Video: My personal exercises for developing baroque improvisation

[Derek]

Since a couple of you seemed to like my baroque improv and expressed curiosity about it, I've attempted to describe some of the ideas and exercises I've used to help me learn the style:

I hope you enjoy them and if you have any questions I'll try to make more videos and post them at the top of this thread.

For newcomers, just so you know where I'm coming from, you can listen to what I've managed to do thus far in the style here, here, and here.  August 8, part 4 actually has a nice example at the beginning of anchoring the middle voice, something I had a little trouble demonstrating in the above videos.

ajspiano has joined me in this thread in a video discussion of sorts exploring baroque improv.
[ Invalid YouTube link ]

Anyone else interested in doing the same is welcome!

[ajspiano]

I havent watched yet but I'm already salivating with excitement.

This may possibly be the BEST thing I've seen posted here, ever (at least as far as what I like to read). And I've seen a lot of piano-street.

I can't wait to watch these properly, i rather feel as though I'm on my own as far as figuring out this topic. I have plenty of ideas of my own, -but seeing someone else's thoughts is going to be AWESOME. There is literature on composition, but improvising seems like its going to be a lost art unless you can nut it out on your own..   maybe i'm just looking in the wrong places.

I was planning to do a HUGE post soon in the students corner as a kind of "how to improvise on the motifs found in the bach inventions" thing - I will surely have to delay that now as I will be fuelled with many many more ideas.

Thanks for posting these.

[ajspiano]

Ok, so now I've watched! Excellent thoughts. I'm glad to say I've had some similar ideas and also some different ones. I'm terrible at executing them in fluent improv and I think your harmony exercises will be really great for that.

 I haven't at all delved into trying to manage 3 voices yet and really liked your idea for begin to manage more voices by 'fixing' the internal voice.

I wanted to make comment, or perhaps add to what you've posted in some way. Maybe I'll be treading on your toes a little - you've clearly spent a great deal more time practicing this than I have... Perhaps you've thought about these things a great deal and just haven't mentioned them here.

So, pretty much all your exercises/thoughts related to vertical harmony.

You talked a bit about being unsure of where 4th and 5ths fit in relation to that harmony.

...

Basic rule would be make everything super 3rdy and 6thy. This was definately the basic thing I noticed when first analyzing. But they throw in these other intervals and incould never replicate the way they worked well. Mine never sounded great..  Anyway.. Here's a vague largly untested theory as I've really only begun trying to nut this out quite recently. I may be totally wrong..

Key harmonic points - strong beats, 1 and 3 - must be Thirdy

Weak beats can be Thirdy, but they are not as important so can be perfect or even perhaps dissonant, allowing for a tension release relationship with the strong beats and their nice harmonies.

The rules can be thrown out the window in passing if required (perhaps preferably on a weak beat) to follow a motif

Motifs can be melodic OR rhythmic OR both.

You can break harmony rules with motifs because they are predictable by the ear, they can still sound right even when the harmony is dissonant.

Practice idea - set a motif, and make the last note variable, and on a strong beat.. Forgive my use on the ever persistent invention 1..

CDEFDEC (G) - last note is the variable key note, the following motif and underlying chord are defined by the G

Other ways to approach G using the motif..

FGABGAF (G) - doesn't sound so good because of the tritone width.
AGFEGFA (G) - inverted

So you can reduce the invention to a framework, the last note of every motif. And then rather than using bach's motif variations you use your own approaching the key note in different ways, I started by doing inverted beginning a second above the key note for every instance.

Some of them sound a bit funny (but this much is only meant to be very preliminary in developing the skill) and you have to replace the non motif harmonies in the other voice with something a bit more consonant..

Evidently Bach chose really well which variations to use and constructed them to form whole phrases.. I find that easier to do if I change the harmony.. So alter the key notes and go off in a totally different direction. I observed the different ways Bach approached the key note, he certainly hasn't exhausted all possible combinations, and so then I used different ways he'd used at different places..  So open the work with CDEFDEC(B) for example.

I liked this particular variant for the start as an example, without writing out all the harmonies around the motifs
RH - CDEFDEC (A)
LH.                 FEDCEDF (E)
       C major.   F major.   C major
       Motif.       Inversion of first motif

[Derek]

I enjoy analysis of bach and other baroque composers as well, at least as far as I've gone with it informally reading and playing pieces. However, the problem with analysis, whether written by a theorist 100 years ago or an informal forum post by us, is that it produces a very intimidating list. It produces the illusion to those who might aspire to improvise in the style that one has to think about all those details at once while improvising. The truth is, one does not. My goal in developing these exercises was to spend as little time thinking about analysis *while* improvising as possible. That isn't to say that analysis isn't informative---or at least affirming of one's own generative techniques.

One is tempted, and I certainly was for years, to check myself at every juncture while improvising: WWBD? (what would bach do?) Problem is, I hadn't enough experience with playing actual baroque pieces or analysis for this to be effective. In fact---in my case, I've only actually played a handful of baroque pieces. For me, it was an absolute necessity to come up with some simple generative exercises to develop the style. I think it is a misconception for many that the only way to learn baroque improv is to play dozens of Bach's pieces. Of course that can't hurt---but it isn't necessary, either.

