. This isn't a standard in scanlation. I'll cover how to level an image in much more detail at a later date.

Stage 01 - Step 03

Now it's time to prepare the layers for step 02.

Resize the image to half of it's existing size. I'll explain why this important at a later date.

Rotate the pages so that they are straight by using the ruler tool.

Select the ruler tool and then find a line that should be straight, such as the edge of a panel. Use the edge of the page if there are no printed straight lines. This isn't ideal as the edge of pages are where the scan in most likely to be distorted. Then click straighten layer in the top bar. Do this for every layer.

Stage 02 - Step 01

Copy the first page and create a new document.

Crop the new document to fit the page. This can either done with the crop tool (hotkey c) or the rectangular marquee (hotkey m).

Delete the layer currently in the new document

Move the layers in the existing document into the new cropped document. Either by copypasting or by using the layer duplication method

Use the move tool (hotkey v) to position the pages.

Stage 02 - Step 02

Don't forget to save the new document

The next step involves some light cleaning.

Check all of the pages for dust or any other contaminates. For white areas either use the paint brush or the dodge tool set to highlights. For dark and halftone areas use the clone stamp tool (hotkey s) with a hard brush. I'll cover how to use some of the tools in more depth later.

Once the raws have been cleaned, they can be saved to individual files ready for a release.

There's nothing particularly difficult about this method, it's just time consuming. There is no use of batching or auto tools. Everything is done manually.

Tips for Typesetting

Center the Text in the Bubble

Text layers can be moved like normal layers by using the move tool (hotkey v). While the move tool is selected, you can nudge layers by using the arrow keys.

Use a appropriate font size.

It's possible to do this easily by using the scale tool

Edit > Transform > Scale

Hold the shift key while dragging the corners of the box to keep the aspect ratio.

TEXT BOXES ARE FOR FAGGOTS

While using a text box is quick and easy, you will get better results by shaping the text manually. This can be done with the enter key. The aim is to have an aesthetically pleasing shape for the text that corresponds to the length and width of the bubble. The only exception to this would be large speech bubbles featuring a lot of text. Perfection isn't always possible. It's worth asking the TL or the script edit if changes can be made to the script to make typesetting easier.

Font Variation

I'll cover font selection in a lot more depth some other time. For now I'll start with the basics.

Use different fonts for strong expression of emotion, regular speech, sfx and internal monologue.

Also remember to use bold and italics where necessary

Use Warp, Vector Effects and Transform Tools

Warping Text

<Right click on text layer > Warp Text

Transforming Text

Edit > Transform

Vector Effects

While a text layer is selected, click the FX button at the bottom of the layer window. Here you will find stroke, outer glow and gradient effects.

Examples

https://exhentai.org/g/1024953/ac2215b314/

https://exhentai.org/g/916106/65a8ea966c/

https://exhentai.org/g/1048280/654d34b787/

Further Advice

Use the typography workspace

Window > Workspace

This will not loose your current work space setting. Selecting Essentials will take you back to your usual work space.

In time you will develop you're style of typesetting. Use the tool presets in Photoshop for easy access to favourite fonts and styles.

>>43589

Mind sharing how to make outlined text like this too?

Dealing with Magazine Raws

Raws taken from Japanese magazines usually suffer with several problems caused by the quality of paper used. The most noticeable issue is speckle. This is appears in scans as noise in shadow and halftone areas. The common way of dealing with this is simply to level the fuck out of the raws. Don't do this. There are several other ways of removing speckle that do not involve destroying dynamic range. I'll explain what dynamic range is and why it's important in another post.

For speckle in highlight areas, use a combination of the dodge tool set to highlights and the paint brush.

Speckle in shadow areas is much more difficult to remove. How this is done exactly depends on the raws themselves. No two raws are the same due to several factors. Such as different publishers using better or worse paper, the scanner used by the raw provider and how the raw provider has processed the raws.

The first step is to remove as much of the noise from midtones of the shadow areas as possible. This is appears as dark grey pixels in areas that should be black.

Inverse Overlay

Duplicate the layer and then change the blending mode of the duplicate layer to overlay and invert. While this is quick, it will result in loss of range and often created more moire problems.

Adjusting Input Curves

Image > Adjustments > Curves (hotkey crtl+m)

Curves are also available as a non destructive adjustment layer

This works in a similar way to the levels panel, only it allows much finer control. As default the histogram is reversed from what is shown in the levels panel, to change this, select Light (0-255). On the line in the histogram window place three points in each range, so one for highlights, one for mids and one for shadows. Then adjust these points by dragging them with the mouse. Remember to check the whole image. Targetting unwanted greys in one area, will effect those same greys over the entire image.

Brightness and Contrast

Image > Adjustments > Brightness and Contrast

Also available as a non destructive adjustment layer

This is fairly simple, use the sliders to adjust the brightness and contrast. Again remember to check the entire image.

Exposure

Image > Adjustments > Exposure

Also available as a non destructive adjustment layer

This is another simple slider panel. It's not usually necessary to adjust the exposure and offset. Use the gamma correction to darken the unwanted greys. Don't forget to check the entire image.

Shadows/Highlights

Image > Adjustments > Shadows/Highlights

Another panel similar to levels and curves, this allows for even finer adjustments. Click the show more options box.

The PA Method

This is time consuming and labour intensive, so I don't actually recommend doing this, unless you want to spend between 2 and 18 hours cleaning a single panel. However there are some techniques that can be taken from this method.

Paintbrush

If all else fails to produce an acceptable result, it's always possible to simply paint over the shadow areas with the paint brush (hotkey b). If you don't have the wrist control to trace over more complex line work, such as hair, you can also use the pen tool.

Underpainting

Rarely used in scanlation, but it can be very effective at restoring range and reducing moire. In PA I did this by hand using the blur tool with the dodge and burn tool. A similar effect can be achieved much more quickly by using filters.

Duplicate the layer you are working on

Place it underneath the working layer

Switch off the working layer to reveal the duplicate layer

Apply a filter, either gaussian blur or median

Adjust the levels of the layer. This can either be done with the dodge and burn tool or by using the levels panel

Blend the working layer and underpainting layer by adjusting the opacity of the working layer.

The Dodge and Burn Tool

Shadow areas can be cleaned by using the burn tool. It's use is similar to the paint brush in that you will need to "piant" over these areas with a brush. Because the dodge and burn tool targets certain ranges, you don't have to be as careful.

In the second gif it can be seen how I have used the burn tool to clean the black areas and have created a gradient that merges the lower part of Yuzuki's jacket with the black area.

Advice on Comfort

Sitting hunched in front of a desk for hours is not healthy. This is especially true when using Photoshop. The pressure put on your hands and wrists can cause compression on the ulnar and median nerves. This can result in chronic pain and a loss of function in your hands. It is worth taking seriously.

Make sure your desk, monitor and chair are at the correct height, so that you don't have to hunch downwards to look directly into the monitor.

Get a mouse pad with a wrist rest.

When using hotkeys do not hang part of your hand over the edge of the desk. Move the keyboard further up the desk.

