Roger Hare Posted November 8, 2016 Share Posted November 8, 2016 (edited) There have been a couple of threads here in the recent past which mention the automatic generation of chords by (amongst others) ABCExplorer, eg: http://www.concertina.net/forums/index.php?showtopic=19008&hl= I have finally got around to looking at ABCExplorer and have tried this chord generation facility on a few tunes (which I then play back using EasyABC, which is my preferred program). Does anyone have any comments on the usefulness of this facility? I am not enough of a musician to be able to decide whether the generated chords are altogether suitable, though they sound 'OK' and are making life a little easier for me when playing a tune back for learning purposes. I think I'm right in saying that at least one tunebook available here has used such facilities in the past. Thank you. Roger Edited November 8, 2016 by lachenal74693 Link to comment Share on other sites More sharing options...
Paul_Hardy Posted November 8, 2016 Share Posted November 8, 2016 I have used automatically generated chords in the past as a starting point for adding chords to bare melodies that I'm putting in my tunebooks.I've tried the facilities in a few programs, and my favourite is the old ABCMus from Henrik Norbeck. This seems to do a better job that ABCExplorer, butstill is not very good. I'd be pleased to hear of any better algorithms available. Link to comment Share on other sites More sharing options...
Roger Hare Posted November 8, 2016 Author Share Posted November 8, 2016 (edited) I have used automatically generated chords in the past as a starting point for adding chords to bare melodies that I'm putting in my tunebooks. I've tried the facilities in a few programs, and my favourite is the old ABCMus from Henrik Norbeck. This seems to do a better job that ABCExplorer, but still is not very good. I'd be pleased to hear of any better algorithms available. Thank you. It was your tune books I had in mind when I composed the original post. I think the key phrase in your reply may be "not very good"... I'll investigate ABCMus, though I see it's not been updated for a long while - I wonder if it will run on my 64-bit machine... Thanks again. Roger Edited November 8, 2016 by lachenal74693 Link to comment Share on other sites More sharing options...
JimR Posted November 9, 2016 Share Posted November 9, 2016 ABCmus works on my Win10 64bit machine. Play doesn't work, but I thinks that's my problem. Link to comment Share on other sites More sharing options...
Paul_Hardy Posted November 9, 2016 Share Posted November 9, 2016 ABCMus works OK on my Windows 10 64-bit PC, including play. Link to comment Share on other sites More sharing options...
Jack Campin Posted November 9, 2016 Share Posted November 9, 2016 You need to be able to tell your chord generator what is actually possible with the instrument you're writing for - you don't want impossible keyboard stretches, out-of-range notes or multiple simultaneous notes on the same string. There is an auto-harmonizer in Athanasius Kircher's Musurgia Universalis (1655), using sliding rods with chord pitches on them. Intended for voices, but the problem is that as you get lower the vocal lines have to leap around more and more to fit the rules of harmony. The bass has a really tough time with this one, though it sounds fine: X:8 T:Tristrophon hendecasyllabum Armenicum T:Melothesia Armenica secundi toni S:Athanasius Kircher, Musurgia Universalis (1650), lib.VIII p.136 N:If I understand the Latin correctly he got this from the N:Jesuits and harmonized it using his mechanical rules. N:The second G in voice 3 bar 1 is E natural in my notes; N:dunno if I transcribed it wrong or Kircher had a typo. M:8/2 % C| L:1/4 Q:1/2=60 % guess W: Garrachat parraz nuruche smidaz luhs W: Ansdieziet ariechagan luhs dzachia huchus W: Pirchieziet puguriz sesa puta pirchel. % W: Splendor gloriae renovat mentis lumen W: Increati itaque solis lux orire huic animae W: O Redemptor Universi hanc festina redimere. V:1 V:2 V:3 transpose -12 V:4 transpose -12 K:G dorian [V:1] d2 dd B2 d d2 cAG c2d2 |B2dc B2c B2A AB A2 A2|G2GG A2A B2B GG ^F2G2 || [V:2] G2^FG G2=F F2_EDG F2F2 |D2DE G2A G2F EG E2^F2|D2DD F2F G2F _ED D2D2 || [V:3] B2 AB G2 A D2 GFB A2B2 |d2Bc d2f d2d ^cd2 ^cd2|B2BB c2c _e2d cB A2G2 || [V:4] G,2DG _E2 D B,2CDE F2B,2|B2BA G2F G2D A,G, A,2D2|G2GG F2F _E2B, CG, D2G,2|| I nominate that as one of the first-ever pieces of "world music". Kircher got tunes from Jesuit missionaries and harmonized them himself. (There's another one using an Arabic Christian chant where Western-style harmonization was an even more improbable project). Terrific tune and I don't see any reason why it couldn't work for concertina band, though. Link to comment Share on other sites More sharing options...