Possibly the biggest challenge when learning baroque is to somehow set aside the intimidation that analysis creates, and the intimidation of our worship of Bach in our culture long enough that one allows oneself to cut loose--but yet somehow explore the style in a practical way.  I'm sorta hoping to communicate that in these videos.

I want to add, there's plenty more I could talk about related to the practical exercises and less of analysis---maybe I'll add a third or fourth video in the coming weeks.

[m1469]

Derek,

I've watched your first video and appreciate your ideas and clear delivery very much.  I have only one question so far, and that is when you are doing 2nd and 3rd species counterpoint, are you still trying to "snap" to 3rds, 6ths, 10ths, etc., when the points between hands meet?  I should be able to hear that or see that, I suppose, but I don't automatically.  Thank you for posting these and as I can, I will be viewing more and putting it into practice.

[Derek]

Yes, on the whole I am trying to snap to 3rds, 6ths and 10ths even as the species become more complex. It is true that occasionally in the moment one might land on a dissonance or a perfect interval---this usually provides an opportunity to defer the "snap" til a little bit later. It's only if you keep banging on the perfect interval or dissonance that the ear will react like "hey, where's my next third or sixth? I was looking forward to that!".

Analysis attempts to address this issue with discussion of what intervals happen on what beats, strong or otherwise. I've found during improvisation it to be a bit counterproductive and difficult to think in terms of a strict time signature. In fact sometimes during even a baroque improvisation I'll change whether I'm using groups of 2's or groups of 3's or what have you.

I think of the use of perfect intervals as being things that should not move too much (except in that 6/3 voicing I discussed in the video). They should either come from or progress to another third or another sixth (or perhaps a diminished or augmented interval depending) for change and variety. They're just too noticable. Especially on a clavichord or harpsichord, I might add. On a piano, fifths sound nice and meaty. When you play these intervals on a clavichord, it is almost as though the other voices disappear completely. It is no wonder to me that the common practice composers chose to emphasize thirds and sixths because of the instruments they were using.

[ajspiano]

You are definitely right in saying there is a tendency to be a bit overly analytical. That's not really what I meant though. I meant using the Bach as a model to delevop a basic rule set to follow. I didn't much like the overall compositional approach as it's so detailed that it would be totally impossible to manage in an improvisation. I meant more that the sum of thought may ultimately be this..  For one particular style/idea

Voice 1 - fixed motif | key note | 3rds/6ths

Voice 2 - 3rd/6ths.   | key note | fixed motif

Where "key note" always lands on the 1st or 3rd beat, and sets the general harmony for the next 2 beats. And where the idea of interval quality and parallel motion are largely ignored ( which should be fine if we use 3rds/6ths on the beat, and free intervals on offbeats.

This is obviously really difficult without a lot of work/practice though :/ atleast it is for me.

Another pattern, curtousesy invention 8

Motif             | counter motif
Counter motif | motif

Where "counter motif" is comprised entirely of chord tones (eg FACAFC) - this should allow alterations in the direction of the motif without do much concern for vertical harmony, especially if the primary motif is heavily defined by it's rhythm
....

The other thing I've started doing is improvising in canon - starting really simple, such as no though to harmony, rhythm all in crotchets, one voice a single note ahead of the other. Then moving to more notes ahead, and some fixed rhymic variation.

This seems to really improve my ability to process horizontally rather than vertically. Meaning that I can begin to make better musical decisions because there is a comprehension of what one voice will be doing in the future without having to entirely fix it, rather than thinking only in the moment.

[ajspiano]

Sorry for my incessant rambling.. Just trying to better gauge how to better practice this.. Hoping you can maybe point out any flaws and why it may ultimately not work in florid improv..

To better illustrate what im trying to get at, and make it more accessible - Suppose we start with a really simple fixed motif to make it easier, and allow it's inversion, and exchange between the hands on the beat using thirds or sixths.. Like this.

Voice 1 - cdef(g)     fgab(C).     gfed(C)
Voice 2 -        edcb(A)    ef#ga(B)

(just throwing that combination out quickly without much thought)

So we only really have to process the harmony points for the overall direction? And then we use the vertical harmony skills from your exercise in between..

Maybe you've done this kind of thing already.

[Derek]

I haven't done much conscious thought on how inversion and imitation works in baroque improvisation. However, I can say that I am certain it happens in my improvisations, at least simple imitation at the octave, like in many of the inventions. What's really interesting is that as you improvise a melody sometimes you can find it will fold back on itself and snap to thirds and sixths----on itself, as the imitation is playing. It's hard to describe, it's like a recursive application of the simple principle of wanting to hear thirds and sixths, without any conscious thought. I will try to make more videos to illustrate this, sometime in the next couple of weeks perhaps.

It's certainly not for me to say that any approach will or will not work. I just know that I am not capable of thinking about very many details at once, so my methods are designed to train the unconscious part of my brain. If thinking about a lot of harmonic details at once works for you then that's great.

[ajspiano]

its not that it necessarily works for me.. i'm pretty terrible at it (though I have a sense that I'll be able to do it in time)

I think its just an early exploration into how the hell Bach did it with such complexity. There has to be a method that reduces the amount of thought (like what you're suggesting) because the amount of thought that would go in to producing an good improvised fugue the way a composition text describes would cause my brain to fry..  and in my experience that doesn't mean bach was a genius so much as he must have thought about it differently, and practiced different things.

