What's new

Divisi Strings with Spitfire Libraries

kinginknyc

New Member
I own both Mural and Sable, but i'm not sure the best way to set up the Sable libs for divisi.

For example the Mural Vln 1 section has 16 players, Sable's Vln 1 section has 4 players. So my question for you all is how are you emulating an 8 player section ?? Obviously one Sable Vln1 instrument gets me a 4 person 1a and 1b section, how are you getting the other 4 ??

Thank you in advance !!!
 
Don't get too caught up in the player arithmetic when layering libraries. It doesn't really work like it does in real life. Consider this: in real life, when you start with 8 violins and add another 8, what happens to the sound? It becomes thicker, more lush, less detailed. But what happens when you take Mural's 16 player V1 and layer it with Sable? It actually becomes more detailed, because you begin to hear the nuances from Sable. In the real world, these would get subsumed (unless you mic and mix them separately of course).

So do what sounds right. Listen to an 8 player section and replicate what it sounds like. I'd indeed start with a straight Mural V1 + Sable V1 and work from there.
 
Unless you are writing for a real orchestra, what should matter is mainly the sound that you get. When using sound libraries, it doesn't matter if you end up with 30 cellos and 15 flutes in your mock-up if that sounds right. Use your ears to get as close as possible to the sound that you want to achieve.
 
I do write for real orchestra .... my mock ups will be performed by an actual group, so I can't have a HUGE string section mock up that dwarfs the real thing. I'm not getting overly obsessive with numbers just wondering what other have done ....
 
  • Like
Reactions: bun
I do write for real orchestra .... my mock ups will be performed by an actual group, so I can't have a HUGE string section mock up that dwarfs the real thing. I'm not getting overly obsessive with numbers just wondering what other have done ....
I would say that even if you write for real orchestras, the point of a mock up is to sound as close as possible to what you want the orchestra to sound like. So again, as @tack said, if layering Mural V1 & Sable V1 sounds more like a real divisi sound than stacking Sable V1, I would advise you to start from there.
 
My best guess is one of the four answers, but it would be nice to hear what Spitfire say:
  • For Mural divisi use Sable (4,3,3,3,3) but layer or double the samples up by pitch shifting (8,6,6,6,6)

  • For Mural divisi use Albion 2 Loegria (the strings are 8,6,4,4,3)
  • For Mural divisi Somehow do a quasi divisi with Mural - by lowering the volume of the samples and adding a little more leader microphone.
  • Avoid Mural altogether and use Loegria and Sable layered. (Loegria 8 + Sable 4 + Sable 4) then when doing divisi use Loegria half-patches and only one sable (Loegria half 4 + Sable 4)

If I had all the libraries I'd tried it out - but alas I only have Sable at moment.
 
My guess is kinginkync was just referring to https://www.spitfireaudio.com/blog/bml-mural-symphonic-strings-announced/ (this blog post) where Spitfire writes:

MURAL's Symphonic String lineup has been designed to exactly match our SABLE chamber strings library, so you can easily use Sable as the divisi sections for Mural.

I don't recall seeing Spitfire say technically how to divisi. But, as always, just do what sounds right. :)
 
Very interesting, thanks for the info.
I’d need some clarification regarding divisi, in particular, using Spitfire Audio Chamber (SCS – the new Sable) and Symphonic Strings (SSS – the new Mural).
I’d like to know also @driscollmusick ’s opinion, since I’ve read in other threads that he uses SSS and switches to SCS for divisi.

1) First of all:
MURAL's Symphonic String lineup has been designed to exactly match our SABLE chamber strings library, so you can easily use Sable as the divisi sections for Mural”, as written by Spitfire and quoted by @tack .
I’m working on a track where, in the same instance of SSS 1st violins staccato in Kontakt, these 1st violins often play two different melodies at the same time (never the same notes of course, at least not in the same octave). This would obviously be impossible in a real orchestra without splitting these 1st violins into two groups of divisi (in this track the same happens with 2nd violins, violas and cellos: each of them often play two different stacc. lines simultaneously).
So, also according to Spitfire’s post here above, should I use SSS only when the 1st violins stacc. play “alone” (i. e. without a second line of 1st violins stacc. overlapping them) and I want the richer sound of SSS 16 1st violins, and, when there’s a second line of 1st violins stacc. overlapping them, switch them from SSS to SCS and use SCS also for the second line? Or use SCS for the second line only?
Sometimes in this track there’s a melody played by the 1st violins legato together with a different melody played by 1st violins stacc.: like above, should I use two instances of SCS 1st violins (one leg. and one stacc.) or SSS for the leg. line and SCS only for the stacc.?

