virtual instruments

Preview: VSL Synchron Bösendorfer Imperial

I collect sampled pianos. And I put them on a fight, when a new one arrives in my collection. As soon as I've purchased the Standard version of the Bösendorfer Imperial, I have therefore immediately compared it to their old Vienna Imperial (released in 2009).

They are really different instruments. The way to approach them is different. Bösendorfer Imperial seems to be smoother than Vienna Imperial, whose dynamics I've always found to lean toward the forte. The position of the piano relative to the player is different, with the Bösendorfer Imperial being farther, more into the space than under your fingers.

Vienna Imperial is a stronger piano. A studio piano, with a marked desire to be first, forward, in your face. Bösendorfer Imperial is more laid back. It has a larger dynamic range, but it dominates it better.

The room mics in the Standard version are fine for me. To be true, I immediately switched from the Room Mix (including the surround mics) to the Decca tree, containing all the room information I need. Others will want to enjoy the 3D feeling of the additional room mics, that can even be used for the new Auro 3D spatialization system.

As for me, I would instead like to have the Tube mic, only included in the Extended version, for that slightly far-from-the-hammers sound; but the included Condenser mic seems to have the right balance between brightness and smoothness, never being harsh. And, mixed with the Mid 1 pair it gives a perfect blend, incisive and full.

As much as I would like to focus on a single piano, I understand that the Bösendorfer Imperial is a much more "classical music" piano, whereas the Vienna Imperial is still a more "jazz club" piano. So, both seem to be useful tools in production. And both are incredible fun!

Incommunicable microtuning

Not many DAWs, notation programs, players, virtual synths and sound libraries allow for alternative tuning and microtonal accidentals. Some allow them, but then are unable to communicate to other software the altered notes.

For example, a DAW or music program may allow for microtonal accidentals, but then send out a message that most sound players can't understand (like VST Note Expression vs MIDI Tuning Standard).

Alternative tuning is useful for music in non-Western standards (Post-Minimalism meets Ancient India). Or for ancient music (for example, tuning a harpsichord to match a lute). Or for experimental music (Harry Partch-inspired microtuning scales, or the mystical Pythagorean tuning).

Microtonal accidentals are used in much contemporary music or hybrid music, but also to transcribe as finely as possible folk music. Some experimental rock/EDM is using clusters, and going over the walls of the Equal tempered system.

Yet, not all the DAWs, notation programs, and sound generators can communicate their alternative tuning and exoteric accidentals. A world open to the world in theory, but less in practice.

A missing standard for keyswitching

The lack of an universal and advanced standard for keyswitching makes me crazy. I’m one of those who prefer not to insert keyswitches in the score, nor use separate tracks for playing techniques. I want a meta-code to drive my technique changes.

What I did, in making my articulation sets for Logic, was to first create my own personal articulations/techniques map, starting from a Spitfire Audio UACC map repeated two times (UACC s 1-128, Logic 1-256). This means that all my maps will have the same articulation types at the same ID. Selection messages will start from those fixed positions.

Unfortunately, not all libraries are coherent in how they map their articulations/techniques, so I'm still using too many articulation sets and expression maps. With VSL VI libraries I built my own presets, all organized in the same way. But this is not possible with all libraries.

Composing contemporary music in the age of sound libraries

I’ve been trained to write music on paper. My main teacher, one of the best composers of his generation, was also a copyist, and insisted on good and accurate calligraphy. Writing on paper seems like the most obvious way, when dealing with a very rational type of music, based on proportions and semi-automatic processes. It’s also the fastest way to notate some tightly integrated music gestures, made of a bundle of pitches, articulations, and expressions, that would be impossible to quickly notate with notation programs (an example: a violin playing a starting pitch, fading into a jété and gliding up to an uncertain pitch).

At the same time, I’ve always felt the need to feel the music under my fingers. Neither my mind, nor the calculated music coming out of a computer when listening to the results of what I was writing, has given me a satisfactory connection with my music. My mind, a powerful generator of music, is only a part of my body; and my ears, receiving the waves of pressure from the loudspeakers, are only a part of the sensorial experience. I need a tactile experience of my music, together with the auditory experience.

When very young, I composed at the piano. Sometimes, I sat at the piano, and went on experimenting with stravinskian overlapping chords, bartókian hammering rhythms, schoenbergian piercing intervals and misty outburst of notes; some other times, I just checked at the piano what I was writing on paper. I had a physical connection with my music.

