Search Gear
 

Master Class: Virtual Orchestra Virtuosity

June 1, 2008
share

Long gone are the days when professional composers would need a wall of fully loaded hardware samplers just to provide a few hundred megabytes of memory in which to realize a virtual orchestra. Today's computers can stream dozens of gigabytes of symphonic samples to provide an almost limitless collection of instruments and articulations. Recent developments in software interfaces allow unprecedented control of performance parameters for heightened realism. Orchestral music on the computer has never sounded so good.

However, all of this power comes with a price. You have to be willing to shell out the bucks for the latest in computer technology and have speedy hard drives for the massive amount of samples required. The better orchestral libraries are not cheap, and the time that must be invested in becoming intimate with the contents, mastering the enhanced performance techniques, and maintaining a working system is greater than ever before.

The goal of this article is to document the experience of working with several high-end libraries in order to realize a single piece of music. This is not a shoot-out between sample libraries or a comprehensive review of the featured products. My intention is to demonstrate the differences in capability and character between four top-notch libraries while explaining what I had to do to create each realization. Overall, I want the libraries to speak for themselves.

For a test subject, I orchestrated a short introduction to a song that will be featured on my upcoming solo CD (see the sidebar “The Score”). Next, I sequenced the piece entirely within a single library, using only the sounds included in that specific collection. I then repeated the process from scratch with the next library, and so on. This resulted in multiple realizations of the same piece of music that ultimately sounded quite different from each other, allowing me to compare and contrast the libraries within a common musical context. (To hear the audio examples, see Web Clips 1, 2, 3, 4, 5 and 6.)

Piano is also featured in this orches-tration, as I am a piano player. Although some of the featured libraries contain pianos, none compare with a dedicated piano library. I leveled the playing field by using the same exact piano performance and sound with the same tempo map for all the renditions: what you'll hear between the different versions is a change in orchestral instruments, while the piano part stays the same. For the piano, I used 2008 EM Editors' Choice Award winner Modartt Pianoteq 2 ($380), an excellent physically modeled software instrument.

I also decided to do whatever it took to get each sample library to sound its best in realizing the orchestration, even if it meant tweaking one library more than another. I wasn't interested in a General MIDI file approach that forces each library into an existing construct. Rather, I wanted each library to sound its best and to document whatever process was required to achieve that end. Obviously, some libraries are more complex and capable than others, which is reflected in the amount of time and work necessary to realize the orchestration to the fullest capabilities of that particular collection. However, I was primarily interested in the final results, which, in my opinion, are what matter most.

Additionally, I decided to create a version that would employ what I consider to be a “best of” from the various libraries. The reason is a simple and pragmatic one: each library has its own sound as well as its own inherent strengths and weaknesses. Mixing and layering samples from multiple libraries as we see fit is exactly how composers work in the real world. In that spirit, the final version (see Web Clip 5) includes instruments from all four libraries as well as from other libraries not featured here, because the point of that rendition was to make the music sound as good as possible with what was available to me, and to let you know how I achieved the final sound.

MEET THE ORCHESTRAS

There are many orchestral sample libraries on the market, but only a few met my criteria for this article. For inclusion, each library had to contain a complete collection of orchestral instruments: strings, woodwinds, brass, and percussion. That immediately eliminated several otherwise good libraries, because there would be no fair way to compare an incomplete collection (a strings-only library, for instance) to the comprehensive libraries within the context of a full orchestration. Also, the quality of the included libraries had to be at the highest level, which eliminated several inexpensive and midlevel collections. (A couple of manufacturers never responded to inquiries about inclusion in this article and were dropped from consideration.)

I eventually settled on four libraries: East-West Quantum Leap Symphonic Orchestra Pro XP, MOTU Symphonic Instrument, Soni-vox Sonic Implants Complete Symphonic Collection, and Vienna Symphonic Library Vienna Instruments. Each of these libraries sounds great, but they also sound quite different from each other, with some lending themselves to certain styles better than others.

