So, with full knowledge that expression map support is still incomplete, I gave Dorico 2.2 a shot trying to drive orchestral VIs and ran up against several dynamics problems. I mention these with the understanding that this area is due for improvement in the hopes that this feedback will be considered:
A) The Dynamic Curve playback system is simply insufficient to do the job we need to do - get our VIs to behave well over their recorded range. Whatever its mathematical elegance, the curve fails to:
allow us to work only in say only ppp-fff range in the score but reach 1-127 on the CCs.
deal with the fact that the sample libraries themselves sometimes have non-linear volumes, which means there is no possible value of a sigmoid curve that will correct it.
deal with the fact that many libraries have transitions between recorded dynamics that we’d rather not rest upon, in order to avoid the doubling/out-of-phase crossfade sound. E.g. if we know there is a transition at 90 we might want to situate mf clearly below 90 and f above.
The right answer is to just let us set the target values for each dynamic and interpolate between them. This is what everyone else does and it is vastly better than dynamic curve. It is also best to allow a separate set of mappings for velocity.
B) The Volume Dynamic binary switch between Velocity and a CC is also inadequate. There are many sample libs that e.g. drive dynamics by CC whilst also requiring specific values of velocity in order to control accents/onsets/legato/portamento etc. When in CC mode we seem to be locked into a velocity of 100
C) There seems to be a bug in accent velocity boost - it seems to set an absolute velocity rather than a relative one? E.g. at p, I get non-accented is 44, accented 100; at f 83/100; at fff 111/100 (note the inversion).
I sincerely hope you seek and get input from people with lots of experience with orchestral sample libraries (other than Steinberg’s) as you work to improve this area of the program, as it currently seems to be insufficiently informed by practice and need.
Thanks for the feedback, Rich. Our plan is indeed that in the future we will allow you to override the dynamic curve on a per-expression map basis, and to provide a means of adjusting both a controller and velocity for dynamics.
I was most interested in this reply on dynamics. I am finding that fast passages in Violin Solo Combi from the default Halion Sonic SE play back at a low level, even when the passage is marked Forte or louder. Example: in Beethoven’s SQ 15 at the beginning, the first violin plays a slow intro pp followed by a fast passage f in 16th notes. There is no increase in volume, even if I change the fast passage to ff. The assigned expression map is HSO Violin Solo. There is also no difference, if I isolate the first violin music in a separate project. As implied, this volume effect is very marked in string quartets, where the first violin (who typically gets most of the fast material) almost disappears in fast passages and re-emerges in slow passages. I am running Dorico on a HP Omen series tower running W10. I also tried using the 4 core setting in Halion but no improvement. What is the best route to getting more realistic playback in string quartets?
Incidentally, I was fascinated to see some large orchestral works in your example videos. How did you get these into Dorico in reasonable time? Even with a good quality PDF string quartet score from the Net, a scanner like Photo Score leaves plenty of errors. You must have incredible patience.
Inputting music into Dorico is pretty quick, so it’s no great hardship to input large-scale works, particularly once you’ve mastered the various time-saving tricks to save time, like using Lock Durations after copying music to another instrument, multi-paste, Explode, and so on.
Re: the violin dynamics, the issue is that when you have fast passages, particularly those with slurs, the rise time of the legato violin patch is too slow to allow the notes to speak properly. Take a look at this earlier thread.
Thank you for the reply. The fast passage that I cited from Beethoven Quartet 15 (Op 132) was in fact legato; however, it is quite typical for the violin. The earlier thread that you cited above was actually a violin range problem; I have also noticed this type of problem, but with the cello. Strange that no one has mentioned the violin fast passage problem, since it is really obvious in string quartets, both classical and modern. A similar problem can be heard with Flute in the Cygnus the Swan example.
Looks like it is time to try Note Performer, which Anthony recommended highly. Should I expect an improvement in violin playback, or do I need to look towards more expensive sample libraries to get more realistic playback?
If you use an instrument with a slow attack, you will probably hear reduced volume in fast passages from just about any library unless their expression map switches them to a a quicker attack in those passages. (This is basically what Daniel said–if I interpret his post correctly.)
If I slow down the solo violin attack by rotating the Halion SE instrument control clockwise, fast passages become silent, i.e., they do not play back at all. Note that these “fast passages” are very typical for the violin and occur everywhere. As far as I can tell, the Halion SE solo violin attack can not be improved. You cite NotePerformer and GPO, which I have not yet tried; are these instruments better in this respect? I have not seen “rise time” cited as an indicator of instrument sample quality, and yet it really is crucial. At least I can audition NotePerformer for free, and it should work right away.
Re the problem with low volume in violin fast passages that arises from slow rise time in the samples. NotePerformer 3 fixes the problem and gives incredible clarity in string quartet scores, as well as Dorico’s example scores. Many thanks.