Later, when I could afford one, I switched to computers. I tried to simulate “real” music and sounds. However, notation programs were unable to make my notes sound as I wrote them; I wrote them as music, the computer insisted on playing them back as arithmetic expressions. The sounds I could feel when playing on the keyboard were not what they were named after: the piano lacked hammers and resonance, violins lacked wood and body, brass did not explode, woodwinds lacked breath and click. The computer was great for electronic music, not for simulating music made with real instruments.

But acoustic sound libraries improved over the years. VSL was a revolution. Other libraries appeared for specialized types of sound. For what I was looking for, VSL and XSample offered great support to my music. At first, I had something resembling realism in the libraries that came with the Native Instruments package (a taste of VSL, then the realistic chamber strings of Session Strings Pro). Then, I could finally afford the solo instruments of the XSample Library with their extended techniques, and the accurate orchestral sounds of the VSL Special Edition. I had a powerful, realistic sonic arsenal under my fingers. I had all the sound tools I could need.

So, I could compose at the computer again. But how? Notation programs continued to be cold as a grave. Wallander’s NotePerformer added life to my Sibelius scores, but still more the life of a lemur than that of a living body. And composing by patiently writing and sculpting pitches on the staff looked like underusing the tools I had. What I had was basically a glorified piano – the same keyboard on which the greatest composers of the past loved to improvise and compose, the same white and black technological interface with music they loved to spend time imagining and feel their music – but capable to really *play*, and not only suggest, a full orchestra.

Is composing at the keyboard legit? With Logic, I can keep the score, pianoroll and controller pages open, in a mix of traditional notation, evolving texture and cluster graphic notation, oscillator and modulator diagrams. I can move the input cursor where I have to insert a segment, and start recording from there. I can step-input pitches as I would do in a notation program. I can have rather accurate notation and realistic sonic rendering at the same time, write notes on staves, and later export a MusicXML file to refine notation with a dedicated program. Logic can also assist me with some serial-based elaboration, unless I want to cut and paste pitches generated by OpenMusic.

What I feel is that, by composing at a DAW that lets me easily have accurate control on the piece’s micro- and macro-structure, I can really go further, and maintain a better control on the piece’s macrostructure and evolution. I can create a general structure and time signature map in advance, insert motives and focal points as placeholders, use the track arrangement space as a blank wall where to attach post-its, and create my piece by going from the general image to the finer details, gradually filling that wall. And always keeping contact with the actual sound, not simply a mental image of the sound.

Isn’t this a lot like the old Maestros sitting at their klaviers, one hand on the keyboard and the other hand writing on a music sheet?

(The above is an old reflection, made in February 2017.)

Sketching libraries and quick composition

Are “sketching libraries” really useful to speed up composition? I’m not a great example of quick and fast productivity, but when writing tonally I like to start playing the piano. Considering how good modern sampled pianos are, I find it liberating to just return to my first instrument and let it help me drawing at raw lines my music.

The new tempo follow metronome in Logic is a revolutionary innovation for me, since I have all my creativity killed by playing on a metronome click. Having the metronome follow me is letting creativity flow freely, as one could do in the age of acoustic pianos, pencil and paper.

The piano sketch is a good place to work on melody, harmony, form. It's just two lines of music and some textual annotations. When done with the basic matters, you can start propagating your music to the other instruments, by copying&pasting or playing idiomatically the new lines.

But I admit that sketching libraries are also great. Not as straight as a piano, they let you write down more information on the first pass. I like Albion One for more booming music, Vienna Smart Orchestra for more classical. The Berlin Orchestra Inspire series should also work great. Orchestra sketching libraries are however already forcing one to follow their pace. String attacks may be too slow, and while you can adjust this with a controller (later or in realtime), here are you already and again facing detailed editing - the thing you were trying to avoid.

On the other side, sketching orchestras may offer you inspiration. I can't wait to try British Drama Toolkit.

The horrors of the ideas about horror music

Jump into a discussion about sound libraries like Spitfire Audio's Albion IV "Uist", Sonokinetic's Espressivo or 8Dio's CASE and CAGE, and you see that they are automatically associated to horror music. Some also associate Spitfire's EVOs, with their intrinsic instability, to horror music.

To be honest, some sound library manufacturer does nothing to prevent this automatic association. Native Instrument called their dedicated library, developed with Audiobro, "Thrill". 8Dio is not hiding this is the intended destination. And it is true that their CASE and CAGE library are very dedicated to the genre.

In my view, however, some of these, like Uist, Espressivo and the EVOs, are simply great tools for modern classical music. They are not "effects", but "words" or "phrases" typical of a particular modern language. In particular, Spitfire's Evolutions are more on the subtle side, so I would exclude them from the "horror effects" category.

