http://img385.imageshack.us/img385/4206/958206pcgames0930shiro0ty5.jpg

i managed to get my hands on the artbook in the photo, inside the book there's a symphonic rain photo (the original holding hands boxart) i'd like to print out in an enlarged size, if possible medium poster size. how do i do that?

help me, and i'll contribute my scan to the symphonic section of this site, which really needed the boxart in my opinion as the poster collection here is so complete, with only 1 missing. help me and i'mm get it scanned and give it to the site owner. my current place of residence is canada.

We have plenty of people here who can help. Looking forward to seeing the scan!

*buy a scanner
*put artbook on scanner
*scan
*post result
*post scanner settings

would be the very first step
of course best scan result can only be achieved if you debind the artbook somehow
if the pic is smaller than the artbook and fits without bending, youre even luckier

other than that, wheres the prob?

Well, I can tell you how I scan. I have a Epson Perfection 4490 photo scanner, I use it in professional mode which gives me control over all it's functions. I scan a @ 600dpi minimum and the file output is jpeg but some prefer png for better image correction in photoshop or other image editing program. I use the histogram adjustment and tone correction functions to adjust light, dark, contrast gamma etc etc. Histogram has input functions which are eyedroppers in my program. For dark, grey and light tones of the image. For dark I usually pick the darkest area I can find usually near the eyes, the pupil or eyebrows but sometimes any dark area will do. It's just a matter of trial and error to find the right look. Light is the same method, but finding the lightest area and normally I will pick either the highlight in the eye or the white of the eye. But again any white spot might do and again it's trial and error. Tone correction has linear, lighten, darken, flat contrast, high contrast, open shadow and a few of my own settings. Again it's a matter of trial and error to see what you like and don't like. Then it's a matter of adjusting your scanning area to minimize the amount that may have to be cropped out later. I also have unsharp mask and descreening filters that attempt to minimize the moire effect from scanners. Some folks don't use the descreening filter and let their image editing program deal with it. My suggestion to you is scan it, post it, and let the photoshop pros do the editing.

I also should mention that where jpeg might put out a 20mb file, it's png equivalent will be a lot larger, maybe up to 3x as much depending on the image being scanned. And you might be better off to just scan in your scanners basic mode at least until you learn what it can and can't do then tackle the more in-depth stuff. And there are tutorials out there and I do believe admin2 has one, or, that might be a debinding tutorial ...

i'd like to print out in an enlarged size, if possible medium poster size. how do i do that?

Difficult to do without losing quality.

Though it's stingy of me to say this. I would prefer and recommend you save it as PNG to preserve quality since jpeg have lack the quality that PNG has.

From what I know about printing, a smaller file will not turn out as good as a larger file will. That's using a jpeg file. Other formats such as RAW and TIFF will probably turn out better printed matter over jpeg but are very large in file (in mb) size. It's as aoie says, jpeg is a compression format and gets rid of certain information to get the smaller file size. So if you want to print a jpeg file it'll need to be very large file size as in 4000 X 6000 as opposed to 800 x 600. Also, the compression ration of the jpeg file will play a role too. Professional printing places might be able to do more with a file than the average Joe can. And they might have stipulations about the picture you want printed.

Here's a link http://www.scantips.com/basics09.html
And Wikipedia's take http://en.wikipedia.org/wiki/Graphics_file_format
And you can check this out http://www.allgraphicdesign.com/scanning.html

For my scan, I go with 1200dpi in PNG, no descreening ( my scanner really doesn't seem to descreen anyway... ) . Then I use Noise Ninja to remove most of the artifact. After, either it's a small Gaussian Blur ( 1px) or a Median Noise ( up to 3px ) if more artefact and finally some sharpening.

I don't do any color manipulation except converting color profile. Some software just doesn't support PNG's gamma correction.

if your scanning and plan to work with the scans, TIFF is your ONLY friend, DO NOT save it as jpeg ABOSLUTLY NEVER, and PNG is not to prefer either, Why? Because all those format you lose quite a lot of important information while you work and edit your scans, TIFF is made for scanning pretty much. PNG and JPG work fine as an end result, but NEVER use them while you work with the pictures! NEVER.

Also don't go complete overkill on to high DPI, most cases on printed books you will hit the max around 300-600 DPI where the rez just become so high that you start to cell the ink cells and at this point is kind of ueless to scan higher, and the quality if the picture kind of get gross by smoting it out to much after here on to. Some books you can scan higher but it depence on the quality of there printing.

But yeah this is just my preference, I'm mostly just been working with black/white ( aka Doujinshis ) rater the color art, over 900+ to be exact )

the whitepaper book above could be color it could also be black and white as that circle seems to release both types quite often. Love there art thought :P

If it's black and white I can write a ton of more advice on how to clear up pictures, make them look like they where original ink drawings while not destroy to much of the shadows and os fort.