..and while composition can be mechanical and follow enormous rule set I doubt that's how Bach (or anyone else) did it, more likely they formed an improvised idea of sound, and use theory and rules to tie up loose ends.

The rules were devised based on the compositions, not the other way around..  originally composers can't have been following the rules because they didn't exist.

edit:
I think these kind of things probably need to be explored through all modes/keys etc to create a sense of familiarity with common processes. Then they get called from the subconcious as necessary without thought..  so the thought is just about sound.

[ajspiano]

Analysis attempts to address this issue with discussion of what intervals happen on what beats, strong or otherwise. I've found during improvisation it to be a bit counterproductive and difficult to think in terms of a strict time signature.

Its obviously fairly difficult to communicate, because thats not really at all what I meant by bringing up the use of strong/weak beats.

There's still not necessarily a need to be driven by the exact time signature..  its case of "snaping" to thirds in a more spaced out way, and with rhythm, because the rhythm defines the music far more than the harmony does (at least that's how I see it)..  I wouldn't go about counting it out..  more like..

snap to 3rd/6th --> go on your merry way doing whatever for a short time --> snap to 3rd 6th

so perhaps musical phrases start finish/restart in harmony.. as opposed to in time, and that tends to happen on strong beats.. or at least the feel of the underlying harmony changes tends to happen on strong beats, so that's where you'd place the focus on the harmony rather than the melody.

..which is pretty much what you are doing in your exercise right? just that you didn't go so far as to break it up in bars or half bars or further..  you're putting thirds on every beat, rather than third/x/y/b|third ...or third/complete phrase/third/complete phrase.

........

I don't mean to be argumentative (or post a hole book of inexperienced drivel) but my main issue with your exercise (method?) for me personally is that it appears so heavily focused on harmony, and I'm not sure what the deal is or how to get to improvising a musical melody that fits into the context of a phrase within a section within a piece, rather than 2 meandering lines that harmonize. Which is why I'm inclined to think that having a motif (or several) - and being able to manipulate them freely - is very central to providing structure to both the direction of a melodic line and the overall structure of the improvisation, along side the basic harmonic constructs. Trouble is I guess that it very quickly becomes brain intensive, you start needing harmonic layers..  such as general harmonic structure, as well side 3rds/6ths against a motif.

My view is also fairly confined to study of specific baroque pieces rather than baroque as a whole - so I'm probably making some fairly obvious errors, and will continue to until I look at other examples.

[Derek]

Communication about something this complex is very difficult especially in isolated posts on a forum. I did talk about vertical harmony some in my videos but the whole point of the exercises is to begin building fluency with the horizontal aspect, as you're describing. I will try to explore these topics in future videos. I think all of your ideas are probably quite sound. Given my peculiar way of learning and thinking though, I may not be able to provide much input other than clarification of my personal methods.

I'm certain I improvised a handful of passages which fit the description of third/complete phrase/sixth etc. etc. in those videos---I probably didn't do a good job yet of pointing out that was what I was doing. I may have referred to it as "cool stuff." I'll try again in another video at some point.

[ajspiano]

Given my peculiar way of learning and thinking though

Perhaps not that dissimilar from my own, only perhaps we've gone of on our separate tangents.

I've spent time with exercises VERY similar to the ones you describe only not in the context of baroque music, rather jazz.. so a fairly different set harmonic/rhythmic guidelines..  and minus the level of polyphony..

Interesting too, to see how the experimentation with the baroque improv makes certain aspects of the jazz style seem hilariously easy. The counterpoint does wonders for general improv.

I'm certain I improvised a handful of passages which fit the description of third/complete phrase/sixth etc. etc. in those videos

Probably - since it works musically, I can't say I observed closely enough to know. That's also I guess what potentially comes from your point about having spent significant time playing through them. Time/practice results in finding what sounds good, rather than just mechanically choosing intervals at random.

[Derek]

I think learning any rich style is probably very challenging. I don't know the first thing about jazz. I have a christmas fake book I like to use about one month out of the year, and reading some of those complex chord symbols is quite tricky for me. Going from there to smooth jazz voice leading would be tough for me, not being a big fan of the style.  I doubt baroque is necessarily any harder than any other style---it just has this stigma of being the sole purview of super geniuses, when I think it ought to be accessible to a down to earth pianist who simply loves baroque sounds. Like I said in the video I kinda doubt I'll ever be improvising a triple fugue with augmentation, diminution, retrograde, imitation at the 12th yada yada yada all at once....my personal ultimate goal would be to improvise convincing 3 part (single subject most likely) fugues that sound like some of the slower fugues from the well tempered clavier perhaps, but which might be a little rough around the edges.

[ajspiano]

In jazz I used to do a fair bit of this kind of exercise..  though I've written this to lend itself more toward a contrapuntal work with lines switching between hands, rather than being all in one hand with LH block chords for example..




