Help talk:Scanning

First comments

A very good start. Some additional points:

• I usually scan to PNG or TIFF, and if that becomes too large, I convert to JPG afterwards. PNGs can also often be (losslessly) shrunk quite a bit using optipng.
• Use the "remove moiré" function of the scanner, if it has one.
• I've seen (elsewhere) the advice that scanning at an angle may reduce scanning artefacts. Never tried it myself, though.
• The black sheet as a background must be mentioned. I should've thought that trick was so old and well-known (I've used it long ago, when all we had were Xerox copiers...) that it was not necessary to mention it, but the Darwin DR proved me wrong on that one.
• Maybe add a few external links such as scantips or halftone scanning and others?

I'm sure this page will evolve into a truly useful guide. Lupo 15:02, 4 March 2008 (UTC)

I've added all but the second and third in - the second because I don't think it's a particularly common software feature (and, in any case, software varies so much that it's hard to say much about what they might or might not have in their scanning software); And the third because rotating the image after scanning at an angle might well produce as many problems as it avoids.

I still have to see a scanner that doesn't have a "remove moiré", "descreen" or similar function. Maybe that just speaks of my limited experience. :-) Lupo 09:39, 5 March 2008 (UTC)

P.S.: See [1] or [2]... both state that most scanner software includes such a feature. Lupo 09:51, 5 March 2008 (UTC)

Oh, descreen. Sorry, I don't actually do much work with half-toned images, and that's what those functions are mainly for, so I largely can ignore them =). Adam Cuerden 12:07, 5 March 2008 (UTC)

Please consider removing the two subordinate phrases in the first sentence. "if you know what you are doing," is unnecessary and condescending. "if you have an interest in history" likewise. A potential contributor wouldn't be here without interest and interests can extend beyond history. Regards, ... PeterEasthope (talk) 15:48, 3 April 2022 (UTC)

Resolution

This is an awesome start. I look forward to seeing this expand.

About resolution. Can you say if beyond a certain point it's a waste of time and pixels? My scanner scans up to 1200dpi, but the files are so large that my image editing software struggles to cope (eg with cropping or rotating). Even with 600dpi it struggles. (This was using Irfan Image Viewer.) So I gave up and went 400dpi and it was still plenty huge of course. I just wonder how much difference there is going from 400 to 600 to 1200. pfctdayelise (说什么?) 05:57, 5 March 2008 (UTC)

It's occasionally useful for very small, and/or very detailed works of art. I think that Image:William Hogarth - Gin Lane.jpg was at around 600 dpi or so. But those resolutions are, in fact, largely intended for the scanning of film, slides, and microfilm. Adam Cuerden 11:47, 5 March 2008 (UTC)

Perhaps the suggested minimum scan resolution ought to be changed from 300 to 600 dpi: "Grain aliasing is observable in the 300 dpi scan because the Epson apparently scans every 8th position, without anti-aliasing... Grain aliasing was less obvious in 600 and 1200 dpi scans." http://www.normankoren.com/Tutorials/MTF8.html --Traal (talk) 19:41, 15 July 2019 (UTC)

Formats - DjVU

Does anyone have any experience with creating DjVU files from scans? I have a feeling it is recommended for pages dominated by text, although one site said it could be used for line drawings as well. Recommendations? pfctdayelise (说什么?) 06:02, 5 March 2008 (UTC)

I don't think that most scanner software supports it anyway. If your scanner does, I'd try experimenting with it yourself. Adam Cuerden 06:34, 5 March 2008 (UTC)

PNG vs. JPG

In this section of the article, it says:

"In any case, it's usually best to scan to a lossless format, such as PNG, TIFF, or, if you have to, BMP first, as you lose quality going from JPG to PNG, but not the other way around."

Wait a minute... surely that should say:

...you lose quality going from PNG to JPG, but not the other way around.

