An object which draws and optionally plays a trill ornament on the staff.
Paste the following snippet into NWC 2.75a, then go to Tools > Manage Objects to install the object.
!NoteWorthyComposerClip(2.75,Single)
|Clef|Type:Treble
|TimeSig|Signature:4/4
|User|Trill.ms|Pos:9|AccStyle:2|Span:1|WhichFirst:Auxiliary
|Note|Dur:Half|Pos:1|Opts:Muted
|User|Trill.ms|Pos:9|AccStyle:3|Accidental:Flat|LineType:Jagged|Span:1
|Note|Dur:Half|Pos:1|Opts:Muted
|Bar
|User|Trill.ms|Pos:9|AccStyle:2|Accidental:Sharp|LineType:Jagged
|Note|Dur:Half|Pos:2|Opts:Muted
|User|Trill.ms|Pos:9|Accidental:"Double Flat"|LineType:Jagged
|Note|Dur:Half|Pos:0|Opts:Muted
!NoteWorthyComposerClip-End
For details on the object's parameters, refer to the context-sensitive help on object's properties page.
Looks good, Mike. Thanks. Now if you can just make it play....
One comment: it's a bit difficult to see the accidental in the second trill in your example. I think it's more common, when an accidental is required, to place it above the beginning of the wavy line, where it can be seen better, rather than between the "tr" and the wavy line. If you keep it between, perhaps you could give the accidental a little more breathing room.
But what style ? I talked with Mike about this a while back. I believe (I stand to be corrected) that there are Classic, Romantic and Baroque styles. There also seem to be a multitude of rules for each type. ie where they start, above, below, how far, how they end, should there be a turn at the end, should the trills be tripletised etc etc. Trills can even vary by composer as I understand it. ie a classical trill for one composer may not necessarily be the same as a classical trill found in a score of another composer.
So, Bill, how about laying down some rules for different types of trill and how they should be played. I know this would help Mike and it may help many others too (myself included).
Thanks for joining the discussion, Rich. As I may have mentioned in other threads, I am not well versed in the types of trills, so most of my knowledge on their appearance and playback are from Alfred's and other online resources. It will probably not be difficult to implement most features that are needed, but we will want to keep the 80-20 rule in mind.
Since this is still a 'test' object, we can use this thread for further discussion and suggestions. If anyone would like to post relevant screen shots or URLs with more information, I think that would be helpful.
One thing I wanted to mention here is that I took a slightly different approach to how I draw the trill, compared with the Arpeggio object (in some ways, drawn trills look a bit like 'sideways arpeggios'). I am using characters from the NWC2 font for the 'squiggle', rather than drawing them from scratch using Bezier curves. This means it should be possible for SwingDings fans to have appropriately styled trills. I might revisit this approach for Arpeggio at some point, since I wasn't aware that there was a suitable character present in that font, that I could have used instead of the curve approach I chose.
Thanks,
Mike
I am far from being an expert on trills. It's my impression, though, that differences in how they are notated from era to era matter less than differences in how they are played in music of different eras and different styles. In other words, how a trill looks in one of Bach's manuscripts may be quite different from how the same trill looks in a modern edition of the same piece, but they are to be played the same way. So I think all that is necessary, at least in the beginning, is to produce something simple and clear that matches basic contemporary notation practice.
The main things a player needs to know about a trill are (a) when it starts; (b) when it ends; and (c) whether to trill a half-step or a whole step. Sometimes it is also necessary to know whether to start the trill on the upper or lower note. I may get flack from the purists for this, but I think anything beyond those parameters should be left for the performer and/or his or her teacher to mark in, using their own understanding of the musical style they are working in.
There are still, of course, some practical considerations. Does the "tr" go over the note, or does it come before the note so the wavy line can start over the note? I've seen it both ways; I think that's an editorial decision, not a fixed practice. Choose one and stick with it. Does the wavy line even exist for trills that are only one note long? Usually not, but it might be useful to be able to choose (it could be that you would want one for a trill on a breve played against a long melismata, for instance). How are accidentals to be notated, if the trill is outside the scale? The usual practice, as I noted earlier in this thread, is to use an accidental above the beginning of the wavy line, but it's also fairly common to place the accidental, as you have, between the "tr" and the wavy line, and I've sometimes seen it indicated as a stemless quarter note in parentheses directly following the note to be trilled, on the pitch to be trilled to, under the line. All of these are pretty clear: again, choose one and stick with it. How do you indicate which note to start on? Current convention starts trills on the lower note; if the composer wants it to start on the upper note, he or she will usually place a grace note before the trill. That is easily done by hand, although it might be nice to have it included in the object. Someday.
I don't think you should get tied up in nuances, Mike. All that is really necessary is something that tells the performer clearly when to trill and which notes to trill on. Just my 2¢ worth....
Bill
As for the sound of the playback....well, anything you do will be better than a hidden staff with slurred 32nd notes.
Thanks again Bill. Some of the things you mention may already be supported in the test version; I'll do some testing tonight when I am at my NWC computer. As for the several "choose one and stick with it": I am fine with that, but I am hoping that others here can help me with the choices, based on what is most common/prevalent. Some of my plugins have succumbed to "feature creep" (e.g., GuitarChord) and I realize that that can be confusing to first-time users who see a dozen options in front of them. Perhaps I need to find a way to hide the "advanced" features, but still make them available to users who need to use them on occasion.
Thanks,
Mike
To that end, I have attached snippets of my wife's flute/piccolo music for her next concert.
Thanks very much for those examples. Here are a few starter questions:
- Several of the examples have a stylized "tr", while others seem to be just italicized letters. I am hoping the latter will work for our object.
