Creating small, text-editable PDFs?

Question

I have many documents from school that I'd like to scan, convert to PDFs and then compile together into one PDF file. The problem is that whenever I scan a document, I can't compress it down enough without sacrificing quality, so I can't have a small PDF file. I have a few eBooks, and, looking at their properties, somehow the creators of those eBooks were able to compress their PDFs down enough without sacrificing any quality to the documents whatsoever. Here's an example:
one eBook that I have is 696 pages, the filesize is 4,159,863 bytes, and it looks like it was simply scanned and put together, because the quality of the eBook matches the book that it came from (which I happen to own). True, the images in the eBook are black and white (some of the images in the book are in colour), but there is no noticeable depreciation in their quality. Whoever created it was using the application PScript5.dll version 5.2, the PDF producer Acrobat Distiller 6.0.1 (Windows), the PDF version is 1.5 (Acrobat 6.x), and was NOT optimized for fast web view. How did the author get the page size so small (kb wise) yet retain quality? If one were to do the math: 4,159,863 bytes divided by 696 pages equals 5976.81, then divided by 1000 equals 5.97681. Almost 6 kb per page? How is that possible?
The smallest I could compress scanned images was about 20 kb, and the quality was terrible. In fact, the only way I could shrink them to that size was to scan them in Black and white (1 bit), and then distill them as much as I dared. The vast majority of documents that I'm scanning are black and white text with very little in the way of graphics.
Here are the programs that I have at my disposal:
Jasc Paint Shop Pro 8 and HP Director (for scanning with my HP PSC 1210 all-in-one printer, which has OCR technology)
Microsoft Word 2000 (9.0.2720)
Adobe Acrobat 5.0.5 (10/26/01) (yes, it has paper capture)
Acrobat Distiller 5.0.5.2001101100
It is driving me absolutely crazy that I cannot figure this out. Even creating my own job option in Distiller and adjusting the settings didn't help.
It should be possible, with the programs that I have, to get the results I want. For those of you who believe I should upgrade to something beyond Adobe Acrobat 5, here's another example for you:
There used to be a link on www.iartonline.ca (which no longer exists) to a PDF that is 13 pages long, created with the application Microsoft Word 8.0, the PDF producer is Acrobat Distiller 4.0 for Windows, the PDF version is 1.2 (Acrobat 3.x) and has a file size of 79,719 bytes. In contrast, I scanned two school documents (using OCR) into Microsoft Word, corrected any errors, and converted it to an Adobe PDF from there. The resulting size (for TWO PAGES) was 25 kb.
So, in conclusion, it HAS to be possible for me to scan my documents, retain their quality while compressing their size, and compiling them into single file PDFs.
I apologize for the lengthy explanation (this is why I couldn't ask the question on WikiAnswers), but hopefully this will save others interested in answering from asking irrelevant (to me) questions.
Thank you for your time.