.......

or with a 3rd voice..  not sure how this sounds, no way to listen (other than in my head and I'm not great at managing multiple voices on sight, so it might not be what i meant ) - here's hoping I haven't created any strikingly obvious problems..  Maybe i'll put something like this through all keys later today and see what I discover



anyway, I'm obviously very much inspired and looking forward to seeing any other videos you put up. Thanks again for sharing.

[Derek]

I added a third video (also in OP), hoping to elaborate a little on the initial exercises.

[pankrpec]

A very nice insight into what you're doing.

I have tried some of these exercises, but it's extraordinarily hard for me. I must go really slow in order to be able to snap the thirds and sixths, even with predictable base. I have no problem snapping the dissonant intervals though. 

What I would like to know is, if you have voices that you play with both hands. What I mean is when you would improvise in three voices, you could play two voices with one hand, as you have shown in one of the videos, and the third voice with the other hand. But have you tried playing the 'middle' voice with both hands somehow?
Perhaps it is only a matter of snapping to thirds I guess, hmm. If you have three voices, to which do you snap? Can you change to which voice you snap and will it still sound good? Damn, I would try it for myself, but I am not competent enough.

[ajspiano]

I'm really enjoying this thread, was great to see another video.

I'm interested to hear how Derek answers but will have a go myself (and I'm hoping to post a video aswell later tonight about what I've been doing in relation to dereks thoughts)

The stronger I've been getting at this (not very strong :p ) I've been practice more and more complex patterns. So rather than using a fixed scale or broken scale I will improvise a free melody and snap to that - which needless to say is a fair bit more difficult. Even more so, which I'm finding hard is frequently changing which voice leads. So the fixed voice is in the RH and the left hand snaps to harmonies, then it switches an the LH leads with the RH snapping having practiced this I then fiddle with the same idea between three voices.

I'm still having a fair bit of trouble controlling moments where I want to send both (or all) voices off in different directions at the same time. I'm also finding that I'm conceiving much better melodies in the RH than the LH and need to work a lot more leading with the LH

Ive been conceiving lots of what I guess are extensions to what Derek has described, each intended to facilitate grasping whatever concept I found difficult.. I'm not sure whether they will be too easy or too hard for anyone else though.

[ted]

Although I have no particular desire to closely imitate baroque, I am keeping an eye on this interesting thread from the viewpoint of general improvisation. Few improvisers in any styles seem to think through the left hand. The good news is that the lack is just habit and is easily corrected with work. In general improvisation I find in recent years that I am much less consciously aware of which hand, or which finger, is playing what. I increasingly employ shared figurations and crossed hands everywhere. The issue of ideas going on at the same time, however, I have not yet fully grasped, in the sense that in many striking sections of my recordings I can hear three, sometimes four, clearly defined ideas progressing simultaneously and independently. It is a very nice effect, but I have absolutely no idea how I do this, least of all while actually playing. This suggests to me, and it is only a guess, that the assumption that contrapuntal improvisation is vastly accelerated contrapuntal composition might be quite wrong, and that we might be looking at two different mental processes which give aurally similar results.

[Derek]

A very nice insight into what you're doing.

I have tried some of these exercises, but it's extraordinarily hard for me. I must go really slow in order to be able to snap the thirds and sixths, even with predictable base. I have no problem snapping the dissonant intervals though. 

Give yourself time. I know it took me several years of very simple two voice playing against a simple bassline to get used to this. Before that, I had gotten very familiar with all major and minor scales and chords through freer romantic/modern improvisation. It is possible that spending some time with that would make it easier to practice these exercises.

What I would like to know is, if you have voices that you play with both hands. What I mean is when you would improvise in three voices, you could play two voices with one hand, as you have shown in one of the videos, and the third voice with the other hand. But have you tried playing the 'middle' voice with both hands somehow?

Yes, I have experimented with juggling the middle voice between both hands. If you've practiced anchoring the middle voice on the "upper voice" of the left hand and the "lower voice" of the right hand, passing it between the hands becomes fairly straightforward. I've practiced anchoring the voice in the lower voice of my right more than the upper voice of my left, so I'm not that great at it yet but I'm making progress.

Perhaps it is only a matter of snapping to thirds I guess, hmm. If you have three voices, to which do you snap? Can you change to which voice you snap and will it still sound good? Damn, I would try it for myself, but I am not competent enough.

What I've noticed as I practice anchoring the middle and upper voices is that different options present themselves at different times. And---as you gain experience with breaking up the other voices rhythmically and melodically, you can expand these options. Before long, snapping to thirds and sixths becomes automatic.

To clarify--snapping to thirds and sixths is probably just my personal term for "making triadic harmony as full and satisfying as possible." When lots of movement is going on, typically we will in fact hear a lot of thirds and sixths. When there's less movement or passing movement, you may hear a fifth---but you care the most about the third and whether it sounds full with the other voices. Interestingly---the clavichord makes it easier to understand voice leading, because the other voices almost seem to "disappear" when you are not treating the third carefully. When I reach a good harmony on the clavichord, the sound seems to almost "light up" in my brain. That might be why I say "snap." On a piano, fifths sound so meaty that this experience is diminished. But the experience of having played the clavichord makes it easier to hear good voice leading when playing on the piano, I think.

I'm really enjoying this thread, was great to see another video.