- Is the "end trill" text common? I.e. is it worth including in the object, or can users add that via standard text in the cases when it is needed? (I can see that for playback, the object still needs to know to stop trilling at the start of the tied note, but that is a different matter.)
- Your examples all use the "wavy" squiggle, as opposed to the more jagged one I am currently using (I have seen both styles in my music searches). Are both of those currently in use, or is it a matter of the era in which the sheet music was engraved?
Regarding the sharp/flat symbols, I have also found that Lilypond uses the alternate notation of a parenthesized stemless cue notehead to indicate the alternating note. So it probably makes sense for the object to have this option as well.
Thanks again,
Mike
You are welcome. A few answers:
IMO, the chars in NWC2STDA.ttf for trill and wavy line are fine for general use.
The "end trill" text is uncommon.
I concur with Rick on both counts.
"End trill" is only useful when a note is to be held beyond the length of the trill. On the rare occasions that I use it, I don't consider it part of the trill symbol, just a courtesy to the player.
I'm not sure it's necessary to include alternate means of indicating trill accidentals. I suspect players can read any of the common methods without much trouble. If someone were trying to duplicate the exact appearance of a score, he or she might need that option, but that's really not what NWC is about. I did notice in Rick's examples, however, a cross between placing an accidental between the "tr" and the wavy line and placing it above the beginning of the line: in Topic 9201 rg02, the accidental is between the two parts of the symbol but is elevated in the manner of a superscript. Not sure I like it, but it is another possibility to choose from. Here is yet another:
tr(#)~~~~~
OK, I'll shut up now....
Bill
Thanks, it helps to know this.
It occurs to me that it should be fairly easy to implement all of these choices via a single "style" option. Or a subset of the styles if one of them should prove difficult.
No worries at all, I am glad for the discussion and feedback. Sometimes "design by committee" can be a disaster, but I think things are proceeding well so far.
Mike
They could be numeric and spun. Span is not needed for trills. In the absence of other instructions, trills continue for the duration of whatever you are calling the "principal note."
The is one occurrence I've seen that doesn't seem to have been discussed.
Unfortunately I don't have an example to show, but imagine a semibreve with the trill symbol appearing later in the bar than the note, OR a pair of tied notes with the trill appearing over the second note OR partway along the tie.
It seems to me they all mean to hold the note steady for a while (half the duration?) indicated by the location of the trill symbol and then start trilling.
I had included Span to accommodate tied notes, but I suppose it is also possible to add detection for ties. For now, I will remove Span, but other logic can be added later.
Parameters
Style and
ExtendToBar have been added, additional accidental types are supported, and the logic for
MinMarks has been adjusted.
Let me know what you think.
Mike
!NoteWorthyComposer(2.5)
|PgSetup|StaffSize:18|DurationPadding:N
|StaffProperties|EndingBar:Open (hidden)
|RestMultiBar|NumBars:6|PrintOnce:N|WhenHidden:ShowBars,ShowRests
|Bar
|RestChord|Dur:Half|Opts:Stem=Down,HideRest|Dur2:Whole|Pos2:0^
|Text|Text:"`_~~"|Font:StaffSymbols|Pos:7|Placement:AtNextNote
|RestChord|Dur:Half|Opts:Stem=Up,NoLegerLines,Muted,HideRest|Dur2:Whole|Pos2:-6z
|Bar
|Note|Dur:Whole|Pos:0
|Bar
|Boundary|Style:NewSystem|NewPage:N
!NoteWorthyComposer-End
Rick is technically correct, but I think it's better to run the wavy line the entire length of the trilled note. It eliminates any possible confusion. It also, from a practical standpoint, will probably make it easier to control the playback when you implement that. IMO, you should either put "span" back in or include detection for tied notes.
It seems that you left Span in, but it no longer has any effect.
It's been removed from the object's spec table, but not from the sample NWCTXT I previously posted, so it may still appear when viewing object properties. I also see now that I missed removing one reference to t.Span. However, as Bill mentioned, there may be instances where it would be easier to indicate a trill that 'spans' multiple tied notes, so I will most likely be restoring the parameter.
As long as we are on the subject, I haven't seen any comments on the various parameters, and whether they are providing "meaningful" trill renderings. Specifically, I'm referring to the new parameters (Style, ExtendToBar), and whether it makes sense to have a parameter to specify a minimum number of marks after the
tr. The majority of your examples didn't include a wavy line at all, but the ones that did showed the line extending to the end of the measure. I guess I would like to review a few more "rules" to know how the line of marks, when present, should be drawn under different circumstances. Perhaps the object could have more intelligence and simply be told to draw or not draw the markers, and it would know how far to draw it, based on the principal trill note's position in the measure: i.e. if it is the final/only note in the measure, draw markers to the bar line, otherwise draw marker up to the next note/rest. Also, I am sensing that some of the people here prefer the "wavy" rather than the "jagged" marker character, so maybe that should be configurable.
Since there will probably continue to be stylistic discussions, I will move on to the playback portion, using Alfred's as a guide. Any other references would be appreciated, especially if Alfred's "has it wrong".
Here are the options I am planning to implement:
- Play - true/false, default = true. As for tremolos/arpeggios, it will require the actual notes be muted.
- Speed - I am thinking of a drop-down with note durations (i.e. 8th, 16th, 32nd, 64th). Is that enough? Too much? Triplet or dotted needed?
Also, I assume that the trill should always be an even number of notes - like a 2-note tremolo.
And one final question: is it ever possible for a trill to be applied to a chord? If not, and the object does encounter a multi-note chord, should it play nothing, or modulate the top/bottom note of that chord?
In piano and flute music, if the notes share a stem they are trilled (this requires more than one flute).
