Dynamic change during a sustained note?

Can a sustained note (e.g., soprano voice) respond to a dynamic change, whether that be via a hairpin or a curve drawn into the dynamics automation window (in Play)?

For some reason, this does not happen for me (using Halion in Dorico 3.5.12).

it depends on whether the dynamic controller is velocity or CC. With the latter it will work and the former not, irrespective of library.

Thanks, dko22.

In my project, I have a lane for velocities, another for “dynamics”, and a third for midi cc (cc#1):

Despite drawing in a decrescendo (to complete silence) for the last note of the exercise (see red arrows), I hear no change in the dynamics there.

The score itself has a decresc. hairpin (that originally went to pp, but which I deleted, thinking it might be interfering) there, but - even so - that isn’t reflected in playback:


I also tried lowering one of the voice’s velocities to almost nothing, there, but no luck.

Any idea what I’m doing wrong?


the lane for dynamics usually corresponds to the primary dynamic controller in the Expression Map. When I just try quickly to replicate a choral work using the Halion Sonic SE, I find the diminuendo does work as expected. Try first to reapply the default template. CC1 for instance will work provided it is listed as one of two controllers. If still stuck, then pls just post the relevant section of the project so we can check out the Expression Map allocations and settings.

Could you maybe also check what happens if you have the two voices on individual lines and use “all voices”? Dynamics can be ignored if not applied to the correct voice.

Thanks, dko22.

I’m fairly ignorant when it comes to this end of things. . .

As in starting afresh, with a new project? I just created a very simple 5-bar soprano voice line, in sustained notes at a slow speed. Rather than create any velocities for those notes, I instead drew a long decrescendo in to the Dynamics midi cc lane (with “upstem voice 1” selected), and listened to the results. There was a definite decrescendo with each new, long note, but I couldn’t detect any decrescendo within each sustained note on its own.

As in here?

I just uploaded a very short passage (similar to the above), with each sustained note receiving a dynamic automation line from 127 to 0 . . . but no evidence of any decrescendo within these individual notes.
Dynamics test.dorico (664.0 KB)

Thanks again, dko22 – I really appreciate your help!

I’ve not looked in the Halion player itself (I never use it for my own projects and it’s not necessary to look at it for the purposes of this exercise).

You can play this three different ways according to the settings in the screenshots

all voices active as in your settings

showing all voices but not specifically activated. This is the normal Dorico setting

just voice 1 as originally

they give different results but all appear to me to do exactly what they’re supposed to. For instance the one showing CC1 follows your curve exactly – you you maybe check that and see if at least that’s not working. I’m afraid I’m mystified just now how you can be getting something different at the moment as I see nothing untoward with what you’ve done. Perhaps I need to sleep on it or someone else could check in the meantime!

Thank-you, dko22.

Not sure what’s going on, but I experimented with two new ‘projects’, and came up with differing results.

First, a few bars of one long note, on violin. I drew the dynamic curve into the midi cc lane, in Play mode, and it did playback according to that curve.

I then changed the instrument, from violin to contralto. Results weren’t as clear: I could hear some changes in dynamics, but – even tho’ the notes were tied/sustained – each had its own attack.

The strangest results were after I exported the above to audio: the violin (as an mp3) played back perfectly, but the contralto shows almost no dynamic range!

I can upload the files, if anyone wishes to see/hear them.

Thanks again!

I went back to your file just now and this time no dynamics worked. I then reapplied the Halion template (Play menu -->Playback template → Halion HSSE + HSO Pro and it works fine again. Either you have changed something in the configuration or there is something buggy in the Halion library itself. But please do what I did and check if you don’t find it now working OK. At least then we’ll know we 're getting the same behaviour. Particularly the middle screen using CC1 as this is the most common way to set things up (if not using a grand staff)

Hi, dko22 – thanks, again, for spending time on this!

I just re-opened the original file, and followed your instructions on opening the playback template. The playback, unfortunately, remained the same as yesterday: that is, dynamic automation drawn in (in the midi cc panel) does have an effect on separate notes – that is, it will change the dynamic of any newly introduced note, relative to previous notes – but it has no effect on a sustained note alone.

To further test this, I added a single sustained note to the end of my Flow 1, and drew in a dynamic curve that goes from highest dynamic to silence, but I couldn’t hear any such decrescendo there, upon playback.

However, in the process of creating more exercises for my students, today, I found that I was able to draw in automated dynamic curves that do playback (relatively) accurately – tho’ some instruments (bowed strings) tend to respond better than others (human voices).

If you wish, I can easily upload the above files – just let me know.

Thanks, again, for your time and help!

by all means do upload the files -perhaps they will shed some more light

Thanks, dko22. Here you are:

*I believe that this is the original one I sent – I’ve since added that one sustained note (bar 7, Exercise 1) that I mentioned:
Morris, 1-to-1, Appendix, Set A.dorico (918.9 KB)

*And here’s a little project that I did today, in which a cello sample (Halion, again) does seem to respond to drawn-in dynamic curves (see, especially, the cello in bars 16 and 17):
Creating ‘A’ structure that is similar - cello.dorico (914.5 KB)

@thinkingMusic.ca hello,
If you are using Halion SE + Olympus Choir for playback, then you should check your Endpoint
configuration in Play Mode, as shown in the screenshot below:

For me the dynamics work fine for slurred notes.

Best wishes,

Thank-you, Thurisaz!

Unless I’m missing something (which could be!), though, I think my settings are the same as yours:

However, as mentioned, the sustained voice notes do not seem to respond to drawn-in gradual dynamic changes.

your settings are the same as @Thurisaz and correspond to the defaults for this library so I see nothing wrong here. Your cello and piano piece also responds the way I would expect.

