Trills

[Michael Willis]

I have memories of trills being used to great effect in several pieces of music when I played clarinet in school band, but I've always been disappointed with how they turn out in my own compositions.

I think part of my problem is that a trill is two tones rather than a single tones, generally a half step or whole step, depending on where it is in the current scale. Any time I think "This would be a great place to put a trill!" I end up unsatisfied with the resulting dissonance. Are there any tricks to getting a trill to cooperate with the surrounding harmony?

Then there's the matter of mechanics. Should trills be considered something like a uniform sequence of 64th notes? I remember just fluttering one of the keys on the instrument as quickly as my finger would move, with no expectation of being in sync with other players. Given a sound of a flute section or clarinet section, is there any way to get a convincing trill sound? Certainly the individual notes shouldn't be tongued, so it seems that a legato patch would be in order, but aside from that I'm not sure what to try.

[Mattias Westlund]

I dunno, I think trills are pretty effective for creating a sense of either threat or expectiation, depending on how you use them. They don't even have to be strictly scalar, as a half step trill over a major chord (e.g. C major with a C/C# trill in the high register) can add a wonderful feel of unrest.

As for the mechanics, I don't think trills are normally locked to any rhythmic note value. Just like tremolo in the strings strings it's basically a question of playing [alternating] notes as fast as possible to create a big, turbulent sound.

And no, there is no way to create really convincing trills with midi. If you don't have sampled trills then it's better to just forget about it. You can sort of fake it in woodwinds if you set separate solo instruments up with monophonic retriggering legato and layer them. But for sections, it's a lost cause. It will sound worse than not having any trill at all.

You CAN however achieve a similar feel by playing close intervals with section samples. A C and a C# in the same octave will with a bit of masking (a tringle and/or timp roll is good for this) will give a trill-like feel due to the two notes beating against each other. Not totally convincing but better than trying to play trills live on your midi keyboard.

All that said, trills are best used very sparingly. Even if you have good sampled trills. Overdo it and it becomes a predictable gimmick

[Michael Willis]

(11-09-2016, 02:11 AM)Mattias Westlund Wrote: And no, there is no way to create really convincing trills with midi. If you don't have sampled trills then it's better to just forget about it. You can sort of fake it in woodwinds if you set separate solo instruments up with monophonic retriggering legato and layer them. But for sections, it's a lost cause. It will sound worse than not having any trill at all.

Aww, so sad 

What about a solo woodwind part playing a trill on its own, rather than trying to fake a section trill with layered solo samples? I imagine that a solo melody could finish off with a trill and it could sound alright. Any opinions on that?

[Otto Halmén]

Well, the trick is... to cooperate with the surrounding harmony! For example, sustained chords, seventh chords, and ninth chords can be voiced with minor and major seconds. That can be a great place to try out a minor or major second trill. Just be careful with the timbres and voicing of your work, as trills tend to fill up the sonic space more than regular playing.

Trills can, as Mattias said, be effective for dissonance, but they can certainly work with consonance as well.

[Samulis]

Trills originated as simple ornaments. When we are reaching a V - I cadence (e.g. G Major -> C Major in the key of C) at the end of a piece or section, that the most important note is the 'leading tone' or the 3rd of the chord (B, in the case of G Major). The B then should always proceed to the C. If a D descends to a C at the same time, it increases this motion, and if an F descends to an E, it creates a particular dissonance that accents the motion.

The most basic thing to do then, as a solo instrument, might be to play D then B over the G chord, then proceed to C. Another option would be to start on a C half-note while playing the G chord, then descend to B in the 2nd half note, then return to C when we reach the C chord. This we call a 'Neighbor Tone', and is a sense, the ancestor of the trill.