IMO, it need not be implemented for playback.
Keep in mind that I am not much interested in playback. You might consider just setting the number of notes that are to be played and make them fit the trill span. Even numbers would start above the principal note and odd numbers would start on the principal note. If you were real ambitious, negative numbers could cause termination from below. Or, a fractional part could specify how many semitones below the principal note would be used for termination.
I am interested in playback, and I have a somewhat different set of suggestions from Rick's.
I agree with his observation that a trill over notes with a shared stem requires all notes on that stem to be trilled, but I would observe that trilling more than two notes at a time on just about any instrument is impossible, and most can only trill one note at a time (keyboards are the principal exception, although a really good string player can pull it off). I think if users want more than one note trilled simultaneously, though, they should write them on separate staves, layering if they really want the finished music to be on only one staff. If the object encounters a chord, I would lean toward having it trill just the top note and sound the others normally. Others may have a different opinion, and it might be necessary to allow a choice between top and bottom notes.
On speed, I like your idea better than Rick's. However, two questions remain: (1) whether to start the trill on the written note or the trill note, and (2) whether to trill up or down. I think these should be handled as options. Default should be to start on the written note and trill up. The method I'm familiar with to indicate starting on the trill note is to place it before the principal note as a grace note or acciaccatura. Trilling to the note below is uncommon, and I know of no standard way to indicate it, although I don't do enough performing to be sure there isn't one.
HTH -
Bill
Rick gives the example of two flutes playing a pair of trills, so that might make the case for being able to handle a trilled "chord". However, this manner of playback would not be much different from tremolo chord playback, which has already been done, so I think it would not be that difficult to do here. However, it might be more straightforward to just play the top note as trilled. (Although playing the other notes "normally" might be more of a challenge, given that the entire chord needs to be muted for trill playback to work.)
From my limited reading, I thought that the optional accidental symbol gave the cue on how to play the trill:
- No symbol = play up one "note" from the principal note, in the current key (i.e a half step for 'mi' or 'ti', a full step otherwise)
- Accidental = the auxiliary note is the note above principal note with the accidental applied. (Again, the current key needs to be known)
Any other step size, including downward, seems to require the stemless cue notehead in parens, And, that style can be used for upward trills too, I guess.
Thanks everyone for these ideas. I'll try to have something with playback to test soon.
Mike
You might want to broaden your reading with: The Trill in the Classical Period (1750–1820) (http://www.oldflutes.com/articles/classicaltrill/index.htm)
If I had to redo the second and third examples from above, I might do something like the attachment. The
tr is 120%, the accidentals are 58% and the tildes are bold at 80%.
Well, you can probably ignore the downward trill. However, the choice of whether to start the trill on the upper note (the trill note) or the lower note (the principal note: the note that is written in the score) does need to be handleable. If the playback recognizes a grace note or acciaccatura before the trilled note, it can probably be handled that way. Better would be if the object had an option for which note to begin the trill on, and would draw the grace note or acciaccatura itself.
This is becoming a bit of "information overload" - I am learning more about trills than I ever thought I would. I'm feeling like I would need several additional parameters in order to handle all of the special cases in the article Rick has linked. One of them might need to be "Year of composition" :) Perhaps it's time to invoke the 80-20 rule...
Configuring the starting note to be upper or lower won't be difficult; probably just a checkbox. However, concerning a leading grace note/acciaccatura, could you answer me this: if that note will be playable anyway, then why not insert a "real" one when needed, and let it play, and then let the trill play the rest of the notes? Automating the placement to include the grace note and its slur can be done via a selector. (I'm not saying that automatically drawing a 'fake' grace note can't be done, but then extra space has to be reserved for it, the slur would have to be drawn, etc.). Also, if needed, the trill object can check for the presence of the grace note and adjust its playback accordingly.
These are all great suggestions; thanks again to everyone who has joined in.
Mike
I think having the user insert a "real" grace note/acciaccatura would be fine, providing it won't interfere with the playback of the trill.
Re the different nuances of historic playing styles: I think it's safe to ignore them. I'm sure my brother Jack, who's an early-music guru at the University of Louisville, could completely overwhelm you with details about how trills should be played in the music of various centuries; but, while things like that are crucial to performers, the performers themselves are expected to know them and to adjust their styles accordingly. For playback in a notation program, they're really irrelevant. Those who want period-authentic playback can do what we've always done: write it out by hand and hide the staff.
I do not think that the key needs to be known.
The upper auxiliary is always a note occupying the next staff position higher than staff position occupied by the principal. The accidental applies to that position. The key and prior accidental state are therefore irrelevent.
I do not know if a trill can alter the subsequent accidental state. I would listen to it both ways and play what sounds best. Not much help if you are trying to program it. Sorry.
Some interesting reading (http://societyofcomposers.org/pipermail/scimembers_societyofcomposers.org/2012-February/003791.html) from a programmer for another notation program:
Yikes!
I think Rick is wrong re the relationship of trills to the key the music is in. I think a trill is always in key unless indicated otherwise. A quick search found this, on a flute trill fingerings website [http://www.flutetunes.com/fingerings/trills.php (http://www.flutetunes.com/fingerings/trills.php)]: What you have to do is to rapidly alternate between the written note and the next note-name above it, in the key signature. Not exactly definitive - this isn't the Harvard Dictionary of Music - but indicative.
Of course, much modern music doesn't use a key signature. In such cases, Rick is correct - the trill is to the next line or space above.
Continuing the reply above....I found this in A New Dictionary of Music (Arthur Jacobs, Penguin, 1963): "TRILL, a musical ornament....consisting of the rapid alternation of the written note and the note above. Whether this note is a whole-tone or a semitone above depends on which of these notes occurs in the scale in use at the moment - unless the composer directs otherwise."