Do not rest your elbows on the arm of the chair. If necessary use supports and place them underneath your forearm.

More on Typesetting

Here is a short pixiv comic that uses a variety of typefaces and sfx. I've already cleaned off the jpeg artifacts and done the redrawing.

>>44817

Missed out the first page

Stage 01

If you're unsure of what text formatting to use, as I am here, then it's best to place the text on the image first. This can be done by copy pasting the translation onto where you want the text to be. It can also done by typing out the text manually. Either way, you will still need to read the translation to check for any errors.

EDIT: grammar

Now it's time to experiment with fonts and formatting.

Refine the layout further, this will help with font selection.

Text layers can be easily moved around the canvas by using the move tool (hotkey v)

Text in speech bubbles is always centre aligned. Look to the original Japanese typeset for how to align boxes.

Text can be quickly resized by selecting the text layer you wish to resize and then selecting scale from the transform menu

Edit > Transform > Scale

Hold the shift key while dragging the corners of the box to keep the aspect ratio.

To rotate text select the text layer to be edited and select rotate from the transform menu

Edit > Transform > Rotate

Drag the points on the transform box to spin the text round.

To add stroke to text click the FX button at the bottom of the layers menu.

Now it's time to add different fonts.

Notice that I am using the typography work space.

By default when you first type into Photoshop, it will use the font that was last used. Depending on what you're typesetting this may be appropriate or not. The typeface and default kearning on the current font is too wide for most of this comic, so I will need to choose a more fitting font.

Look at the original when deciding on fonts.

To warp text, simply right click on the text layer and select warp.

Choose the style of warp from the drop down menu and then adjust the sliders accordingly.

First version of the first page

A good way of saving time when applying vector effects (text effects in the FX menu) is to use the copy layer style feature.

Right click the text layer that has the effects you want to replicate

Select Copy Layer Style

Then right click the layer you wish to apply to same effects to and select Paste Layer Style

Done.

Everything here was achieved with the above methods. Considering this was only a short pixiv comic it has taken quite a while to do but then having to switch fonts and use vector effects will take time.

>>44826

After TL check

Dealing with Low Bit Depth Raws

I wasn't planning to cover this for a while but raws like this are something I don't see very often. Thanks to SHiN for allowing me to use his raws for this demonstration.

Every page, except one is saved in 1-bit. Bit depth or colour depth determines how many colours there are in an image. I will explain bit depth and the relationship it has with colour space and file size in another post. Basically bit depth sets how many colours it is possible to display in an image.

Look at the difference in the four images. Bit depth is in the file name.

2-bit images are fairly easy to clean. There's no dynamic range loss to worry about and most raws of this type will be black and white already. The issue is going to be redrawing. The usual tools such as the paintbrush aren't going to help here. However there are tools in Photoshop that will do the job.

Pencil

Found in the paintbrush menu. Right click on the paintbrush in the menu bar. Hotkey b

This is used just like the paint brush. It has very hard edges and will not produce any grey pixels when painting in black

Eraser

Hotkey e

There is a pencil mode for the eraser

Pen Tool

Hotkey p

In older versions of Photoshop there was an option to turn off anti-aliasing when using the pen tool. Unfortunately this option is no longer available in CC. However there are several ways around this which I will explain later.

Clone Stamp Tool

Hotkey s

There is no pencil mode for the clone stamp tool, so it will need to be used with care.

Redrawing in this bit depth is rather like creating sprites and Photoshop can be a very powerful spriting program when you know a few tricks.

Using the Pencil Tool

Draw with it as you would the paintbrush. When redrawing line work, look closely at the existing line. Your aim is to mimic the original as closely as possible. Look at where the redrawn line meets the existing line and make sure that it matches. Also look at the spacing of the progression in a different direction.

Using the Eraser in Pencil Mode

It removes pixels as it does in brush mode but this time works like the pencil.

How the panel looks when finished

>>45010

Forgot image

Select the pen tool and set mode to shape.

Place the points at the start and end of the line you want to redraw.

Right click on the pen tool in the tool bar and select the convert point tool. The cursor will be a white arrow.

Place the cursor over the point you need to convert. The cursor will change. Hold it and drag to "split" the point.

Two lines with balls the end will appear out from the point. These are the beziers.

To manipulate the beziers click and drag on them. Pulling them in different directions will alter the curve of the line.

Set the colour and width of the line by using the top menu bar.

You will now have visable line.

Click off onto a non-shape layer to view the line without the bezier curves.

There are two ways of proceeding. There may be more ways of doing this that I'm not aware of.

First Method

Duplicate the shape layer and rasterise the new layer copy. I colour the shape layer as red and the rasterised layer as orange to help with layer organisation.

Right click layer > select colour

Select the rasterised layer and then apply a "sharpen more" filter.

Filter > Sharpen > Sharpen More

You can now press crt+f to reapply the layer as many times as you need to remove the antialiasing.

The line doesn't match the existing line work, it's too thick. Trim it down with the eraser tool for a perfect match.

Second Method

It appears that this only works with shape fill and not with stroke. Even when used with a filled shape it doesn't work as well as the above method.

Proceeding on from the end of the first method.

The line we have currently looks too basic it doesn't fit well with the rest of the line work, so a few extra details need to be added. This is where it really helps the have a good knowledge of anatomy.

Have a really good look at Yurika's left arm as you will need to mimic that style exactly. A major component of the art of perfect redrawing is to able to mimic the drawing style of the artist.

Before doing any work with the pen tool, it may help to roughly draw in the extra detail with the pencil.

Using the Pen Tool in Path Mode

For the upper crease on Yurika's arm, I'm going to use the pen tool again. This time rather than creating a line that is of equal thickness all the way a long, I'm going to draw a line of variable thickness that more closely resembles drawing with a real pen or pencil.

Select the pen tool (hotkey p). From the top menu bar, click the drop down menu that currently says shape and select path.

As you are not using stroke, you need to place three points to make a complete path.

Manipulate the beziers to make the shape that you want.

Make a new layer.

With the pen tool selected, click Selection in the top menu bar.

Choose your settings in the window.

Fill the selection by using the paint bucket (hotkey g).

Like before, it may not be perfect and will need further adjustment with the pencil and eraser.

How it looks when finished

Something that I find mildly annoying in the work of other cleaners, is that they do not align their pages. What I mean by this, is that when working with a doujin or manga that has framed edges, the placement of this frame differs by page. It's really easy to fix and I don't understand why more cleaners don't do it. Pages without frames, generally do not benefit from alignment. It's possible to align the page numbers if they are centre aligned on every page but this is too autistic even for me, and it's usually better to prioritise preserving the amount of artwork shown in the cropped image.

Here's how it's fixed

Start from Stage 02 - Step 01 >>43244

Correctly place the first page.

Switch on the second layer for the second page, leaving the layer for the first page visible underneath.

Change the opacity of the page 2 layer to around 50%

Select the move tool (hotkey v)

Zoom in (hotkey z. Hold alt and click to zoom out) to the border

Use the arrow keys to nudge the page 02 layer to align with the page 01 layer.

