Unison And Divisi Considerations For Instruments In Either Single Or Ensemble Instances In Spe...

[DaddyO]

I am aware that historically VSL has purposely avoided separate sampling for multiple instances of single instruments (two oboes, for example), even sections (second violins, for example). This post is not an attempt to rehash that, though I admit I personally find the conclusion curious. Still, who am I after all, and it's hard to argue with the success of VSL. What I really hope to do is learn how different people approach the reality in this area that VSL presents. I want to keep the Dimension series out of the discussion for now, because as I understand it those products seem to me to address some of my concerns, but they remain outside my budget for the present. Anyway, other than the Flute, which for some reason comes in Flute 1 and Flute 2, in the Special Editions and the individual Instrument and Ensemble products, I am presented with single sample sets for each instrument. I get solo samples for single instrument instances, and I get ensemble samples for ensemble instances. The Horn ensemble is 4 players (a4), the remaining brass and woodwind ensembles are 3 player samples (a3). If I want a unison line for multiple instances of an instrument, I have two choices. Either use multiple instances of the single instrument patches and implement the transposition trick to avoid phasing, or a single instance of the ensemble patches. If I want to write an a2 Clarinet line, for example, this have no choice but to use multiple instances of the same single instrument samples. If I want a divisi (or chord)line, I have two choice. Again, multiply instrument instances and use the transpostion trick, or use the pre-made ensemble patches. BUT (!), if I use the ensemble patches I then have 3 or 4 instruments, depending on the ensemble size, playing each note of the divisi or chord. This, of course, is not the sound I would be looking for. So as I see it, I really only have the one choice, multiple instances of the single instrument. So in the end, am I correct in seeing that the only use for the ensemble patches is when I want the sound of exactly the number of players pre-sampled, and even then only in a unison line (because a triad for a 3-player patch would result in a 9-player sound)? How do people actually approach this issue. I imagine some might say that the difference between a 3-player ensemble and a 2- or 4-player ensemble is negligible and irrelevant. Perhaps they are right, but why then did VSL make the distinction of a 4-Horn ensemble and 3-player ensembles for all the rest, if indeed there is no significant distinction? Please understand, I am not trying to gin up any controversy. These are honest questions, as I am in the process of setting up my orchestral template. I want to learn from this community by seeing the different responses to my questions. (FYI, I put paragrahps in my post, but they don't seem to be reflected in the presentation of it.

[Conquer]

You can sometimes avoid unison sample-phasing by using different articulations of the same instrument - for example, sus and perf trill.

Re. the ensemble question, my personal view is that If a MIDI arrangement is going to be used in the mix (as opposed to a mock-up intended to be replaced by real players), it doesn't matter if it departs from reality in places. If you like the sound of a 9-player flute triad, go with it!

[William]

There are 2nd violins in all the ensembles now. Also, solo violin has a 2nd violin.  It is a transposed version of the same samples but works very well.  Also, for something like clarinet I've done two things - one is to transpose the part up a half step and pitch shift down (which generally sounds better than the reverse). The other is to use the ensemble clarinets on one part, and the solo on another.  You might object to that for the number of instruments, but it actually sounds very rich and good in an orchestral context as long as that number remains constant. In other words, you cannot keep on adding ensembles which thickens the sound, but having one ensemble that is doubled with a solo when in unison and then splits from the solo sounds totally natural if that numerical relationship is maintained.   

Another thing I've done and really like for oboe is to use the similar instruments - oboe d'amour for second oboe, etc.  Also, on woodwinds or brass one can often use separate sample sets for specific articulations.  For example, short portato on one intrument,  staccato on another works in most musical contexts.  Similarly, you can use the non vibrato articulation on sustain/legato, combined with the vibrato or progressive vibrato.  That might sound questionable, but often in orchestra  one player will have more vibrato than another in an ensemble and this is a way of increasing the complexiity of the overall sound.  On solos in brass I really like the sound of combining the progressive vibrato trombone with the ensemble trombones.  Again, the exact numbers of players does not matter if the initial number is maintained.  I have been using ensemble trombones for 1st trombone, solo trombone for 2nd, and bass trombone for 3rd and it works extremely well, even though the numbers are larger than most live orchestras.   This is of course  ignoring the Dimension Brass and Strings which completely eliminate the situtaion.  The main point is there are many options for getting a 2nd player that work very well and are not difficult to implement, especially if you set up a template.

[noldar12]

Depending on how much of SE you have, there are some other options as well.  If SE vol. 2 is included, you have two oboes: French and Viennese (personally, I would use Viennese on the subordinate parts).  You also might be able to get away with using the regular clarinet for a lower part and the Eb clarinet for the higher part.  As you mention, there are two flutes, and two bassoons do exist, though the 2nd bassoon is only available as a full download library.  If you don't currently have SE vol. 2, adding it will significantly increase your instrument flexibility.

For my own purposes, I use the transposition method, and having the older MIR SE, will then use two different character presets to differentiate the sound between the two instruments.  As for divided strings, the old "traditional" VSL method is to layer orchestra and chamber strings, and then split out the chamber strings for the second divided string part.  As for the 2nd violins, as above, I use the transpose function coupled with a different character preset.  Note that I tend towards being rather precise with the articulations, and generally prefer similar articulations in the different parts, so I don't normally use the different articulations on unisons method.

Long-term, depending on the size of the ensembles you are creating, saving up for MIR 24 might not be a bad idea (assuming 24 parts are enough for what you are doing).  Given its limitation of 24 tracks, one can get by with a less powerful computer than what one would likely need for MIR Pro.

[DaddyO]

Thanks for the comments. Any more are welcome.

[didger]

This pretty much echoes comments above, but there are indeed two flutes, multiple types of oboes, and two bassoons. If you ever need three of those instruments playing independent lines, chances are good you can orchestrate the 3rd part using some of the same samples but never overlap the same notes. And yes, I have at times, as you suggested, just accepted that 3 clarinets doesn't sound THAT different from 2 and used those samples. As long as it's in the mix of a big orchestra, it doesn't matter too much. For the most part, however, I do whatever I can to have different samples playing for each player whether in unison or not so I can a) composer more naturally, b) not change the character of the "players" and c) play or randomize each part separately, so my flutes don't hit every note at exactly the same time.

But the main reason I'm responding to this post is to once again plea for a Clarinet 2. :) The different types of clarinets don't work nearly as well as substitutions for a second instrument as the different oboes do. For me, a second clarinet would complete VSL's woodwinds.

Brass on the other hand is much more difficult until you can afford Dimension Brass. If you need 4 independent horn lines, for example, then I guess you have no choice but a lot of transposing. The Dimension release was an absolute Godsend for me.

And a quick answer to one of your questions, as to why VSL chose 4 horn players versus 3 of everyone else, I'd imagine it's because that's what you'll find in real orchestras. Four horns is the norm. The other brass and woodwinds are harder to claim a standard for and two could arguably have been a better choice. Three woodwind players in unison is rare in the whole spectrum of orchestral music, but it was probably just considered the best bang for our bucks and for the software and computer power at the time these were recorded, back in the early days of VSL. Any composer who knows how a real orchestra works would rather have a full set of samples for every player that would be in an orchestra, but back in the days of GigaStudio running on your Pentium computer with 1GB of RAM and 200GB of hard disk space, it was inconceivable.