All considered, we often consider "horror music" the soundtrack assembled by Kubrik for his movie Shining. But these were, in origin, modern classical pieces from Bartók, Ligeti and Penderecki. The ones to whom the finest of these libraries are inspired.

Presets, templates and the good practice

All considered, making an universal preset schema that can work for all your sounds, and arranging the sounds in a template that will make your ensemble sound coherent, is not dissimilar to tuning the instruments of an orchestra in the old times.

With digital instruments, we no longer need to spend time tuning, unless you have the old-fashioned idea of calling in some acoustic instrument. In the pre-digital era, you had to accurately tune everything, including some analogue synth.

Renaissance and Early Baroque – a bit earlier than the birth of the opera, and at the times of Monteverdi – you had to choose a tuning among the many experimental ones. Bach had not yet come with his propaganda for completely equal tuning, good for all the scales. Meantone tuning, for example, was one good for great sounding major thirds and decent sounding fifths, and you had to think to the most used ‘modes’ to choose the best sounding intervals. You had to patiently tune by ear by listening to beatings between intervals, and ultimately following your taste.

Creating presets and templates is our preliminary work on what would lead to great sounding music. You have to think to the music that you will want to do, and decide which articulations, which layout, which selection controls and which space to use. It’s nothing more than what our ancestors did with a tuning fork and good ears: tuning their instruments to make them their immediately collaborative musical partners.

Interlinking a score and its sounds

I confess: I’m (mostly) a serial composer. Hybridized with a spectral one. With no little influx from ambient and brutal noisemakers. I’m one of those kids who compose not only with notes, but also with everything happening to them. Odd bow movements, overblowing, punching unexpected parts of the skin with unintended tools.

As a consequence, there is this schizophrenia for which there are times when you have to accurately listen to the resulting sound, comparing it to the one you were imagining, and put it to the test of reality. And other times, when you need absolute silence, to prevent the wrong sounds to interfere with your creative process.

A discussion in the forum dedicated to Dorico, the notation program, offered an interesting perspective. You can easily and completely separate the score from the sound generator. Dorico can drive external players, like NI’s Kontakt or VSL’s Vienna Ensemble Pro. It can also switch to a silent player.

Therefore, you can build your orchestral template as accurately as possible. Make all the noise you like. When you need silence, just run Dorico without first running Vienna Ensemble Pro. No waiting time, no wrong sounds. Just you and your inner musical daemon.

Outdated pianos – is the new trend loving something old?

So, the new trend seems to be loving something old. VSL have just release their antique piano, a magnificent Blüthner Series 6 dated 1895, and perfectly restored without modernizing it. Incidentally, not the oldest new release they had lately, since a reduced version of their Historic Winds was introduced as a Special Edition (SE) volume at the same time.

VSL are not the first one to give us antique pianos. Pianoteq has a whole library of antiques. In that case, however, one can question if you are really listening to a piano, or its synthetic avatar. Native Instruments gave us a rare Bechstein from 1905, with an old modernized sound. Galaxy Instruments has another Blüthner, this time a baby Model 150 from 1929.

From 1895 to 1929 the piano evolves, but it is still technically the same. Older means sounding more aged. Newer means that it can still sound new. But the base technique is the same. The great inventors of the end of the Nineteenth Century have already imagined all. Cast-iron frame, crossed strings. Untile the next step in materials.

Cinesamples has released a Steinway Model D from 1949, found in the Columbia Studios, and Embertone a later Steinway from 1955 owned by John Q. Walker for his Zenph project of revived piano rolls. Two world wars in the middle, and metallurgy has changed. Music-making has changed, with the after-war pianos meant for bigger halls and a less sentimental music.

The Embertone Walker sounds modern, but with a veil of antiquity, or savvy calm. It is not the angular, sharp piano we can listen to in modern recitals, or in newer sample libraries like the Synchron Yamaha and Steinway from VSL, or the aged after-war Steinway from the MGM Scoring Stage at Sony Pictures Studios in Los Angeles produced by Cinesamples with their CinePiano.

Antique is the new new. Traveling in time is a way to find inspiration. Even if you are not after an historically informed virtual performance, there is something disturbing and evoking in worn sounds. They come from some unusual place - if not in the space, at least out of our time.

Writing for the real, AND virtual players

Let's be honest: when writing a piece for a real ensemble or orchestra, we know that it will probably remain in the realm of virtual instruments. Our piece will never, ever be performed by a real orchestra.

