LASS in Kontakt 3.5. In addition to retuning the library, new version 1.1 provides access to all the helper scripts right on the user interface. The first chair (solo) cello legato program in the first slot is open to the legato page, and among other things you can see what happens at which velocities. In the second slot is a spiccato articulation with the delay and humanizer page open. Next is the staccato program showing the Auto Rhythm Tool. You program these patterns and recall them with keyswitches. Finally, you use keyswitches to tell the trills programs what key you're in.


Audiobro LA Scoring Strings
The first of a new generation of sampled strings raises the bar way up.

By Nick Batzdorf

Man has been trying to invent the perfect sampled string library since he lived in caves back in the early '80s. It never will be invented, of course, and reasonable people disagree whether the results so far are three quarters full or one quarter empty. We happen to be in the three quarters camp around here, but even the other 25% have to admit that in some contexts it's become very difficult to tell whether the strings on a recording are real.

Meanwhile several developers have announced next-generation string libraries. The first one out - Audiobro's LA Scoring Strings (LASS) - is a significant improvement over what's come before it. LASS has quickly become the gold standard-both for what it does do and almost as much for what it doesn't do: sound synthy in any register. It's extremely versatile, alive, detailed, and above all expressive.

Now, there are 50 billion ways to play and record just a single note, so every existing top-end library still offers something unique - never mind that you can combine any of them with LASS and/or each other very effectively. But due to a new multi-section concept, very solid programming, and clever Kontakt sampler scripting, LASS is able to pull off some string parts that have never sounded quite right before-including those elusive flowing expressive melodies.

Sections

In addition to its programming and scripting-and judicious restraint in tuning and processing the samples - probably the most distinctive thing about LASS is that it was recorded in multiple small sections with different players in each one. These really are different players, not the same ones overdubbing, and the multi-section approach results in a realistic and lively section sound.

Each string instrument was recorded in four groups: first chair, A, B, and C. Groups A and B are small sections of the same size: four violins, three violas, three cellos, two basses. Then group C is larger: eight violins, six violas, four cellos, four basses.

LASS also includes manufactured second violin sections A, B, C, and first chair, created by pitch-transposing the first violins but taking care to make sure they don't cause phase problems when played with the first violins. You'll be pleasantly surprised if you expect that to sound like total hiney, because it works very well.

But if you prefer a different sound, you could throw in one of the first violin groups to disguise what's going on. You could also make a smaller second violin section than your first violins.

Recording

LASS was recorded in a decent sized scoring studio, but it isn't intended to be used without reverb. It has short enough reverb to allow the real recorded legato and portmento transitions without the first note ringing unnaturally, yet the instruments start with a good room sound to build upon.

A selection of tailored reverb impulses by convolution boffin Ernest Cholakis of Numerical Sound is included. There are several early reflection programs and several tail programs to mix and match. Having the tails separate allows you to use just one for several instruments, and it also lets you add predelay. Predelay can make strings sound huge without losing definition, since the reverb tail is separated from the string attack.

LASS also works very well with the other reverbs you already have in your arsenal, of course. The large scoring studio programs that come with Audio Ease's Altiverb sound great, for example.

System

The roughly 25GB 24-bit LASS library-duplicated as a 17GB 16-bit version that lowers the disk-streaming strain by a third-comes in a Native Instruments Kontakt 3 player. Among other improvements, this new NI player has the big RAM access feature first introduced in the full Kontakt 3.5 sampler (they're on version 4 now).

Running LASS in the full Kontakt 3.5 or 4 lets you access the sample mapping and other lower-level details, but you can still adjust the most important scripting options in the player's user interface. Audiobro is about to release version 1.1 (see sidebar), which among other tweaks will make virtually all the parameters accessible or more accessible than they are now.

Whether you use the player or the full Kontakt sampler, the age-old restriction of somewhat under 3GB of loaded sample starts per computer is gone. Windows users can now access all their installed RAM under a 64-bit version of the OS; the Mac OS X Kontakt version is still 32-bit, but it uses its own memory server that runs outside the host sequencer and automatically starts a new memory allocation when it fills one up. Depending on what you load, the library takes up maybe 3.3 GB of RAM at the default buffer settings.

For the ultimate performance Audiobro does recommend two computers if you're going to run everything in the library at once, however. The memory access battle has been won, but heavy Kontakt scripting like this can beat up on the CPU(s), and disk streaming performance is still a very finite resource-although solid-state drives are on the cusp of improving the situation.

Having said that, it's not at all like you'll be unhappy running the library on a single modern machine. We ran it mainly on a 2008 8 x 2.8GHz Mac Pro with 14GB of RAM installed, but also installed it on a "legacy" 2.8GHz Pentium 4 Windows XP machine as an auxiliary.