As for “Morris 1-to-1” please have a look at my screenshot which show the piece the way I would expect for checking.

There you can see the primary dynamic (f), the velocity and the selected CC channel which is here CC1.

Look at the Expression Map screenshot

The primary dynamic is velocity and the secondary is CC1. They will both be active and should match each other. The reason for the small discrepancies is in Playback options you have left the humanise settings to default rather than switching them off. Entirely your choice and won’t make a significant difference in this example.

Here, I would expect to see a clear hairpin on the last long soprano note – dim. within the note is after all what we’re testing. You won’t hear much difference during notes in bar 5. This partly because the notes are not that long and party because the dynamic range of these choirs doesn’t seem to be that great to begin with anyway (you can control the CC v velocity scaling from within Halion but I haven’t really tested it and that’s not the issue here). Tie bars 6 and 7 in sops and draw a CC1 line from max to min like this

Surely you hear a difference now?

Many thanks, dko22!
I hardly even knew about these Expression Maps interfaces!

I think I’ve now set everything up as you’ve shown (including tying the last two bars of soprano, and extending the diminuendo hairpin further, to a pppp), but . . . I’m afraid the aural results aren’t much different. I recorded only the last three bars – soprano only – with three different settings:

  1. With ‘humanize’ still ‘on’ (dynamic curve set to ‘2.5’), and drawing in a midi cc decrescendo that begins at 127 (ie, much louder than the dynamic that immediately precedes that, but identical – I hope – to what you showed me) and dim’s to 0. Here’s a screen capture, followed by the mp3:

    . . . you can hear a few problems: 1) while there is a small diminuendo, it never progresses anywhere close to the “0”, or/ “pppp” , that is sought; 2) even tho’ the last two soprano notes are tied, we still hear an attack on that second note. 3) of course, the beginning of the midi cc diminuendo is far too loud.

  2. The same, but with humanize off, and the dynamic curve set to “1”:

    . . . the diminuendo is a little better, but it’s still far from complete, and we still hear that last soprano note’s articulation. Of course, the beginning of the midi cc is still too loud.

  3. The same as 2), but with that midi cc diminuendo now dove-tailing with the level of the preceding dynamic:

    . . . because of the lowered beginning dynamic, the diminuendo is less noticeable – but, as in all of the above, the final dynamic is still way louder than the desired “0” / “pppp”. And we still hear that last soprano’s articulation, even tho’ it is within a tied note.

I’m not sure why mine isn’t playing back properly, but I’ll understand completely if you don’t want to pursue it any further – you’ve already been super-generous with your time! :smiley:

of course the beginning of the CC1 dim. is too loud – it was designed to get as much of an effect as possible for testing purposes :smiley:

Anyway, I made a little edit to my previous post which you may not have seen and followed up on the CC v velocity scaling as below

I have put both sliders all the way to the right which means that the dynamics are controlled almost exclusively by CC1. They appear to have more effect than whatever is in Dorico itself. As we know that velocity cannot change the dynamic within a note, this appears to be the best approach as there is definitely more difference. However a complete morendo you do not get and I’ve just found from this thread Does anyone like/use Soundiron Olympus Choir Elements? | VI-CONTROL that the light versions such as the Micro version supplied with Dorico lack seem to lack the dynamic layers of the full version.

In other words, you’re probably stuck – there is simply a limit to what’s possible.

Thanks again, dko22!

As you can likely tell, I don’t know much about expression maps and various aspects of automation, but I was wondering about this, even before you pointed out these dynamic sliders (between velocity and cc as controlling the dynamics). There seems to be a similar duality involved in the Expression Maps “Base switches”:

. . . These (Volume dynamic and Secondary Dynamic) seem to offer some choice over the method of controlling dynamics – in fact, it offers both, but I have no idea how “both” works (e.g., if ‘primary’ volume dynamic is “note velocity” and “secondary” is cc#1, does it mean that note velocity is over-ridden whenever the mod wheel is used, or vice versa . . . or something else again?).

At any rate, I experimented with the combinations available there, and nothing seemed to make any difference.

Regarding the relative paucity of dynamic layers in this choir library: even so, shouldn’t the mod wheel be able to control relative volume – objective volume, if 0 -to- 127 is specified – regardless of the number of sample layers available?

(And yet another question) Of course, another way to achieve that complete diminuendo-to-silence would be to simply use the mixer. BUT: I can’t seem to find any way of recording live mixer information. Is this not possible in Dorico?

Again, my apologies for my very limited knowledge of these things!

If there are Dorico video tutorials that cover these topics in detail, I’d sure welcome a link or two!

Thanks again.

Note velocity is not over-written as a rule – what happens is that it and the CC are given the same value (unless the min/max range is scaled but let’s leave this for the moment) but how that is interpreted will vary according to the individual library’s programming.

Spot on!
In fact it was very remiss of me not to test both the MIDI default dynamic controllers (the humid heat is getting to me…). Although every library has it’s own preferred control method, the MIDI standard is that CC1 = dynamics in the sense of the tone of the sound and CC11 = is the instrument volume level. If I draw the line in CC11 instead, it does indeed fade to zero. That doesn’t always work – you basically need to read the VST documentation – but does seem to here as well as some other libraries I use. I hope you find the same if you redraw the dim. in the CC11 lane.

The area of playback dynamic control is extensive and rather complex because of the different methods used for different libraries. I haven’t found any tutorials which go into sufficient detail in this sort of thing though looking at the easily-found YouTube videos on Expression Maps for instance might help.