Keep in mind that you may need to alter how you originally cropped the document. This can be done by using canvas size.

>Image > Canvas Size

Do not use image size as this will change the resolution of the layers as well as the document.

For this particular doujin, I need to align the white stripe.

When done, change the opacity back to 100%

Switch of the page 02 layer and repeat the same actions with the remaining page layers

When this is done, every layer will be aligned with the first page.

The result doesn't have to be perfect, one or two pixels of misalignment won't be noticeable. There also may be reasons, such as distortion when scanning that prevent perfection. Doing this doesn't take much time and adds a level of polish to raws.

EDIT: quoted post I didn't intend to

>>46768

>>46769

>>46768

>>46769

>>46770

Site software uploaded the images in the wrong order

gif shows misaligned pages

>>46772

>>46773

Dealing with jpg

It's 2018 and this is still an ever present issue. There are many filters made to deal with jpg artifacts and some work quite well for colour. However the best way to deal with it in greyscale is with the dodge and burn tool.

Here's a page from a raw I've been asked to clean.

The issues are as follows.

>poor quality scanner used

>low resolution

>jpg artifacts

The raw provider hasn't leveled these at all, so the dynamic range hasn't been completely destroyed.

Duplicate the layer as this is the layer that you will be working on. Leave the original layer as a comparison.

Apply an anisotropic diffusion filter

Fade the diffusion filter

>Edit > Fade

Exactly how much to fade the filter depends on the raw. An anisotropic filter will make a raw appear blurred and lines take on a liquid appearance like they're melting. The aim of fading the filter is to remove this while still keeping the increase in dynamic range. The mild blur effect will also remove some of the jpg artifacts.

>>46806

The next step is to remove the artifacts from the highlight areas outside the panel. This can be done by using the dodge tool set to highlights and exposure set to somewhere around 2-5%.

>>46807

Working gif

>>46808

Now I'm going to correctly level the image. Due to the jpg degradation this is going to be a complex task.

Step 01

Before touching the image with the leveling tools, look at the artwork and decide how to level the image. Does the line work need to be completely black? Are the highlight areas sufficiently leveled already? Is there a lot of midtone?

Here the artwork uses a lot of midtone, so the aim of leveling will be to preserve this while darkening the shadow areas to bring out detail and contrast.

Step 02

Open the levels panel (hotkey crtl + l)

Under level the image without destroying any detail. It may help to zoom in to an area such as eyes where there is a lot of fine line work. If you are having difficulty doing this, you can always undo (crtl + z) and redo it. It may also help to use a nondestructive levels adjustment layer.

Step 03

Zoom in to the image and look at what ranges still need to be correctly leveled. Here I need to darken the shadow areas a bit more.

Open the curves window (hotkey crtl+m)

How to use this is described here >>43706

While the artwork inside the panels is now correctly leveled, the panel edges and speech bubbles are still not black, so this will need to done manually with the burn tool. Usually this done with the burn tool set to shadow with a 5-15% exposure. However on this raws I want a much more gentle effect, so I will be using the burn tool set to midtone with a higher exposure of 30%.

How it looks when finished compared to the raw.

The change is subtle and the majority of people will not notice or appreciate the difference.

Once detail has been lost, it doesn't come back. Photoshop can only alter what is there.

Remember to always save your work to png.

Using References in Redrawing

When redrawing you're going to need at least a basic understanding of architecture and human anatomy. Before you can redraw, you need to visualise exactly what you are going redraw. Buildings and people are something we all see every day, so how body parts and windows look is fairly imprinted on the mind. Due to the majority of yuri taking place in Japan, you're going to come across things that you don't see in real life. In Japan infrastructure, landscape and flora are all different from how the appear in other parts of the world.

Recently this what I came across in the doujin that SHiN had asked me to clean. This doujin features at least three different types of lilies.

>>46853

Botany is not something I have a particularly in depth knowledge of. Knowing that this in a Aikatsu doujin and Aikatsu takes place in Japan, an image search for Japanese lilies is the best place to start. Once you've identified the name of the lily, image search the name again to find more images of it.

First lily

This is either Lycoris sprengeri or Lycoris longituba

Second, third and fifth lily

This looks like the lily that is very often depicted in yuri which is of course Lycoris squamigera

Fourth lily

Initially I thought this might have been a star gazer lily. Upon further research I discovered that these aren't native

to Japan, being a commercial American hybrid. It's actually Lilium auratum.

>>46854

First Lily redrawn.

By looking at images I can now visualise the parts of the lily that need to be redrawn. Any time I'm unsure of how to proceed, I can refer back to the images that I have found. The result isn't perfect but it's above average.

>>46905

Forgot image

Step 02

Look at images of the lily. Using this as a visualisation aid, roughly redraw what needs to be done using the paintbrush (or as is the case with this doujin, the pencil).

Step 03

Replace the roughly drawn lines a line at a time with the pen tool.

For this petal edge I am going to be using the pen tool in shape mode. This is a quicker for line work.

When deciding the width of the line, look at the other lines drawn by the artist. For example, I'm going to use 2px for this line. By default the width of the stroke will be in points (pt), to change it to pixels (px) simply delete the pt at the end of the number and type px.

As a shape is a vector, it will need to be rasterised to be altered, either with filters or brush tools. Right click on the shape layer and select Rasterize layer. The make the line like the ones drawn by the artist, I need to apply a Sharpen More filter to harden the lines.

Finish the layer by using the eraser to shape the line more like lines drawn by the artist.

There are times where it will be better to use the pen tool in path mode. This is usually when you need to draw something other than a line. It's entirely possible to do with the pen tool in shape mode using the fill option. Using the pen tool in path mode is quicker but it requires more practise, as it's more difficult to step back. Once the path has been converted to a selection, there is no way of going make adjustments to the points or beziers. If you're less confidant, keep using the shape mode. I will cover how to use the pen tool in much more depth when I have a more appropriate doujin to use as a demonstration

Look at the rough layer redrawn with the paintbrush. Use that as guide as to where to place the points.

Split the points and manipulate the beziers to create the desired shape.

In the top menu bar click make selection and use the default options in the window that apperas

Fill the selected area by using the paint bucket (hotkey g)

Then proceed as you would if using the pen tool in shape mode. Using the pen tool in path mode skips the rasterisation step.

>>46930

How it looks when finished

>>46932

Dust and Scratches on Raws

I've been cleaning some raws for the admin of fancomics.org and something that keeps reappearing is dust and scratches, mainly on the covers of doujin. It's not possible to eradicate this completely but is possible to minimize the problem.

>>47039

How the above cover looks like when cleaned. Notice that I have not only removed the dust and scratches, I've also correctly colour balanced the image and restored the dynamic range.

>>47040

Scratches

I've already given an explanation of what causes scratches to appear on scans. This is what I posted when I encountered the problem with a secondhand Noir doujin

>This is taking me longer than I expected. The Noir doujin is from 2002 and the cover is scratched to fuck. The scratches are not visible to the naked eye but will show up when scanning. This is because glossy cardboard is coated with the thin film of clay and transparent adhesive. When the light source from the scanner shines light onto the cover, the light is diffused through the top layer of clay and middle layer of adhesive. Any scratches or pits will interfere with the even distribution of light and this will be reflected back into scanner sensor.