Articulations

Breaking from the "dizzying number of articulations" concept most orchestral sample libraries employ, LASS relies on real-time controllers to get expression out of a relatively small number of articulations. That makes it easy to deal with.

The programs in LASS are: expressive long notes (featuring real recorded legato, recorded portmento [with different speeds you control with a slider-a great feature], and glissando transitions); non-vibrato long notes (no legato transitions); spiccato and staccato short notes; tremolo; sordino; pizzicato; and violin and viola trills (with a script that uses keyswitches to keep the trills diatonic). At least one of the recorded sections of each string instrument includes harmonics, and all the sections other than the first chair have sordino strings. For now there are no recorded legato transitions in the sordinos, but they do sound wonderful.

To make things simple and also to save memory, there are "full mix" programs already pre-combined for you; while the individual sections are the heart and soul of the library, the full mix programs might be useful just for writing. There's nothing wrong with these programs, but they turn LASS into a standard string library-a good standard library, but it's not the same thing as having the individual sections. Audiobro has announced an LASS Lite library with them alone, along with a First Chair library that only includes the first chair instruments.

Big string section samples tend to sound somewhat homogenous, and the sectional approach solves that problem. Each of LASS' sections sounds like a section with individual players, rather than a single sampled string section. That's party due to the quality of the recordings; more than partly because you can randomize the sections' timing and tuning; partly because you can position the sections in space to control the depth and width of your mix; and partly because layering samples is just a very effective technique. The results here would indicate that a combination of the above is what makes live string sections sound the way they do (not counting the live players).

Beyond all this there are both practical and creative advantages to the sectional approach. The obvious practical one is that you can use the small sections intact or combine them to create larger ones. Each section's first chair player was also recorded solo, so LASS can do anything from a string quartet to a pop or traditional TV studio sound all the way up to a large string orchestra, with realistically sized divisi sections.

First chair not solo

There's a reason the first chair samples aren't called solo strings, and that's because the attitude is a little different. First chair players are listening to the sections they're leading and trying to blend with them, while soloists are going to make all their gestures larger.

However, the first chair programs do work very well as solo strings-you simply ride the expression controller a little more radically to create the expression. The only very minor caveat is that occasionally you'll hear two notes sounding during a crossfade between legato transition samples and their destination note. You don't notice that when they're leading a section.

Economy

Since not everyone has the most recent computers available, LASS includes memory-saving versions of the long note programs. These programs do without the legato transitions, or just include one portmento speed, and so on. (As it happens, LASS was developed on "legacy" Pentium 4 PCs.)

Contrary to what one might expect, the 16-bit duplicate of the library doesn't actually save memory, it reduces disk streaming traffic and lets you play more voices. The idea is that since all the recordings are normalized-which raises the low-level details represented in the 17th through 24th bits above the bottom threshold that 16-bit audio is able to reproduce-there shouldn't be an audible difference in sound.

If you are able to hear a difference between the two bit rates-in all honesty I wasn't-you could combine, say, the 24-bit A and/or first chair sections we'll describe next with the 16-bit versions of the remaining sections. It's hard to imagine how anyone could tell the difference in that scenario.

The set-up

LASS' multiple sections are a new concept, and today's sequencers aren't necessarily set up to deal with it yet. So before achieving bowed bliss it's first necessary to figure out how you're going to set the library up to be able to play up to four different string sections at the same time.

It helps that the library is programmed so all the sections respond the same way to the same MIDI data. That means you don't have to re-program parts for different sections when you copy MIDI tracks or change the sections on a part.

But how are you going to cut out half the fiddles when you get to a divisi part? The way you work with the library is going to depend on which sequencer you're using and on your own preference.

It's all a trade-off. You can put each recorded section on its own track, but then you're going to have eight violin tracks for each articulation. If you put more than one recorded section on a MIDI channel, you can't shut one of them off very easily...although version 1.2 will let you remap MIDI controllers to provide independent level control over each section even when they're on the same MIDI channel.

Another possibility is splitting each instrument across computers. That will spread the load around and also give you independent control when you're playing divisi. Not all sequencers make it easy to send a MIDI track to multiple destinations, for example Logic requires some fancy stuff in the Environment to do that. So you may need to use multiple tracks for each part.

This is not a serious issue, just one you have to work out. In the meantime Audiobro has an excellent private support forum for its users, and there are people on there sharing some very sophisticated programming. One fellow posted an automatic divisi Kontakt script, there are Kontakt Multi set-ups that switch articulations by program changes...all kinds of goodies.

Up to speed

Anyone who's used a library before knows that becoming familiar with it involves getting the sound of the articulations in your head. But with LASS you're performing the dynamics using MIDI controllers, so learning the library for the most part means finding the sweet spots in the velocity layers and figuring out section combinations and set-ups you like.

