User Tools

Site Tools


scanning

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision
Previous revision
scanning [2015/12/02 07:21] sbwscanning [2023/08/16 13:22] (current) sbw
Line 1: Line 1:
 ====== Scanning SBW magazines ====== ====== Scanning SBW magazines ======
  
 +Given the first stage of the project is largely complete, this page is more for information. It may help other organisations undertaking digital history projects.
  
 ===== 1. Remove the magazines from the springback binder ===== ===== 1. Remove the magazines from the springback binder =====
  
-The magazines are all stored in springback binders, one year (usually 12 issues) per binder. Springback binders are basically like a big bulldog clip. By bending back both halves of the cover, the folder with the magazines inside can be carefully removed.{{ :processed_7004.jpg?nolink&100|}}+The magazines are all stored in springback binders, one year (usually 12 issues) per binder. Springback binders are basically like a big bulldog clip. By bending back both halves of the cover, the folder with the magazines inside can be carefully removed.
  
-{{:processed_7005.jpg?nolink&200|}}{{:processed_7006.jpg?nolink&200|}}+{{ :processed_7004.jpg?nolink&100|}}{{:processed_7005.jpg?nolink&200|}}{{:processed_7006.jpg?nolink&200|}}
  
 ===== 2. Remove staples from the magazine ===== ===== 2. Remove staples from the magazine =====
Line 16: Line 17:
 ===== 3. Scan the magazine ===== ===== 3. Scan the magazine =====
  
-Scan the pages one at a time, into a multipage PDF. +Scan the pages one at a time, into a multipage PDF. A high end office scanner found in most large corporates is ideal - modern one scan quickly and at high quality.
  
 I use the following settings: I use the following settings:
-scan to PDF +  * scan to PDF 
-resolution: 300 x 300dpi +  resolution: 300 x 300dpi 
-paper size: A4 (if you leave it on Auto, sometimes the machine will get it wrong) +  paper size: A4 (if you leave it on Auto, sometimes the machine will get it wrong) 
-Black and white text (don't use photo settings, as you will get speckles; don't use greyscale) +  Black and white text (don't use photo settings, as you will get speckles; don't use greyscale) 
-If you have a density setting, somewhere in the middle is probably best - too high and you will start to see speckles, too low and the text will be a bit faint. +  If you have a density setting, somewhere in the middle is probably best - too high and you will start to see speckles, too low and the text will be a bit faint. 
-If you have a contrast setting, higher is generally better.+  If you have a contrast setting, higher is generally better.
  
 High end office scanners normally default to PDF these days. Normally you have about 60 seconds after you have scanned the previous page to scan the next page, in order to keep it as one document. I find that it takes about 15 seconds/page once you get going.  High end office scanners normally default to PDF these days. Normally you have about 60 seconds after you have scanned the previous page to scan the next page, in order to keep it as one document. I find that it takes about 15 seconds/page once you get going. 
  
-It may be possible to use the automatic document feeder for more recent magazines. The main risk with using the automatic document feeder is that if something goes wrong, it may damage the pages. However, most of the magazines (anything prior to 1984) are quarto sizenot A4, and these ones will need to be scanned manually.+It may be possible to use the automatic document feeder for magazines with more modern paper. The main risk with using the automatic document feeder is that if something goes wrong, it can seriously damage the pages. The ADF tends to work better when the documents are fed in lengthways (rather than sidewayswhich is probably the default on most machines) - this may need configuring on the scanner.
  
 Before you start scanning in earnest, check that the quality of the scan is as high as possible. Spend some time trying various settings to see which works best. Any areas that are not text/pictures should ideally come out white in the scan. Any speckles or colouring reduce the ability to subsequently OCR the documents, as the below examples show. Before you start scanning in earnest, check that the quality of the scan is as high as possible. Spend some time trying various settings to see which works best. Any areas that are not text/pictures should ideally come out white in the scan. Any speckles or colouring reduce the ability to subsequently OCR the documents, as the below examples show.
  
-1Black and white – text/line art setting +Also note that the best quality for OCR may not be the best for other purposes eg photos.
- +
  
-Result of OCR+**1. Black and white – text/line art setting** 
 +{{:bw_line_art.png?nolink|}} 
  
-the lap of luxury. +**Result of OCR**
-IYOu know", said Frank, "a tourist's life won't be bad." "No", said Snow, "just on the tarn." +
-But we were all looking forward to it!+
  
-2Black and white – text only setting +  the lap of luxury
-  +  IYOu know", said Frank, "a tourist's life won't be bad." "No", said Snow, "just on the tarn." 
-Result of OCR+  But we were all looking forward to it!
  
