I attempted a transcription of the beginning of Andrew Hill's solo on "Dusk" from the album of the same name. It looks a bit ridiculous on the page, but that was part of the experiment. I wish I had time for more ... maybe another time.
Some quick thoughts on what I found:
The composition “Dusk” itself is only the top line at A and the two-bar bassline that happens under all the written material. Is remarkable for its subtle ambiguities. The bassline and melody on their own produce an ambiguity of modality. Whether we are in D major or F# Phrygian is unclear, though that it is this collection of pitches is never in doubt. At various points, Dmaj or F#phryg seem more prominent, but one never fully wins out. There is an unfinished or open quality to the composition, in this respect. There is also an ambiguity of strong vs weak beats. The piece can be felt as notated; however, just as sensibly, it can be felt shifted one beat to the left, so that the first two notes of the bassline are a pickup. The phrases and bassline are of a nature such that they can be felt either way. Finally, there is also an unpredictability of phrase lengths. The first phrase seems to be about four bars, though perceptually, it is more complicated. The first sub-phrase is 1½ bars, starting in the middle of bar 1. Then there’s a clear 1 bar sub-phrase. And then another, which spills over one 8th-note into the next bar, (and also has a trail in the middle voice in the expanded version to emphasize the unpredictability here). The next phrase is 1½ bars again, but it starts on beat 4, perhaps a pickup to the next bar, perhaps not. We then have another 1 bar phrase, but it sounds like perhaps a 1½ bar phrase again because the final phrase starts in the middle of the following bar. These ambiguities keep us from falling into any obvious pattern, while on a macro level the composition stays in an entirely predictable place.
The arrangement for three horns adds a few more interesting layers. Most prominently is that the three sections, with their developing and building harmonic and rhythmic information, imply an overall formal arc for the piece. The form is vamp-AA-open-BB-open-CC-open-AA-vamp. The improvisations between the sections serve as bridges to get to and from the written sections. The band clearly plays with this awareness. For example, between B and C, they slowly make the 16th-note subdivision more and more prominent, anticipating the importance of 16th-notes in the C section. Another obvious addition that the arrangement makes is to harmonize the line. This is done using many closed position triads, and stacks of 4ths (oftentimes including one augmented 4th). The expansion, like the melody itself, is never entirely predictable in its approach. In B, a slight rhythmic displacement is made, and some inner voices are altered, which travel outside of the mode for the first time. These excursions away from the mode are prepared by Andrew’s piano playing from the opening of the track, which often alludes to the triads of F major and Eb major, as a kind of gravitational pull back towards D major / F# Phrygian. Notes from these same two triads are chosen as the alterations.
The level of listening on the improvisation within the band is astonishing. They are always in time, yet they all, particularly piano and bass, play around the beat, sometimes returning to a phrase more clearly in the beat to show each other that they haven’t gotten off. There are two main ways, in my analysis, that Andrew and Scott get “off” the time. This does not happen in the classic Ornette fashion of playing a phrase that is on, then off, then on the time, like a switch. Here, rather, there is the sense of gradations of playing on and off the time. One way that this is accomplished is by playing very complicated groupings in the triplet subdivision. Examples in the piano are bars 2, 4, and 10-12. The drums almost never play in a triplet subdivision, which helps make the piano and bass triplet rhythms stand out. Another method of seemingly “off” time playing is a subtle speeding up or slowing down—like a kind of phasing—and then getting back on. This is a rubato in the true sense of the word, meaning to “steal” or “rob” time, and then, of course, give it back. These are represented with irrational tuplets of ratios near to integers (e.g. 11:6, 8:7, 11:10, 9:5). These examples all appear in the transcription, whereas ratios that are further from the overall pulse (e.g. 10:7, 8:5) do not appear. (The exception is 12:7 which is another way of writing triplets inside 8:7—in other words, in 12:7 moments, Andrew or Scott is essentially playing sped up triplets, before returning the stolen time).
Another angle on the band’s amazing listening is their magical lining up on the moments that Andrew plays a low fifth in the left hand. These moments come when Andrew seems to have given enough information in the main right hand voice so that the music needs a moment of relief. The band and Andrew seem to want these anchor points from which to spring. As a way of grounding things, he plays these D-A fifths in the LH. Either Scott, Billy, or both of them seem unbelievably aware of when these moments will come. Examples are bars 7 (and/e of 2), 10 (beat 4), and 16 (and of 2).
The process of transcribing made apparent to me two curious things about Andrew’s piano playing. One is about the duration of many of the notes. It seems as if each one has an intentional duration. Very few notes are truly staccato, but most are short. It is almost as if every note has a tenuto, as if his hand is very conscious of going down into the key to produce the sound, though the notes are also short. Very few notes are touched lightly—the touch is always deep into the keys. Also, there are many ghost and other very quiet notes. I have put these into the transcription without accounting for dynamics (only for lack of time). These notes are not necessarily perceived in real time as part of the phrase that he’s playing, but they’re an essential part of the sound world he’s creating, I believe. They cause the listener to lean in to try to hear more of the subtly, and they add to the rhythmic complexity.
Lastly, it is interesting to note how close to the modality both Andrew and Scott stay. Very little of the interest of the improvisation here is gotten through harmonic means, superimposition of key areas, or anything like that. There is the occasional F and Eb major triad allusions, which is important. But most of the interest is made through the use of the mode, and of course, more importantly through the rhythmic play and band interaction.