Here's the difference between jpeg and TIFF. I scanned an image of Yoko from Gurren Lagann that is 5x3.5 @ 600dpi and 1200dpi. Both images came out at 2987x2074 @600dpi. The jpeg topped out at 3.45mb while the TIFF topped out at 17.8mb. And at 1200dpi the images came out at 5974x4148 but the jpeg was 34.1mb and the TIFF file was 70.9mb. My scanner doesn't output to PNG so I can't show the difference there. My old one did but not this one. So output is dependant on how much drive space you have. If you have lots of hard drive space, or don't plan on keeping them for very long, by all means save them as TIFF's. Or if you plan on doing image manipulation TIFF would be the way to go. Then again, it's all up to the person doing the scanning ...

MugiMugi said:
if your scanning and plan to work with the scans, TIFF is your ONLY friend, DO NOT save it as jpeg ABOSLUTLY NEVER, and PNG is not to prefer either, Why? Because all those format you lose quite a lot of important information while you work and edit your scans, TIFF is made for scanning pretty much. PNG and JPG work fine as an end result, but NEVER use them while you work with the pictures! NEVER.

It would be much more helpful to say what information PNG is losing, then to use a lot of caps.

Also don't go complete overkill on to high DPI, most cases on printed books you will hit the max around 300-600 DPI where the rez just become so high that you start to cell the ink cells and at this point is kind of ueless to scan higher, and the quality if the picture kind of get gross by smoting it out to much after here on to. Some books you can scan higher but it depence on the quality of there printing.

Well, you're not seeing "ink cells", you're seeing the screening patterns--and that's what you want to get.

If the screening/dithering/halftoning resolution of the media is at 300 or 600 DPI, then it's probably useful to scan at 600 or 1200 DPI, so the screening patterns are as clear as possible--so they can be removed cleanly with your favorite descreening tool. If the screening is blurry or has a moire pattern, it's probably not going to descreen as cleanly.

Of course, since the actual data being represented by those dots is lower than the resolution you scanned at, descreening is restoring the image to that original data (approximately)--so there's no point in leaving it at that excessively high resolution. If the source is 300 DPI, and you scanned and descreened at 1200, you have something like 300 DPI worth of data, so scale it back down a bit.

I'm also interested in the information lost when compressing to PNG over TIFF, considering PNG is lossless, and TIFF out of a scanner also. ICC Profile are embedded, 24bit depth color so for quality and color, it's the same as TIFF.

To edit image also it's perfectly fine, in my opinion, to load the PNG in Photoshop then save in PSD for all the edition you might add ( Vector, Layer )

This appears to have a lot of information about PNG. http://www.vias.org/pngguide/wrapnt_gamma_correction_and_precision_color120.html

PNG is not to prefer either, Why? Because all those format you lose quite a lot of important information while you work and edit your scans

. Facepalm.psd , you know...PNG is a lossless format.

The information you lose for saving it in PNG is mainly information about your scanner and some color information. Pixel per pixel is nothing lost saving it to PNG. It's some minor information but it helps fixing color problems.

But most mayor difference between TIFF and PNG while working on it, is that it takes awfull long time to save/load PNG files due the compression it use, hence working with PNG is quite slow.

While working with pics in general 100 70MB pics today is nothing and easy for whoever to store on almost any HD right now to work with, it's not like you keep the TIFF pics when your done with all editing that you required.

I looked at the information stored in both my TIFF and PNG scan:

TIFF: http://img161.imageshack.us/img161/2261/screenshotosx22zf8.jpg
PNG:
http://img246.imageshack.us/img246/4199/screenshotosx21ch2.jpg

And so as to understand the information, the specification of the TIFF 6.0 format:
http://partners.adobe.com/public/developer/en/tiff/TIFF6.pdf

From what I understand of the spec, most addition in the TIFF say that it's a 24bit RGB picture. Nothing much that is better for resolving color problem. Other than that, both store the ICC Profile which is the real deal for color.

For the time taken, yep it's longer but I'm a bit limited by storage space ( laptop ) since I need to leave some free space so as to prevent free-space fragmentation ( bad for my swap )

The initial saved image always should be either TIFF, BMP or other high quality format to preserve all information saved when the initial scanning starts. But all raws should be saved in a date preserving format like PNG, instead of jpeg, since Moe don't support that many file format. (I acutally tried searching for it on the site, but Couldn't find what moe supported >.<)

I know what I was scanning my drawings for editing, It sorta almost didn't matter cause it was simply line arts (since I was gonna reconstruct it on photoshop anyway), so a jpeg format could've been used, but I still saved it as a BMP to save quality.

I'd save as PSD, since I'd probably scan directly into PS, and PSD is fast and stores all information Photoshop receives. Depends on your tools.

PNG (and PSD and TIFF) can save as 16-bit. I've figured that would be pretty useful for scanning, though nobody seems to use it. It'd make the output quality better if you do any adjustments (levels, curves, brightness/contrast), I think.

You mean the 16 bit mode in Photoshop? This would mean that each color channel would be coded on 16bit ( 48bit in total ), but LCD screen and scanner are mostly 8 bit per channel ( 24bit in total ), or even 6 bit + dithering for the quick response time display. I'm not sure it's worth it..