While it's impossible to eradicate the problem of scratches, there's no need to make it worse than it already is. Simply take care of the doujin you've bought. Make sure that your shipping proxy adequately packages doujin during transit. When they arrive keep them in their plastic cover for as long as possible. Store them in folders or box files. When debinding, use a rubber craft mat, as this will stop the doujin getting scratched on your desk surface and it will protect your desk from heat.

Using equipment intended for photographers is a fair choice and will be relatively inexpensive.

Equipment

>cotton negative handling gloves

This prevents grease from your hands getting on doujin pages and smearing on the scanner platter

>lens blower

Enables you to blow dust from the platter and doujin pages without touching them

>disposable lens wipes

A wipe pretreated with lens fluid that can be disposed of after use

>lens fluid

A spray bottle of optic cleaning fluid

>goats' hair brush

A soft brush that will not scratch the platter surface. Synthetic versions are available.

>microfiber cloth

A soft cloth that won't scratch the platter surface for use with lens fluid

The major problem with this set of equipment is that none of it is antistatic. What this means is that in use cloths and brushes will pick up dust and transfer it to the scanner. To avoid this, they must be frequently cleaned and eventually replaced.

If you have money to spend, it's possible to use professional grade optic cleaning equipment. This can be prohibitively expensive and items will need to be ordered from specialist outlets.

Equipment

>antistatic arm guards

Stops clothing fibers from contaminating doujin pages and the scanner platter

>antistatic gloves

As with the cotton gloves above. Has the added bonus of not picking up dust.

>compressed air

A quick blast will remove dust from the platter without using recycled air

>antistatic optic fluid

When sprayed onto the scanner platter, it will remove dust and smears as well as prevent dust from settling.

>disposable lint free tissues

Can be used with optic cleaning fluid. Throwing them away afterwards prevents dust from being reintroduced to the scanner each time a cloth is used

>antistatic brushes

These brushes do not attract or store dust

>antistatic plastic cleaner

Helps prevent dust from laying on the outside of the scanner

This is the cleaning equipment that I personally use. I find that the investment has been worth it. I spend less time cleaning my scanners, as well as less time cleaning dust from raws.

How to Level an Image Quickly

Sometimes for what ever reason, as a cleaner you will need to be able produce results quickly. Being with a scanlation team that allows you to put in the time and effort to produce your best work is the ideal but reality is often different. There a groups that value quick releases over exceptional quality. Here I will show you how to level and image in one action and disprove one of the biggest bullshit myths in scanlation.

Step 01

Open the image in Photoshop and duplicate the layer as this will done destructively. Zoom (hotkey z) in to an area of high detail that needs to be preserved.

Step 02

Open the levels panel (hotkey ctrl+l). Notice that the cursor has turned into the dropper. Click around the highlight (white) area until you have found the darkest area, this will typically be caused by paper texture. Exclude obvious spots of dust. Look at the colour panel and notice that the black percentage changes as you click in different areas. Adjust the highlights until the black percentage is at zero.

Step 03

Repeat the same process with the shadow (black) areas take it as close to acceptable detail loss as you can.

Step 04

The image in it's current state is now over leveled. The whites may white and the blacks may be black but you've lost too much midrange and the image will appear too dark. Adjust the midtone to lighten the midrange to an acceptable level. Notice that when you do this, the shadow and highlight will also move. Carefully readjust the shadows and highlights to get the best compromise of detail preservation and removal of dust, paper texture and speckle. Any remaining contaminants can be quickly removed with the dodge and burn tool or the paint brush

>>47420

This method produces quick results that are more than acceptable to the untrained eye. It does however destroy the dynamic range, which can be seen in the second gif showing the histogram. There is also some mild detail loss in the midrange.

How to Clean Difficult Raws

Sometimes there are raws that are awkward regardless of how they were scanned.

For the past three days I've been working on a Mira doujin and the raws still look like shit.

>be in a clean enviroment

>handle the doujin with antistatic gloves

>debind it with a professional heatgun

>clean scanner with antistatic equipment

>scan with professional scanner in 1200dpi

>raws look bad

>spend 3 days cleaning

>raws still look bad

So what to do about it?

The first step is identifying why the raws look bad. For these it seems to be that the glossy coating is not only on the outside cover but on the black and white pages inside. What I'm seeing on the raws as dust is infact pits in this coating and it would also explain why the shadows are so noisy. The coating is reflecting too much light back into the scanner.

My usual processing methods have failed to produce satisfactory results, so it's time to try something different and come up with an alternative method.

>>47905

One possible method is to simply level all the problems out. It's quick and easy. However this destroys detail and dynamic range which defeats the purpose of scanning in 1200dpi. With such high resolution scans there are options available to a cleaner that are not on raws scanned in smaller resolutions.

The main problem here is noise. I've already dealt with three forms of noise in this thread.

>>43706

>speckle in magazine raws

<solution

Careful leveling in multiple steps

>>46804

>jpg artifacts

<solution

Blended anisotropic diffusion

>>47039

>dust and scratches

<solution

Blended surface blur

Second image is what happens when I combine the above methods.

Third image is a gif that shows how much detail and dynamic range has been preserved compared to just leveling away the problems.

Now all that's left for me to do, is to repeat this 82 times.

The Importance of Choosing a Scanner

I now have four scanners sat on my desk. These range from the Canon flagship 900F to an all in one I bought for $30. So exactly how much does the type of scanner matter? Is it really necessary to spend a lot of money on a scanner if all you're going to do is upload porn to the internet?

Let's find out

Scanner Models

Canon PIXMA MG2550S

Sensor: CIS (contact image sensor)

Light source: RGB LED

Cost: 25 GBP

This is cheap all in one printer/scanner for home office use. I've never used it as a scanner before so I'm not sure what to expect. My previous experience with all in one scanners has not been positive.

Canon LiDE 220

Sensor: CIS (contact image sensor)

Light source: RGB LED

Cost: 50 GBP

This is a popular scanner model for raw providers. I was very familiar with it before I bought one for myself. It's a brilliant scanner for it's price. However it's still a CIS and suffers from everything that CIS scanners do.

Plustek OpticBook 3800

Sensor: CCD (charged coupled device)

Light source: Cathode

Cost: 170 GBP

I've owned this one for years. It's by far favorite scanner. Compared to the LiDE and the 9000F, it's incredibly dated. It's also big, noisy and slow. To me the results are worth it but will it stand up to the 9000F?

Canon CanoScan 9000F Mark II

Sensor: CCD (charged coupled device)

Light source: White LED

Cost: 170 GBP

The top scanner in the Canon range. It's capable of scanning everything from negatives to book pages. This will only be the second time I've used it.

>>48583

For the colour tests I will be using pages from the Selector Visual Collection as the deck colours in Wixoss are a good way to show the three channels in RGB colour space.

