I'm playing Bach's Partita No. 2 in C Minor.

Generally, I lean more towards romantic and contemporary music and when I've been to masterclasses and lectures, I've tended to go to ones on later repertoire. This is, I suppose, the first major early work (before the Classical period) that I have taken on.

Now I have a question about trills and how they should be played in this style. It's actually quite a specific question, but I guess it would apply to other pieces too:

In bar 29 of the Sinfonia, at the end of the first section - the bar directly before the abrupt change to the fugue - the trill in the right hand starts on F#. It is preceeded by an A and followed by a G.

Should the first note of the trill be a G, or the written F#?

I know its written as F#, but I have a vague feeling that a G should be played at the starting note of the trill despite that, as a matter of style/convention, because the note directly preceeding it is the Ab. Is this right or wrong? Or is it purely a matter of taste and what you like the sound of? Does it even really matter? (I suspect it may!)

Thanks to anyone who has some insight,
C x

P.S.: I'm using the Henle Verlag edition.

I don't know this particular piece, but when playing other Bach I have been taught to go from the upper note, so I would play the G as the starting note of the trill. I don't see that it would differ between different Bach pieces, I thought that it is the same for trills up untill the early 20th century, although i may be wrong.

Is it a rule to go from the upper note? Depending on the situation, I sometimes go from the lower note, if that sounds better.

In this case, it sounds good from the G or F#, only that it is the F# that is written.

I have Peters ed and am struggling to line this up.
Partita No2 in c minor, mvmt 1, has 3 sections
1. Grave (French overture style double dot or not decision? and whether to arpeggiate the some chords)
2. 'Andante' (don't know it this is an editorial marking)
3. untitled fugato section

the bar before the fugato is 29 in my edition - there is no a-flat. both 'trills' are marked a mordants, but I think both are usually and should be played as trills starting on the upper note. In such cadential cases the upper note is dissonant to the harmony and starting/dwelling on it emphasis this.

I don't know the piece in question but the general rule for pre-classical music is to start the trill from the note above. I think there is an exception where you start the trill on the note (possibly if the note before was the note above?), but that doesn't happen very often.

The PURPOSE of a trill in Baroque keyboard music was:

1. Create emphasis (on Harpsichord or Organ where varying the volume is not possible)
2. Prolong a note on an instrument of limited sustaining power (Harpsichord, Clavier)
3. Decoration

Emphasis can be created by a momentary dissonance, hence the convention of starting a trill on the upper auxiliary note. For other purposes it is just as effective whatever note it starts on.

On a modern piano trills are no longer necessary for reason 1. A note is easily emphasized by playing it a little louder. So there is no good musical reason for a rule that says always start on the upper note. The reason we might choose to use such a rule is to create an authentic Baroque 'feel' to the music, and that is related to the expectations of the audience based on what they have heard before, what they have been told, and accepted as truth, and historical research into what performers actually did back in the late 1600's and early 1700's.

It all depends on what you, the performer, are trying to achieve in your interpretation.

You would lose the dissonance if you played it the other way round.

Which on a modern piano:

- does not matter if the purpose is purely emphasis, as you can use extra volume for emphasis
- does matter if emphasis through dissonance is part of your concept of a "correct" baroque style

These are very good points, and give me something to think about - maybe it doesn't need to start on the upper note, as a piano does not use trills for emphasis, but rather for decoration and perhaps a Baroque "feel".

As I am playing the trill because I think it sounds prettier with it, that suggests I'm playing it purely for decorative reasons, meaning that maybe it is not imperative to start on the upper note?



My mistake, sorry yes - b.29 is the bar directly preceeding the fugue section and the note is an A, not an Ab.

I wrote the post late last night and wasn't reading straight.

I've correctly it in my original post.

As a general rule - start on the upper note of a trill up until about Haydn. This does mean that if you choose to start on the lower note, to anyone who knows the piece it will sound wrong, it also doesn't always fit right.

It is worth getting the book Keyboard Interpretation by Howard Ferguson.

I should probably mention also (which I completely forgot about when I made the post but having now played it I remembered) that another reason I'm doubtful about starting on the upper note in this instance is that the trill starts on a chord - does that affect it or make a difference in the decision at all?

With reference to the original manuscript, under pain of death, you must start on the upper note. The ornament Bach's used here is the one shown in the bottom right of:

http://icking-music-archive.org/scores/bach/ornaments.jpg

As you can see, it's not just a normal trillo. The vertical line to the left of the the ornament, extending upwards, specifically indicates that the ornament begins with the upper note held for a short amount of time. As such, it's (usefully) more precise than the usual 'put a twiddle here', which raises the sort of discussion you can see above. I'd encourage you to grab a copy of the Bärenreiter edition of the Sechs Partiten, which preserves these ornament symbols - if you're putting the practice time into these works, it's worth putting just a small amount of your time into becoming familiar with the range of ornaments Bach used, so there's less sitting between you and his intentions.

