I’ve added some slides and trills to the horn parts, and added a tuba on the bottom. Of course no Prent Rodgers piece would be complete without finger pianos, so I added those as well. They are enveloped to sound more like an organ. Here’s the function table as input to Csound:
f283 0 1025 6 0 384 .5 384 1 128 .5 128 0 ; e15 one hump towards
There are some strange things going on in the last stanza of this piece.
Here are the ratios I used for chords 53 and 54. Relatively small numbered ratios, but also very otherworldly.
I ended up with 15 different notes to play with in the scale. Here they are from B upwards in 72-EDO and ratio approximations.
I’ve completed all 65 chords of the chorale. My goal in choosing pitches was to use as many small number ratios as possible. I need to do some more work on some of the chords. For now, I’m happy with the first 40 chords.
I’ve represented the first chord like this, from bottom to top. The ratios in the first column are from the key of B to this note. The next column to the right are the ratios between the 1st to the 2nd, the 2nd to the 3rd, and so on. The next is the ratio of the 1st to the 3rd, 2nd to the 4th, etc. The last is the 1st to the 4th. This gives me a good overall view of consonance.
The 9th chord in the piece looks like this, with a flatted seventh:
This is a chord not in use during Bach’s time, as far as I know. The 21/16 in the bass is an 8/7 from the key of F#. But it’s a passing tone in this use, and hardly noticeable. In the 15th chord, the first one in the second stanza, it’s in the soprano part and held for a quarter note, so it sticks out. I’m sure Johann would have slapped my hand on that.
When we get to chords 45-58, pictured here, I go way out on a limb, then saw it off.
I picked some weird notes for the D# in chord 52. I’m going to have to go back and redo that section.
This is one of the chorales in the St. Matthew Passion by J.S.Bach. I used to sing it as a soprano at St. Peter’s Church in Albany, NY in the 1960’s. They would wheel the boys out when it was our turn to sing, then wheel us back in as the rest of the numbers were performed by the adults. This is just the first 17 chords, played straight up. Let’s see what it sounds like when it gets the drunkard’s walk treatment. But that’s for another time.
I’m just about finished with this hymn. At this point I am just generating new versions until I find one I like. The piece is scored for French Horn, Trombone, Tuba, Finger Pianos, and vibraphone. The tuning is 72-EDO, using the following notes:
This version goes forwards and backwards in the piece at different times as much as 10 chords at a time, plus much use of the drunkard’s walk markov chain where it’s random whether the next chord is forwards or backwards in the list. This iteration includes chords 1-14 plus 36-42. You never hear 15-35, since the piece is only 4 1/2 minutes long. There seems to be a bug in Csound tempo operation that is limited to 100 tempo changes in a piece. I’ll try to track that down with the developers.
In the mean time, enjoy iteration #6.
I’ve flattened some of the sevenths to make them 2 72-EDO steps lower. That makes the Eb sound better as the seventh of an F Major dominant chord in measures 13, 17, 19, 20, 24, 27, 37, and 41. The rest of the Eb’s are a 4:3 above the Bb major key, and they don’t sound so hot with a flatted Eb. I don’t care for the 21/16 that results from that discord.
Here is another pass, with some proceeding in chronological order, and some as a drunkards walk. Tempos are slowed down a bit. Due to a bug in my code, it was running for an hour, until I discovered what I’d done wrong. The Csound T0 function for setting tempos has a maximum number of points that I was exceeding. I’ll have to look into that.
Recall that the Markov Chain Drunkard’s Walk algorithm is one where the next step is either to left or right in a sequence. In this case, I’ve applied it to the Rock of Ages Hymn. The preprocessor chooses to start with one of the many chords in the hymn, and goes either forward to the next chord in the sequence or back one. Repeat for a while. This was hard to do, but I think the effect is interesting. I might try to use it more, now that I finally figured out how to code it.
I’m starting out a new Hymn. This one is also known as Toplady, after the Reverend Augustus Montague Toplady, who wrote it in 1763. He reportedly conceived of it while sheltering in a rock crevasse or cleft during a storm. Go figure.