Setting up for this article was harder and more time-consuming than I had imagined. Just getting all the libraries installed, authorized, and running properly on both my desktop and laptop systems was a major endeavor. The libraries contain huge amounts of data comprising the instruments and articulations, and the only way to learn each library was to put the time in and go through it thoroughly. In addition, I had to learn three different software interfaces: Kontakt, MachFive, and VSL's proprietary instrument (more on all this later).

Because this is a master class, I will not touch on the basics of each program. Details about the included instruments and interfaces can be found on the manufacturer's Web sites.

FIG. 1: EastWest Quantum Leap Symphonic Orchestra Pro XP.

FIG. 1: EastWest Quantum Leap Symphonic Orchestra Pro XP.

EASTWEST QUANTUM LEAP SYMPHONIC ORCHESTRA PRO XP

The EastWest Quantum Leap Sym-phonic Orchestra Platinum Complete bundle ($1,195 [MSRP]), which includes Platinum Pro XP, takes up 138 GB of drive space. EastWest (soundsonline.com) offers several smaller collections that will suit some users. Currently, EWQLSO requires Native Instruments Kontakt 2.2 or later (see Fig. 1). However, EastWest says its custom PLAY software interface will be available for this library soon. (PLAY will be 64-bit compatible and tailored for enhanced performance control of the company's sampled instruments.)

The Platinum Pro XP edition provides an extensive collection of instruments and articulations. The samples are phase accurate and can be mixed and matched to create the desired ambient balance of stage to hall and/or surround mixes. EWQLSO sounds especially suited for bold, epic pieces and movie scores, but it is very demanding on computer resources when utilizing multiple perspectives at the same time.

The library samples are recorded with three mic perspectives: Full, Close, and Surround. I decided to use mainly the Full microphone placements in my final mix because I was not creating surround mixes in the context of this article, and using all three perspectives can be a bit much in a stereo mix, especially if you feel that the particular hall the instruments were recorded in isn't right for the type of music you're doing. I tried using all three positions simultaneously and they worked great together: there were no phasing issues, and EastWest seems to have thoroughly matched and mapped every corresponding program and sample properly. Kudos to the company for executing this feature so well.

Layering a touch of the Close mics with the Full perspective for certain instruments is just the ticket for getting a specific melodic line to jump out or to add clarity to a part, much like bringing up a few close mics on a recording date. Just copy the part to a new track, change the instrument to the Close mic version, and add to taste. This feature works very well and really makes EWQLSO stand apart from the other libraries in this regard. You can hear it employed in several of the woodwind lines in my piece.

Obviously, the tax on your computer setup is much higher if you want to hear all three perspectives at once, because the number of voices is doubled or even tripled when layering two or three patches per instrument. Most users seem to work with the Full set and add the other perspectives toward the end of the recording process. If you do want to render all three perspectives and mix them together at the final stage, it's not that hard to duplicate your setup with the alternate programs and render the various mic positions separately. However, it's a rather time-consuming process, as care must be taken to match each corresponding program exactly.

The muted strings in the opening passage sound great, although I had to edit the level of the release samples in several of the Mod Wheel crossfade patches because the level would really jump out on key releases. EWQLSO does not provide separate collections for first and second violins, and only the 18 Violins patch contains con sordino (muted) violins, so the same patch had to be used for both sections and panned differently. There is only one patch of sustained con sordino per instrument in the string section, with no variations for runs, trills, and so on, but they sound good in this restrained context.

Although the list of included ensemble string articulations is not exhaustive, I had no problems re-creating everything asked of the string section in the orchestration. The runs in bar 14 and the bold, melodic expressiveness in measures 15 to 18 were a piece of cake for EWQLSO. The articulations speak nicely over a wide variety of dynamics, which is hard for many libraries to do. But this one does it effortlessly.