I'm interested to hear how Derek answers but will have a go myself (and I'm hoping to post a video aswell later tonight about what I've been doing in relation to dereks thoughts)

The stronger I've been getting at this (not very strong :p ) I've been practice more and more complex patterns. So rather than using a fixed scale or broken scale I will improvise a free melody and snap to that - which needless to say is a fair bit more difficult. Even more so, which I'm finding hard is frequently changing which voice leads. So the fixed voice is in the RH and the left hand snaps to harmonies, then it switches an the LH leads with the RH snapping having practiced this I then fiddle with the same idea between three voices.

I'm still having a fair bit of trouble controlling moments where I want to send both (or all) voices off in different directions at the same time. I'm also finding that I'm conceiving much better melodies in the RH than the LH and need to work a lot more leading with the LH

Ive been conceiving lots of what I guess are extensions to what Derek has described, each intended to facilitate grasping whatever concept I found difficult.. I'm not sure whether they will be too easy or too hard for anyone else though.

What I'm finding for myself is that by practicing a wide variety of broken and unbroken melodic movement, anchoring the lower, middle or upper voice, is that it actually makes it easier to improvise against a melody and then juggle it between the hands. I have made numerous attempts to improvise fugues so far, with a fair amount of success. The neat thing is---if you've practiced anchoring a stepwise, stairstepping, chord-arpeggio, etc. etc. in lots of different rhythms and in all voices---it will have trained your reflexes against a fixed voice such that you can come up with any fugue subject you can imagine and harmonize it effectively in the other voices---that's the ultimate goal of these exercises.

This suggests to me, and it is only a guess, that the assumption that contrapuntal improvisation is vastly accelerated contrapuntal composition might be quite wrong, and that we might be looking at two different mental processes which give aurally similar results.

That's quite possible. I have a book by Handel on continuo playing and also the Langloz Manuscript: Fugal Improvisation through figured bass. These books basically give full compositions that just don't have all the parts filled in. For me, I want everything to be spontaneous, from the melody I create to the resulting composition that I improvise around it.

[Derek]

Ted Greene baroque improv on guitar (note to newcomers, this Ted is not the above Ted)

I can't remember when I found this video or whether it influenced my ideas. But it's neat to see someone out there independently have similar ideas. Though---I think this guy thinks about chords a lot more than I do. Probably how we describe a process is less important than how we experience it and how much variety we inject into our playing in order to learn to play spontaneously.

I also wanted to say, its funny how many different attitudes there are out there about learning. I decided around 2001 or 2002 that I wanted to learn to improvise fugues no matter how long it took me. Over 10 years later, I'm getting very close to that goal (in fact I could say I've achieved it, with simple fugues anyway). Not too long ago I found a forum post by somebody saying that they wanted to learn to improvise fugues "this summer." I could be wrong---but I don't think anyone should expect something like that of themselves in a summer. I definitely don't expect anybody who sees these videos to just get it immediately or something. It's more to show you a way you can begin building if it is something you enjoy as much as I do.

[ajspiano]

What I'm finding for myself is that by practicing a wide variety of broken and unbroken melodic movement, anchoring the lower, middle or upper voice, is that it actually makes it easier to improvise against a melody and then juggle it between the hands. I have made numerous attempts to improvise fugues so far, with a fair amount of success. The neat thing is---if you've practiced anchoring a stepwise, stairstepping, chord-arpeggio, etc. etc. in lots of different rhythms and in all voices---it will have trained your reflexes against a fixed voice such that you can come up with any fugue subject you can imagine and harmonize it effectively in the other voices---that's the ultimate goal of these exercises.

Yes that seems likely - I've always found that if you're unfamiliar with these kind of processes you need to start fixed and as you get better then branch off. I suspect that I'm a little impatient with it because I have done very similar things before but with a different stylistic rule set. I've done a fair bit of work with known chord progressions, and improvising with the main targets being 3rds and 7ths above the chords root.

Anyway, I made a video last night covering some of what I've explored in relation to the exercises you presented - as well as some of where I've branched off in my own direction making little discoveries here and there - such as a few different ways you can think about it. I've been fooling around with a different approach to the fixed line one, that being that you can approach it from a "type of motion" perspective, as another early exercise I guess, I talk about that in the latter part of the video.

Its difficult for me to present anything particularly well I feel, I'm not sure enough in my own execution yet..  the other side of it is that I can see/hear a giant mountain of musical possibilities that I can't get my hands and rational brain to put into practice.

For example, in the latter part where I talk about getting to close harmonies, I'm only dealing a few possibilities but I've explored many more and how they work. There's also some obvious harmonic movements presented by your exercise that I didn't approach in the video using the different thought process, but then I go on to use them when I actually improvise a (very) short musical figure toward the end. But they are not concious choices, they just happen - which isnt necessarily a bad thing but it kind of means I've lost control of the improv and its just going in a "comfortable direction/pattern"..   bla bla.. 

something I plan to work on in the near future is also modulating to "near keys" by with I mean, relative minor (or major) and any key that has either 1 more sharp or flat..   such as that if the key is C major, I'm going to develop a range of exercises that result in modulations to A, E and D minor, as well as G and F major. These should turn out to be really valuable, particular the dominant and subdominant ones, as they will be able to work around the cycle of 5ths/4ths to act as a "all keys" improv exercise..

