Bach chord voicing proof reader
Hi folks,
I'd like to share a new plugin I published yesterday to get thoughts and feedback.
In a nutshell it allows one to adhere one's SATB score to SATB chord voicings used at least once by the great J.S. Bach
it is rough so...
all feedback apreciated:
https://musescore.org/en/project/check-harmony-against-actual-bach-chor…
Thanks!
Comments
I feel really ambivalent about this, but feedback was asked for. I have developed QML code, and I know how difficult it is and what an achievement it is to make something that works at all, including the necessary understanding of MuseScore data structure to do this, so it is proper to congratulate you.
On the other hand, I’m profoundly knowledgeable about Bach’s music, and his harmonic and contrapuntal techniques, and spend a great deal of time teaching them here, and I have a responsibility to the Thomaskantor not to let errors and stray directions be perpetrated in his name.
In short, this plugin (whose code I have examined), measures chords in midi semitones, and sees if any such stack of three midi-keyboard distances ever appeared in the chorale corpus as subjected to the same analysis. It doesn’t even know the difference between the minor third C to Eb and the augmented second C to D#, and I assure you, Bach certainly did and treated them very differently.
There is zero useful knowledge or direction in the data you have accumulated, nor can one learn or imply anything about the harmonic technique of Bach and the Baroque from it. Understanding of Bach’s chorale style involves (let alone knowing the difference between a D# and Eb) the sequences of harmonies, transition mechanisms between them, and positioning and joining these gestures in the context of theological rhetoric.
Knowing that your stack of midi semitones doesn’t include any not used by Bach, or knowing that it has some, tells you absolutely nothing, especially about Bach. This is like learning painting by measuring how much of each tube of paint an artist used. Enjoy making such a thing, but as an expert in this subject, I can tell you it has no value in learning and understanding.
Bach was not a brand name on a set of rules, or several, or statistics, but one of the signal creators the human race has ever begotten, whose musico-dramatic-theological-technical-emotional language has yet to be equalled.
This is not musicology, but an elementary programming exercise with some humor. Learn about harmony, counterpoint, and the vocabularies of early music, and find a more productive research path.
Sorry, but you asked.
In reply to I feel really ambivalent… by [DELETED] 1831606
BSG,
Now that was feedback!
No need to apologize. it seems your criticism has its place. I do afterall want to offer something usefull to the community and its users, of course!
I will continue to learn the plugin framework and music theory as you suggest.
Like I admitted, this concept is fresh and perhaps, with open discussion here, we can develop something useful somewhat in this same spirit.
Thanks
In reply to BSG, Now that was feedback!… by bliston
Thank you for your extremely kind response, considering my tone. Why don't you learn about "hidden Markov models", and analyze statistically how chords progress to each other over an extended stretch (of course, you will have to consider real tonal pitch-names and intervals, not mere midi distances!). Then you might have something that actually reveals structure. The present concept is a dead end. Good luck!
In reply to Thank you for your extremely… by [DELETED] 1831606
Okay, thank you!
In reply to Okay, thank you! by bliston
I have more thoughts on this. "Neat chord voicings" are an absolutely essential feature of jazz, esp on guitar where all notes aren't readily available, and books of such neat chords and voicings can be had. And one can and should learn from them. And depending upon the style of the piece and the performer, the voicings of adjacent chords in a piece either relate to each other in important ways or not.
In classical music, esp of the Baroque, chord voicings are completely determined by counterpoint aka "voice leading", how the notes of the same "voice" in adjacent chords make independent sense. Although great composers arrive at some neat "voicings" (the concept is not irrelevant), unlike jazz, voicings on their own mean nothing, and the fact that Bach never used this one or that one says nothing about whether it works in a given piece. In jazz, if you stumble upon some great voicing of an Emaj-min-7th ("Hendrix") chord, the fact that it can't be found in the works of Joe Pass or the late Bucky Pizzarelli means nothing. A statistical assay is not a comprehensive dictionary.
In reply to I have more thoughts on this… by [DELETED] 1831606
FWIW, while guitar does indeed impose some particular physical constraints that can result in particular voicings being chosen independently of voice leading, I wouldn't go so far as to say voice leading doesn't matter in jazz. That is, "unlike jazz" in the above is a bit misleading - voicings on their own mean little in jazz in most cases as well. On the other hand, I would also say that there are some respect in which voicings on their own do make sense to consider in all genres. Not so much to say, "this voicing is good even though voice leading is bad", but more to say "even though the voice leading is fine, the overall voicing is not so good". Which is to say, it's more about rejecting (or at least calling attention to, so the composer/arranger can give further thought) particular voicings here and there.
Obvious examples that come up in any beginning SATB-writing context are things like the advice against intervals of more than an octave between two adjacent upper voices. Not a rule set in stone of course, but definitely something one might want to think about. But of course, probably somewhere Bach does this, so simply using a black/white determination based on whether Bach did it ever is not very useful. Pointing out that a particular voicing was not commonly used, though, could be interesting. Then you might look at it in context and ask yourself if there is any actual reason it might not have been used much and whether there is some principle at play and whether there is some alternative to consider.
Anyhow, interesting to think about as an intellectual exercise, not so sure I'd personally recommend it at a practical level.
In reply to FWIW, while guitar does… by Marc Sabatella
All granted, but at very least voicings ought be catalogued measured in real intervals, not midi distance. I said voice leading matters or doesn't matter in jazz depending upon the particular stylistic context.
All good points, though.