EWQLSO has a decent selection of brass that is consistent from patch to patch and sounds pretty natural, but not without some compromise to the process. The biggest problem I have with the provided brass is that you are stuck with either one or two solo instruments or fixed larger ensembles.

For instance, I wrote for four French horns playing individual parts — not an uncommon configuration. However, there is only a single solo French horn or a 6-piece ensemble to choose from in EWQLSO. Using the same solo instrument four separate times for each part ends up sounding unnatural. But playing four separate notes with a 6-player patch results in 24 players!

The compromise was to back off the Full perspective on the French horn sections and use mainly the Close programs. This gave them a little more immediacy, with the chorusing of multiple players acting more as ambience. I actually tried going the other way by making the ensemble patches more distant with the Surround mics, but that just got muddier and more indistinct.

The same problem exists with the trombones: you get one solo trombone or a 4-person section to choose from. For the three trombones in the score, I used the same solo trombone for all parts. In neither of these cases were there ideal solutions. However, there are two separate solo trumpets, so I was able to use keyswitching and multiple programs per instrument to create better-sounding, fluid lines there.

The woodwinds in EWQLSO suffer from the same limitation of offering only a single solo instrument or a fixed ensemble. However, enough patch variations are available in the solo woodwind instruments that two different programs can usually be found to mimic two different players, or the same patch can be used without it being as obvious as it is in the brass instruments. In general, I would have preferred to have had at least two complete (and different) solo instruments for most of the brass and woodwinds so I could build up more-convincing 3-part — or even 4-part — ensembles by staggering the two players.

FIG. 2: MOTU Symphonic Instrument.

FIG. 2: MOTU Symphonic Instrument.

MOTU SYMPHONIC INSTRUMENT

MOTU Sym-phonic Instrument ($295; motu.com) weighs in at a mere 8 GB of samples and includes the fewest instruments, articulations, and ensemble configurations of all the products in this article. It is also more limited in its ability to manipulate the samples in sophisticated ways.

But it's the sound that matters most, and SI is a capable orchestral collection — among the favorite libraries of several top composers. It can be operated as a plug-in, or the instruments can be loaded into MachFive for additional parameter control and integration into a preexisting work-flow environment (see Fig. 2). Built-in traditional and convolution reverb are both included. It is by far the easiest to use of the four libraries featured, as well as the most affordable.

Because SI doesn't include con sordino strings, I used softly played sustains instead. It does provide separate first and second violin sections. For the most part, they sound like different samples altogether, so they can be layered. A few notes in the “sus f” patches (G above middle C and the high D) clearly utilize the same sample because they phase when played together on the same MIDI channel. (Using separate MIDI channels and a slight timing offset usually solves the problem if first and second violins play in unison.)

SI suffers from the same basic limitation in the brass and woodwinds as EWQLSO, in that the user must choose between a limited number of solo instruments — usually only one — and preconfigured ensembles. In the case of the French horns, two sections are provided: 4 player and 8 player. This is preferable to me, as the 4-player section is fairly dry and can pull off a 4-part orchestration without sounding too big, although it's still not an ideal solution.

A separate bass trombone is supplied, which helped counter having to use the same solo trombone for the other two parts in the orchestration. Four solo trumpets are available (Trumpet 4 is muted only), offering a nice amount of flexibility.

Of particular note are the two harps. I used Harp 1, which is smaller and more distant, and also more appropriate for the context of this arrangement. However, Harp 2 is a big, gorgeous instrument that would handle soloistic work very well.

Unfortunately, SI doesn't provide a lot of dynamic changes to the samples other than what is available via Volume and Expression or any sample switching using Velocity that is already programmed into the patch. So it's mostly a WYSIWYG library. (An Expert mode allows you to tweak some layering options for crossfades between patches.)

I wish MOTU had left out the saxophones (unnecessary for a symphonic collection) and maybe even skipped on the choir stuff (although it sounds pretty good) and delivered more content on the meat-and-potatoes orchestral sounds. In spite of the limited number of programs in such a small footprint of data, SI offers quite a bit of detailed orchestration due to the fact that the samples themselves sound really good.