2) As regards the “transposing trick”:
For Mural divisi use Sable (4,3,3,3,3) but layer or double the samples up by pitch shifting (8,6,6,6,6)
In my case, when I have two lines of 1st violins stacc. playing simultaneously, would this mean I should use two instances of SCS 1st violins stacc. and apply this trick to both of them? So that, if I got it right: 8 (SCS, first line) + 8 (SCS, second line) = 16 (as the total number of 1st violins in SSS)?
By the way, is this practice really necessary (even if the two lines of 1st violins stacc. don’t play the same notes)? I’m not sure I can hear any phasing issues when they play together.

3) I guess it’d help to adjust the three mic positions (close, tree and ambient mics, offered by both SSS and SCS) in a different way for each instance (besides their volume)?
For example, in my case, using C, T and A in one instance of 1st violins stacc. and only C and T in the other?

Thanks in advance for any advice!
 
Last edited:
The goal here is just realism. If you have a real-life Violin 1 section that has 16 violins (SSS, Violin 1), each note you play with them will sound as though 16 violins are playing it. The problem is if you play more than 1 note at a time (unless they are reasonable double-stops--usually not) in essence you suddenly doubling the size of your ensemble from 16 to 32 violins (48 if you play a triad or 64 players if you play a seventh chord!). This often sounds fake (a "phantom" ensemble that grows and shrinks willy-nilly), though not everyone cares about that...

But at any rate, SCS Violin 1 section is the sound of 4 violins. So imagine if you have a section of 16 players in real life and if you want to have them split in half playing 2 different lines, you need to create the sound of 8 violins per line. To achieve that in SCS, you'd have to lay out two separate SCS Violin 1 sections playing each part (4 players per section x 2 sections per line x 2 lines = 16 total players = 1 SSS Violin 1 section). To avoid phasing, you should do the transposition trick to the 2nd of each pair of SCS Violin 1 sections, as in this case you are having the same VST play the exact same line. Honestly, it's a good idea to do it all circumstances because even a single note shared by the same instrument can stick out badly.

You can then switch between the SSS 16 players and the four SCS sections and it should sound realistic. It goes without saying they should never play at the same time--otherwise you are just recreating the same *phantom* strings section problem. This can all be further complicated, too, by how you want Violin 1 to interact with a Violin 2 section and whether or not you are even trying to do the traditional violin section setup (vs a film score setup with a larger group of violins playing a single part).

I admit that I've been cheating with only a 2:1 SCS to SSS divisi, though technically I should be doing the 4:1 I described above. Maybe I'll work on that...
 
The goal here is just realism. If you have a real-life Violin 1 section that has 16 violins (SSS, Violin 1), each note you play with them will sound as though 16 violins are playing it. The problem is if you play more than 1 note at a time (unless they are reasonable double-stops--usually not) in essence you suddenly doubling the size of your ensemble from 16 to 32 violins (48 if you play a triad or 64 players if you play a seventh chord!). This often sounds fake (a "phantom" ensemble that grows and shrinks willy-nilly), though not everyone cares about that...

But at any rate, SCS Violin 1 section is the sound of 4 violins. So imagine if you have a section of 16 players in real life and if you want to have them split in half playing 2 different lines, you need to create the sound of 8 violins per line. To achieve that in SCS, you'd have to lay out two separate SCS Violin 1 sections playing each part (4 players per section x 2 sections per line x 2 lines = 16 total players = 1 SSS Violin 1 section). To avoid phasing, you should do the transposition trick to the 2nd of each pair of SCS Violin 1 sections, as in this case you are having the same VST play the exact same line. Honestly, it's a good idea to do it all circumstances because even a single note shared by the same instrument can stick out badly.

