Hi all - About a year ago, several of you answered questions of mine about notating a film score. I reached a stopping point with the first cue and learned a bit more about git, so I have the first film cue from the score up on github now:
https://github.com/tunesmith/TheForgivingSea (pdf of score at: https://github.com/tunesmith/TheForgivingSea/blob/master/1M2/pdf/1M2.pdf ) The score was originally prepared (in Sibelius) a few years ago, under the tutelage of a professional hollywood film composer, Hummie Mann, and was performed/recorded by a volunteer orchestra, so it's been through several revisions to tighten it up. This project was to re-notate it in lilypond. It's also a simple, short score, so it might be useful for various learning purposes. I'm also more than happy to receive any suggestions or "pull requests" on how to improve the score or simplify the lilypond coding. Appended below is part of one of the README's - the "unexpected difficulties" I had with lilypond, as in the areas that didn't seem to work easily right away. (Note that I started this re-notation project as a complete lilypond beginner.) A big one here is "giant time signatures", which are common in film scores but don't seem to have a graceful solution in lilypond. Please note that these are just the (minor) complaints, and that overall I was *extremely* happy with the experience and output. I did this all in eclipse (elysium) on a 15" MacBook Pro and loved the "coding" experience. Thanks for all your help, I hope someone finds this useful! Curt ~^~^~^~^ Unexpectedly hard parts of creating this score (all specific to v2.16): - General spacing and staff sizes. I believe Lilypond by default puts everything too close together for music that is read by instrumentalists, particularly sight-readers. The spacing commands are easy to use, but difficult to find and look up if you don't already know them. - I make liberal use of "tagging" for part extraction. It appears this is the best way to handle minor differences between parts and full scores. - Hairpins are surprisingly difficult. Most instruments do not have a natural decay, so hairpins don't necessarily start or end right at the note boundaries. It's necessary to use "fake voices" in these cases. Even with this, it didn't support having a decrescendo end at the Fine bar - I had to make it end at a note value before the Fine bar. And if you have ties over these fake voices, you have to know about \set tieWaitForNote = ##t - Header text elements are a bit bearish to configure. Our instructions were to put the instrument name in the "upper left" of each part; I ended up using the out of the box "poet" slot, and then later reconfigured all of bookTitleMarkup to reposition "instrument" when it became clear I'd need the "instrument" slot for later pages. It also could be easier to put a simple newline in, for longer instrument names. - The alignment of the flat sign in text markup like "Clarinet in Bb" is difficult. I gave up on this one because the approach to make it look right felt too hard-coded. - The "\sustainOff" right-alignment looks bad to me. It should end at the barline or at the rest you stop pedaling; not right afterward. Pedaling usually implies you pedal for the duration of the note, but not longer. - It was difficult to figure out how to create a percussion staff where someone switched from a pitched instrument to a rhythm instrument. Also, I'm not quite convinced on the choice for a percussive half note- I've seen open-heads in these situations before, but I found it impossible to override the notehead in \drummode. - It was extremely hard to specify a subito dynamic right after a hairpin. This is a relatively common use-case, but I had to pull in a pretty complicated scheme function, and modify it, to make it work as expected. This one requirement probably took around six hours. - Making courtesy/cautionary accidentals show up in just the parts was a more verbose process than it needed to be, it seemed. I wasn't able to do this reliably unless I tagged the entire measure. The programmer in me wanted to just tag the cautionary accidental alone. - Fermatas were often misaligned, too close to or colliding with slurs. Manual padding was necessary. - In film scoring, it's common to include the information of the SMPTE timecode of when a last note in a cue gets cut off, for the instruments that are playing at that time. It was not possible to make a \markup element right-align with the final barline. This eventually required a few overrides to Score.RehearsalMark - not too bad, but it felt a bit hackish. - Specifying an arbitrary bar number (like after a long multi-rest) is not supported out of the box, but I found a lovely, concise snippet to help with this at http://lsr.dsi.unimi.it/LSR/Item?id=431 - Eyeglasses are sometimes used in the bottom right of a page to remind an instrumentalist that there really is another page. I had to override rehearsal mark in a few ways to get this to work. - One interesting semi-bug is that top-markup-spacing doesn't seem to apply to 2nd pages (and later pages) of scores, even if they have the instrument name at the top of the page. When I got to two-page parts, I had to rejigger my formatting to use a larger top-margin, introduce top-system-spacing, and reduce top-markup-spacing. - It would be nice if, in a PianoStaff, you could invoke "sustainOn" in the upper staff (for instance if you're in a melody-only section) but still easily have the pedal markings show below the lower staff. - Figuring out large bar numbers was difficult, and it actually required some alpha code that is included in an issue in the lilypond issue tracker. The mailing list was *great* at pointing this out, thanks Nick! - Giant time signatures are actually somewhat common for film scores, but difficult to create in any notation system. Best option I came up with was to jack up the font size and assign them to staff groups. _______________________________________________ lilypond-user mailing list lilypond-user@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-user