SI's built-in reverb can eat up a lot of processing power, so I chose to leave it off and use a generic room while sequencing. Some of the reverb programs actually pegged out the CPU on my dual-core 2.2 GHz MacBook Pro with only one instance activated! However, the included reverbs are excellent, so I edited one of the Concert Halls for the final mix.

Because I typically mix and match instruments from many different libraries, I usually use an external reverb to unify the ambience and glue everything together. But MOTU's reverbs are a nice inclusion, and if you have enough processing horsepower, they sound great. I imagine they would be killer in a live setup that uses soft synths.

FIG. 3: Sonivox Sonic Implants Complete Symphonic Collection.

FIG. 3: Sonivox Sonic Implants Complete Symphonic Collection.

SONIVOX SONIC IMPLANTS COMPLETE SYMPHONIC COLLECTION

I really love the overall character of Sonivox's Sonic Implants Complete Symphonic Collection ($2,995; sonivoxmi.com), and the room where the samples were recorded is a defining part of the sound. To my ears, there is just enough room sound to make the samples come alive, yet you can easily douse them in more reverb without the results turning to mush. Sonivox really got it right on the recording end. In my final mix, I didn't add any additional ambience, as it sounded great already.

Originally designed exclusively for Tas-cam GigaStudio 3, CSC requires 80 GB of storage and has recently been scripted for Kontakt, providing sophisticated performance control of the samples (see Fig. 3). The hallmark of CSC is the near-perfect blend between the close mics and the ambience of the hall where the samples were recorded and the in-place location of the samples. Release samples are also available to maintain the proper sense of space. This is a very detailed and musical library with a well-thought-out collection of useful articulations, and the con sordino string samples are among the best anywhere. CSC is a great combination of affordability, musicality, versatility, and ease of use.

The Kontakt programming is consistent and well executed. Liberal use of programs with Mod Wheel crossfading for dynamics, vibrato, alternate bowing, and timbre changes makes it easy to breathe life into a score. There is a lot of animation to the samples themselves, so I found it easy to get the score sounding good quickly with this library. I particularly liked the space surrounding the glockenspiel and the timpani rolls, which you can swell with the Mod Wheel.

Again, when it came to the brass and woodwinds, I faced the same challenge. Using the single solo French horn for four separate parts sounded unnatural. Sections for two, four, and six players are provided, so I opted to go for the 2-player sections as a compromise. Oddly, there is no solo tenor trombone, only sections of two and three players. There is a solo bass trombone, which is included in the score, but the range doesn't go high enough to cover all the tenor notes. So again I compromised and used the 2-player ensemble against the solo bass trombone. That resulted in five players rather than three, but that's not easily discernible in the ensemble mix.

The library includes both a solo Bb clarinet and a solo Eb clarinet, which can easily cover the two clarinets written in Bb, so I had no problem there. In addition to the 2-Flutes programs, there is a single solo C flute and a solo alto flute. The alto flute range is high enough to cover the second flute part, so I used it just to add variety between the two players.

The harp is from a separate product, Symphonic Collection Harp. Although it is sold as an add-on ($259), it is very much a part of the CSC library and is an excellent instrument in its own right and capable of solo work.

A good, but sparse, collection of sordino strings gives the first eight measures the color I was looking for. The liveliness in the samples sounds very realistic; I just wish there were more variations to the muted strings as there are in the regular ensembles.

The attention to bowing makes it easier to create realistic-sounding lines, as in measures 24 through 29. Because CSC employs separate sets for both first and second violins (sampled in performance position), and the naming structure is consistent between all programs, getting a good section together initially takes less time than with some of the other libraries.