Answer 1

Well, back to square one. IrfanView seems like a good program, but I'm running into the same issues I had before (sacrificing quality for a more compact file size). Using IrfanView to scan the sheet music opens up HP Director (since I have an HP printer it makes sense that this would be the default program). With HP Director, my output type options are as follows:
1. Millions of colors
2. 256 colors (8-bit)
3. 256 colors (web palette)
4. 256 colors (system palette)
5. 256 gray shades
6. Black & White (1-bit)
Choices 1 and 6 are out automatically. Since the output dimensions (area to be scanned) is 7.86 x 10.37 inches, at 200 DPI, the image size is as follows (using the remaining choices from above):
2. 3.19 MB
3. 3.19 MB
4. 3.19 MB
5. 3.19 MB
The scale is 100%. After selecting choice 2 (256 colors - 8-bit) and scanning to IrfanView, the Image Information is as follows:
1571x2075x8BPP, 3.11 MB
So I'll save this as a TIF file (using ZIP compression) and the resulting file size will be 532 KB. Opening this file with AA5 and saving it as a PDF will result in a file size of 512 KB. The quality is definitely there, but the file size is way too big.
If I go through all that again and use JPEG compression instead of ZIP, the resulting file size is 1,836 KB. After opening the file with AA5 and saving it as a PDF: 354 KB. Once again, the quality is there, but the file size is too big. It appears that JPEG compression is the better choice in this particular case.
So what can I change? Going back into HP Director, I adjust the scale (size) to 50%. Output dimensions: 3.93x5.19; DPI: 200; output type: 256 colors (8-bit).
In Irfanview: 788x1035x8BPP, 797.5 KB; quality is a little bit less. Saving to TIF with JPEG compression: 592 KB. Opening with AA5 and saving to PDF: 125 KB.
Interesting side note: I decided to perform a Paper capture on this 125 KB PDF file. In the Paper capture Preferences, I chose 'Searchable image (compact)' as the PDF output style, and chose 'None' for Downsample images. Result? 267 KB.
Trying that again, this time with the PDF output style of 'Searchable image (exact)' produces the same file size (267 KB).
Mind you, when Paper capture was performed on the PDF, it straightened the image, which may have contributed to the larger file size. I noticed that IrfanView did something similar: if I rotated the image slightly (in this case, 1.10), IrfanView changed the image's BPP from 8 to 24 (but I changed it back to 8 BPP by going to Image>Convert to Grayscale).
Tried scanning it at 150 DPI instead of 200; quality was still there. TIF file size: 911 KB, PDF file size: 201 KB.
100 DPI. TIF: 468 KB, PDF: 109 KB. Starting to get a little blurry, but that could have been due to making the rotation change and then converting back to grayscale.
Tried scanning at 75 DPI, 256 gray shades and half the scale (size). Quality was HORRIBLE.
IrfanView has no "Print to file" option, so using Acrobat Distiller is out.
I've noticed that, if you rotate an image in IrfanView, sharpen the image, then convert it back to grayscale and save it as a TIF file, it will be blurry when it's opened in AA5.
HP Director doesn't always straighten the scanned image, even though there's an option to do so (and I always check the box for that option). I've re-scanned the sheet many, many times, but it's only at random that HP Director automatically straightens the scanned image.
I did a scan with 75 DPI, 256 gray shades, 100% scale. It looks great in IrfanView (602x775x8 BPP, 458.17 KB), but when I save it as a TIF (JPEG compression) and open it with AA5, the quality had considerably diminished. Even redoing the scan and saving it as a TIF with ZIP compression, it's still the same.
I tried Packbits compression, same thing; LZW, ditto.
So I upped the DPI to 150 and tried again, this time using JPEG compression. TIF: 1,077 KB, converted to PDF it was 231 KB. The quality was there, but the file is TOO BIG.
All in all, I've tried a few different approaches and it seems to me that I can have one or the other: quality or compact file size, not both. But I refuse to believe that. It MUST be possible.
I have an eBook that is 736 pages long and is 6.75 MB (7,077,900 bytes). The PDF producer was Adobe PDF Library 5.0 and the application used was Adobe InDesign 2.0.2. The quality of the eBook almost exactly matches the actual, physical book (I know this because I've borrowed the book from the library).
Did Adobe PDF Library 5.0 and/or Adobe InDesign 2.0.2 create such small PDFs from scans that were responsible for such a small eBook, or was it some other program? And, if it was some other program, then what's the name of it? Is it freeware?
You can download the eBook in question for free here: http://www.deliberatedumbingdown.com/MomsPDFs/DDDoA.sml.pdf
Bill, in HP Director, after clicking on "Scan document", the user is presented with 4 choices:
- Text & Graphics as image
- Text as image
- Editable text
- Editable text with graphics
The settings for these 4 choices can be modified. For the first two choices, resolution can range anywhere from 75-19,200 PPI, and the output types can be Millions of colors (24-bit), 256 gray shades (8-bit grayscale) or Black and White (1-bit).
The settings for the latter two give less options: resolution can be either 300 or 400 PPI, and the output type does not include the Black and White (1-bit) option. (This bit about HP Director is mostly for your information to give you a better idea of the program's capabilities.)
It IS possible to scan a document with HP Director, do OCR on it and save it as a PDF. The file, however, is very large.
I now understand that a document can have OCR performed on it OR have AA5 do a paper capture, but not both.
I really appreciate the time you've taken to address my questions, Bill. As I've stated before, this will also serve as a guide to others in the future should they have similar problems/questions.