IMSLP Forums

I'm tearing my hair out trying to keep the pdf size down. Which factor is most important, the size of the tif I'm converting to pdf, or the dpi OF the pdf I'm printing to file. I keep getting 15 plus MB files for a 25 page score. Help! Thanks.
Neil

The most important factor is to control resolution, color depth, and compression of the tif files. A recommended configuration is 600 dpi, monochrome, CCITT G4 compression. In these conditions a typical score page takes around 100 - 200 KB, so a 25 page score would be around 2.5 - 5 MB. When you have all tif files of this size you can combine them into a pdf file without any further change of resolution/color depth/compression using free tools like these.

Max

The information from one of the TIF's I have gives the following:

600 dpi
LZW Compression
Resolution unit 2
391 Kb (the tifs in this batch range up to 800KB)

1108 x 1417 pixels (All images are the same in this batch)

So, without getting too technical, do I simply "re-size" the images in a photo editor? And then, when it asks what DPI to use to print to PDF, what setting should I use. Thanks for your help

Neil

Size and resolution are OK, I think that you have to use a photo/image editor to adjust color depth to monochrome and compression to CCITT G4. Given the current size, I guess that they are monochrome already, however compression is LZW.

I wouldn't recommend to use a virtual printer then, because it may change resolution, color depth and compression to different settings than those in the tif files, thus jeopardizing all your efforts to reduce them to 100 - 200 KB per page (this may depend on the virtual printer you use, though). You'd better use a program that simply "assembles" multiple tif files into a pdf "envelope" like the one that I suggested, without touching the tif settings. Anyway, if you stick to the use of a virtual printer, you should set it for 600 dpi resolution as well.

Max

Where does this initial compression take place? I don't see that my Brother scanner gives any options.
If I understand correctly, virtual PDF printers are different than the converters like PDFill, which, by the way, I can't get to work for love or money. I can't get more than a couple of images into it at one time. I have another program called PDF Architect which doesn't recognize TIF, but will take .png and the rest. Any harm converting to png?

Okay, I have to find software that lets me use different compression during scanning. Any freeware that works?
Neil

IMO, the key requirements for software are:

- interface to catch data directly from the scanner;
- format conversion functions to adjust resolution, color depth and compression when saving scanned files;
- capability to run format conversion in batch, in the case there is a large set of files scanned already to be converted altogether.

I use a very old version of Paint Shop Pro, which includes all the above functions, however it is not freeware. I'm pretty sure that modern free image/photo editing tools can do the same, however I cannot suggest one. Can anybody suggest? Maybe it's also time for me to replace my decrepit software tool...

Max

So, based on what I've been reading, here's what I'm working on:

600 dpi B&W scan from the scanner
Conversion to CCITT 4 Compression with Irfanview - That really reduced the size of the image.
ScanTailor workover
Hand work to clean up
Conversion to PDF with PDFill

It'll be another couple of days. Should be good.
Neil

Hi Neil!

I just wanted to share a couple of observations. First, I'm confused by the dimensions of your scan. If you scanned at 600 dpi the resulting dimensions of your image (1108 by 1417 pixels) seem too small. I'm under the impression that if you were to scan a 9 by 12 inch sheet at 600 dpi the resulting image would be 5400 by 7200 pixels (9 x 600 and 12 x 600). Given your dimensions it would appear that the true resolution of your image would be around 125dpi (assuming a 9 by 12 inch original). At any point in your processing of the image did you reduce the dimensions? Secondly, if you are going to use ScanTailor you can skip the previous step of using Irfanview to convert to CCITT 4 compression because the output tifs from ScanTailor will be compressed using LZW. There doesn't seem to be any options to control this. If the software you use to create the PDF file cannot compress the output images from ScanTailor using CCITT 4 then you could run those through the batch conversion in Irfanview to compress them.

Cypressdome

Cypressdome,
Yes, I'm confusing myself! As you say the original scans were quite large. Also, now I see why after ScanTailor, the compression is LZW again. To make things even more confusing, in between I experimented with re-sizing using a photo editor. I'm going to discard everything and start with my original scans.

So the order should be:
Scan
ScanTailor/hand work
Compress to CCITT 4 with Irfanview
PDF

Does that sound right in order to keep the quality good and the size down?
Neil

That looks good to me. Looking forward to your submission!

Cypressdome

Here it is. http://imslp.org/wiki/Double_Bass_Conce ... r_Albin%29

A five-star scan in my opinion. Nice work!

I would bet that the work never had its copyright renewed in the U.S. Hopefully, I can confirm that and get it marked as being public domain in the U.S.

Thanks,
Cypressdome