However, I wish Sonivox had different sizes of string ensembles to choose from. As it stands, each string section is offered in only one size (eight first violins, six second violins, six violas, five cellos, and four basses). I love to layer smaller sections of four to six players with some solo instruments on top of the string sections. Unfortunately, no small configurations or solo string instruments are included in CSC. The variety in the performance you would naturally get by layering tracks is missing when using this library exclusively, which results in having to edit the timings, note lengths, and Velocities of the strings in greater detail, as they must stand completely on their own. This is most noticeable in the opening four measures, where the strings play unaccompanied.

CSC's strength lies in its basic sound. If Sonivox decides to add more to this collection using the same development team, it would be a very welcome addition.

FIG. 4: Vienna Symphonic Library Vienna Instruments.

FIG. 4: Vienna Symphonic Library Vienna Instruments.

VIENNA SYMPHONIC LIBRARY VIENNA INSTRUMENTS

Vienna Instruments is sold in small instrument collections for as little as $445 and in bundles for as much as $18,000, with a lot of options and volumes in between. I don't own everything Vienna Symphonic Library (vsl.co.at) offers — for this piece I used Symphonic Cube, Appassionata Strings I and II, and Harps — and I am already up to more than 410 GB of samples on my hard drive. This library is the ultimate in sampling, practically adhering to the philosophy of “just sample everything.”

Innovative developments in the custom software interface have recently made navigating the cutting-edge performance capabilities of this library easier than before. However, if you've worked with high-end orchestral bundles such as Vienna Instruments, then you know they are time-consuming to install and take dedication if you want to get the most from them. In the case of VI, the choice of articulations and methods of exploiting them are mind-boggling, but that's the price you pay for playing at the top of the mountain: VI is the undisputed benchmark of orchestral libraries (see Fig. 4).

A common conversation among VI power users goes something like this: “I'm trying to get the strings to do X.” Response: “Well, did you try doing Y in the Performance control page?” or “Why don't you do Z in the Matrix Editor?” There are so many choices and so many ways to do things that it's easy to feel like you're never finished with a piece. In fact, as I write this, I am still tweaking my sequence in VI, always dangerously close to scrapping it all and starting over again, which would not be the first time. Of course, the upside is that using VI is very much like playing an instrument; no two people are going to sound alike on it, because the choices are just too numerous.

At the heart of VI is the proprietary instrument interface, which employs a sophisticated blend of keyswitching, controller crossfades, analysis of player speed and velocity for determining articulations, and matrix switching, for some elaborate real-time control of instruments. Initially you will spend the bulk of your time just learning the library and getting your head wrapped around the various ways you can approach the sampled material.

Although all of the samples are recorded ambience-free, they do not have a claustrophobic, anechoic-chamber feel. There is enough air around them to sound open and natural yet neutral. The user is responsible for creating the desired ambience through processing, and the library can sound close and intimate or big and ambient with equal ease.

VI provides separate first and second flute, oboe, and English horn, as well as an alto flute, Bb and Eb clarinets, and single bass clarinet, bassoon, and contrabassoon instruments. Woodwind ensembles are in trios, which I didn't mind because I was able to accomplish my instrumentation easily with the supplied individual woodwinds.

VI comes the closest to getting it right in the brass section. Although not providing separate first and second players along the lines of the woodwinds, there is a separate piccolo trumpet, trumpet in C, bass trumpet, and cornet (oddly, no trumpet in Bb). Slides are covered with an alto trombone, tenor trombone, bass trombone, and contrabass trombone. Three-player sections of trombones and trumpets, as well as a 6-player trumpet section, are also included.

I used a combination of three different trumpets, depending on the range and sound in various sections of the score to pull off my 2-player instrumentation, and the alto, tenor, and bass trombones made a nice 3-piece section throughout the score.

Unfortunately, it was the French horn section that, once again, was underrepresented for the way I like to write. A single Vienna Horn and Triple Horn are the choices for single players, yet the two do not jell together well enough to stagger them into a 4-piece ensemble section. The fixed-ensemble choices are 4 player or 8 player, and I opted to go with the smaller for this piece — again, a compromise.

