Making a book with PDF music and lyrics - workflow questions & file size trouble

Previous topic - Next topic

santiclear

Hello everybody,

I'm quite new here. I was an occasional Adobe InDesign user (on shared hardware I had access to), but for a variety of reasons, including future-proofing my work, I decided to switch to open source.
I am working on a 200-page songbook, using as source material PDFs exported with MuseScore: I would like to add numbering to the individual songs, blank pages, occasional images, occasionally put the content of multiple PDFs on the same page, introductory pages, indexes etc. Basically I'd like to create a book with all the music I've put together.

Last year, while still using InDesing, I made the first attached file - and achieved it with this InDesign workflow:
  • Placed/linked (doesn't matter which) individual PDFs (approx. 40-60 kb for each page)
  • In the master page added a second layer on top of the existing one, with song/page numbers
  • Added the pages containing text
  • Added vector graphics to the first and last page
  • Exported the final PDF, which has searchable text in all pages, including those containing the MuseScore generated  PDFs (size: 1.5 Mb, including all cover graphics, see attachment)

I tried to do the same in Scribus 1.6.1 on Windows 10, beginning with just 15 pages/15 MuseScore-generated PDFs (40-60 Kb each, I attached a sample of one of them), but:
  • I had to import the PDFs as vector graphics - if I try to import as text, Scribus wouldn't show neither the music nor the text, despite me having both the Edwin (text) and Leland (music) OTF fonts which Musescore uses installed systemwide (Libreoffice sees them and uses them normally). Scribus got a bit lagging, switched the "Background" layer with the PDFs to outline view to speed things up.
  • I did the numbering using the scrapbook feature on a different layer (numbering will be more complex than 1-2-3-4-5 ecc.) - figured it out with no problems.
  • Saved the file - the *.sla is 33 Mb in size :o
  • Exported a PDF - it turned out a 141 Mb behemot :o , with no searchable text.

What am I doing wrong?

AdmFubar

Using Scribus 1.6.1, openSUSE 15.6
Advanced hobbyist

santiclear

MuseScore supports export in PDF, SVG, PNG, MusicXML and MEI. However, MusicXML, while XML based is more of an interchange format for interoperability with other music notation software, and is not quite reliable in layout preservation.

PatJr

it looks to me like everything gets put into a separate frame and the musical notes are combinations of characters
preserving the text in an editable, searchable form looks like it will be very very difficult and time-consuming

idea, import the music parts as vectors and the cut and past from your MuseScore program for the text?

santiclear

I wrote all the lyrics in MuseScore so that I would'nt have to do more editing in a separate program. One of the basic principles behind it was 1 page = 1 (or more) songs, sometimes 1 spread = 1 song. No page turning during performances. All those files have been written in a couple of years (which means that it would be quite time consuming to remove and then rewrite and redo the layout for all the lyrics), and were used and printed individually.

All I really wanted to do was put together the individual sheets, combine 2 or 3 songs on the same page where this is possible, number them and add introductions and indexes.

Aleks100

Pdf is container for everything (and anything).
Have you tried to export in svg format, or even png in high resolution - grayscale 450dpi for example?

Nermander

Quote from: santiclear on April 29, 2024, 11:18:15 PMWhat am I doing wrong?

If I am not wrong, a PDF can be added to a Scribus document in different ways. Did you try to import the PDF into an image frame?

Because I think otherwise the contents of the PDF is imported as individual objects.

But I am not sure if importing it to an image frame causes it to be rasterized.

PatJr

Quote from: Nermander on May 02, 2024, 04:47:51 PM
Quote from: santiclear on April 29, 2024, 11:18:15 PMWhat am I doing wrong?

If I am not wrong, a PDF can be added to a Scribus document in different ways. Did you try to import the PDF into an image frame?

Because I think otherwise the contents of the PDF is imported as individual objects.

But I am not sure if importing it to an image frame causes it to be rasterized.

I tried it, the pdf imported/opened as a vector graphic. I looked OK but all the text was outlined and you couldn't used a text search on it. I think that's the end goal here for the op, text searching the exported pdf after it's assembled in scribus.

Aleks100

For start in "A5 Venite fedeli (Stefani-Wade) (accordi).pdf" colors are wrong.
They should be 100% Black - they are 300% mixture of CMYK.
In booklet every page with notes have same mistake.
Pages made in Scribus are fine, 100% Black as they should be.

Does all text have to be searchable?

Below is pdf converted to grayscale

santiclear

Quote from: Aleks100 on May 02, 2024, 10:29:24 PMFor start in "A5 Venite fedeli (Stefani-Wade) (accordi).pdf" colors are wrong.
They should be 100% Black - they are 300% mixture of CMYK.
In booklet every page with notes have same mistake.
Pages made in Scribus are fine, 100% Black as they should be.

Does all text have to be searchable?

Below is pdf converted to grayscale
I had not noticed that. In fact, all the PDF output from Musescore has registration black instead of K black.  :'(

Hi-res pngs work fine, of course (they are large); have still not tried what Scribus does with MuseScore's SVGs, which are quite heavy as well.
Unfortunately, these are all workarounds which do not allow me to keep text searchable.

Aleks100

Not registration black - 4x100% ink.
It is a rich black for extra black offset printing.
I have also tried to import individual pdfs as curves (loosing all text) in Scribus and it looks fine, but I have not tested it in Acrobat.

santiclear

Isn't registration the sum of 100% inks that can be used (in this case C100%+ M100% + Y100% + K100% ?)
In any case, do you think I should report this to the MuseScore developers? In my opinion it should be treated merely as K100%, just as typical black text from word processing /desktop publishing software. But again, I am no expert in printing processes, so feel free to correct me.

As for my initial question, I seem to understand that there is no straightforward solution to my problem - I can achieve what I need only with workarounds, at the expense of file size and/or searchable text, correct?

Nermander

Yes, registration black is usually 100% of all 4 colors, because you want the registration marks to show on all 4 plates. That is kind of the whole point of the registration marks.

https://en.wikipedia.org/wiki/Printing_registration