FR: Contextual stem direction on the middle line

I guess the algorithm is based on the following note (whole notes ignored)

but perhaps the algorithm should be ‘tuned’ to look at the previous notes too as an option? At X1-X3 I think a stem down would be better (especially since there are previous notes on the middle line with the stem down). All easily fixed (hopefully we will get a Find/Filter function like in Sibelius) by flipping the stem, but perhaps an option for a future version?

(Adding similar bars to bar 2 after bar 2 makes the X1 flip down, so the algorithm is quite clever, looking forward only(?) 1 bar. Dorico is indeed a very clever program.)

There is a setting to have this always default down if that’s what you want. I’ can’t remember at this moment if it’s in notation or engraving options.

Engraving Options > Notes > Stems > Stem Directions
I almost always turn off “Determine by context” as the algorithm produces these kinds of wrong results about as often as not using it results in awkward down-stems.

1 Like

Now I understand why this “contextual” stuff never felt good for me.

I was raised in an “always stem down” culture in Austria, and those oddities made it so much harder to understand why anybody would want that :smiley:


I’ve been having stem direction problems since D5. If I have an A below the B in the middle of the treble clef and I move it up using the up arrow the stem stays up all the way to the top of the staff and vice versa.

Absolutely +1 from me too. I blame Gould for reintroducing this as even an acceptable option. We’ve had 50+ years of unanimity from notation manuals and style guides all saying while stems up was once acceptable here, now only stems down should be used. I don’t think this would even be under discussion without Gould saying stems up is ok.

Virtually all publishers and style guides other than Gould are in the stems down camp:

Boosey (extreme duress LOL)


Kurt Stone

etc. …


Thanks as usual for the citations! Gould over-represents certain English house styles, especially OUP, that to me seem somewhat provincial. But I found this to be true of Sibelius as well, many years before Behind Bars was published.

At least we can set the option as default.

1 Like

Would you prefer 1st or 2nd bar (in organ music, not vocal)?

When a slur is present I think the 2nd is a bit “neater”, but I cannot say for sure it is more correct;

What is your recommendation? Should the middle stem always be down, no matter what?


1 Like

First bar for me, always. Music notation “rules” aren’t necessarily set in stone. They evolve to meet the demands of musicians, and historically it seems like most of the time they evolve to facilitate sightreading. Every little additional unnecessary eye movement, even if it’s just 1/10 of a second, slows down the sightreader.

A fixed inflection point on the staff is consistent and reliable. I don’t even need to read the note on beat 3 in the first bar of your example because I already know it’s on the third line due to the stem flipping directions from the previous note a 2nd lower. In bar 2, because this stem varies from the common practice of the last 50 years, my eye sees the note, then verifies the location with a quick scan. Obviously we’re talking about fractions of a second here, but for a sightreader they add up.

One of the things I think the best copyists/engravers/publishers do is eliminate unnecessary eye movements on the page, especially in music that will be read at performance. (For a keyboard piece that will be practiced until memorized, I’ll grant this is less important.) Clear layout and reliable, predictable notation facilitates this. The reason that this upstemmed style went out of practice IMO, is that it is not reliable and predicable and thus does not help the reader. I’ll prefer the first bar here every time.

1 Like

Fascinating! I was first trained in the Canadian conservatory system (in the '80s) and we were taught the context rule. So, bar 2 seems correct to me. I didn’t realize that this was one of the many British influences on the Canadian system–although we had no truck with crotchets and minims and whatnot.

A few more citations to back up the down-stem rule. Many of these acknowledge the older way of using the middle line as neutral, and say that’s no longer the accepted style.

Ross (1970) pg 83

Read (1979) pg 64

MPA, Standard Music Notation Practice (1993)

Rosenthal, Practical Guide to Music Notation (1962) pgs 3-4

1 Like

These are from a hand copying perspective, but a couple more …

Ken Williams, Music Preparation (1980) pg 10

Clinton Roemer, The Art of Music Copying (1985) pg 17

Thanks for the post and citations. Then I assume I turn OFF the context rule;

and the looks will be more “classical”?

1 Like

I use “Use default direction” and also have the next setting set to Down as well.


Thanks for the expert advice. I have the same settings now.
The Amadeus system never had “context” (and it was made by a German…) so I guess I fell into the “Gould-trap”. Resetting all my scores now in Dorico, and it looks much better.

1 Like

Fwiw, Mats, I very much preferred the “wrong” options you offered.

You have to remember that organ music is an oddity, and the exception to the rule. Voices of dense counterpoint will retain their stem directions regardless of where they fall on the staff: high or low; in other words, these rules do not apply equally to organ music as they do to non-organ music.

To my eye it is quite ugly to only have a single note stemmed the other direction under those slurs. Then again, my eye has been trained to think that any more can look perfectly fine up or down, depending on the context.

Here’s a little Barenreiter Bach, for example.

I don’t see the place in the music. Except for the pedal and the last beat in LH, it’s 2 voices and for 2-voice the stem direction is strictly regulated and can never flip. I might miss something in the music though.

1 Like

I’m not seeing it either. All the voice stemming looks fine to me. This accidental looks pretty sloppy though:


My point was just that it’s perfectly acceptable for the middle line to be stemmed up in organ music, because you can stem up clear to the rafters.

Contextual stems I think are in need of a “consequence report” or “result report” in Dorico (and perhaps many other similar settings). Changing the setting can have dramatic effects on the layout in a score and you have no clue where things change, unless you re-read it bar-by-bar. I think this is a crucial function that has been overseen by the dev. team. The idea to change such a setting late in the engraving process is daunting. Changing a hairpin line width is a minor change as it does not change much in the layout, but flipping stems and slurs etc. are on another level.
It would be nice to have scripting, then clever functions could be developed to help a lot for engravers that care about the highest visual appearance of the music. Possibly a “Compare scores/version” function (also very welcome by many I assume) could catch this as well.