for anyone that wants to watch, the video is nearly 20 minutes long.. apologies for not being more concise (and not knowing much about the topic generally )

When I play it back its a bit out of sync, I seem to be having this problem with youtube and I have no idea why.. can anyone tell me if its like that on their end? If you do have problems you may be able to use the below link to fix (takes a bit of fiddling around but certainly works)
https://youtubeaudiosync.tumblr.com/#watch=https://www.youtube.com/watch?v=gCphbR9lq7M

[ajspiano]

The issue of ideas going on at the same time, however, I have not yet fully grasped, in the sense that in many striking sections of my recordings I can hear three, sometimes four, clearly defined ideas progressing simultaneously and independently. It is a very nice effect, but I have absolutely no idea how I do this, least of all while actually playing. This suggests to me, and it is only a guess, that the assumption that contrapuntal improvisation is vastly accelerated contrapuntal composition might be quite wrong, and that we might be looking at two different mental processes which give aurally similar results.

I'm really getting the feeling that this is the case..  that there is actually just one thing going on mentally, but that it can be percieved as many things aurally..   difficult to explain...   or perhaps there is two things (or more) going on mentally, but in the same way that in non improvisational tasks we learn to put 2 hands together as one task..  mentally we learn to manage several improvisational ideas as one task..   this is probably true in non-contrapuntal situations as well.. 

..if I'm playing block chords with my left hand and a lead in the right, and I improvise the chord changes and the melody for the improvised changes then I'm managing two tasks..  but it feels like one once its working. Perhaps this is no different, its just a slightly more difficult task (or perhaps its just newer to me) with a higher degree of indepentant coordination required..?

[Derek]

ajspiano,

I got a chance to watch your video last night, thanks for posting that! In one part you said you wondered why I say "snap" instead of just "play." Well, I think the reason is that....during an improvisation, I don't really know when exactly I'm going to be describing a third or a sixth. That particular vertical harmony might be displaced in time because of broken rhythm for instance. "snapping" is sort of a vague goal. It doesn't even mean play both tones of the third or sixth at the same time. It just means that it "attracts" simultaneous voices. The exercises you've devised for yourself I'm certain will be just as effective. I can't imagine there's any "wrong" way to approach this.

However, as Ted points out probably the hardest thing to really communicate when discussing improvisation is exactly *how* something new can come out. Baroque presents an interesting challenge in that patterns can easily become locked in physical memory and be hard to break out of (you touched on this in your video ajspiano).

It's possible the simple answer is just to continue to inject randomness amongst order. When I started out as an improviser, the only real "order" I had was knowing some scales and chords. Beyond that, it was almost all "injection of randomness" that led to the experience required to create something musical and interesting on the spot. With baroque improvisation, I've sort of come full circle. I had to create some order (fixed anchor voices or patterns or whatever else), and then inject randomness. It's still a challenge: On some days I'll feel hopeless and just be locked in these well known patterns. On other days, it feels far more malleable and I'm able to break up the rhythm and timing enough to make the familiar patterns sound new. To me, that's essentially what Bach's art was all about. He uses the same patterns and devices all over the place, but each piece is unique and interesting because of broken up melodic and rhythmic timing and displacement of multiple voices.

*edit* To add an idea to the above paragraph---one thing I've sort of learned in recent playing is, with baroque improvisation we are not locked into a time signature. As you begin playing, even in the first few notes---you don't yet know if you're playing with groups of 4, or 3, or 6, or 8, or 12 etc. etc.  So you can literally do almost anything you could imagine, land on any dissonance you can imagine---but if you keep that vague goal in mind of hearing thirds and sixths----you may find with experience and practice with spontaneous play that something new will come out and an arrangement of thirds and sixths will be described (more or less) in a new displacement in time that you hadn't yet thought of before. Not sure if that makes sense.

To put it another way--when we practice fixed patterns, or fixed scales, or fixed anything---the temptation is to NOT break it up. But if we practice taking a leap and messing with the rhythm more than we thought should be possible--it can create startling surprises.  I think of several spots in Bach's sinfonias I've been playing recently where in the middle of the phrase there's some very ugly dissonance but soon there's another full harmony. As an improviser---it doesn't matter when this happens. All the ear hears is a succession of full harmonies---some time. It does not count 123412341234 or 123123123.  What I'm attempting to communicate is that---once one gains enough experience, baroque, paradoxically, will seem and feel just as free as any other improv style. Sounds crazy?

[Derek]

I had another thought. The thing is---when one injects chaos into one's playing---the thing that is so hard to communicate or describe, and even understand for many, is that this cannot be a conscious process.. The only thing our consciousness can do is be informed from past experience as to what opportunities have opened up having just inserted X random phrase or rhythm or melody. The fixed patterns, fixed anchor voices, etc. will not teach you improvisation---they can only give you a framework.

This is the classical problem of teaching improvisation. One always wants to demonstrate a series of "things," or "whats," but actually explaining the "how" of generating something new is quite confounding indeed.