Right? - dcljr 06:00, 18 March 2008 (UTC)

I think I phrasesd that oddly - what I meant was that JPG has already lost quality that won't be regained by a switch to PNG, but a JPG made from a PNG is the same or better as a JPG made in the first place. Adam Cuerden 06:20, 18 March 2008 (UTC)

Category:Scans

is now available. Should we mention that in the text? --11:50, 10 October 2008 (UTC)

Warning at 5?

I don't think we warn about uploads in excessive of 5MB anymore, do we, as the text claims here? Dcoetzee (talk) 22:42, 13 March 2009 (UTC)

Why 95-100?

In better image editing programs, you will get the choice of quality vs. compression for your JPEGs. In general, if the scale is 1 to 100 (with 100 being best quality), don't go under 85 or so (and keep it between 95-99 unless file size issues force you to go lower), and check the image at full resolution before uploading to make sure it still looks okay.

Why does the Help advise users to keep the JPEG quality between 95 and 99? I always use 100. I would prefer to see Commons advising users to use the highest quality possible for the original image, i.e. 100. Later, other users can easily make lower quality derivatives, for efficient thumbnails maybe. -84user (talk) 18:22, 31 March 2010 (UTC)

No response after one year, so I reworded that part to advise uploaders to use the maximum quality possible. -84user (talk) 09:48, 11 October 2011 (UTC)

Dealing with bleedthrough

When scanning pages printed on both sides bleedthrough from the back side can be a problem. It can be effectively cancelled by backing with a completely black card as now all the back is the same black color and any bleedthrough will be totally uniform and unnoticeable. It can be used to prevent bleedthrough even if the page is not printed in the back but we are using some book to keep the page flat. A black thick card works better that a sheet of white paper. I always keep a large, black card sheet with my scanner for this purpose.GS3 (talk) 00:54, 11 October 2011 (UTC)

The black paper advice was already at the end of the General advice section, but I added more mentions further up the page. -84user (talk) 10:00, 11 October 2011 (UTC)

Bleedthrough in pre-scanned material ?

Any suggestions for post-processing to deal with bleedthrough, if all I have is scans made by somebody else & no access to the originals? At the moment I'm interested in some 19th century images from books scanned by the Internet Archive, which unfortunately suffer from quite marked bleedthrough.

Given that the scans of the reverse pages are also available, is there any software that might intelligently remove the bleedthrough, perhaps in a similar way to Blind Signal Separation for audio? Does anyone know of such code? Jheald (talk) 17:22, 2 March 2012 (UTC)

Various research groups have certainly worked on this, from going back 15 years right up to the present (eg [3], [4], [5], [6]); and some high-end commercial vendors of scanning solutions appear to offer it. But at first sight there doesn't seem to be an accessible free toolchain that I could find. Obviously, if anyone does have such a thing, I'd be very interested. Jheald (talk) 22:11, 2 March 2012 (UTC)

Possible matlab material: [7] Jheald (talk) 22:29, 2 March 2012 (UTC)

TIFF

It's not a "very strange" format! It is very useful but not well understood by most people. the Wikipedia page on Tiff does a fairly good job of explaining that it is a container which can use different compression schemes (like WAV). I agree that it is better suited for archival purposes (and that is why it is used by academia) than for online display. One advantage is that it allows multi-page per file and using CCITT FAX 4 compression it is unsurpassed for archiving B&W documents. It will produce smaller files than PNG and hold many pages which PNG cannot do. It also allows annotations with programs like Imaging which PNG cannot do either. It is a very useful format for certain things but probably not really needed for displaying images online where other formats are better suited. GS3 (talk) 01:16, 11 October 2011 (UTC)

I removed the strange "strange" and reworded to encourage use of lossless compression. 84user (talk) 10:16, 11 October 2011 (UTC)

Color "correction"

I have seen some instances of people taking images of art or documents and subjecting them to color "correction" using editing software, then re-uploading the images. The Editing section of Commons:Image_guidelines discourages "Unnecessary or inappropriate use of artistic filters and effects.--Larrybob (talk) 22:35, 12 December 2012 (UTC)"