Paul_Hardy Posted November 10, 2016 Share Posted November 10, 2016 You need to be able to tell your chord generator what is actually possible with the instrument you're writing for - you don't want impossible keyboard stretches, out-of-range notes or multiple simultaneous notes on the same string. I think you have moved the goalposts a long way if you want multi-voice single-note harmony, or require that the chords are played on the same instrument as the melody. This thread was asking about generating simple chords to accompany melody, usually for a guitar player strumming chords to accompany a concertina. That's a much simpler task, although still not easy. It's also not unique or definitive, as adding different allowable chords to the same melody can change its feel considerably. Link to comment Share on other sites More sharing options...
Jack Campin Posted November 10, 2016 Share Posted November 10, 2016 OK, I was thinking this might be targeted at the Duet or EC. Agreed that's tougher. Link to comment Share on other sites More sharing options...
Roger Hare Posted November 11, 2016 Author Share Posted November 11, 2016 (edited) ...This thread was asking about generating simple chords to accompany melody, usually for a guitar player strumming chords to accompany a concertina.That's a much simpler task, although still not easy.... Yup! That's correct, that is basically what I was asking about, with the underlying idea that the chords could be played either on an accompanying instrument, or on the 'tina itself.. However, JC's contribution was (as always) interesting and thought-provoking. Thank you, Paul and Jack... Roger Edited November 11, 2016 by lachenal74693 Link to comment Share on other sites More sharing options...
Roger Hare Posted December 4, 2016 Author Share Posted December 4, 2016 There have been a couple of threads here in the recent past which mention the automatic generation of chords by (amongst others) ABCExplorer... I'll quote my own OP as a preamble to asking what is (I hope) the final question relating to this automagic chord generation: ABCExplorer allows the generation of one, or two, chords per bar. I just noticed that sometimes the chords generated are completely different, depending upon whether one has selected one or two chords per bar. Is this something I should 'worry' about? Thank you. Roger Link to comment Share on other sites More sharing options...
Paul_Hardy Posted January 6, 2017 Share Posted January 6, 2017 ABCExplorer allows the generation of one, or two, chords per bar. I just noticed that sometimes the chords generated are completely different, depending upon whether one has selected one or two chords per bar. Is this something I should 'worry' about? Yes and no. The algorithms are trying to find a 'best fit' chord against a given set of melody notes, in a given key. For one chord per bar, the chord has to work against all the melody notes in the bar. When they do two chords per bar, each chord has to fit to a smaller set of melody notes, and hence there may be more and different acceptable choices. It's a trade-off between local fit and less chord changes. Indeed, a human arranger may well try and fit a single chord to a complete melody phrase across multiple bars. Certainly when I used the computer chords as a starting point, I then got ABCExplorer to play the tune to me several times, trying different tweaks and in general removing unnecessary chords to simplify the accompaniment. It would be good if some academic organisation could fund one or more PhD theses into researching algorithms for automated chord accompaniments. If I were younger, I'd find it an interesting challenge! Regards, Link to comment Share on other sites More sharing options...
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now