I keep the Jacobs next to my computer keyboard at all times, but - funny story - it wasn't the first place I looked to answer this question. I went first to the Harvard Dictionary of Music, then to other musical dictionaries and style books I have, then to the Web and to skimming a few scores....until I finally thought, "I wonder what Jacobs thinks?"....and there it was. I think this IS definitive - and very clear.
Please note that I was referring only to those trills that specify an accidental. I stand by my statement.
Having been away from my computer for the last several messages, I haven't been able to Bill & Rick's replies and counter-replies. I see what Rick is saying, but I was thinking more in terms of knowing whether the aux. note was itself already sharped or flatted (per the current key signature), as well as determining whether the subsequent interval would be.
I am currently wondering if it might be simpler for now for the object to have a parameter that tells it the number of steps to use as the interval (most likely defaulting to 2), separate from the accidental symbol. At a minimum, this is how I will do it initially, so I can actually release something that people can play with.
Thanks again,
Mike
Rick, here's your original quote:
If you're following this, and you see an E-F trill in music with a D key signature, how do you play it? The way I read it, you would trill E-F natural unless the composer indicated that a sharp was called for. But that's not the way composers think. We assume you will trill to the sharp if there is a sharp in the key signature. If we want an E-F natural trill in music with a sharp in the signature, we'll indicate a natural in the trill. In this case,
the natural is the accidental.
In trills where the accidental is indicated, you are correct that the key signature and prior accidental state are irrelevant. In trills where no accidental is indicated, though, the key needs to be known to play the trill correctly.
Incidentally, Mike, in regards to whether the accidental in a trill applies to the following notes in the measure: no. The following notes revert to the key signature unless otherwise indicated. In those cases, though, I usually include a courtesy accidental for the player to make sure he or she knows what to play.
removed by author. See the attachment:
Rick, I included your complete quote. The only thing I left out was the following bit about whether a trill accidental affects the notes following the trill in the same measure - which has no bearing on the question of whether or not the key signature matters to the trill. So how is it taken out of context?
OK. I see what you're saying. You "do not think the key needs to be known" in this case. (Italicized phrase added). I agree. But I wish you'd said that. Lacking the clarifying phrase, there is not enough of Mike's quote included to contextualize what you said, at least for me.
I apologize for the misunderstanding. And let's all remember that, in this forum (or in any case where our sole contact is the written word) we have to take pains to make sure what we say is completely clear in its own immediate context.
Alright, everyone please calm down :) As it turns out, I am "cheating" on this for now; please read on for more information.
An updated Trill.test object has been uploaded, with significant overhauls:
- Span has been put back, and has a dual use: when set to 0, it suppresses the wavy line; when it is non-zero, it specifies the extent of the trill line, as well as the playback duration.
- ExtendToBar and MinMarks are gone. The wavy line now automatically extends either to the next bar, or to the note or rest that follows the Span region, as appropriate.
- I've added a fourth accidental style: above the tr.
- It plays. The start note is specified via parameter Play Which First, and the trill speed is specified by Playback Note Type (current choices are Sixteenth, Thirtysecond and Sixtyfourth. More can obviously be added if needed)
- To specify the trill interval, I have added a parameter with the awkward name of Auxiliary Note Offset, which specifies the number of half-steps. It can be positive or negative, so downward trills are also possible. It can also be 0, in which case it produces a rather odd-looking single note tremolo. (I realize this might be a bit of a "cop out" but this was easier than adding logic to determine the current key. Perhaps I'll give that a shot later.)
I've also tweaked the draw code a bit to improve the appearance. By popular demand, it now uses the "wavy" character to draw the line.
As always, feedback and suggestions (and bug reports) are appreciated. (Foremost among suggestions are better names for some of the parameters)
Thanks, Mike. Looks good. I haven't tested it extensively yet, but it seems trouble-free so far. I appreciate the offset ability for the end of the slur. Not so sure about the wavy line (I'd prefer thicker and a bit more angular), but I'll probably get used to it.
Thanks for the feedback. Actually, it isn't difficult to add yet another option, allowing for either the "wavy" or "jagged" character. I know that Rick prefers the wavy line, and for some reason I though you did as well.
Actually, I decided to go ahead and do it. Give version 0.4 a shot and let me know what you think. While comparing the two types side by side, it seems to me that the "jagged" character was a bit too big compared to the "wavy" one, so I've included an internal scale factor. Let me know if it needs adjusting.
Mike
Hi Lawrie,
I'd forgotten to reply to your message, so I am doing so now. In the examples and "specifications" I've seen for trills, I haven't encountered any examples of the first or third types that you mention. The second type is handled by the current test object, as shown below:
!NoteWorthyComposerClip(2.75,Single)
|Clef|Type:Treble
|TimeSig|Signature:4/4
|Note|Dur:Half|Pos:0^
|User|Trill.test|Pos:8.5|Span:2|AuxiliaryOffset:1|LineType:Jagged
|Note|Dur:Half|Pos:0^|Opts:Muted
|Bar
|Note|Dur:Whole|Pos:0|Opts:Muted
|Bar
!NoteWorthyComposerClip-End
If you aren't worried about a version that plays back correctly, it could also be done by inserting a trill before the beginning semibreve, and using Start Offset to push the start of the trill to the right by the desired amount.
Let me know if the above choices would fit the bill.
Mike
Hi Mike,
I only brought them up for completeness. I've only seen them very, very rarely. So rarely that it's probably good enough to simply use the old way with hidden playbeck and text trill marker.
