Part of Scores of Scores.

Notes for Transcribers

Welcome!

Thanks for participating in this very exciting encoding project.

In this project you’ll find a wonderful corpus of songs … or rather just the melodic line of those songs, desperately in need of piano parts. That’s where you come in! Please encode the piano parts to complete the songs.

Basically this is a normal OpenScore encoding project, except that here some of the job has been done for you! Simply:

  1. Pick an available score from the spreadsheet.
  2. Update the spreadsheet to assign yourself (add your MuseScore username in a comment).
  3. Find the template score via the “Template score URL” link. The score’s description will include a link to a recommended IMSLP source score (a PDF).
  4. Open the template score in MuseScore and encode away!
  5. Upload your finished score to the Lieder Corpus. Update the spreadsheet (via a comment) to request a review.
  6. Rinse and repeat!

FAQs

Why are we doing this repertoire?

This is an important candidate repertoire for transcription for many reasons:

  • The repertoire needs to exist in multiple keys to support different singers’ ranges. Encoded formats give full flexibility.

  • It is a key repertoire for the teaching of music theory for various reasons (about which see also our parallel work-in-progress Project 2, Cut Outs, not least of which being that it’s a great way to improve the representation of women composers on the syllabus.

  • The encoding of vocal lines and text is already well underway thanks to another, previous para-academic project which has provide the starting point here.

Where have these vocal lines come from?

We have inherited this corpus of vocal lines from Leigh Van Handel et al. who created these partial scores in another file format for a research project. Leigh has kindly given us permission to use them as the basis for our present goal of creating full encodings (thanks Leigh!).

Are they perfect?

Unfortunately not. Converting file formats can be hazardous business. I have used some great existing code for converting these files (thanks Myke Cuthbert, Craig Sapp, et al.!), written some more to fix certain problems unique to these pieces, and fixed a lot of little errors, but it’s hard to get a 100% success rate when transcribing, or when converting, let alone when doing both! Look out for the occasional:

  • Odd lyrics, e.g. with } or a number.
  • More odd lyrics: with accents that should / shouldn’t be there.
  • Ghostly second voices that shouldn’t be there.
  • Errors with grace notes.
  • Bars rest missing (very common).

Are the vocal lines complete?

Often not. Please do encode the rest of the melodic line and text if you’re happy to do so. If the song is long, then perhaps focus on another, shorter piece first so we can get more songs finished. If all of the music is there but not all of the verses, then that’s fine for now – thanks! We’re aiming for complete music.

Why aren’t there piano parts?

That simply wasn’t part of the Van Handel et al. study. Anyway, we had to leave some of the fun for you!

Is there an Editorial ‘policy’?

Absolutely … of a sort. See below.

IMSLP issues

Your job is to copy the given IMSLP edition for which an URL should be provided. If there is no URL, or the URL is for the wrong piece, then we’ve messed up: please accept our apologies and get in touch with your designated driver … I mean reviewer.

If the URL is fine, but you simply don’t like the edition chosen then either 1) grit your teeth and bear it, or 2) suggest an alternative edition that is also on IMSLP and in the public domain. We will be happy to consider making the change except where the work is part of a wider collection and transcribing work has already begun. In that case, changing edition would cause too much disruption to other transcribers.

Can’t I just do my own edition?

Definitely, but not for this project. For various reasons, we need the Lieder corpus songs to match the listed IMSLP edition as closely as practically possible. We hope and expect that these transcriptions will be used for more scholarly editorial work down the line, but that is not the present task.

What key should I use?

Often, IMSLP will have the same song in multiple keys. We’ve tried to upload scores in the original key and direct you to a page that matches this so you can transcribe in the key of the IMSLP source. If there’s a mismatch then please get in touch.

This score is corrupted / unacceptably full or errors.

Sorry about that. Let us know the offending score’s name and we’ll see if it’s redeemable / speaks to a wider problem. In the meantime, there are plenty more!

Got any timesaving tips?

I sure do! The main one is to look out for any repetitions, either exact or partial. If a whole section is repeated then that’s obvious: copy and paste! Often you’ll find more equivocal cases. For instance, many of these songs have a repeating accompaniment pattern; it may be worth copying one over and changing the pitches if the pattern is exact, the voice leading between notes is close, and you are careful to work through systematically. Otherwise copy and paste might cause more problems than it’s worth.

Is this good for me?

Absolutely! Apart from helping the music community, encoders benefit from the task of copy-editing music itself – musicians have written out scores as part of their education for as long as records exist. The story of Wagner copying out Beethoven’s Ninth Symphony twice stands as a famous testament to this. If you want to take the pedagogical side further, then check out these suggested exercises.

I eat 1,500 songs for breakfast, when are we going to do the rest?

Wow. Ok. A few thoughts:

  • Let’s knuckle down and do these ones first please.
  • We are definitely very happy to accept other French and German 19th century lieder that are not on the current list if they’re on IMSLP in a public domain edition.
  • We don’t mind you uploading other public domain lieder, but we might not have time to review them right now.

This project is so awesome, I want to help in other ways

Great! Please get in touch if you want to help review, or if you’d be happy to help in other ways, like uploading to IMSLP public domain editions of scores which are not currently on that site, but which we have in the repository. Finally, anything you can do to help publicise this would be really great, not just among other transcribers but among everyone that stands to benefit from knowing about this corpus.

Editorial ‘policy’

‘Policy’ is in quotes here because we are simply aiming for a pragmatic solution with minimal adjustments to reach a plausible editorial level. That means following the given IMSLP source as exactly as you can for the new piano part and leaving the vocal line as it is. Here is a summary of what you will find in the vocal line (and should do in any new encoding):

  • Separate stems (no beams) for notes with one syllable (though leave beams if they’re already in and make musical sense).
  • No melisma continuation lines required (though again, leave them if they’re already in)
  • Umlauts and all other diacritical marks are in but we use ss rather than the Eszett ß.
  • We’ve used the ‘en-dash’ for the literal dash in French lyrics (e.g. ‘veux-tu’) to avoid it being read as the kind of dash used for ‘syl-la-bic’ joins during the file format conversion, though these can be corrected to dash in MuseScore using ‘Ctrl+Hyphen’ if you’re feeling very editorial! Don’t worry if you don’t follow.

We hope you find these acceptable at least for the time being. Most of them we can automate a fix for if we choose to later (e.g. after the main encoding drive). If you are sure you’ve found something that doesn’t correspond to this policy then please do make the change.

If in doubt, refer to OpenScore’s general Tips for Transcribers or get in touch with your reviewer, particularly if you discover consistent errors or editorial policy decisions that should be on this list!

Who’s Who?

The wonderful team of Lieder Corpus Project Managers includes:

The Supervisors are:

  • Peter Jonas (shoogle) - MuseScore / OpenScore Manager
  • Mark Gotham - Principal Investigator

Many thanks for being a part of this initiative. Here’s to a great encoded corpus in the making!