I had the honor and pleasure of having some of my pieces performed by real players. Very skilled musicians, sometimes among the best players for that instrument. Yet, this is something that can only happen in some special events for most of us.

The rest of us will, most of the time, continue to listen to their works from virtual performers and instruments.

That means that we have to find a balance between writing for real instruments, and at the same time make accurate prototypes, that will have to be considered an alternative form of the final piece. Our piece has to sound great both when read by real performer, and when performed by our samplers.

We write for virtual orchestras. This is no longer to be considered a second choice. Virtual orchestras very often go into feature films released in major theaters. Virtual orchestras are a real instrument, even if the human content is just that of the musicians who recorded the samples, and the composer that created the virtual performance.

A tale of two sample players

Both Spitfire Audio (SA) and Vienna Symphonic Library (VSL) have recently introduced their new players. The move from SA is easier to understand: they need to separate from Native Instruments and their sample player, Kontakt, on which all SA libraries where based.

With VSL things are a bit more complicate, because they already had their own player – Vienna Instruments (Pro) –, and this is a second player to maintain. The more apparent reason for this new software is that the new Synchron libraries, with their multiple mic sets, can't be managed very well with the old player.

Synchron Player is also easier to understand for a new user. Together with the easier-to-use new libraries, this might be a move towards becoming more commercial!

A quick comparison of the two VSL players is this, in my personal view:

- Synchron Player, based on a coherent metaphor of folder hierarchy, could be more immediate to grasp for the new user. How do you reach a particular sound? Just follow the path (as the Rabbit was telling Alice).

- Vienna Instruments' matrices are totally reconfigurable, and you can choose the number, size, position of the cells everywhere. Apart for you, nobody else will immediately be able to read a custom preset. And reading factory presets is not easier.

- Reaching a sound in VI can be a lot faster than in SP: point at that cell in the matrix. With SP, you have to go through the full path everytime.

- Crossfading between more that two layers is the real bonus of SP. Being able to fade between non vibrato, vibrato and molto vibrato, or between sul tasto, normale, and sul ponticello is something I've always wanted in VI. Will be there further development for VI? Please, add a third column of slots in a cell!

- Controller assignment can be much quicker in SP. You don't have to use CC1, you can use a meta-control that you can globally assign to CC1, or to any other controller. With VI you have to reprogram all matrices. (This is another thing I would like to see in VI: meta-controllers!).

- The standalone version of VIP allows for nearly-quick replication of presets from existing presets. A true life-saver.

The fast changing pace of articulation switching

What happens, when you play a piece with one of the modern, rich sound libraries? The computer takes your music symbols, and converts them into commands to select articulations. Here is how a solo violin looks in VSL's Vienna Instruments Pro, in a short video I assembled over an orchestral prototype I'm working on. Isn't it a bit like the Nineteenth Century Iron Age's mad utopia of musical automata?

Sound libraries and extended techniques

Grown in the European academia, I always feel a strong need for extended techniques. That is, those strange sounds going over the usual round technique used for crafting beautiful melodic lines, or the smart spiccato with which you push your most frantic rhythm ahead. Scratchy digging, feeble harmonics, colliding multiphonics.

These aren't very commonly included in sound libraries. I’m grateful to the handful of houses making sample collections including at least some of them.

While testing my new VSL-based orchestral machine, I moved in territories very dangerous for a humble sampler. And tried to replicate the mad violin of Sciarrino’s Capricci. Nearly impossible to play with a real instrument, virtually impossible to make with virtual instruments.

Yet, there are some interesting findings in exploring the extremes. And trying to recreate a, so to say, naturally produced sound can teach a lot on the nature of instrument virtualization.

Sound maps extending over 128 entries

Keyswitching can be done better with a common reference map, allowing for easy exchange of the same code between different sound libraries. The same Expression Map or Articulation Set can then be used or easily adapted to the various libraries in your arsenal.

My personal maps for libraries like VSL and Spitfire are modeled on the UACC map. Spitfire is not always coherent with their own map, and it is easy to understand why, thinking on how little conventional are some or their libraries like Tundra or Uist.

VSL has no reference organization system, but they have tried to standardize their presets over the years. Collections from different generations have a similar system to organize the many articulations, but these systems different between the different generations, and the different instrument families. The very flexible Vienna Instruments allows however to create your own presets, and set them as they better fit your workflow.

It may seem strange, but the 128 slots allowed by the Spitfire's UACC map are not always enough. My personal map contain many nuances, going from the basic sustain vibrato or non vibrato, or espressivo, to things like molto sul pont., with heavy pressure, or various degrees of measured tremolo. Overthinked and overworked, maybe, but an effective tool to avoid thinking to the mechanics behind the libraries when actually making music.

