How to Prepare Orchestra Parts from a Full Score

The problem with the traditional process

You are programming a concert. You find the score as a PDF - maybe from IMSLP for a public-domain work, maybe from the composer, maybe scanned from a library collection. Now you need to hand out parts to thirty players.

The conventional options have not changed much in decades. You can print the full score, photocopy individual pages for each player, and manually collate them - a process that takes a librarian hours and produces results that vary in legibility. You can try to obtain rental parts, which may not exist or may cost more than your budget. Or you can re-enter the entire score in notation software, which is an enormous amount of work and guaranteed to introduce errors.

For contemporary works and commissions, composers often send a score PDF but do not have time to typeset individual parts. For historical repertoire outside the standard library, published parts simply may not be available. In both cases, you are stuck doing the work yourself.

Unscore Me is built for exactly this problem. Give it the full score PDF; it gives you one clean PDF per instrument. The entire process - upload, review, download - takes minutes instead of hours.

Step by step: from score to parts

1. Get a clean PDF of the score. The cleaner the better. If you have a digitally typeset score from Sibelius, Finale, Dorico, or Lilypond, that is ideal. Scanned scores work too - aim for 300 DPI with good contrast. IMSLP PDFs are usually sufficient for detection purposes, though older engraving can be lower quality.
2. Upload to Unscore Me. Drag and drop the file, or click to browse. There is no account required to start - you will only need one to download. The detection pipeline runs automatically. For a 30-instrument, 40-page orchestral score, expect it to finish in 2-3 minutes.
3. Review the detected segments. Click Edit segments to open the editor. You will see boundary lines drawn across the score pages, one per instrument row. Scroll through a few pages to confirm the detection looks right. On cleanly typeset scores it is usually accurate; on scanned scores, some boundaries may need adjusting.
4. Label every instrument. In the sidebar, type the name of each part - "Flute 1", "Flute 2", "Oboe 1", "Oboe 2", and so on down the score. Take the 5-10 minutes to do this thoroughly: the label becomes the instrument name printed in the footer of each part PDF, which is what players will see when they open the file.
5. Handle doubled instruments. Some instruments share a row (e.g., Flute 1 and Flute 2 on the same staff, or Clarinet 1/2 in unison passages). If the two players need the same content, label the row with both names. If you need truly separate parts for colla parte passages, you may need to handle that portion manually.
6. Check grand staves. Harp, piano, and celesta all use two staves with a brace. Unscore Me detects these automatically. In the editor, verify that each grand staff instrument appears as a single row rather than two separate rows. If detection split one, merge the rows manually.
7. Save and generate. Hit Save in the editor. On the score detail page, click Generate parts. The system assembles all parts simultaneously - a few seconds of processing regardless of ensemble size.
8. Download individually or as a ZIP. Download each part separately if you are distributing electronically, or grab the full ZIP to print everything in one pass. The ZIP file contains one clearly named PDF per instrument.

What the parts look like

Each part PDF is built from strips of the original score image - the rows of staves that belong to that instrument, taken from every page and assembled in order. The result looks like a proper engraved part: the original notation quality, continuous across pages, with no content from other instruments visible.

The footer on each page shows the score title and the instrument name you typed in the editor. Page numbers start from 1 for each part. If a rehearsal mark or tempo marking appears above the top staff on a given system, it will be included in the top instrument's strip - this is the standard behaviour for extracted parts from a full score.

One practical point: rehearsal letters and tempo markings that appear only at the top of a system will only be visible in the topmost instrument's part. This is standard in any professionally prepared orchestral part - conductors expect players in the lower staves to identify these marks from context or rehearsal. If you need those marks visible in every part, you would need to annotate the PDFs after the fact in a PDF editor.

Tips for large orchestral scores

• Start with a page check before labelling. Before you type any instrument names, scroll through the full score in the editor to spot detection errors. It is faster to fix boundaries first, then label, than to re-label after moving boundaries around.
• Use the thumbnail strip for navigation. Large scores have many pages. The thumbnail strip along the bottom of the editor lets you jump to any page quickly - useful for checking that boundaries are consistent in places like the end of movements where system layouts sometimes change.
• Check the first and last system of each movement. Transitions between movements sometimes have different system layouts (e.g., a condensed system at the end of a movement). These are the most common places for detection to need manual adjustment.
• Transposing instruments keep their transposition. The extraction is purely image-based. If the score is written in concert pitch, the extracted parts are in concert pitch. If the score is written in transposed notation (which most published orchestral scores are), the parts will be in transposed notation. You get what the score contains.
• The ZIP is useful for your librarian. If someone else is handling printing and distribution, share the ZIP. All the files are clearly named, and each opens correctly in any PDF viewer.