How can I ...

• May 28, 2015 - 19:15

... import a .pdf of a "song" into MuseScore? Or am I just dreaming.

The thing is, I found a .pdf of "Modern Major General" for voice and piano - great! - but I need to be able to edit it! I wrote some additional words (long term Gilbert and Sullivan tradition) and I want to be able to reprint the score with the words for my (long suffering) accompanist.

yes, I could photocopy the score, clip the systems apart, erase the existing lyrics that I don't want, type up the new lyrics, print them, cut the page into strips, glue the strips in place, re-photocopy the new score, give it to the lovely young lady who's agreed to play for me, then try to fix all the "sort of worked but not really" bits.

In time to sing it this weekend.

Now, she's very good, and will fake it wonderfully if I ask her, but I thought it would be "friendlier" if I could produce a copy she can read.

(Mind you, I haven't even thought about the potential for the new score to be in a different key!!!!!)

So if anyone has any ideas, I'd be very grateful - and stand a much better chance of the wonderful young accompanist agreeing to do something with me again!

Roger Priddle
Tiny, ON, CA


In reply to by Roger Priddle

Not sure what you mean about "windose" - this is a web service, OS shuld have nothing to do with it. What specifically went wrong when you tried?

Finale and SIbelius probably do provide some sort of PDF import feature, and it probably works very poorly there too. It's just a way more complicated problem than people realize. The tehcnology for computer understanding of printed music is basically about where voice recgnition technoogy was a few decades ago. Imagine if someone tried building "Siri" in the 1970's and that's probably about right.

In reply to by Marc Sabatella

Thanks - didn't realize it is a web service. It was asking me to update something (Java? don't remember now...) so I assumed it was windows that was unhappy.

In the mid-80's, there was a machine in use at the university library for the blind that could read text from a page and convert it to voice. I don't suppose it would run on my 286 at the time, but OCR software has improved a lot, and processors have gotten faster, so I sort of assumed that reading and converting printed music would have advanced too.

Not that I'm complaining - if I weren't so lazy I could enter the music into MuseScore myself and edit the lyrics as I go.

But imagine if I'd done that, THEN asked and found out that MuseScore had it covered! Not only would I have wasted a bunch of time, but I would have been mercilessly (and rightfully!) mocked.

Oh well, can't blame me for wishing...


Basically, converting a musical score *to* pdf is usually accomplished by one of two methods.
1. Scanning the score as an image file - i.e. a 'picture' - then saving as pdf.
2. Creating the score using a scorewriter (like MuseScore), then saving as pdf.
To reverse the process and import the pdf score into a scorewriter is not so simple. MuseScore can import XML, so there are ways of converting pdf into XML.

To convert a pdf which was created by method #1:
A scanned pdf (picture) image has no information as to content - a 'picture' can be of anything - so an optical music recognition (OMR) program must identify *all* the score elements - staff lines, notes, clefs, lyrics, etc...
What's even worse is when the scanned original is very old, or a scan of a previous scan, or the pages were skewed when placed on the scanner glass! Imagine telling a machine to identify horizontal staff lines - much less notes, clefs, hairpins, etc. - from such an image. I believe there has been some impressive progress on this front - the OMR Audiveris comes to mind.

To convert a pdf which was created by method #2:
A pdf generated from a scorewriter contains discrete elements. Those elements can be used to 'reconstruct' the score as XML. PDFtoMusic is such a tool. It will reject pdf files created from scanned score images - i.e. from the first method. This type of conversion more often has good results.

So, to determine whether a pdf was made from an image, open it in Adobe and click/drag the mouse over an area of the score. The colored area will show as a solid rectangle if the score is a scanned image:

The above shows a pdf created from a scanned image.
To determine if a pdf was created from a scorewriter, open it in Adobe and click/drag the mouse over an area of the score. The colored area will show as a discrete elements if the score was exported from a score editor software:


The above shows a pdf exported from MuseScore.

Depending on the need, both tools are good to have in the toolbox. In fact, uses Audiveris in its pdf conversion service.


Attachment Size
image based.png 236.59 KB
discrete score elements.png 70.17 KB

Do you still have an unanswered question? Please log in first to post your question.