[ajspiano]

Thanks for the reply derek. I think I did understand the meaning of snap - at least in however it relates to me as an improviser. I think the thing I did early on in the video with the suspended notes is a reasonable reflection of the concept. The harmony is there and you can hear where the ear wants it to go even though you are not necessarily playing a 3rd/6th at that moment, opt on the beat. In my brain in often register the harmonic change on beat but resolve to it off beat in my playing - atleast I do when I'm doing something more free than why I did in the video.

I'm finding that I have an significantly improved facility to do this over the last week or so having worked through your ideas and found some of my own.  I still finf that if I just freely improvise I run into "stalls" where I have a bit of a brain meltdown trying to figure out what it is in my head and translate that to the piano.

I've been doing other work with motifs too, last night I tried to have a motif and counter motif work together which resulted in all sorts of mental challenges. It can even be a coordination issue sometimes - things that I'd have no trouble reading and playing but when you have to focus on creating the ideas, and memorizing them without any repetitious practice I guess it drags the brain away from focusing on physical hand coordination.

I'd like to do a bit of a video on that in the future - and also hear from you about anything to do with fugues.. How you go about producing a subject and developing the work.

Edit:
As a side thought, you can add my video to the OP if you like? I know it's your thread so what ever you want to do - but perhaps it can become an index of ideas? I'm sure I'll post again at some stage and it would be handy to have everything in one place... And ofcourse awesome if we had more video contributors.

[Derek]

This is really cool. I'm really happy to have found someone who wants to enthusiastically engage with some of my ideas for learning this style and happy to hear it seems to be helping you. Just in the past few days, I can say that reading your thoughts has helped me as well.

As for "stalls," I think this is a natural part of growth in learning to improvise in this style. What I've found is that as one practices various fixed and broken patterns in different voices, one gets used to using certain ones in a more malleable way. It takes conscious effort to expand the possibilities and actually hit "stalls" quite frequently to grow. I have a bunch of early baroque experiments recorded where I stumble quite a bit. It's only recently I've become a lot more lucky and have been able to make some complete pieces with hardly any of these. It's definitely a challenging style but it's not impossible to learn to improvise in it. I like it because I've never been particularly good at learning to play virtuosically, and while baroque does contain some virtuosic passages, it is by and large a non-virtuosic style (in terms of pure physical dexterity).

Maybe learning baroque improv can be summed up by Stephen Leacock: "I am a great believer in luck, and I find the harder I work the more I have of it."

[ajspiano]

Over the course of my life I've repeatedly received excess praise for my musical abilities, which I don't consider to be that wonderful anyway, even if they were often better than fellow students..  but for whatever ability I do have now (I don't feel that I have any insurmountable technical limitations any more) I'll will gladly attest to it being a great deal more about thoughtful practice than talent. I think Bach is even somewhere on record saying something along the lines of "if one worked as hard as me they would be as good as me".. I'd have to check my references though.

...

I find my stalls occur in certain situations, and it has a lot to do with keeping track of things that happened in the past as I try to add in the imitation element..   as an example, I feel like I should be able to get to the level of improvising any of the 2 voice inventions. Some of which seem distant but accessible..   but No. 2 for example, seems VERY difficult.. the length of the canon is too much.. I can play in canon (and remain in harmony), but only a few notes ahead, and with pretty much zero melodic direction, and certainly no rhythmic variation.

The problem also works in reverse..  when working with a set motif..

for example.. suppose motif ends (in that one moment of processing) on G#. So with the other voice I then begin the next inverted motif (inverted - because I'm that brave) on the B below, because its a 6th..   now while I'm trying to harmonize that new motif I'm also having this difficulty trying to look into the future and figure out either where its going to land at the end (if its fixed) so that I can adjust the upper voice accordingly..  or figure out where I want it to land to fit where I want to upper voice to begin. This also presents processing problems with situations where in advance I know I want the last note of my motif to be "x" which means the first has to be "y", or some other note if its inverted..

I'm sure it would be possible to just be in the moment and do what ever.. But I have a musical idea circling around in my head as is.. I can hear the whole progressing in a particular direction but I don't know where that is on the piano fast enough..  so then I stall. :/

..I've had some success producing short passages with themes/motifs but I've usually practiced my motif in a very linear fashion before trying to improvise on it..  There's another level that's still driving me nuts as to how i'm going to be able to manage it mentally..  that just creating a theme of any reasonable length on the spot and being able to develop it throughout.

anyway, I'll stop crapping on now..  save any more rampant posting till I have something more concrete to contribute.

[Derek]

Improvising canon would surely be very difficult and the difficulty would only increase with the length of the melody, I think. The existing literature we have from the baroque era from what I can tell has lots of information on continuo playing and fugal improvisation, but I am not certain I've seen any information from back then on improvising canon. As an amateur with other hobbies and a time consuming career I don't have time to explore composition on paper very much, so I want to see how far just improvisation can take me. I'm pretty certain it can take me farther than most people believe it can! I think fugues are probably more achievable than canon, at least simpler sorts of fugues with a single subject. They seem freer and there may be more options as one plays than other contrapuntal forms. My goal in this style is to achieve a personal transportative state rather than to go extremely far with compositional technicalities. It already does that, though!

[ajspiano]

It certainly really difficult, - I've been doing a bit of work on (this baroque improv work has been so scattered) just remembering larger improvised subjects, and repeating them immediately either at the octave or the 5th.