Or, possibly like this:
!NoteWorthyComposer(2.75)
|PgSetup|StaffSize:18|DurationPadding:N
|StaffProperties|EndingBar:Open (hidden)
|Clef|Type:Treble
|TimeSig|Signature:4/4
|RestChord|Dur:Half|Opts:Stem=Down,HideRest|Dur2:Whole|Pos2:0^
|Spacer|Width:200
|User|Trill.test|Pos:8.5|Span:2|AuxiliaryOffset:1|LineType:Jagged
|Note|Dur:Half|Pos:x-1|Opts:Stem=Down|Visibility:Never
|Spacer|Width:200
|Bar
|Note|Dur:Whole|Pos:0|Opts:Muted
|Spacer|Width:400
|Bar
!NoteWorthyComposer-End
That's tricky - I tried to create that option, but it did not occur to me to use an A double-sharp to prevent the broken tie. (The hidden note should probably be muted for playback)
Thanks,
Mike
All (especially Lawrie),
I remarked earlier that one of the reasons I picked the "text only" approach, and use of the StaffSymbols font, for Trill.test was to allow the alternate NWC2* fonts in the PardyPack to be usable by this object. However, in my testing, I've found that those other fonts don't work, because they don't have shapes defined for the characters used for the tr, "small space" and "squiggle" characters (` _ - ~). However, the accidental symbols I am using are rendered correctly in those fonts, although that is a small consolation.
I've noticed that tr followed by 2 squiggles, as well as standalone 2- and 3-squiggle clusters, are characters in BOXMARKS2 and several of the other PardyPack fonts, but those combined characters don't really lend themselves to programmatic trill drawing. I'm not sure how much effort would be involved in creating the separate characters in the alternate NWC2* fonts so that these sorts of trills would work correctly. Or, perhaps I have overlooked the separate characters in the PardyPack suite?
EDIT: As Emily Litella would say... "Never mind!" Turns out I had older (1.00a) copies of the NWC2 fonts installed. I've updated to 1.1, and everything works fine!
Thanks,
Mike
Hello all,
Other than Bill's reply, I haven't heard any feedback about the Trill 0.4 posting. I am assuming (hoping) that no news is good news, and that it's working as intended.
I'm working on a change, per a few private suggestions from Eric, that should allow the object to more automatically tell which interval to use for the second note. If that works, it may change the number and type of parameters that are needed. In any case, if you have any feedback or experiences (positive or negative) to share, now would be the time to share them, so they can be incorporated in the upcoming update.
Thanks,
Mike
Sorry to be so late in weighing in on this topic. I haven't been doing much notation lately, and don't feel I have the knowledge to experiment with user objects, etc. I do have a few comments about various aspects mentioned in this thread so far:
I've never seen it in real life but Rick's wife's snippet illustrates what Bill says. "end trill" can easily be added as left-justified text over the last note, so it doesn't
need to be an option.
Trills are sometimes used with short notes, sometimes several in a row. If the squiggle is used, then it may make sense to put the accidental above it to allow room to squeeze the
tr expression, the accidental and the squiggle into the space above the affected note.
I've never seen this. It's rare enough to allow a workaround, perhaps with a layered staff with invisible notes, with the trill in the underneath staff starting where you want it.
Trills do sometimes extend over several measures, but if you just put a trill on all the notes within that "range," you'll need to know if the notes are tied or separated. I think that justifies a span direction of some sort. A trill covering multiple bars of tied whole notes must be played differently than the same number of untied whole notes, each of which, on a wind instrument, would be tongued and individually trilled.
Going back to a question I posted a couple of years ago, using this example:
Since the composer's wishes for the trilled A# and B# are ambiguous (read, "
confusing to many musicians"), a solution is to show the upper note of the trill as a bracketed target note over the beginning of the trill squiggly. I think the script or LUA (
I'm easily confused) should allow for such a ghost-note as well as an accidental.
Thanks David. You posted a lot of quotes, and it probably doesn't make sense to respond to each one, other than say "I agree" and "The current object should allow this." As you might have seen, I got rid of a couple of parameters and put Span back, because it seems to provide a simple way to control the extend of a slur, including the marker line. As for this:
Another notation I've seen is a stemless cue notehead in parens, positioned on the staff line/space with optional accidental, in lieu of the accidental symbol above/next to the "tr". I am planning to support that in the next version.
Thanks for the feedback,
Mike
Good! That's exactly what I was trying to describe.
The Trill.test object has been updated with the following changes:
- Parameter "Auxiliary Note Offset" has been replaced by "Auxiliary Note Interval", which has a default setting of "Auto", which will automatically determine the pitch of the auxiliary note based on the current key signature and effective accidentals on the staff, as modified by the "Accidental" parameter. For instances where this does not work, you can choose an explicit interval between -5 and 5.
- Parameter "Accidental" now includes "Double Sharp" as a choice.
It should be noted that the Auto setting described above can result in a zero-interval trill for certain notes and accidental symbols (i.e. in the key of C major, a trilled B with a flat accidental). I'm assuming such cases are "nonsense notation".
Also please note I haven't yet implemented the "cue notehead in parens" accidental style. It remains on my ToDo list.
Please let me know if you think the accidental/trill logic needs adjusting.
Thanks,
Mike
The Trill object has been updated, with the following changes:
- Graduation from ".test" to ".ms". Use the "Change Object Type" user tool to update any scores which contain the .test object.
As previously noted, "cue notehead in parens" accidental style has been reserved for future implementation.
Thanks,
Mike
After I found this nice object, I thought a little about playing trills better. I did have some ideas here, mainly from having arranged quite a number of trills for crank and band organs - so I set out to describe them, and even write an experimental program that would create trills that are "more musical." The result of all this can be found here: "Some thoughts on automatic playing of trills" (http://harmonic-musings.blogspot.de/2015/11/some-thoughts-on-automatic-playing-of.html).