-the lap of luxury+**2Black and white – text only setting** 
-"You know", said Frank, "a tourist's life won't be bad." "No", said Snow, "just on the turn." +{{:bw_text.png?nolink|}}  
-But we were all looking forward to it!+**Result of OCR**
  
 +  the lap of luxury.
 +  "You know", said Frank, "a tourist's life won't be bad." "No", said Snow, "just on the turn."
 +  But we were all looking forward to it!
  
  
-The margins are not even, so take this into account when you are scanning, and shift the pages one way or the other, otherwise you may lose text off the sides. This is primarily an issue for older magazines which were printed on quarto paper, which is wider than A4. Newer ones are A4 and the text should fit on an A4 page regardless.+ 
 +{{:processed_0688.jpg?nolink&150 }}{{:processed_0689.jpg?nolink&150 }} The margins are not even, so take this into account when you are scanning, and shift the pages one way or the other, otherwise you may lose text off the sides. This is primarily an issue for older magazines which were printed on quarto paper, which is wider than A4. Newer ones are A4 and the text should fit on an A4 page regardless.
  
 I suggest scanning the various cover sections (front cover, inside front cover, back cover, inside back cover) separately. This makes it cheaper and easier to OCR. For the cover (and any advertisement pages), I use a low density setting, as otherwise it comes out very dark. Again, some trial and error may be needed to get good results. I suggest scanning the various cover sections (front cover, inside front cover, back cover, inside back cover) separately. This makes it cheaper and easier to OCR. For the cover (and any advertisement pages), I use a low density setting, as otherwise it comes out very dark. Again, some trial and error may be needed to get good results.
Line 58: Line 61:
  
 Save the magazine with the following names: Save the magazine with the following names:
-main part - <yyyymm>_<issuenumber>.pdf eg 195401_230.pdf for January 1954, issue number 230. +  * main part - <yyyymm>_<issuenumber>.pdf eg 195401_230.pdf for January 1954, issue number 230. 
-front cover - eg 195401_230_cover.pdf+  front cover - eg 195401_230_cover.pdf
  
 In a lot of cases, the same inside covers and back cover are used over a year. You only need to scan it once. Name them: In a lot of cases, the same inside covers and back cover are used over a year. You only need to scan it once. Name them:
-1954_inside_cover.pdf +  * 1954_inside_cover.pdf 
-1954_back_cover.pdf +  1954_back_cover.pdf 
-1954_inside_back_cover.pdf+  1954_inside_back_cover.pdf
  
-More recent magazines may not have issue numbers, in which case you can leave it out +More recent magazines may not have issue numbers, in which case you can leave it out eg 198701.pdf
-eg 198701.pdf+
  
 I can compile the various sections back into a magazine for the website. If you make a mistake (eg a page scans badly, page out of order) make a note of it and just keep scanning. For badly scanned pages, either rescan the page straight away, or scan it separately afterwards. I have software which can manipulate/insert/remove pages after the fact. This is usually easier than starting again. I can compile the various sections back into a magazine for the website. If you make a mistake (eg a page scans badly, page out of order) make a note of it and just keep scanning. For badly scanned pages, either rescan the page straight away, or scan it separately afterwards. I have software which can manipulate/insert/remove pages after the fact. This is usually easier than starting again.
  
-Zip sets of PDF files together (max of about 10MB per zip file) and email them to me at tomb@ozultimate.com. Include a note explaining any errors that need to be corrected or pages rearranged.+Zip sets of PDF files together (max of about 10MB per zip file) and email them. Include a note explaining any errors that need to be corrected or pages rearranged.
  
 ===== 5. Finishing up ===== ===== 5. Finishing up =====
  
 Insert the magazines back inside the springback binder. Don't restaple them – we'll look for a better method of archiving that doesn't mark the magazines. Insert the magazines back inside the springback binder. Don't restaple them – we'll look for a better method of archiving that doesn't mark the magazines.
 +
 +===== Other notes =====
 +Technical notes - to come back and fill in more detail on:
 +  * Double sided scans via newer photocopiers
 +  * Use of OcrMyPDF (including Tesseract) for OCRing
 +    * + OnlineOCR
 +  * Centering magazines
 +  * Use of GhostScript to reduce file size
 +  * Use of PDFSAM to merge double sided files (not needed if double sided scanner is used)
 +  * Use of iconv to remove some non-printing characters
  
  
scanning.1449001270.txt.gz · Last modified: 2015/12/02 07:21 by sbw

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki