Scanning Publications

[mgilvey]

Greetings,

I have a client that needs scans made of some magazine and news articles that were written up about his business. The final output is to a downloadable PDF.

Is it possible to get this done in one scan or do I have to scan the text and the image seperately and if so...well you know where I'm going with this. Give me a plan/workflow for scanning printed matter both photos and text that have to be in the same file but only for onscreen viewing.

My experience tells me to scan the text at 600 dpi line art and the photo with descreening but it's such a task to have to combine them back together especially when the article has to be scanned in two parts because it is too big for the plattin. Isn't there an easier workflow than this?

[Gregory C]

I've been working on the same thing; scanning articles from newspapers and magazines; and I've realised the same thing as yourself, that for a perfect result, text; or more precisely CMYK-native mono-colour 100% text (usually black); needs to be scanned at 600 dpi while photos and screened text need to be descreened. it would be great if SilverFast had an option where the scanned area could be subdivided in separate frames with varying settings and automatically combined and output as a single file. unfortunately, this is not possible at the moment.

in a recent project, I first scanned the article at 300 dpi (600 would of course be better) without descreening and then again with descreening. I then opened the files in GraphicConverter and selected/copied the descreened photos into the 300 dpi file and saved the combined result.

note though that to scan and output the article at 100%, the descreening density/resolution value has to be set to around double the actual density of the article. the result is about the same as using the actual density value and sometimes better.

if someone knows of a better workflow, I'd love to hear about it. I have many articles waiting to be scanned.

regards
Gregory

[mgilvey]

Ok I did a test. This test is from a magazine, looks like 300 dpi to me. Here goes.

Scanning Notes: Scanning Printed Matter

1. Prescan: 36-24 Bit RGB, no filters. Make sure you include a step chart.
2. Use the Color (C) button and then pick highlight and shadow with eyedropper tool.
3. Set percentage to 100% @ 600dpi. For some reason, at first I could not get it to 100% unless I moved the DPI slider down to 133 but then I got it. I don't know if this was enabled or not but when I did get 100% @ 600dpi, the Output locks were unlocked (didn't really notice if they were locked or not previously).

TEST ONE
Page is from Photo Pro, looks like 300 dpi.
This scan was without descreening (17mb JPEG 12). After the scan was complete, I opened it in Photoshop, cropped (5.28 x 10.76 inches, just cropped, no recuction), rotated and spotted. Saved file(10.6mb JPEG 12). This is my working scan.

a. Then saved a PS PDF (1.6mb JPEG 5). Image looks terrific!

b. Then I went back to the working scan and saved another as a PS PDF but this time at a JPEG quality of 2. The file went down to 1,012 k and still looked terrific.

c. Then next test I did, I went back to the working scan, adjusted the levels so the white point contained 0 dot. Then converted the image to Indexed color, 128 colors, no diffusion or dither. Saved the file as a GIF (5.5 mb) It looked perfectly acceptable to me, especially for on screen viewing.

d. Opened the GIF version in Photoshop and saved as a PS PDF. Because it was indexed color, the JPEG compression option was not available so I left it at the default "Zip" and saved it. The file size actually increased to 6.6 mb.

e. Opened the working file again and converted it to a 64 color GIF. 4.4 mb.

f. Saved the 64 color GIF as a PS PDF (zip). 5.6 mb

It appears that the best results from above would be save a PS PDF at JPEG 2 from the working file which put the file just under a meg.

TEST TWO
Same image accept this time I scanned it at 300 dpi. My goal is to get the scan time down and get the best text I can at the smallest size.

The scaning time was about a quarter of the 600 dpi scan, or so it seemed. Just like the scan above, I cropped, rotated and spotted and saved as a JPEG 12. 4.8 mb.

a. Saved a PS PDF at JPEG 2. The file was 392 k and still looked good.

TEST THREE
Same image accept this time I scanned it at 150 dpi. My goal is to get the scan time down and get the best text I can at the smallest size.

The scaning time was about a minute or less. Just like the scan above, I cropped, rotated and spotted and saved as a JPEG 12. 1.3 mb.

a. Saved a PS PDF at JPEG 2. The file was 192 k but the text was not so good.

b. Saved a PS PDF at JPEG 5. 248 k and the type still looked bad but you could read it.

c. Saved a PS PDF at JPEG 5 again but this time I set the whited to 0 dot. 244 k and the type still looked bad but you could read it.

d. Saved a PS PDF at JPEG 8 with white point at 0. 384 k and the type looked soft but you could read it.

e. Saved a PS PDF at JPEG12. 1.4 mb and the type didn't look much different than example e above.

RESULTS: The 300 dpi scan that was saved as a PS PDF with JPEG 2 was only about 10 k larger than the 150 dpi scan saved at JPEG 8/12 and looked much better. The 600 dpi scan that was saved as a PS PDF JPEG 2 looked the absolute best but it was almost overkill.

The image looked perfectly acceptable on all scans, even though I didn't use any filtering. I think I got lucky.

I will try again tomorrow (if I can get the time) on newspaper articles where the line screen is lower and there is a gray paper base that I will want to get rid of.

Here is a link to the PDFs I made:
http://homepage.mac.com/megilvey/FileSharing2.html

[mgilvey]

Ok, I did some testing this eveing with plain old news paper. I found a Le Monde in my recycle pile and did the following tests.

B&W
1. Prescan: 16 Bit Grayscale, no filters. Make sure you include a step chart.
2. Pick highlight and shadow with eyedropper tool on the step chart.
3. Set percentage to 100% @ 600dpi.

NOTE: Make sure the save options are set to save as a TIFF. Grayscale JPEGs will not open in PS.

TESTS
Page is from Le Monde, looks like 160 dpi or so.
I made three scans for each test:
1. Points were set to the step chart-TEST ONE
2. Points were set to the paper stock and the masthead logo-TEST TWO
3. Scanned in 1 bit line art, DPI automatically defaulted to 400 dpi-TEST THREE

I felt that setting the points on the paper and the blackest part of the text did not give the image a nice appearance. Setting the black point also made the mid tones block up which I didn't want to have to spend extra time to open up. I threw this out.

Scanning as 1 bit line art was also not a good idea. If I had to reproduce this in print, then yes, scanning the text in line art is good but not the image. I threw this out.

So, setting the points on the step chart turned out to be the best. All tones were nearly an exact match. I think I would like to get the paper tone down to about 3% and I would be happy. I scanned this way at 600, 300 and 150 dpi. Then I tried making index color versions, they were smaller in size but not as small as the PDFs i made. I also made PDFs that had JPEG compression of quality 1 and 2. The 300 dpi and 600 dpi versions looked acceptable, even at JPEG 1. The file sizes in this case were a little larger than the RGB versions I did yesterday but only because the dimensions were larger.

FINAL RESULTS
Next time I have to scan Color or Grayscale publication for on-screen viewing I will do the following.

Color:
1. Prescan: 36-24 Bit RGB, no filters and include a step chart.
2. Use the Color (C) button (I have calibrated my scanner, hense the "C" in the one button) and then pick highlight and shadow with eyedropper tool on the step chart.
3. Set percentage to 100% @ 300 dpi.
4. Scan to JPEG Quality 12. Since the final output is a PDF, I'm not concerned with artifacting, especially at 300 dpi.
5. Scan the other portions of the article if it is too big to scan in one shot.
6. Open all parts and combine in Photoshop, adjust tone and spot if needed and flatten file.
7. Save as a Photoshop PDF with a JPEG compression of 1 or 2.

Grayscale:
1. Prescan: 16 Bit Grayscale, no filters and include a step chart.
2. Set highlight and shadow points with eyedropper tool on the step chart.
3. Set percentage to 100% @ 300 dpi.
4. Scan to TIFF since Photoshop will not open a grayscale JPEG.
5. Scan the other portions of the article if it is too big to scan in one shot.
6. Open all parts and combine in Photoshop, adjust tone (set paper stock to 3% and leave black alone) and spot if needed and flatten file.
7. Save as a Photoshop PDF with a JPEG compression of 1 or 2.

Live long and prosper.