You can then switch between the SSS 16 players and the four SCS sections and it should sound realistic. It goes without saying they should never play at the same time--otherwise you are just recreating the same *phantom* strings section problem. This can all be further complicated, too, by how you want Violin 1 to interact with a Violin 2 section and whether or not you are even trying to do the traditional violin section setup (vs a film score setup with a larger group of violins playing a single part).

I admit that I've been cheating with only a 2:1 SCS to SSS divisi, though technically I should be doing the 4:1 I described above. Maybe I'll work on that...
That's how the math works, but it rarely works that way in practice in terms of recorded sound.
 
Thank you very much for your explanation, @driscollmusick !
I’ll try to take your advice since my aim is to sound as "realistic" as possible.

As you know, it works out arithmetically for 1st violins and violas, although not perfectly for 2nd violins and cellos:
- 2nd violins (3 SCS players per section x 2 sections per line x 2 lines = 12 total players = nearly 1 SSS 2nd violins section, which instead has 14);
- cellos (3 SCS players per section x 2 sections per line x 2 lines = 12 total players = nearly 1 SSS cellos section, which instead has 10).
But I don’t know if such a small difference could be noticeable (maybe to super-trained ears?).
Also considering that both Spitfire and others suggest to use SCS for SSS divisi, like “The Sound Architect” site, which writes:
Having smaller sections also allows you to use multiple instances of a section to write divisi (dividing up each section into multiple lines). Additionally you could also use SCS to add definition to parts where you’ve used larger sections.

That's how the math works, but it rarely works that way in practice in terms of recorded sound.
What would you suggest then, @jbuhler ? Any advice is always more than welcome, thanks!
 
Last edited:
The math is misleading. When you layer two 16-violin patches from a single library, for instance, yes that is technically the sound of 32 violins. But they are 16 violins placed exactly in the same space as the original 16 violins. You're not adding 16 physical seats, since the second 16 are seated in the exact same spots as the original 16, all the sound holes and microphones of the second line are occupying the exact same physical coordinates of the first line. So it's not technically a physically "larger" section.

so instead of:

x x x x x x x x x x x x x x x x + x x x x x x x x x x x x x x x x (16 + 16)

it's more like

X X X X X X X X X X X X X X X X (just a thicc 16)

...If that makes sense. And then think about group size, and how as groups get larger, the effect becomes harder to identify. Sure, stark difference between solo and a2. And between a2 and a6, etc. But the difference between a2 and a6 is going to be more noticeable than the difference between 22 and 26, and even more still than the difference between, say, 42 and 46, even though the change in the number of players is the same in each case. Diminishing returns or something like that. So 48 "virtual players" is not necessarily 3 times as big/powerful as 16.


So how does this help you? Eh, it doesn't. I guess my point is, just looking at the arithmetic of "virtual player" counts alone is not enough info to make a judgement. I hate to be the "just do what sounds right" person, so I'm glad someone else beat me to it.
 
Last edited:
What would you suggest then, @jbuhler ? Any advice is always more than welcome, thanks!
[/QUOTE]
I think math is rarely reliable as anything but the roughest guide when it comes to samples. I find SCS works really well as 1/2 section for SSS despite the fact the proportions shouldn't work. (Then too you can make SCS sound much bigger than it is.) I think you can layer many libraries and it doesn't sound at all fake or like the 40-50 player (or higher) first violin section that the numbers suggest it should be. I think it's weird that sometimes the math seems to work (especially with low player counts) and sometimes not really at all and it's often hard to predict. Alchemy. Magic. Whatever.
 
I tried SCS for divisi with SSS. I found the result not convincing. It sounds like two different libraries and that's what it is.

Best results: using a similar or the same library for the divisi notes and reduce the volume a bit. And only for monophonic legatos. For staccatos: play two notes.

Don't count sample voices. A simple dynamic crossfade doubles or triples you players and nobody cares.
 
Best results: using a similar or the same library for the divisi notes and reduce the volume a bit. And only for monophonic legatos. For staccatos: play two notes.

Don't count sample voices. A simple dynamic crossfade doubles or triples you players and nobody cares.
I do this sometimes too, though usually with SCS rather than SSS and find it generally works. I haven't found the technique of reducing volume especially convincing for SSS however. If you are having luck with it, I'll have to go back and try it some more. I should add that when I use SCS as half section I layer SCS and SSS for full section then cut to just SCS for divisi.
 
Top Bottom