One can have more fun than this! Quickly alternate between C and B over the G chord, then resolve to the C rather than (boring!) hold them each a half-note, and you have created extra tension and motion. Motion around cadences adds excitement and contrast- when the cadence resolves, motion ceases or slows, creating a 'sigh'-like feeling, like fluttering butterflies in your heart calming. Because the trill was originally essentially an elaborate ornament, it typically had a phase of rapid, but in-time ("measured"- either 16th or 32nd notes depending on speed of the piece) alternation, between a note in the chord, and one outside of the chord (typically a half step or a whole step), then a small ornament. Trills in many cases start on the note that is "out" and rapidly "resolve" to the note that is "in", but there are no fixed rules that are clearly visible across all periods- for example, in early music, it is common (even in a preliminary sense, a 'rule') that the trill will start on an upper note, while in modern music, trills in many cases start on the lower note.

https://youtu.be/2s31pmhyVh4?t=246

Basically, if you had a G chord in 4/4, you might play in the Baroque period, as sixteenths:
CBCB CBCB CBCB CBAB | C--- ---- ---- ----

In some cases, the first note will be a leap from the previous measure, and thus emphasized and/or accented, as it is an Appoggitura.

Nowadays the ornamenty bit at the end is often left out, and trills can occur starting both on the upper and the lower tone. The largest change is that the trill no longer so much performs a harmonic function, but rather an emotional function. Examine, for example, Golden Age Hollywood scores, or the works of threater/Broadway from that time- endless use of major 2nd flute trills.

In addition, we also find the use of "unmeasured" trills and tremolos, where the repetition is not normal and is simply the fastest alternation that can be done or that fits with the mood of the work. Historically, a trill or tremolo was almost always "more or less" a 16th-note pattern (or in the case of a particularly fast piece, 8th notes, or particularly slow, 32nd notes)- although a performer obviously has to "milk" it and play rubato (starting slower and speeding up in most cases) and in a shape that will lead the listener to the cadence.

Increasingly, a trill of an interval larger than 1 step is used, but this is actually not a trill, but a tremolo. You see, back when music printing required engraving or music type pieces, it was very time consuming and expensive to carve, draw, or place many repeated notes. Therefore, an abbreviation of this, the tremolo, was used: draw a half note with two slashes and instantly indicate eight sixteenth notes.
https://easymusictheory.wordpress.com/ta...d-tremolo/

The magic is when you notate two notes with the tremolo between them- this cases them to be repeated between rapidly, but at a measured rate. There is nothing better than drawing two whole notes with tremolo markings and get a bunch of 16ths back and forth when your concerns are with ink, time, and a limited number of (expensive) musical printing press type pieces. The first few pages of the original score for Stravinsky's Rite of Spring is a good example of these measured tremolos in action- the obvious reduction of clutter is enormous.

As for simulating trills, I have had excellent results using WIPS/SIPS with settings of around ~30-50 ms on the length of the legato and length of the crossfade. Play your starting note and hold it while rapidly tapping the lower or upper note you want to trill with. If you have extra care, you can add key clicks. There is also a slight sound made from covering holes that is tricky to simulate, but if you were to sample it, that could be added as well.

EDIT 2: Here is a demo with a few turns and trills simulated with WIPS. You'll hear the plain bit, then the trill. In most cases, with these cadential trills, I've stopped trilling. You will also hear a few ornaments (appoggituras, 'turns', etc.).
https://instaud.io/BUH

EDIT: I forgot to mention, there also existed in the early days of synthesizers and game audio, a form of rapid tremolo between a large interval, say an octave or a 5th. This was because the number of voices was limited and this essentially allowed them to sound multiple notes. This had some origins in earlier jazz and ragtime styles as an extension on the trill.