For the greyscale tests I will using pages from (C92) [Kamotamaza (Kamotama)] Chi no Ka Iro no Ka (Rick G Earth).

All scanners used will be cleaned with compressed air, lint free tissues, Kinetronics Precision Optic Fluid and an antistatic brush. The LiDE, the 9000F and the 3800 are connected to my computer using premium Lindy Chromo USB cables. The Pixma uses the generic cable that was included with it.

For the Canon scanners I will be using the ScanGear software. For the 3800 I will be scanning directly into Photoshop using the WIA protocol.

Canon PIXMA MG2550S Red

Canon PIXMA MG2550S Blue

Canon PIXMA MG2550S Green

Canon PIXMA MG2550S Grey

Canon PIXMA MG2550S Mixed

Canon PIXMA MG2550S

Greyscale scanned in RGB

Greyscale scanned in Greyscale

Greyscale scanned in Black & White

Canon LiDE 220 Red

Canon LiDE 220 Blue

Canon LiDE 220 Green

Canon LiDE 220 Grey

Canon LiDE 220 Mixed

Canon PIXMA MG2550S

Greyscale scanned in RGB

Greyscale scanned in Greyscale

Greyscale scanned in Black & White

Canon CanoScan 9000F Mark II Red

Canon CanoScan 9000F Mark II Blue

Canon CanoScan 9000F Mark II Green

Canon CanoScan 9000F Mark II Mixed

Canon CanoScan 9000F Mark II

Greyscale scanned in RGB

Greyscale scanned in Greyscale

Greyscale scanned in Black & White

Plustek OpticBook 3800 Red

Plustek OpticBook 3800 Blue

Plustek OpticBook 3800 Green

Plustek OpticBook 3800 Grey

Plusttek OpticBook 3800

Greyscale scanned in RGB

Greyscale scanned in Greyscale

Greyscale scanned in Black & White

Here's a section from the full 1200dpi images.

The closest in colour accuracy to the image as it is printed is the 3800. Notice how the difference in light source changes the black level of Yuzuki's hair.

Another section of the same image.

Here look at the halftone in the white circle. The sharpness of the halftone is greatest on the 9000F, the LiDE second, then the 3800, with the Pixma last. The 3800 also fails in this section as the incorrect addition of magenta can clearly be seen.

In this first section of the second image notice how the two CIS scanners have made a distortion in the halftone. The 3800 once again produces the accurate skin tone.

In this section the price difference between the scanners becomes more apparent. Look at the clarity in halftone of the two CCD scanners. The 3800 has the superior white balance, check the colour of the light reflections on the water and see how the surrounding blue bleeds into the white.

On this image there is a slight reversal. Where as before the bright cathode in the 3800 has given it an advantage. On this page that is mostly white, there is now too much light being reflected back into the CCD. The grey in the background in almost completely lost.

The main difference to point out here is the lack of clarity in the Pixma. Ulith appears blurry with halftone distortion. The two professional CCD scanners provide a clear image with the LiDE just behind.

>>48613

Plustek OpticBook 3800 Mixed

How to Correctly Level an Image

I've spoken about "correctly leveling an image" but what does it actually mean?

In scanlation this typically means that black areas are actually black and white areas are actually white. In my opinion this is a naive and overly simplistic approach to leveling. It completely ignores the importance of dynamic range and midtone.

In black and white photography how an image is "leveled" is largely subjective. While it's possible to do it incorrectly, what is correct is down to the photographer. The use of film, aperture, exposure, lens filters and dark room techniques all conflate to give the final image. In scanlation there is far less subjectivity but there will still be variations in what some cleaners find acceptable.

>>48841

Photography is relevant in this instance as how Photoshop handles image leveling is still very much rooted in techniques used in film photography.

Terminology

Highlights, Midtone and Shadow

These terms refer to different parts of the dynamic range.

Highlights - white and light shades of grey

Midtone - the middle range of greys between the highlights and shadows

Shadow - black and dark shades of grey

High key, Mid key and Low key

Images can be described with either of this terms depending on the distribution of dynamic range.

High key - a light image consisting of more highlights than midtone or shadow

Mid key - an image that is neither light nor dark, consisting of more midtone than highlight or shadow

Low key - a dark image consisting of more shadow than highlight or midtone

Dynamic Range

This a term that appears in many fields. It generally means the ratio of difference between point A and point B. In image editing it usually refers to the range of shades in colour space channels. In greyscale, there is one channel. So dynamic range is used to describe the amount of greys between white and black.

Tools and Options for Leveling

In Photoshop there are many different tools, options and panels that all appear to do near enough the same thing. However all them alter the histogram of the image using different algorithms.

Dodge and Burn Tool

It's possible to level an image using nothing but the dodge and burn tool. I wouldn't recommend this as it would be extremely time consuming. The tool is better used to work on problems in specific areas. It's a brush tool and works in a similar way to the paint brush. I will cover the dodge and burn tool in depth at some point. For leveling the most important functions are dodge - highlights and burn - shadows. The term dodge and burn comes from the days when photographs would be developed in a dark room. To dodge is to expose and section while not exposing the rest of the image. To burn is to not expose a section while exposing the rest of the image. This basically means that dodge lightens and burn darkens.

Leveling with the Dodge and Burn Tool

Use dodge-highligts to clean white areas and burn shadows to clean black areas. One of the advantages of using the dodge and burn tool is that it targets certain parts of the dynamic range. Both dodge and burn allow you to set the range you want to either darken or lighten. For example using the brush set to burn-shadow on a black line on a white background will not change the white in any way.

Potential Issues

-it must be used destructively

The changes made using the dodge and burn tool must be applied to the working layer. Unlike the sharpen and blur tool there is no option to apply the effect to all layers. A way round this is by duplicating the working layer, allowing you to delete the duplicate should you not like the changes.

Time

Using a brush to clean the entire image is not an efficient use of time when a similar result can be obtained by using other methods.

Consistency

When using the dodge and burn tool it's highly likely that your brush strokes will be visible. If you look into Sei's eyes, this can clearly be seen.

The Levels Panel

This is by far the most commonly used method of leveling images used by cleaners. It can used destructively and as an adjustment layer. Both have their advantages and disadvantages. The panel itself consists of a slider and graph that is the histogram of the image. The black slider will alter the shadows, the grey slider will alter the midtones and the white slider will alter the highlights.

Leveling with the Levels Panel

When using it destructively, moving the mouse outside of the panel will change the cursor to into the eye dropper tool. Click on a part of the image and the slider in the colour panel (usually on the top left of the screen) will change to display the percentage of black where you clicked. Using this is an incredibly easy to take an area to either white or black. Simply adjust the sliders in the level panel until the slider in the colour panel displays the value you want.

Used as an adjustment layer the levels panel doesn't have the above function but it gives you the advantage of being able to change the degree of leveling at any time, as long as the adjustment layer has not been merged with the working layer.

Potential Issues

The levels panel only offers adjustment of the three broader ranges. This is often not refined enough to preserve detail. Taking the shadows to black and the highlights to white will more than likely remove finer lines in artwork. This is visible in the example image.