Merun said:
You mean the 16 bit mode in Photoshop? This would mean that each color channel would be coded on 16bit ( 48bit in total ), but LCD screen and scanner are mostly 8 bit per channel ( 24bit in total ), or even 6 bit + dithering for the quick response time display. I'm not sure it's worth it..

The point isn't to be able to display it. If you have 8-bit input data, and you do filters on it (like levels and brightness), those filters only get 8-bit data and they output 8-bit data, which results in quantization.

You can see this easily: load a good 8-bit image, hit ^L, press ^1 to only show the red channel. Note that the histogram is smooth. Now do auto-levels (or brightness or any of those filters), and hit ^L, ^1 again. Notice the holes in the graph. That means there are specific, isolated color components which never occur in the image. The filter never output them because none of the 8-bit input values mapped to that 8-bit output value. This is just the most obvious side-effect of quantization; in short, the effect is that it's outputting a 7-bit image.

If you scan and work with the image as 16-bit, the filters have a lot more data to work with. Do the above on a 16-bit image (a real one, scanned as 16-bit; you can't just convert an 8-bit image!), and there'll be no holes in the graph. (Well, there are, but they're on the order of 16-bit data.) Then, when you're all done working with it, convert the image to 8-bit before exporting to the final PNG or JPEG.

This is the same principle as recording audio in 24-bit/96khz for audio processing before resampling to 16-bit/44/48khz, which also reduces quantization error.

Of course, this depends on a scanner that can give you 16-bit data. Many scanners claim that they can do this; search NewEgg for 48-bit/96bit scanners. Of course, some crappier brands may be lying, eg. if they're referring to the bit depth of the "interpolated resolution" (look mom, 19200x19200 DPI!), which is clever marketing (read: lies). Use your judgement when evaluating hardware, as always.

I see your point. Thanks for the information. My scanner can't output 16 bit data so I can't test this however ( it wasn't really intended to scan high quality stuff at first ).

i realized my old clam shell scanner is terrible, i'm just going to see if there are any services that can provide good scanning for my artbook.

please be waiting! very sorry! T_T

And thus I got a scanner. Espon V200, after reading Wraith's rather lengthy post. Using those to my advantage, they did the trick but it did not remove those morie patterns, and after playing around with the tricks I knew in photoshop, I did remove them but the color were less saturated then I added a saturation layer to improve color. Added a brightness/contrast layer too.

Initial image with little filtering - http://aoieemesai.files.wordpress.com/2008/07/otona-no-moeoh-knee-socks-gallery013.png

After filtering and lots of play time with PS - http://aoieemesai.files.wordpress.com/2008/07/otona-no-moeoh-knee-socks-gallery019-copy.png

I'd just like ya to compare them, see if the 2nd one is a good improvement.

aoie_emesai, you really killed the shadow detail on that filtered image. It looks a lot worse. You should stay away from the brightness and contrast controls in Photoshop because they will almost always really mess up your image. Deal with levels and curves instead.

Below are a couple versions of that first image which I did a quick filtering job on:

Sticking as close to the original brightness as possible (to maintain a bit of the washed out from strong sunlight look) with better contrast and denoised:
http://cyberbeing.extra.hu/test1.png

A little darker than the original (removing the washed out look) with better contrast and denoised:
http://cyberbeing.extra.hu/test2.png

Thanks for the advice, Cyberbeing. Testing it out more on photoshop again.

I after more testing on photoshop to remove more morei, I eventually got to the darker parts of the scan which seems really difficult to remove/ clean them out.

How do ya guys clean it out from really dark spots?

This is a cropped version of pretty much where most of the dark areas are located at, but you can look at the sample and see what I meant.

Everything else looks, okay...

Linky - http://aoieemesai.files.wordpress.com/2008/08/girls-of-summer001-copy3.png

The lazy way is to just use Surface Blur (and optionally History Brush afterward to repair any detail loss).

That scan is already very smoothed/blurred already so it may not be a bad idea to apply Surface Blur earlier in the de-noising process as it doesn't get along very well with ultra-soft blurry edges. If you have decent edges Surface Blur can usually make them more refined (sharper), but if they are blurry and faint it will do the opposite and erase them so you have to be careful.

Lazy example with Photoshop Surface Blur:
http://cyberbeing.extra.hu/girls-of-summer001-copy4.png

Lazy example with NeatImage Plug-in (Home+ $50 or Pro+ $75):
http://cyberbeing.extra.hu/girls-of-summer001-copy5.png

Lazy example with NoiseNinja Plug-in (Home Bundle $45 or Pro Bundle $80):
http://cyberbeing.extra.hu/girls-of-summer001-copy6.png

Pah, just use the GIMP with GREYCstoration. Works for me.

Radioactive said:
Pah, just use the GIMP with GREYCstoration. Works for me.

I've tried messing around with that briefly a few weeks ago but it seems insanely slow with no real documentation. What type of settings do you usually use/tweak?

Documentation is seriously lacking.

A quick fix is uncheck Approximation, and select Patch-based instead. Up the number of iterations if you want to trade image quality for better smoothing.

You need to watch out for the 'halo' effect on white areas, especially if you leave Approximation checked.

I don't know if anyone else uses it.