Regarding the note falling on a chord - partly this is irrelevant given the specific ornament used. However, even if it was a Trillo, it would still be okay to start on the upper note - think voices, not chords, and it's an extension of the soprano part that comes before it.

Thanks, this is most helpful.

My Henle edition simply has a "tr" above the note, instead of any symbol, which would have made it explicit exactly what was required.

I was going to say, if you're playing it for an exam, then the examiner would expect you to trill from the upper note, as that would be idiomatic for Bach's music.

Then I remembered playing in a performance of Bach last year, wincing through the first playthrough as all the strings players failed to do that , then later sitting through a baroque workshop for which most of the musicians also didn't know this one. Is it really so little taught, about the baroque trills? A small section of classical musical history maybe, but still a significant one. I know that I'm probably being over fussy, as a baroque specialist, but also cannot remember when I was first taught that the trills were the other way up to classical ones and later.

What I didn't know, until relatively recently, was that very early baroque trills, say from about 1625 or so, are more usually started from the lower note . Arggh!

It's funny because I was always taught trills go from the note above, and it wasn't until much much much later in life, i.e. after my 15+ year hiatus, that I discovered that trills can in fact start on the lower note. Spot the recorderist...
(though now I wonder about any more modern, i.e. not baroque, stuff I played, did they have trills in, and if so was I playing them wrong...)

I don't understand this need to find cut and dried rules from some higher authority in order to do the "correct" thing.

Surely you should study all you can to understand what the composer was trying to achieve and why, what was conventional at the time and why, what is generally done nowadays and why and then, with all this background knowledge, and with a clear idea of what you, as the recreative artist are trying to project, and who is going to hear it, make up your own mind.

Bach didn't indicate phrasing or tempo, and left it to the performer to decide where to play legato and where to play staccato. Also his keyboards were not capable of gradation of dynamics or isolated accents either, but careful use of dynamics can clarify the lines in a fugal style of composition. A Bach score is closer in spirit to the script of a play than an engineering blueprint or a computer program.

In fact there is no almighty power checking up that we all do our best to arrive at the one correct way of performing each piece. You don't even have to play the piece the same way in every performance. You can vary tempo, phrasing, dynamics, just for the sake of variety, and reveal different aspects of the piece each time.

If the performer is to be trusted with the massive difference that those four dimensions can make to a performance I don't see why you should not also play the trill as you see fit for your conception of the piece, on a modern instrument for some particular audience.

In computer programming there is a languiage called PERL that has the motto "There is more than one way to do it". That would be a pretty good motto for interpreters of Bach.

I agree with what you have said entirely Mad Tom. No one has said that there at any time that there is a "need to find cut and dried rules from some higher authority in order to do the "correct" thing".

All that I am doing is finding out what people here think is general convention for this type of trill, as I don't have time over the next couple of days to go down to my music library. And once I have found out what is generally expected of trills of this sort, I will take that into account and play around with it until I find what fits best with me for my performance of the music, which may or may not be what has been described in these posts, or indeed what is the 'norm'.

Hence Knit one, Perl one??

And the comment was not directed at you!

Excellent.

Does PERL also point out that while there is more than one way to do it right there are an infinity of ways to do it wrong?

It leaves you to discover that for yourself. Which does not take long.

I'm surprised you made this slip, Tom - Bach's love of the clavichord is well known, in part because it is capable of dynamics and accents. If you haven't had chance to play one, I'd encourage you to have a go - from my experience, early 18th century clavichords are more than capable of providing a flexible palette on which to base ones renditions of Bach. For further reading, have a look at an article from Early Music:

http://em.oxfordjournals.org/cgi/pdf_extract/XXVIII/1/140


Yes and no. Yes, insofar as interpretation is of course needed; but no in that Bach had less intention for people to Richard around with his music than others such as Telemann; e.g. he writes out many more of his ornaments and cadential passages in full. But then, if your argument is that we can wrap his remains in copper wire, place a magnet at his headstone, and generate electricity by making him spin in his grave, I suppose you can do whatever you want .


Other than examiners.


But then again, unless one runs perl -w, it won't warn when performing string concatenation with undefined variables. As such, it's generally a good idea to run -w at least once in a while; and here, similarly, to subject ones interpretation to closer scrutiny.

Should that read "There is more than one way to obfuscate things"?

Yuck, Perl... I avoid at all costs unless I need to do date+time calculations or something similar bash scripting/hacking FTW

Yes, I have heard precisely that rule too.

Chopin is on record as having criticised Hummel for "the modern habit of starting the trill on the note", which suggests that the convention transferred from the Baroque keyboards to the Classical piano.

Especially, but not only, when the trill is on the leading note at a cadence leading to the tonic, there appears to have been a convention to finish a long trill with a turn to the note below, e.g.

....CBCBCBAB|C--

The turn was typically not notated in the orchestral parts of the Beethoven symphonies, but Liszt added this figure (a "nachslag") to at least some of Beethoven's trills in his piano duet versions of these works, written some decades later for amateurs who, he suspected, would not have known the convention.