The program is my first Lua program, so it might not be the very best style; and it does not at all interface with NWC's objects - it's just a demo program for my trill note computing algorithm. But maybe it might be possible to include part of the algorithm in the Trill.ms object ... if my ideas make any sense.
H.M.
Thanks Harald - that was an interesting post/article. There is obviously much more to trills than meets the eye.
There are certainly a number of things that could be added to Trill.ms to make it more versatile/realistic. However, the object currently has 11 parameters, and there is a hard upper limit of 16 parameters available (although it should be possible to combine a few of them). One easy improvement would be to expand "Playback Note Type" from the current 3 choices (16th, 32nd, 64th), or possibly change it to a numeric value, as in the Arpeggio object. This would give much finer control of the speed than can be had with the various dotted/double-dotted/triplet combinations that are possible with standard notation.
Mike
Suddenly, I get this:
C:\ProgramData\Noteworthy Software\nwc2\UserPlugins\Trill.ms.nwcuser.lua:138: attempt to perform arithmetic on a nil value
Unfortunately, I cannot reproduce in a smaller score thatn the complete one.
Harald
It's very easy to reproduce this: add a trill object to the very end of any score (e.g. with no notes after it).
I recently discovered that several other objects will throw a similar error if there are no notes following them. I'm posting an updated version of the Trill.ms object presently.
Thanks for the error report.
Mike
The Trill object has been updated, with the following changes:
- Fix a bug which caused an exception when there were no notes following the trill object.
Please refer to the parent post for a description of the available parameters and a sample clip.
My new trill that led to the error was not at end, but somewhere in the middle ... but be that as it may, now my trills trill again!
Many thanks!
Harald
Couple of observations:
If there is a fermata delay then it does not trill through the added delay. If the fermata is on the trill staff then it doesn't start until the delay (added on the front of the note) has elapsed. If the fermata is on another staff on a later note then it stops trilling for the length of the delay and then starts again!
If the trill is on a note tied to a subsequent note then the sound stops altogether at the end of the first note.
Not a big deal but it would be nice to have it perfected. I had thought that a work around would be to have a hidden tied note to mimic the fermata but the second problem kills that idea. :(
I'm not sure I am visualizing this exactly. Can you provide a clip that shows what you are trying to do?
This should work correctly if the Note Span property is set to 2, so it plays through both notes. However, please note that for ties, you should only mute the first note, otherwise you'll get a stuck note. Again, if you show me a clip, I can make sure I am giving you a valid fix.
Thanks for the feedback; If I can't find a reasonable workaround, I'll find a way to correct the object so it handles these situations.
Example:
!NoteWorthyComposer(2.75)
|AddStaff|Name:"Staff"
|StaffProperties|EndingBar:Open (hidden)
|Instrument|Name:"Recorder"|Patch:74|Pos:12
|Clef|Type:Treble
|TimeSig|Signature:4/4
|User|Trill.ms|Pos:9|AccStyle:2|Span:1|WhichFirst:Auxiliary
|TempoVariance|Style:Fermata|Pause:8|Pos:9|Justify:Center|Placement:AtNextNote
|Note|Dur:Half|Pos:1|Opts:Muted
|Rest|Dur:Half
!NoteWorthyComposer-End
Thanks for the example Rick.
I'm not sure exactly why this is happening. For the record, you will get the same "first note plays too long" effect if you try to use Arpeggio.ms on a fermata chord. The first nwcplay() has its duration increased by the fermata delay.
I will continue to experiment with this, to see if this is something I can adjust for during object playback.
Tremolo and TremoloSingle both do this too.
It happens because there can be no MIDI events during the pause.
Rick is correct. A performance pause on a fermata (or other eligible tempo variance object) essentially pauses the song for the specified duration where the item occurs. Nothing new can be played during this pause.
Therefore, to achieve 'fermata' trill playback with a custom object, you want to create a tied (or longer duration) note, most likely on a hidden staff. As mentioned earlier, the Span property on a trill would need to be increased to cover both notes.
In this system, the trill stops at the end of the rest under the note, not the end of the note. Also it seems difficult to get nice short trill markings - I'd like tr. then maybe two waves; it seems to want to shoot off for ages.
It is only playing for a half note duration because Note Span is set to 1, and the rest chord has a duration of a half note (the rest duration). If you set Note Span to 2, then the trill will play for the full duration of the whole note.
I implemented the trill marking according to references that said it should cover the staff duration of the notes being played, Setting the end position to a large negative number is one way to shorten the line; another would be to set the object's visibility to Never, and then add your own text "tr~~". If this is a legitimate trill style, I could probably add an option to implement that, that would be easier than entering the negative offsets.
The longer trill marking is a more modern convention. Have a look, for example, at the 1° Violin part, page 4, second system in https://archive.org/stream/lestroqueursopra00hrol#page/4/mode/2up
Nothing to object to your object, Mike, :D but, just in case, you can look here (https://forum.noteworthycomposer.com/?topic=8345.msg58660#msg58660)
I disagree. Looks like a trill.
I stand corrected. I had not seen this style of trill before. Baroque?
No, I think it's working fine. ;) (If it ain't Baroque, don't fix it.)