eg: https://instaud.io/vkn#0:31.9

So, TL;DR-
- You can use trills and tremolo however you want, there are zero rules that matter anymore.
- If you want your trills to mesh nicely, use it sparingly- typically only in your "lead" voice or else some other higher voice.
- Trills should always resolve to one of the two notes that is part of the trill, or the note above or below it.
- The easiest implementation is a half-step trill between "do" and "ti" in a Perfect Authentic cadence (V - i) , or the whole-step between "mi" and "re".
- Trills best fit wherever there is a 'Neighbor Tone' or a 'Passing Tone'- because then you have a resolution ready and waiting for you. (e.g. if you have the line "G - F } E", each being a half note, you could trill between G and F where the F half-note is currently, starting on G, then leading to E)

[Mattias Westlund]

(11-09-2016, 02:52 AM)Michael Willis Wrote: What about a solo woodwind part playing a trill on its own, rather than trying to fake a section trill with layered solo samples? I imagine that a solo melody could finish off with a trill and it could sound alright. Any opinions on that?

That was kind of what I meant with "You can sort of fake it in woodwinds if you set separate solo instruments up with monophonic retriggering legato and layer them". I think that's what Samulis was getting at as well. Woodwinds lend themselves to fake trills better than other instruments.

You could also experiment with mapping a square wave LFO to a non-vibrato'ed woodwind sample and map it to either a half or whole tone pitch modulation. Layer a couple of these and it sounds reasonably convincing.

[Michael Willis]

(11-09-2016, 03:25 AM)Otto Halmén Wrote: Well, the trick is... to cooperate with the surrounding harmony!

Thanks! Sometimes it takes somebody throwing my own words back at me for me to realize that I knew the answer all along

(11-09-2016, 04:59 AM)Samulis Wrote: So, TL;DR-
- You can use trills and tremolo however you want, there are zero rules that matter anymore.
- If you want your trills to mesh nicely, use it sparingly- typically only in your "lead" voice or else some other higher voice.
- Trills should always resolve to one of the two notes that is part of the trill, or the note above or below it.
- The easiest implementation is a half-step trill between "do" and "ti" in a Perfect Authentic cadence (V - i) , or the whole-step between "mi" and "re".
- Trills best fit wherever there is a 'Neighbor Tone' or a 'Passing Tone'- because then you have a resolution ready and waiting for you. (e.g. if you have the line "G - F } E", each being a half note, you could trill between G and F where the F half-note is currently, starting on G, then leading to E)

Thanks! This is just the kind of answer that I was looking for (and yes, I did read it and not just the TL;DR)

(11-09-2016, 02:11 AM)Mattias Westlund Wrote: All that said, trills are best used very sparingly. Even if you have good sampled trills. Overdo it and it becomes a predictable gimmick

Your usage of the Warcraft II midi (in the thread about the retro midi machine) influenced me to give another listen to that soundtrack, it's been years for me. That soundtrack uses woodwind trills like crazy! They're all over the place! I must confess that I love every single one of them, and don't feel like they're a gimmick, but rather an integral part of the style of the soundtrack.

Off-topic, I also noticed that all kinds of rules are broken by that soundtrack, like the harpsichord solos have dynamic panning that shifts from left to right and back to left again. On the other hand, it seems like maybe it's in sync with how low/high the notes are, which maybe makes sense if you consider the harpsichord to be about a meter away from you.

Regardless of "broken rules", samples that are clearly showing their age, and limitations of doing it all within the general midi standard, I still find the whole soundtrack musically brilliant and just as immersive and enjoyable to listen to as it was twenty years ago.

[Mattias Westlund]

(11-12-2016, 12:21 AM)Michael Willis Wrote: Regardless of "broken rules", samples that are clearly showing their age, and limitations of doing it all within the general midi standard, I still find the whole soundtrack musically brilliant and just as immersive and enjoyable to listen to as it was twenty years ago.

Haha, I definitlely agree with that! Listening to those old midi files made me want to play the game again. WCII also has some of the most charming 2D graphics of its era. Sadly, I've always sucked at RTS's and despite its cute look WCII is just as hellishly difficult as other games in the genre.

[Samulis]

Sometimes "what sounds freakin' awesome" isn't the same as "following the rules/realistic".