C P E Bach recommends that the nachslag be notated in full but also shows it indicated by appending a mordant sign to a trill sign (the sawtooth, not "tr"); my opinion is that JSB would have indicated this where he wanted it.

Disclaimer: everything that follows is a statement of my opinion. I may be wrong as I am not a Baroque expert.

In Bach's case, his ornamentation rarely leaves any doubt as to how to start a trill. YAP's screenshot looks like the same one I have at the start of my edition of the Goldberg Variations. Follow Bach's directions and I will not go far wrong, is the principle I adopt.

Trills in general sound better when started on the upper note throughout the Baroque, Classical and early Romantic periods. Baroque trills sound even better when the first and last notes are longer than the intervening ones - a frequent characteristic of Haydn's trills as well - and the quicker notes are measured rather than a frantic wafting of the fingers.

Smashing thread this one, cadence. It is great to read other opinions about this subject.

This way overstates the case ...
... but this ...

Is rather more balanced and seems to sum it the practical choice: Usually starting on the upper note sounds better (but now and again it does not).

Is this one big wind-up? My post was on the topic of the ornament on the last beat in a bar in the first movement of Bach's second partita. Original source clearly indicates an ornament starting with a lengthened upper note - the | before the trillo symbol leaves precisely no ambiguity in this regard. This symbol was not preserved in the original poster's edition, having been replaced with the entirely unhelpful tr by an over-zealous editor, and from thence the confusion arose.

PDs reply highlights something else I'd always done automatically without understanding it: the concept of a degree of expressiveness to a trill, rather than the frantic as-short-a-subdivision-of-the-beat-as-can-be-managed version also often heard; for those trills ending in a turns, that's another place where it works better if the turn is not just dismissively shot through at top speed. Maybe that is part of the problem with modern perception of baroque music - that it is rigid, strictly conforming to the text and lacking in emotion: it is none of these things. Rubato as found in later music doesn't work, neither does an overlaid thick and shapeless vibrato, but that does not mean you can't be expressive with it, and maybe those who dismiss Bach's music as just a lot of notes haven't really heard how it can be done to proper effect.

My teacher refers to the later as "electrified" trills (the result of the musician having stuck his fingers in an electric socket)

This is your instinctive understanding of how to play trills expressively. A good trill is like a good book in having a beinning , middle and end.

A Baroque trill is most likely to start with a long upper note, followed by a measured alternation, finishing on the melody note at roughly the same length as the first.

A Classical trill is most likely to start on the upper note but without the pause. The trill starts more slowly and accelerates, then decelerates towards a turn at the end.

Trills were used less frequently during subsequent years. Chopin was a master of them, and usually used grace notes to tell players how to start and end them.

All of these points are generalisations of more-or-less accepted practices. Of one thing there is no doubt; a beautifully-played trill is a joy to hear.

Wind up? No. Just a comment on "under pain of death"

But you are right (in a different post) that I had forgotten about the expressiveness of the Clavichord - I was thinking only of Harpsichord and Organ

post deleted

Starting on the written note here, though, would amount to completely ignoring an appoggiatura. The symbol used on the score is:



...which is short-hand for an appoggiatura joined onto trill. Whatever ones take is on playing trills, appoggiature cannot just be ignored willy nilly, so one is clearly to start on the note above that written in this case.

Death seems a little harsh. The worst punishment that even I have ever suggested is 8 hours of Hanon daily

One simply has to instill a sense of discipline into people.

And, as we all know, Bach is God's favourite composer - so He could get pretty annoyed by incorrect interpretations - there could be plagues, curses...

Or, heaven forbid, an ego that struggles to get through door frames .

What on earth would a student of Bach be doing with Hanon?

Nothing. He\she pre-dated Hanon.

If you have an uneven floor it can help to stop the piano stool wobbling

aha, so that's what it's useful for . I did wonder...

How about replacing broken tiles on the roof?

Yes, frequently, if the upper note is played immediately before, it sounds much better to begin the ornament on the note itself. This is fairly common practice and - more importantly - was then, by all accounts. See C.P.E. Bach et al.


Which just goes to show that, in this instance, Bach was deliberately aiming for the opposite to the convention noted above.

I thing it still sounds better when the trill starts on the upper note in this situation.

I treat the preceeding upper note as an anticipation of the trill, as in an anticipation at a cadence - hope that makes sense.

'Thing' whatever you like, mate! If you read the whole of my post, you'll see that that is actually what I say about this particular piece!

I think, however, that musicians are often still too preoccupied with the dogma put forth that Baroque ornaments should always begin on the auxiliary, rather than the note itself and, when we hear ornaments not starting on the auxiliary, it is somehow perceived as being 'wrong'.

However, that said, I think that, as YAP has said, Bach's original ornamentation (if indeed it were written by him, but that's perhaps another thread for another day.... ) is quite emphatic that the trill should, indeed, start on the upper note in this particular instance.

Oops. Confusing my thinks with my thingys. Could be nasty.