I have attached two images,
* The first 2 pages of a chart I did in lilypond
* The first page of the chart my Nashville producer created and was used in the recording session
There may be info on the Nashville chart that derives from having listened to a pre-existing version of the song, as well as looking at the lilypond chart.
Here is the recording:
https://soundcloud.com/david-elaine-alt/octagon-arbiter-anthem> Regarding the Nashville system, the ruleset is as follows:
> FORMAL:
>
> Key, time signature, feel, tempo, and title are stated on top of chart.
This could be accomplished using existing header features.
> Every section gets a new line, with the section name on the left.
Can be done with existing mark technology.
> Lines in sections may break after 4 or 8 bars (or however long the
> phrase is)
one can already break, or not, as desired
> HARMONIC:
>
> Chord roots and inversions are given as scale degrees, e.g. 1/3 for a
> first-inversion major tonic.
Slash notation is part of lilypond chord symbols.
If the feature of using numerals instead of note names were implemented, I would assume the slash notation would work, too.
> Chord type is stated as in jazz notation, i.e. 1-maj7 for a minor/major
> tonic.
As is my opinion of all styles of chord symbol notation, you will likely have to customize them to get the right look for your charts.
But that is not a fault of lilypond, it is due to the non-standard and wide-ranging set of practices that is chord symbol notation, such that there is no default that speaks to everybody.
In particular, for your example, I would not expect mixed use of symbols "-" and letters "maj" on the same chart.
But, in general, the lilypond chord symbol features can accommodate anything you may need in this regard.
However, there is one other Nashville rule that would need to be accommodated, which is that for minor, the numbers still follow the relative major.
So, for a song in A minor that goes A- D- A- E7 A- it is not 1- 4- 1- 57 1- as you might expect, but is rather 6- 2- 6- 37 6-
That also seems like something that would be possible if the general problem of using numerals instead of note names were done. In concept, it would just be another transformation, among which many are already going on.
> Measures where no chord is being played (general pauses) are denoted X.
lilypond has a no-chord feature, you can choose whether to print something, or not.
> Modulations are stated by the new tonic as superscript (Key of Eb) or
> relatively (Key: fourth up).
This brings up a lot of thoughts.
In general, I would say that this is something you can accomplish using markup.
This is based on the observation that in your example, this text varies. So, you will be writing some kind of custom text for it anyway, it is not something that it seems that you would expect should be done automatically.
Likewise, I am not sure of your intended use of your charts.
If it is for actual Nashville sessions for popular music,
I would expect modulations would already be occurring at sections,
which would start on a new line with left-aligned section titles, and this info would be added there?
On a more philosophical level, the general adoption of the Nashville system is designed to accommodate key changes for singers.
As such, the musical part of the chart is supposed to be absent of keys, and the mention of key would be header info.
> RHYTHMIC:
>
> Held chords (let ring) are denoted by a diamond around the chord symbol.
>
> Diamonds can be tied.
>
> Stops/breaks are denoted with superscript marcato: ^
Regarding all these, and the anticipations below, these rhythmic elements all seem to exist in a context above the chord symbols.
(I have certainly also seen the diamonds around the symbols, but for now let's just consider the above-the-symbols part, which is necessary anyway, and can also accommodate the held notes diamonds)
So, in terms of concept, I think that what could be feasible
is to take lilypond chord input, and convert it to two things:
* a music _expression_ with rhythmic content that expresses the stops, holds, anticipations and delays
* a chord _expression_ that is the same as the input, except the rhythms are quantized
These are then printed one above the other.
There is a lot more to say about each of these.
I could help with the algorithm,
but would need help with the lilypond coding,
in terms of consuming the original chord content and producing the rhythmic _expression_ and quantized chord _expression_.
Likewise with designing the semantic content of the rhythmic notation, and then rendering the symbols.
I have ideas, but would want the input of others.
I think once we start breaking this problem down into smaller parts,
they will mostly be things others can help with.
> When there are multiple chords per measure, chords of one measure are
> underlined. There are no bar lines in Nashville charts.
In the charts produced for my sessions, multiple chords per measure were separated by slashes, not by underlining.
As per your requirement, this could be accomplished by a third context, that only contains these underlines, below the chord symbols.
So, I would have to amend my above suggestion to having a function that takes chords and produces 3 expressions, rhythm above, chord symbols, and rhythm below.
> If the chords in one measure aren't held for equal amounts of time, each
> chord on the underline receives as many superscript dots as it has beats.
In my examples, more complicated rhythms are notated in regular western fashion.
As per your requirement, this would be another part of the algo that produces the top rhythmic line.
> If the chord change is anticipated by 1/8th, this is denoted by
> superscript <. If the anticipation is by a 16th, the superscript is «.
> This has to be the leftmost superscript over that chord symbol.
>
> If the chord change is delayed by 1/16th (rare), this is denoted by
> superscript ». This has to be the rightmost superscript over that chord
> symbol.
This would also be part of the job of the function to produce the upper rhythmic line.
> One-off measures with a different number of beats are notated in
> parentheses
I'm not as familiar with the time sig features, but I think there are already ways to tweak the time signature presentation, and this could be done using those techniques.
> and the chord symbol within the parentheses receives as many
> superscript dots as the measure has beats.
Another requirement of the algo regarding the top rhthmic line.
> That's it. Maybe this helps gauge how much effort this would be.
>
> Cheers,
>
> Amy
Producer ~ Composer ~ Instrumentalist ~ Educator
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
lilypond-octagon-pages-1-and-2.jpg
nashville-octagon-page-1.jpg