OK–I have installed the May '21 BBCSO playback template and tried to figure out what is going wrong. The following passage plays back very unevenly–almost like the quarter notes are swinging…
Maybe check Playback options under Timing.
Jesper
1 Like
Yes, I have tried a few tweaks, but something seems to be delaying the onset of slurred notes.
BBCSO doesn’t play very well with the built in ‘jitter’ automation. Its too sensitive and picking up on the jitter which becomes obvious with legato’s or slurs. Go into CC11 and add a straight line to the peak of your phrase, and then to the final C. Small changes to these expressive lines make noticeable changes. When you bring up CC11 you’ll see the auto jitter (which works OK for shorts, not longs).
Fortunately we have Write mode key editor which makes this so much easier. I’m content, but if the team were to add automatic phrase detection for CC’s (just pick the low/high points of phrases) that would get i t pretty close.
On my list to see if there’s any tweaks that can improve this. I work with NotePerformer to begin with to not have to mess with it, then switch to BBCSO when ready for export.
1 Like
Thanks, Dan. But I am not sure I see what you are talking about with CC11. Do you mean draw in CC11 as a straight line?
Crude example but to illustrate on a scale
Phrasing generally follows the peaks of the phrase and such, so simple automation like this works well. The other thing for good playback is you’ll see I shorten the note at the top - for a breath, or just natural human phrasing. It’ll sound 10x better.
I get all sorts of weird sound without doing this, the plugin doesn’t react well to jitter CC. Here’s the CC1 jitter with BBCSO template
CC11 is the same
OK maybe one way to avoid this out of the gate, in the expression template
Change the CC1/CC11 for the longs (legato is shown) to 64 min and max. That’ll just give a nice flat value so it’ll sound boring but not bad. Later go in and phrase it. I can’t think of any instrument offhand where jitter in a long makes sense. Jitter only makes sense with shorts I think.
Might be other side effects but seems better, I’ll try it later.
1 Like
When the phrasing is with two slurs to a bar like this at the beginning, I’d often intend a swing effect in fact and in this short extract, I don’t find the playback is so far from what’s notated.
But in general, first thing I’d do is switch off all dynamics Playback options. Libraries which are sensitive to small changes to CC1 (dynamic cross fades or levels) don’t get on with the well-meaning but often destructive “humanisation” features. CC1 is even more important than CC11 but both active dynamic controllers are affected by this. If this is done, the CC11 line will be absolutely straight and jitter-free without needing to do anything else – except when there is a patch change.
There is a certain amount of phrase detection built into the better libraries including the BBC these days so you don’t necessarily want to do much else but if you do you can of course choose between hairpins (particularly when you want to emphasise the phrasing) or drawing an appropriate line in either the CC1 or CC11 lanes. CC1 is tone/expressiveness and CC11 pure volume so it depends on what you want to change.
The other potential problem is that the note entry latency keep changing when moving from a long articulation to a shorter one (or groups of slurs which will briefly reset to “nat”), though the timing doesn’t seem too bad in this short extract and the BBC isn’t among the worst libraries in this respect, I find.
That gives flat line dynamics for longs, very good. I did a test with shorts and it doesn’t affect that which is also good for this case. Maybe because just above there are a bunch of controls for short dynamics.
However I did a test with a very expressive solo line I have, with default (flat) dynamics BBCSO plays it essentially flat with no expression, so phrase detection isn’t working or it doesn’t have it. The legato patch is being triggered. Some CC work as I indicated above makes a world of difference on it.
“When the phrasing is with two slurs to a bar like this at the beginning, I’d often intend a swing effect in fact and in this short extract, I don’t find the playback is so far from what’s notated.”
Um. What? sorry dko we are not on the same page at all!
probably we are – I’ve maybe just expressed myself badly. You asked for instance
That’s the built-in latency which a number of libraries apply to the beginning of a legato phrase. Cinematic Studio, for instance, document very clearly exactly what the delay is, Spitfire don’t (to my knowledge) but that doesn’t mean it’s not there. If it’s annoying, you simply need to either apply an offset in the “Properties” panel of Write mode or draw in the Key Editor.
this is because of the above as there are two separate legato phrases plus the fact that there is often a slight accent or push at the beginning of a legato phrase in the BBC and other libraries. If you wanted a smooth slur across the whole bar then you should have written one.
it is probably working – it’s just the effect is often rather subtle to the extent you may not notice it. I think we’re in agreement that if you want to make the phrasing clear then you should intervene with a hairpin in the score or manual line in the Key Editor,
Exporting and looking at the waveform there does appear to be phrasing, here with my drawn in phrasing
However that can be deceptive, look at the loudness
Essentially flat - which is the most accurate perceptual experience and what I hear. This is the Loudness Display in Wavelab.
So I’d say it’s possibly working, but making no different perceptually so might as well not rely on it in any shape or form. Curious, where do you get the information that BBCSO does phrase detection?
I didn’t specifically-- I just sometimes hear it from the waveform sometimes but agree it’s probably partially subjective. Interesting test, anyway!
This particular line I have is naturally expressive and emotionally you can’t help but respond to it - I recognize that but know it’s inherent in the music and probably not perceptual. The loudness graph shows the reality, so I believe they’re not attempting phrase detection. I’d be unwilling to do so if I was them, too easy to go wrong, and then customers are fighting the VI.
And Hollywood composers are crazy these days anyhow, half the time it seems they’re doing sound design and not music, so they may want a phrasing which goes counter to the music or who knows what weird effect.
The plot thickens, in v3 of the BBCSO playback template there is this change
Somehow, I have v3 of the Playback Template, but my expression maps are still the old ones. This is the updated
Anyhow, this should fix the unevenness without changing the global setting for dynamic humanization, so double check your expression maps.
these settings (ignore default note duration) should have been there from the beginning. It’s irrelevant to me personally as I use the Core library which has no more recent update. And I have made my own adjustments to the EM’s, primarily making the legato patch the default.
I’m confused. When I open up the v3 BBCSO project from Steinberg, all the Expression Maps are correct. When I do a fresh import of the included Playback Template, but open up my customized template, it’s hit or miss. Some instruments have the updates and some do not. Can Expression Maps get stored on in the project?
Edit: Yes, they do! Yikes … I just did a test, they are stored. However resetting doesn’t fix it, apparently this is the way.
And testing it, indeed it fixes legato playback! So obvious.
and now in Dorico 4 we have the alternative of the new Library Manager where you can easily compare the Expression Maps and playback techniques between projects and quickly import the one you want into the current project without guessing which exported backup was used where…
Crud! I didn’t even realize Expression Maps and playback was in there too! Been busy elsewhere in D4, great, thanks for the tip and good job Daniel/et al!
Dang, this makes it sooo much easier …