Best practices for good performance copies?: Things are different now with respect to PDF files and paper size

Advice and Help

Moderator: kcleung

Post Reply
DuoPianist
Posts: 4
Joined: Fri Oct 19, 2007 7:06 pm
notabot: YES
notabot2: Bot
Location: Chattanooga, TN

Best practices for good performance copies?: Things are different now with respect to PDF files and paper size

Post by DuoPianist »

Let me describe some recent adventures I've had trying to produce good performing copies. (From these files FWIW: IMSLP147747-SIBLEY1802.17248.8143-39087012930998septet_vol._1.pdf and IMSLP147748-SIBLEY1802.17248.9897-39087012930980septet_vol._2.pdf)

These files are typical in that they have a series of embedded raster graphics, one for each page. It appears to me that PDF standard allows for specifying that the page is only a window cut from the graphic. This seems to have caused endless grief for the standard Linux tools for cropping. I had to go through many steps to get decent results:

strip the embedded graphics into individual files
deskew them although I could have skipped this
crop margins
reassemble into PDF files

The aggravation of it all led me here to this forum and I was surprised that there seems to be a lot of stuff about PDF that is not current and very little stuff that can help make many things better.

A true statement, yes?: Every single page of a PDF file has its own resolution and virtual page size. The whole focus on Letter, A3, A4, etc., traces back to old versions of ghostscript that needed it.

If you allow the format to do what it does well, then each page can have a different virtual size but will be rendered pretty much as effectively as possible when the setting for print in Adobe Reader or Foxit is set to "Fit page."

Where am I going with this? I'm looking for the best, reliable tools, to margin pages down to minimums. The nicest one I have seen in terms of interface is called krop by Dr. Armin Straub, but it chokes dramatically on the files above. (If you are interested, you can see the input and the failures here: https://drive.google.com/drive/folders/ ... sp=sharing) Linux command pdfcrop is another tool that ought to work but fails. If there are alternatives that work or updates to these that have bugs fixed, that's a great step forward. The conclusion I reach is that the kinds of PDF files we care about on IMSLP have contents that don't play nice with otherwise helpful tools people have written.

While I am on the general topic, there is a feature that I am looking for in some program that prints PDF files. To wit: an exact magnification option but with a specified offset or a specified center point (the same thing, just the way to look at it). Printing at xxx percent will position the lower-left corner of the document at lower-left corner of the paper and may overflow the other 3 margins. What if you could specify the point in the document to match to the lower-left corner of the print? By setting the percent appropriately you could margin all four sides at once to your liking. For someone wanting a lot of scores printed that come out better than the downloaded version, this would be a really quick way there.
coulonnus
active poster
Posts: 1530
Joined: Thu Jul 12, 2007 8:53 am
notabot: 42
notabot2: Human
Location: Nice, France
Contact:

Re: Best practices for good performance copies?: Things are different now with respect to PDF files and paper size

Post by coulonnus »

These files are here: https://imslp.org/wiki/Septet_in_E-flat ... udwig_van) Arrangement for Piano 8 hands (Pauer). If someone wants to research this topic.
coulonnus
active poster
Posts: 1530
Joined: Thu Jul 12, 2007 8:53 am
notabot: 42
notabot2: Human
Location: Nice, France
Contact:

Re: Best practices for good performance copies?: Things are different now with respect to PDF files and paper size

Post by coulonnus »

The vol.1 was sort of locked in the sense that I can't split it into individual pages with pdftk. I worked around with Method 3 of https://imslp.org/wiki/Locked_PDFs But this increases some margins.
Last edited by coulonnus on Wed May 22, 2019 12:11 pm, edited 1 time in total.
coulonnus
active poster
Posts: 1530
Joined: Thu Jul 12, 2007 8:53 am
notabot: 42
notabot2: Human
Location: Nice, France
Contact:

Re: Best practices for good performance copies?: Things are different now with respect to PDF files and paper size

Post by coulonnus »

I have no trouble converting the 36 pdf pages to tif at 400 dpi with imagemagick. Now I would have no trouble cropping the margins with irfanview and assembling the pages into a new pdf file with Procedure No.4 of https://imslp.org/wiki/IMSLP_talk:Scanning_music_scores
coulonnus
active poster
Posts: 1530
Joined: Thu Jul 12, 2007 8:53 am
notabot: 42
notabot2: Human
Location: Nice, France
Contact:

Re: Best practices for good performance copies?: Things are different now with respect to PDF files and paper size

Post by coulonnus »

Read https://imslp.org/wiki/IMSLP:Image_Conversion and you will be able to make the changes yourself.

For your last paragraph: if the pdf size and the margins have been defined correctly there is no need to adjust the size and the offsets. Selecting page scaling: none with Acrobat reader should give a good print.
DuoPianist
Posts: 4
Joined: Fri Oct 19, 2007 7:06 pm
notabot: YES
notabot2: Bot
Location: Chattanooga, TN

Re: Best practices for good performance copies?: Things are different now with respect to PDF files and paper size

Post by DuoPianist »

coulonnus wrote: Sat Jun 01, 2019 4:48 am Read https://imslp.org/wiki/IMSLP:Image_Conversion and you will be able to make the changes yourself.

For your last paragraph: if the pdf size and the margins have been defined correctly there is no need to adjust the size and the offsets. Selecting page scaling: none with Acrobat reader should give a good print.
Of course. But if the pdf size and the margins were defined correctly this thread would be unnecessary. Nor will the pdf size ever be defined correctly for everyone. If optimized for A4 it will likely be suboptimal for US Letter-size or vice versa--the most common issue.

In terms of the latest PDF tools the size is pretty much irrelevant. The availability of fit, shrink, actual size, and scale get us a long way toward the best possible world. The remaining thing missing that I was talking about in my last paragraph is the ability to specify an explicit scaling and a center point so that unwanted space around the music will be clipped on multiple sides instead of flowing off up to threes sides across and down from the upper left corner.

What I am working on for a great deal of the last month is an automated process to identify the music in an image and reliably crop it to just the music. What is difficult is to allow for titles at the top and text at the bottom without including blotches and stains. I don't believe it can ever be perfect but I want to assist with addressing the thousands of scores that could benefit from this.

I will be back when can offer something more.
Post Reply