Curves Panel

This works in a similar way to the levels panel, only it allows for much finer control for targeting ranges within the shadows, mids and highlights.

Leveling with the Curves Panel

By default the histogram in the curves panel will be reversed to what is shown in the histogram window and the levels window. To change this select Light instead of Pigment. This doesn't alter the way the panel functions, it just avoids confusion. In the histogram segment of the panel, there is a line. Place points on the line that correspond with the range you wish to target. You can place as many points as you need. Use the mouse to slide the points up and down. If you need to adjust the range being targeted, slide the point either to the left or right.

Potential Issues

The adjustments you make must be very precise. The curves panel is quite powerful. The fine targeting of certain ranges can produce unexpected results and fuck up an image in ways you never thought possible. It also takes more time than using the levels panel.

Brightness/Contrast Panel

This a simple slider panel that enables you to adjust the brightness and contrast of an image rather like on a television.

Leveling with the Brightness/Contrast Panel

I don't recommend using it for leveling an image by itself. It's not usually possible to take the shadows to black. It's useful when combined with other methods. Clicking the Use Legacy box will make the panel behave as it does in older versions of Photoshop.

Potential Issues

It's unlikely to take shadows to black and taking highlights to white will probably destroy a lot of detail.

Exposure Panel

This is a very powerful tool. The terminology and it's effects will be familiar to anyone with experience in Photography. Any adjustments you make in this panel will be small.

Leveling with the Exposure Panel

Exposure targets highlights and lighter midtones. Offset targets shadows and darker midtones. Gamma will adjust gamma, obviously. Gamma is best described in simple terms as a form of brightness. It's useful for getting greys out of shadow areas.

Potential Issues

You will need to watch the effect on the midtones very carefully. If adjustments aren't fine enough it can cause unexpected results in ranges that were not intended to be targeted.

Shadows/Highlights Panel

This is very useful for tuning the midtones to perfectly level gradients in the halftone. Exactly to how to use it isn't immediately obvious, it almost seems to work in reverse.

Leveling with the Shadows/Highlights Panel

As a tool to level an image it's not that useful by itself. It's more an addition to more finely target ranges. Increasing the amount of shadow will lighten midtones. Increasing the amount of highlights will darken midtones. The tone slider will adjust the tonal width of what midtones fall within the targeted range. Radius will adjust the radial effect to pixels outside of the targeted range.

Potential Issues

Not very useful by itself. It's focus on midrange can leave highlight and shadow areas under leveled.

Picking up where I left off…

It's finally time to work on my own raws.

The use of a high quality CCD scanner and high resolution raws will make this much easier.

Start by placing all the greyscale pages into a single document. This is useful for consistency when leveling. This doujin has 36 pages and takes around 2GB in scratch.

Before you start leveling, look at each page and decide exactly how to level using the above methods. This doujin is mainly midkey with a few highkey pages. The panels are also full of richly detailed halftone that have been fully preserved by my scanning methods. Any leveling here will be light and I will probably need to use the dodge and burn tool quite a lot. Using my own high resolution raws also gives the me opportunity to take advantage of bicubic resampling.

Step One

Use the levels panel to take the highlights as close to white and the shadow as close to black as you can before you start to see detail loss. It will help to zoom in on an area where there is likely to be detail loss, such as the eyes. Take a mental note of the number in box below the shadow marker, as the number is likely to be similar across the pages. This number is the shades of grey that you are cropping from the dynamic range. Think of it like trying to stretch a piece of fabric over the top of an open box. The more you trim the fabric, the thinner you have to stretch it.

After doing this the image should be darker and have a lot more punch to it.

Step Two

Now it's time to start targeting the remaining noise in the shadow areas. Identify the range you wish to target and open the curves panel. Here I will need to target greys in the darker midrange.

Open up the curves panel and place a point on the histogram in the range that needs to be removed. Notice that moving this point will change the tonal line through the histogram and effect the entire dynamic range. To stop this place a point further towards the highlight area. This will limit the movement of the line and narrow the targeted range. Place as many points as you need.

At this stage the contrast of the halftone gradients should really start to stand out. This is partly due to the removal of noise. It is important during the level and curve steps to not over do it as it will destroy the dynamic range and remove the different greys that should be there. Also notice that I haven't removed all of the noise.

Step Three

The final leveling stage is to finely adjust the appearance of the midtones by using the shadow/highlights panel. Judge for yourself if this is necessary on the raws that you are working on.

On these raws I don't need to make any adjustments on the shadow and highlights as this was done during the level and curve step. All I need to do is make a slight adjustment to the midtone.

With the leveling done, the next thing to do is to resize the raws. Their current size is 4160x6030 which is too large. While I release raws in this size, it's not yet plausible to release a typeset doujin in this size. Online readers will resize the raws to fit the screen and anyone reading a direct download is likely to do the same.

Resizing the raws is likely to loose some detail, especially in the lower midtones. To reduce this, I'm going to use the dodge tool set to highlights on low exposure to lighten some of these details.

To downsize these raws I'm going to use one of the standard algorithms built into Photoshop, bicubic resampling. These days there are many more algorithms available, some of these are probably better than what is available in Photoshop. I don't have the same understanding of how these work and can not predict the final result with as much accuracy, so for now I'll keep using what's provided by Adobe.

Notice that downsizing the raws using this method has removed most of what was left of the noise while still keeping fine detail. This is due to how bibubic reduction works. I'll explain more about this at a later date.

After resizing I'm going to look through the pages and remove any remaining noise with the burn tool. This may take a while as I have intentionally been conservative during the leveling stage.

Before I start removing the text and getting the raws ready for typesetting, I am going to use a blended anisotropic diffusion filter. This will smooth out any visible brush marks from using the burn tool. It will also have the added effect of enhancing the already rich gradients.

Select the layer of the first page, apply the anisotropic diffusion filter (Filter > Stylize > Diffuse). In the window there will be four options. Select anisotropic. This filter is fairly heavy on the CPU, so if you have a slower processor this may take a while. The effects of this filter in the preview window may not seem desirable at all, which I why I am next going to blend it. There are two ways of layer blending in Photoshop.

The first method is quicker but is also destructive.

Edit > Fade Diffuse…

Using the slider to set the level of fade.

The second method is non-destructive and probably better to use if you are unsure of the exact amount by which the layer needs to be blended.

Duplicate the working layer and apply the filter. Then by using the opacity slider at the top of the layers window blend the layer to the needed level of fade. Once you are happy with the result these two layers can be merged.

The effect of the blended anisotropic diffusion filter will not be immediately obvious.

Zoom in on an area of gradient. Notice that the blacks are no longer black but dark grey and white areas are now a very light grey. The areas of pure white and black are unaffected. It's a neat visual trick that adds tonal depth to gradients.

This method does take a lot more time than just using the leveling panel as most cleaners do. However I think the results speak for themselves.

With the leveling stage done, it's now time to prepare the raws for typesetting. Before removing the Japanese text, there is a decision that needs to be made, are all the sfx going to be cleaned off and redrawn? In a group setting it would be the head of the group that would make that choice, depending on the skill of his staff and the time available. Another thing to consider is how the sfx are used by the artist. Would cleaning off the sfx remove the integrity of the panels and would the typesetter be able to copy the look of the Japanese text?