Without a doubt, the VI percussion collection is the cream of the crop. The only challenge is to use ambience properly to move the instruments to the back of the room if you're going for that authentic symphonic sound, as the recordings are very prominent and up close. I used an extra bit of early-reflection ambience in the final mix to move the percussion and harp back, and I rolled off some of the extreme high and low frequencies to create distance.

The string sections are versatile in some ways and limited in others. The limitations are that the basic violins, for instance, are a fixed 14-person section. No separate first and second violin sections are provided, so they must be created from the same collection and panned accordingly. However, VI also offers two volumes of Appassionata Strings (larger, lusher, and muted), Chamber Strings (smaller and more intimate), and Solo Strings. Bringing these collections together can create some quite detailed string passages.

In the orchestration presented here, I used various combinations of all the VI string collections to add depth, variety, and animation to the strings. The samples alone are so good that even though I haven't mastered all the available performance techniques yet, the results are very convincing to the ear.


Rob Shrock plays keyboards with Burt Bacharach and has worked with a who's who of artists.

THE SCORE

This piece of music is based on an original song from my current solo project. It was orchestrated in MakeMusic Finale 2007 (see Fig. A) and can be downloaded as a PDF file at emusician.com/tutorials/Orchestral_Libraries_Fig.A.pdf. Finale includes a subset of Garritan Personal Orchestra for notation playback, which is actually quite sophisticated. I''ve also created an MP3 of the unedited output from Finale using this library (see Web Clip B). The piano part is conspicuously absent from the printed score because I had not settled on an exact piano part at the time I wrote the orchestration.

It would be impossible to cover every instrument as well as all of their possible articulations in a single, short piece. However, I wanted this score to be a piece of music that could be performed with a real orchestra in my live show; hence the smaller configuration of woodwinds and brass typical of what is commonly available. (Because the score is not intended for commercial publication, I have not labored over the details of the layout, either.)

The idea was to try to include as much of the orchestra as possible within a short piece of music. Although this is not exactly what I would have orchestrated had I not also been working on this article, there are some specific things to listen for in the score, as I put them there intentionally to challenge the libraries.

The strings employ mutes for the first eight bars. Note the fast runs in the violins, violas, and cellos in bar 14, which is typically challenging for sample libraries, while the basses use tremolo. Starting in bar 15, the upper strings are espressivo, carrying the melody in double octaves for four measures (typical of Romantic music) before BREAKing off into a bit of contrapuntal interaction. Although the notation in measures 21 and 22 indicates bowed tremolos, I actually intended that to be fingered tremolos and was just too hurried to notate it that way. As it stands, fingered tremolos are impossible to pull off effectively in most libraries anyway, in which case I employed bowed tremolo in divisi to create the movement.

The bowed tremolo in bar 24 is inten-tional, followed by more contrapuntal action, with divisi cellos and pizzicato basses. Finally, the whispery ascent in the upper strings tests the delicacy of each library.

The woodwinds were written in pairs for the most part, with only a single bass clarinet and bassoon. I tried to give them a good cross-section of ensemble chords, fast runs, melodic lines, and wide dynamics to get an overall sense of their characteristics in each library.

The same holds true for the brass, as far as what I included in the orchestration goes. I thinned the trumpets down to only two players, while maintaining four French horns, three trombones, and a tuba playing chords (a configuration not easily realized in a lot of libraries, it turns out). Bar 15 sees the trumpet at its upper extreme, which is about as high as you would want to take an orchestral trumpet. In the real world, this voicing would probably jump out a bit too much, but that''s one of the beauties of being able to more easily control the virtual orchestra. A smattering of harp, timpani, glockenspiel, and other percussion rounds out the orchestration.

Show Comments

These are my comments.

Reader Poll

Do play more hardware or software synths?


See results without voting »