And that's a lot of fun. Often less obvious combinations produce surprisingly good results. For example, some of the demos on Audiobro.com combine a couple of legato sections with standard sustain programs that don't have the transition samples; this blurs the legato in a natural way. Another effective combination is standard A and B sections with sordino C.

You may also stumble on combinations by accident, such as the following example of a pizz violin section combined with two bowed ones. This pizz part is one you'd never write for real fiddles. But because the basses and cellos are playing pizz, you don't really notice the extra one, and it adds definition and a nice random factor to the sound.

Advertisement







Performing

LASS is one of those libraries that you can't stop playing when you first set it up; it simply feels great under your fingers and sounds wonderful. Really my only criticism of version 1 is that while most of the tuning is just imperfect enough to sound alive when you're combining sections, sometimes in the legato programs it's a little too much for my taste when you're using the sections alone. However, Audiobro was just about to release v.1.1 as we went live, and the library has been thoroughly retuned.

It's also fair to say that the violins aren't quite as rich and full as the other instruments. They still sound excellent, just less spectacularly excellent than the others. Developer Andrew Kerecestes explains that this was a conscious decision in order to preserve detail; he and his beta team agreed that EQ-ing them to sound darker also made the sound less detailed. However, there are EQ presets inside LASS, or you can use the Kontakt EQ.

LASS' legato programs are the ones you live in the most, since they cover everything but short notes. The first chair legato programs even speak quickly enough to play runs.

You switch between four dynamic layers to control the level using MIDI cc#1 (the mod wheel), plus ccs#7 and 11 are extra volume controls. As on actual string instruments, the sweet spots are generally below the loudest dynamic layer. These looped long tones don't swell enough to get sucked into the track, but some of them do bloom slightly, so you'll definitely want to ride their levels constantly.

The legato programs are controlled by what Audiobro calls their Real Legato script. Playing a note before or immediately after the previous one is released (or holding down the sustain pedal) triggers a recorded legato transition between the two.

At "standard" velocities you get a normal legato performance, soft velocities play portmento transitions, and extremely soft velocities give you a glissando performance. MIDI cc#83 controls the speed of the portmento or glissando by selecting one of three actual samples of the different transitions.

One of the clever things about Real Legato is that it's polyphonic, sensing when you play multiple notes simultaneously so that chords/diads are possible. You must move from one chord/diad to another as a block if you want the second chord to be a chord. Holding down, say, two notes and then playing a third will cut off the first two. Now, you can use this to musical advantage, because it does it smoothly and musically. But if you want to play polyphonic moving lines on the same program, you turn off the legato script temporarily using cc#110.

Or you can use an Expressive Sus program, which doesn't have the legato script. Incidentally, the sustain programs-Expressive Sus, Nonexpressive Sus, and Tremolo-use velocity to control the initial attack, unlike the legato programs.

The legato programs also use what Audiobro calls the A.M.G. (anti-machine gun) script, which alternates normal samples with pitch transposed adjacent ones to vary the timbre. You can turn this off, of course, and/or you can use it only in selected sections to bury the effect. The pizzicato programs also use A.M.G.

LASS' short articulations-spiccato, staccato, and pizzicato-work differently. As with the Expressive Sus programs they use velocity for volume, but here cc#1 adjusts the release times, which is a very useful feature. The spicc and stacc programs automatically round-robin between four different samples.

These programs (including pizz) also use the Auto Rhythm Tool (A.R.T.) feature, which is triggered by playing the sustain pedal before holding a note or chord. A.R.T. uses Kontakt's arpeggiator to play the notes you hold down (synced to the host sequencer's tempo) in programmed rhythmic patterns. You can program rhythms-with velocities-very easily and then call up different ones up with keyswitches. This makes it easy to play parts that would be difficult to play from the keyboard, and of course you can use the feature creatively.

Once you've triggered an automatic rhythm and are holding the keys down, releasing the pedal allows you to play freely over the held notes looping the rhythmic sequence. Another interesting aspect of the Rhythm Tool is that instead of starting at the next beat, each note starts the rhythmic pattern independently at the time you play it. So, for example, rolling a chord of pizz instruments sounds like a mandolin tremolo chorus.

Conclusion

LASS is one of those really exciting libraries in which the concept and the programming just come together to draw you in - to say nothing of the performances and the sound, which of course come first. In a year with major advances in sampling technology, LASS is right up there with the three or four most significant products of 2009.




LA Scoring Strings, $1399

www.audiobro.com

Format: Native Instruments Kontakt Player —stand - alone, VST, AU, RTAS. Can also be opened in the full versions of Kontakt 3.5 or later.

Requires Mac OS X 10.4+, G4 1.4GHz or Intel Core Duo 1.66GHz, 1GB Ram; Windows XP or 32 - bit Vista, Penium or Athlon 1.4GHz, 1GB RAM.

Copy protection: online using Native Instruments utility.


RETURN TO TOP