Actually I was considering a new parameter on the Trill object, "Style", which would have values of "Modern" and "Baroque" (and perhaps there are others). I would have used the subject line "Trill.ms: Going for Baroque" but you beat me to the pun :)
Sounds useful. It'd also allow various other styles to be added later. By the way, working on this score has taught me lots about terrible old conventions of music. Like the transposing natural trumpets labelled using the word nowadays used to indicate [French] horns, but which have to be trumpets because they're very obviously within the natural trumpet harmonic series, to writing the Soprano part using the Soprano clef (a C clef on the bottom line) - so just because something's in an old score doesn't always make it a good idea. =)
You do realize there are Bb horns, don't you? The modern version is called the "marching" horn. Historically, in the days of natural horns and trumpets, horns were made in multiple sizes for use in different keys, and the Bb was known as the "alto" horn. It would have had the same harmonic series as the natural Bb trumpet.
The Trill.test object has been updated with the following changes:
- The object will now render trills which span to successive systems; the extended marking will be marked with "(tr)".
- Other code improvements have been made, resulting in more accurate rendering.
The current speed options available are: 'Sixteenth', 'Thirtysecond', 'Sixtyfourth'.
In some cases I find Thirtysecond too slow and Sixtyfourth too fast. In others Sixtyfourth is still too slow.
Since in both Arpeggio and Acciaccatura Mike used the much more flexible parameter "Rate" (and surreptitiously allowed the fusa ;) ) I'm wondering if the same could be done in Trillo too.
The code could certainly be changed to use a Rate value instead of the specific note types, but of course there would be the issue of backward compatibility. That too can be handled by the "audit" event, which would convert the note type to a corresponding Rate value when the score is opened. I assume you would want values of 16, 32 and 64 to correspond to those three note types?
It seems that you want to have faster trill options, so I assume that 16 would be an acceptable minimum value. Would 128 be a fast enough top value?
Very good guess! :)
The problem is to have a decent trill on a largo.
As far as I know, 1/128 (fusa) should be enough.
By the way: isn't that already the limit for the other plugins' rate?
Oh, yes.
But I was thinking something like: if PlayNote argument is not a number then it should be one of the three options and you can convert it to the relevant numeric value on the fly.
The Trill.ms object has been updated, with the following changes:
- Change the "Playback Note Type" dropdown parameter (Sixteenth, Thirtysecond, Sixtyfourth) to a numeric "Playback Rate" parameter (as is also done for Arpeggio.ms and Acciaccatura.ms). Scores which reference the older parameter value will be automatically upgraded to the rate parameter. (Thanks to @Flurmy for the enhancement request.)
Already done?
Thanks to you for the quinck upgrade, Mike!
I am trying to print the wavy line without the "tr" text. Is that possible?
At present, the object will always print the 'tr' symbol, although the wavy line can be suppressed (for one-note trills). If you do not require playback, you can insert a text object with "~" (wavy) or "-" (jagged) characters, using the StaffSymbols font.
Hi Mike
I was looking at my computer file for user plug ins and objects, and I noticed that I had two versions of a few things, with different names.
For example, Trill.ms.nwcuser and Trill.ms.nwcuser.lua
Can I safely delete the .lua versions (which have an older date on them)?
Thanks
It depends on whether you have Windows configured to hide the extensions of files. When you are looking at the files in Windows explorer, what does the "Type" column show for each of the two files, i.e. Trill.ms.nwcuser and Trill.ms.nwcuser.lua ? You could also highlight each of the files, right click and choose Properties, and then go to the Details tab, and see what it displays for Name.
Mike, I've moved the question here, https://forum.noteworthycomposer.com/?topic=10023.0
as it is related to BAK files. Thanks for your help.
Don't ask me what I think about it... >:(
The Trill.ms object has been updated, with the following changes:
- Added a built-in user tool to toggle trill on/off for a selected note.
- Inserting a new trill will now copy parameter values from a previous instance in the current staff.
I stumped on a problem with the trill object.
If the trill is the upper note of a chord (no matter if the duration of the notes is the same or different) the chord don't plays at all and the trillling note is... the lowest one!
If I use the old tool that makes a trill using the grace notes I have the upper note trilling ok but also the unpleasant result of the chord playing at the end of the trill. What I need is not possible in this way. The only way is an hidden staff.
Maybe the simplest fix could be to make the trilling note the upper one instead of the lowest.
The very best would be to have the upper note trilling and the rest of the chord playing as usual.
I'll take a look if I can fix the object, but maybe Mike can do a miracle...
Gosh! I didn't expect it was so trivial to apply the fix!
About the best solution... That's a different hat.
N.B. The attached file is zipped because I'm not allowed to upload an object of somepone else, and I don't want to steal the honor from Mike.
I hope Mike will not be offended by my "instrusion" and will update the official object.
I just saw both of these messages, and since it's been 3 years since I last updated this object, I needed to refresh my memory a bit, reconfigure my code environment, and then figure out what you had changed in the code. Having now done that, I see that your change was very straightforward and should work as expected.
When I added playback to Trill.ms, I obviously did not try to implement chord playback, which might have been possible, but I can't think of any instruments that can play trill chords. If someone needs to do that, it could be notated using Tremolo.ms (which can also do a single note trill). I chose the lowest note of the chord for playback because it seemed like a safe choice. Playing the highest note is also a safe choice. However, different scores might want different results, so the "proper" way to implement this would be with an additional parameter, to specify lowest vs. highest. For backward compatibility, the lowest note would be the default value for the parameter. Does that make sense?
As a side note, the reason the chord doesn't play during trill playback is because the chord is normally muted. You can unmute the chord and it will play, with the trill playing over the entire chord, but that is probably not what you want. Separate staves will give you the most control for that situation.
No, Mike, I don't want the whole chord to trill! :o Only the top note.
At least that's what's needed by the scores for piano I have. :)
And I doubt that, in case of a trill on a chord, someone wants the lower note trilling. In that case a more specific (and clear) notation is in order! ::)
I feel stupid: I forgot that the chord is muted. Simply unmuting it solves the remaining problem! :-[
Possibly there is a double key-on midi event for the top note (I didn't check), but this seems creating no problem (checked ;) ).