For this doujin I am going to do something I rarely do and that is not remove the sfx. The main reasons for this is because in this doujin the Japanese text is an innate part of the artwork and I will not be able to replicate the exact look in Photoshop. The 8/u/ TL has translated the sfx, so I will be using the typeset technique shown here >>43589.

To get the raws ready for typesetting, I am copy pasting each page back into it's own document. The page is pasted three times. The top and bottom layer are unedited. The middle layer will be edited to remove text. It's possible to use scripts to do this but I prefer to just use the paintbrush.

Debinding Doujin

This is an essential skill for any raw provider.

It's an unavoidable reality that you need to debind whatever you intend to scan. I tried to avoid this by buying a specialist book scanner, which is how I ended up with the 3800. While I don't regret spending the money on it, as it turned out to be one of the best scanners I've ever seen used in scanlation, it failed at the original aim I bought it for. If anyone remembers the Wixoss threads back on 4/a/, I started scanning the art books using the book edge of the 3800. It wasn't successful as the format of Japanese books is different to that of western language books. Japanese books and doujin typically use B paper sizes and do not have a sufficient gutter.

What is debinding?

Debinding is the removal of pages from their binding in the spine of a book or doujin. There are different ways that books can be bound.

>Saddle stitching

This consists of leaves that have been stapled. It's by far the easiest type to debind.

>Perfect binding

The most common form of binding that you will come across. Individual cut pages are glued to the spine of the book.

>Sewn binding

I never seen this method of binding used on doujin. However it can be found in officially published books. Groups of pages or leaves known as sections are stitched together and then glued to the inside of the spine.

Equipment

This will follow what is required for the modified hair dryer method that I use.

Professional Variable Temperature Heat Gun

I prefer to use a heat gun over a hair dryer for several reasons.

Domestic hair dryers do not get very hot. 40°C is around the maximum. A heat gun will go up to 600°C. While such high temperatures are not suitable for debinding, the ability to go up to 120°C makes things much faster.

Domestic hair dyers do not have robust motors or heating elements. If used for an extended period of time, they will over heat. With a professional grade heat gun I don't need to worry about this. I can spend hours debinding doujin or thick official books and manga tankubons without stopping.

A hair dryer will spread hot air over a wider area. One of the reasons a heat gun is faster is because the air flow is concentrated into a narrow direction in front of the nozzle. This means that more heat is directed into the spine where it's needed.

The only negative I can think of when comparing a heat gun with a hair dryer for this purpose, is that a heat gun is quite a bit heavier than a hair dryer. If you don't have the strength in your wrists and forearms, you may struggle to use it for longer periods of time, which negates one of the advantages.

If you are considering buying a heat gun for debinding, it's important to invest in a model that provides true variable temperature control. Cheaper heat guns usually only provide two heat settings.

This is the exact model that I use

https://products.dewalt.co.uk/powertools/productdetails/catno/D26414/

Gloves

I wear gloves for two reasons. Firstly, using a heat gun means I can use a temperature that is high enough to burn my hands. Secondly dust management is an important part of the debinding and scanning process as I explained here >>47041. I wear a pair of Rollei antistatic lens cleaning gloves because the PU coating on the finger tips makes it easier to grip the heat gun and gently pull on doujin pages. On my left hand I also wear a framer's fingerless glove underneath the antistatic glove which provides protection from heat.

Anti-static Sleevelets

This is to stop fibers from my clothing contaminating the area where I am debinding the doujin.

Craft Mat

There are four reasons for using a craft mat. It's possible to scratch the doujin cover while moving it around your desk. The surface of the craft mat is softer and will generally be more responsive to being cleaned with anti-static dust spray. A craft mat will conduct heat back into the spine of the doujin after it has been heated. It will also protect you desk from heat.

Tweezers

I keep two pairs of tweezers for debinding. One pair of plastic Knipex anti-static tweezers and a generic pair of nylon tipped anti-static tweezers. Useful for pulling staples out of saddle stitching and for picking hot glue from doujin pages.

Scalpel

This is needed to cut the stitching in sewn binding. I've tried this with various types of knives, such as Stanley knives and box cutters. I've found that using the scalpel included in a craft knife set is the best solution. This is sharp while being long and light enough to balance in my fingers.

>>50703

https://archive.is/iEln9

Best I can do. I'm sure some other folks have the autism to compile the information here in an easy to share .PDF

>>50677

>>50703

>>50808

When I've finished covering everything I want to cover, I'll write up the information here in an actual guide with links to the resources that I've compiled. This is thread is a bit of mess as I'm covering topics as they arise rather than in any sort of ordered fashion.

It may not look like I'm doing much at the moment but I am sorting through the font collection and upgrading my workstation, as well as scanning and debinding some colour doujin.

Here I'm processing the raws from one of the Wixoss magazines. The paper used is highly textured with the fibers being visible even before being scanned.

The first stage is to colour balance the image in steps. Notice that not only am I compensating for the coating, I am over saturating the image.

The next part is a multi step process of using filters. Here it helps to have OpenCL capable GPU as this will enable Photoshop to utilise the GPU as well as the CPU.

How exactly to use filters will depend on the scanner. This raws were scanned using the 9000F. The CCD in this scanner picks up every detail so I am going to have to carefully select the correct method of filtering to get satisfactory results.

The first filter I will be using is an anisotropic diffusion filter. You will notice that I use this a lot. It's a filter that is poorly understood through out the rest of the scanlation community. However I find it as useful today working on raws as I did fifteen years ago when I was working on scanned photographic prints. It's been a feature of Photoshop since at least 1995.

Anisotropic diffusion is a complex filter and how it works isn't obvious.

All filters in Photoshop work by applying an algorithm to the data in the image. In anisotropic diffusion the algorithm uses partial differential equation to apply Fick's laws of diffusion to the data in the image. What this essentially means is that the pixels in the raster are moved closer to near by pixels that are similar.

Duplicate the base layer two times. Apply the anisotropic diffusion filter to the top layer. Then adjust the opacity slider in the top right corner to blend the filtered layer with the layer underneath. The aim here is to achieve slight removal of the cellulose fibers without blurring the image.

It may seem that I am stopping and starting projects at the moment and that is exactly what I'm doing. I've completely failed in my original aim of getting the C93 and C94 order released before C95. This is for two reasons. The first is health related, the second is hardware related. As this current round of upgrades to my workstation will be targeted at performance gains in Photoshop, this seems like a good opportunity to explain how Photoshop uses hardware and how to build a machine that takes Photoshop into consideration.

>>51916

When considering performance in Photoshop it may be useful to understand the original hardware it was released for back in the late 80s.

The hardware market in the late 80s and early 90s was very different to what it is now, it was a time of invention rather than stagnation. There was no one dominant computer architecture for all purposes as there is now with x86-64. The hardware company that had come to dominate the professional design sector was Apple. They had done this with the introduction of the Macintosh, an all in one that came with a GUI OS capable of running WYSIWYG (What You See Is What You Get) software. How this came to be is rather complicated and would later have consequences for Microsoft as well as Apple.