And my canon exercise.. which basically has a never ending line repeated 2 notes behind, and the leading line has to stay in harmony with the following one.. which is quite a mental stretch right now. I haven't really tried to analyse canonic works to figure out any movement patterns yet.. invention 2 blows my mind when looking at the idea of improvising it...   But at the same time, I was initially only able to handle a one note difference in canon and can now just barely handle 2 so I can see improvement and have no reason to stick a sealing on it. 2 notes goes ok for a while and then it becomes increasingly more difficult, as if my brain is getting tired.

I'm pretty certain it can take me farther than most people believe it can!

I wouldnt doubt that for a second. I'm already experiencing the ability to predict baroque works and read them far more competently. Something which just playing them never really seemed to do..  Its not as if I know whats going to happen for sure, but when I do get there its like "oh, well obviously that comes next" - and I'm constantly spotting things (many things within a single work) that I've practiced in either exercises I've got from your videos or ones I've been fiddling with myself.

[Derek]

I've been doing a bit of work on (this baroque improv work has been so scattered) just remembering larger improvised subjects, and repeating them immediately either at the octave or the 5th.

This sounds similar to what I'm trying to do for improvised fugues. There are several modes that I find myself using when exploring this style (talking with you ajspiano has helped me figure out what these are and how much I value each):

-free mode: there is no theme that must be repeated at the 5th or anywhere else. I enjoy this type the most--and it can surprise me and themes will bubble up and persist anyway. At present, I have most facility with strong bassline oriented improv, reminiscent of French suites and similar genres. But I'm slowly gaining the ability to improvise effectively in 3 voices. All of my example baroque improv posted here is in this mode. (except for one two voice fugue tucked away in a thread somewhere where ted posted a melody and we all improvised around it. it's not a true fugue though---the second entry I believe is on the octave rather than the more standard 5th)

-building mode: This is where I consciously inject more complex baroque rhythms and phrasings and try to get used to the sorts of situations they get me in in multiple voices. I stall a lot in this mode but that is expected. The exercises I describe in the videos are essentially building mode.

-canon (invention) mode: Now that I think about it I do try to improvise in canon in two voices, with mixed success. Ironically, I think this may be the most difficult of all of these, but also the least satisfying (to me personally)

-fugue mode: This mode feels like a mix of building mode and free mode, because I'm trying to consciously stick to a melody once I have improvised it. I'm having more success with this mode than canon mode, but not quite as much as free mode. I enjoy it almost as much as free mode, and I think it has potential into the future.

[ajspiano]

-free mode:
-building mode:
-canon (invention) mode:
-fugue mode:

Reading this clarified things a lot, I cleary observe these different ideas in my playing now - and perhaps deliberately switching between them at times. It was interesting how similar your descriptions were to how they appear to me..

this especially -

there is no theme that must be repeated at the 5th or anywhere else. I enjoy this type the most--and it can surprise me and themes will bubble up and persist anyway.

..

I think i delve into the idea of fugue mode.. but that I'd be being a little arrogant suggesting that what comes out is infact a fugue, or even close to..

[Derek]

There's a pretty extensive tradition of improvised fugue from the baroque era which from what I understand is much looser than the composed fugue. The nice thing is, most of us can't tell the difference aurally. Here are the two best examples of improvised fugue I've found anywhere:

I like the first of these better, the second guy starts out pretty good but after a while it just becomes really bangy and I think he loses all track of the original subject. At least that I can detect.

Even so, watching both of these guys has really helped me out. Just knowing that other people out there can do this in and of itself is encouraging to me. Hopefully within a year or so I can post a satisfactory (to me) fugal improvisation perhaps in this thread.

*edit* I should mention, the fugue subject that Karst uses is a really nice one to practice with for some reason. I've toyed around a lot with it and variations on it to learn.

[ted]

ajspiano,

All the ear hears is a succession of full harmonies---some time. It does not count 123412341234 or 123123123.  What I'm attempting to communicate is that---once one gains enough experience, baroque, paradoxically, will seem and feel just as free as any other improv style. Sounds crazy?

By accident, I found a very peculiar thing about metrical groups. It is that, if I hear a moderately paced succession of notes, having no accents and no particular melodic phrasal groups, my brain wants to impose 4/4 or 2/4 onto it and not groups of 3 or 5 or some other composite time. I found I could imagine and feel it to have groups of three with concentrated effort, but it surprised me how hard it was. It came about when I wrote a computer program to compose and play fugues (not really baroque ones). Why did everything snap to fours and not threes ? I certainly wrote no particular metre into the code. I still don't know any satisfactory answer to this. Triple time seems to need very much more definition in terms of phrase and accent otherwise four is some sort of aural default. Stumps me entirely this one.

[ted]

I like the first of these better, the second guy starts out pretty good but after a while it just becomes really bangy and I think he loses all track of the original subject. At least that I can detect.

I liked it when the second man went berserk at about 5:40. That's the sort of thing I would be likely to do. I would have liked more of that and fewer octaves, and I think the octaves would have been better lighter and unpedalled. As you say Derek, a wee bit thumpy. But, we all have to play precisely what we like best, no good being someone else.