What was the need to mute the note in the score? Maybe the clash between the starting note of the trill and the note of the score?
I think I can live with it (i.e. not using "auxiliary first" with chords)... :P
Well, now I'm completely satisfied of the last (mine...) version.
Thanks a lot.
Every other object I've written whose playback replaces the notation (arpeggio, tremolos, glissando, acciaccatura) requires the notes be muted in order to sound right. In some cases, not muting the notes will still sound okay, but as you have noted, there might be some double or overlapping of the playback. Of course, special care needs to be taken when there are ties, but this is noted in the instructions for the affected objects.
Since you are happy with your fix, I'll hold off on making the change official, in case someone else has a suggestion or other feedback to offer. Maybe I'll make the change for the next release of NWC :)
Hi Mike,
I am trying to understand what was being discussed here under topic of Ojbect, it is better if there is the image/picture, that we can see as the result of using the object, for example can anyone show the pictture of song sheet that used Page Maestro text?
Thank you,
Frank
Well, not exactly a picture, but maybe even better: look at all my recent submissions to the Scriptorium.
A tiny change suggestion: In the attached score of Durante's Magnificat in b flat, the trill should play in eighths.
The piece is notated alla breve, the tempo is about half=105...110. In more modern times (starting around Haydn), this would of course be notated in common time and quarters, but it isn't so here. So, I think the minimum trill rate of 16 should be changed to 8 (I can do that locally, of course; but why not change it for all :-[ )
H.M.
The Trill.ms object has been updated with the following changes:
- If the trill is applied to a chord with playback selected, it will play the top note of the chord, instead of the bottom one (courtesy of @Flurmy).
- The minimum value for the Rate parameter has been changed from 16 to 8, as requested by @hmmueller
... small hiccup: The help text for the Rate parameter still says that the minimum is 16.
Also, there's a playNoteRate array (whose purpose I don't see on my quick glance) - would it now like to heave an Eighth entry?
H.M.
You are correct about my failure to update the help text. I realized that right after I posted the update. It seems a shame to waste another version number just to fix the help text. :P
The playNoteRate array is only used by the audit() method. As you might recall, the Trill object used to have a PlayNote parameter which had a text value of Sixteenth, Thirtysecond or Sixtyfourth. I later changed the object to have a numeric Rate value, which is also used by Arpeggio.ms. The audit() method uses the array to convert the old text value to the correct numeric Rate value. Since the old object never had a text Eighth value, there's no need to add it to the array for conversion purposes. Does that make sense?
It does - and I didnt look there. Aha!
For a version number, the going rate here is quite low - purchasing another one ... well :))
Edit: Also, the complete thread would like to get a subject update to Trill.ms (2.3) - or 2.4? ;)
... I know I'm, hm, getting irritating ...
H.M.
Nope, I'm not wasting a version number on a simple help text edit, since there are no functional changes.
I'm using Mike's trill object for the first time. Hey, it's only been available about 7 years, these things take time.
My flatted trill begins with a half note on beat 2 of a 3/4 bar, tied over a 4/4 time change to 21 whole notes, tied to and ending on an 8th note. I've learned that the only way to display the trill the right number of bars in the edit window or in print preview is to use the Span class setting. (I didn't check playback; I only use playback as a proofreading tool and I'm not there yet with this transcription.)
Where the trill lasts past a system break, I see (tr) at the beginning of the next system just before the wavy line. Perhaps on the next update, that could change to show the accidental too? Otherwise someone reading the chart might wonder if the trill is supposed to change to a normal trill at that point.
That aside, what a wonderful tool, once again. A little late, but thanks, Mike.
(And for anyone writing for bass clarinet, please don't write trills from 3rd space C to Db for 22 bars. It's far too much work for the little fingers, with either normal fingering or the alternate. Grrr. Those keys have very long rods! An octave down or an octave up would be okay.)
Has it really been 7 years? How time flies...
That's a good idea. Generally I used "Behind Bars" as my reference on how to implement objects, and maybe she didn't include that scenario. If you can find a piece of sheet music that does this, that I could use as a sample, I'll make this change.
Thanks for the kind words, and you're very welcome.
This is from a Boosey & Hawkes, Inc. chart, but it doesn't say which publisher engraved it.
Here, the trill has no descriptor at the beginning of each new system. I have no preference, but if there is to be a descriptor, an accidental could help the user.
"Behind Bars" gives two examples: the "(tr)" notation that my object uses, and a plain wiggly line like in your example, but it says the "(tr)" is the preferred method. It would be fairly easy to remove the "(tr)", and a little more work to add the accidental. However, the object supports 4 styles, only 3 of which are mentioned in BB. The 4th style (enclosing the accidental in parentheses after the "tr") wouldn't work very well I think.
I would make the argument that the (tr) means "continue the previous trill" and therefore doesn't require the accidental. If it were going to change, you could start a new trill on that note. But, let's see if anyone else weighs in with an opinion on this before I start making code changes. I will probably end up adding an option to enable/disable this new behavior anyway. Maybe make it 3 options: (tr), (tr) with an accidental (inside the parens), and no (tr) at all.
The Trill.ms object has been updated to version 2.4 with the following changes (to the user tool only):
- Multiple notes can be trilled/untrilled at one time.
- The trill vertical position will default to a nicer value.
- Trill objects will be avoided when a trill note is tied.
- Parameters for new trills will be inherited from prior trills on the staff.
Thanks to
@Flurmy for these additions.
and the trill at the first note will have the "note span" automatically set to the relevant number to show the extender line.