The first computer with a functional GUI was the Alto. This was developed by Xerox PARC in the 70s. The Xerox executives never saw any commercial use for this new technology and it was confined to Xerox's own offices. Apple released the Apple II in 1977. This is probably the first release of what is now recognised as a personal computer. Steve Jobs intention was to create a machine that would find use in the homes of the average consumer. While the Apple II was a commercial success, Jobs saw the CLI as a obstacle to greater adoption. During 1979 Jobs arranged for Apple engineers to visit Xerox PARC in exchange for Apple shares. Working at Xerox PARC were two people who would later become the founders of Adobe.

In 1981 Xerox finally commercialised the Alto as the Star Workstation. Despite it's obvious advantages in the design and publishing industries the Star Workstation was a failure, mainly due to it's astronomical cost. Two years later Apple would release the Lisa, the first personal computer with a GUI OS. This included many of the features originally developed by Xerox, such as windows, the folder directory system and the use of a mouse. The Lisa, like the Star Workstation before it, it was not a commercial success. Steve Jobs had been removed from the Lisa project a year earlier and instead began working on the Macintosh Project. This is also when he begins Apple's involvement with Adobe, initially offering to purchase the company from it's founders. They refused but did allow Apple to own a portion of the company's shares.

After lessons hard learned with the release of the Lisa, Apple launched the Macintosh with was what is now an historic advertising campaign. The 1984 Super Bowl commercial featured the Macintosh, shown as a woman in bright sports clothing, breaking through the grey monotony of Apple's then main rival, IBM. Bundled with the Macintosh was the System 3.0 OS, along with a small number of programs including MacWrite and MacDraw. The Macintosh was hugely successful, yet it still didn't met Apple's expectations. The reason for this was identified as being the lack of third party software support. The Macintosh didn't have a stand out application that justified the cost of purchase. Software companies did not know how to write software for a GUI and didn't think it was worth the investment to train staff.

What saved the Macintosh project from failure was Apple's relationship with Adobe. 1985 saw the release of the first LaserWriter printer. In a situation that mirrored what happened with Xerox, Steve Jobs observed the development of laser printers at Hewlett Packard and then took these ideas to Apple. Since it's founding Adobe had been working on the commercialisation of PostScript, a vector design language used in printing and WYSIWYG publishing software. This was a further progression of technology from Xerox PARC. Jobs negotiated with Adobe for the LaserWriter to become the first PostScript enabled printer. This new and superior technology convinced software companies to invest in writing software for the Macintosh.

>>52133

If anyone wants the Hennojin copy of the Xenogears Doujin.

>>51919

I would actually change it again, if you can. It looks like Yuri-ism is gearing up to release the linked book in the next few days. Pic related. If you are able, I would change to one of the two linked below, which haven't been done yet, to my knowledge.

First Kiss:

https://exhentai.org/g/1166896/f77ada33f5/

Catch Love 2 (21 pgs):

https://exhentai.org/s/861469f0d2/750067-67

>>52133

>>52134

Oh fug, what a strange turn of events.

You should probably try visiting the board's irc from time to time and ask what we're currently working on or plan on doing to prevent something like this from happening again…

Illustrator was released by Adobe in 1987. The concept of a computer drawing program was nothing new, Microsoft had released the initial version of Paint in 1985. What made Illustrator revolutionary was the integration of PostScript and the introduction of the pen tool. It was also the first implementation of Adobe's scratch disk system, which is a persistent feature of Adobe programs to this day. The original Macintosh was equipped with only 128k of non-upgradable RAM. To get around this limitation, document changes were written to disk. Virtual memory was not a feature of Mac OS and wouldn't be until System 7.

In the same year Apple would address some of the short comings of the original Macintosh by releasing the Macintosh II. Unlike the first iteration, the Macintosh II was targeted at the professional desktop publishing industry. It was able to display 8-bit colour graphics with an upgradable video card and could take up to 8MB of RAM. This would forever create a split in Apple's product catalogue. Low cost and often gimped machines would be made for the home and education market, while their best and most expensive hardware was aimed at professionals.

Photoshop would appear a year later. Contrary to popular belief Photoshop was not initially developed by Adobe. It was created by two brothers using a Macintosh SE/30 and was first commercialised as bundled scanner software. It was then licensed by Adobe. After further develpoment Photoshop 1.0 was released in 1989. Like Illustrator, it wasn't that the concept of a raster editor that made Photoshop unique, it was it's user interface and unique tool set.

Demonstration of the first Adobe version of Photoshop by it's creator

>>51919

>>52236

Yuri-ism just released it. Sorry man.

https://yuri-ism.com/2019/01/04/release-576-back-to-the-office/

>>52563

>Do obscure stuff

Poached

>Do stuff that's been untouched for awhile

Poached

>Het Tank that I had my eye on because it looked decent and hasn't been touched in 3 years.

Poached and fully translated on Christmas.

Good news is that my two main ones weren't touched. Yet. They should be uploaded up to Sadpanda hopefully soon.

I've got 2 more small ones being worked on but they're bloody obscure that I don't think anyone else has the Raws for them as I had to find them myself. I've been wrong however and if Yuri-ism release it before I do, I don't know what I'm gonna do. There's also another small one from the Yuri Nishin that I wanted.

Also anyone got any recommendations for translations? I'm out of personal stuff and looking for something interesting.

(C93) [Nedaore (Ayane)] Nihaku Mikka wo Present (Puella Magi Madoka Magica) [English]

&

(C94) [Nedaore (Ayane)] Itsumo no Tomoe-san-chi (Puella Magi Madoka Magica) [English]

https://a.uchi.moe/aanfdr.zip

https://a.uchi.moe/pqwqfd.zip

[Ruri Hozuki, Sakura Shio] Tsunderella (Yurimm) [English]

>>53473

[Ruri Hozuki, Sakura Shio] Hanzel, Gretel And The Witch (Yurimm) [English]

>>53474

[Fukuroumori] Kitsune no Yomeiri (2D Comic Magazine Yuri Ninshin Vol. 3) [English] [Digital]

>>53475

>>53476

>>53477

>>53478

https://mega.nz/#!d1cgyCCb!mK_gJ2X2y9ANi2mtTOPXZnGGIipI5mXeXgUaifIdeUo

https://mega.nz/#!p41kXaLI!9RWiTQefFlEiGaIas6J43UZelsR4Ry0vgGqSMhuOTZw

https://mega.nz/#!c4VmzY5a!Fnztey9Xa9v3LH6CO6xukl1N3a6UAipx9IHTIvJdTO4

>>54062

I mean in regards to typesetting. And not just the simple wordless kind of moans, which can often be left in the Japanese without being too intrusive if they cause too much difficulty, but also more coherent things like pic related not from /u/ material.

I can't seem to find the right typefaces or distortions, and it ends up looking like something I pasted over top of the page. Even if that's exactly what it is, it shouldn't conspicuously look like it.