Something I've done in Logic to get more articulation slots is to duplicate the first 128 entries to the second group of 128 in the Articulation Set, to fill all the available 256 slots. This way, any variation to the base articulations can fall in a slot mirroring the ones in the first group. For example, there are two basic Longs at #1 and #2 in the UACC map. If you need two more, you can place them at #129 and #130.

Blending Spitfire Audio and VSL strings

Spitifire Audio's London Contemporary Orchestra (LCO) Strings are a fantastic addition to a contemporary composer's box of tricks. Open tuning, raw and sincere bowing, an exploration of the strings from bridge to touch (and back).

However, they lack true legato. You can try to fake it with the Release control. An Attack Offset control would have been a nice touch, but it is not there.

So, I did some tests with combining LCO Vivid Violins with some legato patches from other libraries. "Vivid" are the (nearly) ordinary articulation of this library. These are the LCO Vivid Violins alone:

Together with the six violins of the Vienna Symphonic Library (VSL) Chamber Strings, they resulted in a too big ensemble (despite the much lower volume I mixed the VSLs in):

On the contrary, only two of the VSL Dimension Violins worked fine, with the right ensemble size and the LCO timbre prevailing in the mix:

The VSL Dimension Violins alone, with the performance trill patch selected; there is some apparent phasing, but this is typical of this particular library, exhibiting its full beauty only when everybody is playing together:

All of them placed in the Teldex wide via MIR. No patch changes, only the raw ones.

Orchestral prototypes

We all agree that there is no substitute to a human player – a skilled one – playing contemporary techniques on real instruments. As for me, this is still the final goal of composing: making music for a community of people wanting to enjoy music, both as players and as listeners.

The use of computers and sound libraries is a help for composition. What once was done with a real piano, or even with the family or friends reading your freshly written music in the evening, is now done with this sophisticate keyboard instrument that is the sampler.

There is a growing opportunity of making nearly credible prototypes with the current tools. The libraries from IRCAM are explicitly made for contemporary music, and Xsample libraries contain several of the same techniques. Some smaller houses, like Soniccouture, Sonokinetic or 8Dio, have other useful sounds. VSL, Orchestral Tools, Spitfire Audio are adding more and more of these techniques, even if their core business is not mainly this niche, but the modern composer for media in advanced markets (like the sophisticate world of film and TV production in London, Paris or Berlin), where using the most recent vocabulary is allowed and even required.

My impression is that a composer is no longer forced to just imagine his/her music, but can put it to test with the available compositional tools. And if these tools can open a way to other new techniques – either in the world of acoustic or electronic instruments – music will gain something new.

Instruments and their colors

For my virtual instrments, I like to use earth-inspired color codes
Read More…

Logic's undervalued plugins

Logic plugins are often undervalued, just because they are free
Read More…

VSL Vienna Imperial vs. Embertone Walker 1955

Stiil looking for the perfect sampled piano, today I did another comparison. I pounded some Bach and Mussorgsky on the Walker (v1) and the Vienna Imperial. Both set at 64 samples of buffer. Both read from an external SSD and controlled by a VPC-1. Playback via a pair of Mackie HR-824 mkI.

The pedal, first of all: while neither the VSL nor the Embertone feature half pedal, the pedal release in the former is more gradual, and can approximate at least the effect of continuous release of the pedal. Not so the Walker, that is immediately cut. It is as if one has a long release in the sample, the other lacks it. Maybe the variable release sampling made by VSL also regards sustained notes.

The on/off activation of both the damper pedal and the soft pedal seems better balanced in the Imperial. With the Walker I can hear a change in volume when pressing one of the pedals. The Imperial only changes timbre. Also, I feel the pedal change noise to be too strong with the Walker. While missing some important features, I feel the pedal behavior of the VSL more natural, with even a hint of repedalling (that shouldn't be there, but can be clearly noticed).

The Walker seems to do ribattuto notes better. This is surprising, considering that it seems to remain behind during normal playing. Play big chords in the "Pictures at an Exhibition", and the sound comes a little after you expect it. It also seems to miss some notes sometimes. Not so with the Imperial, always perfectly in time.

The Walker lacks a little on the fff side of dynamics. The Imperial on the opposite side, with ppp always sounding a bit too loud. Both have a gorgeous sound, an excellent representation of the original instruments – rich and well blended, much on the wooden side, the old Steinway, clear and focused, a bit steely, the Bösendorfer.

Still, I find that the most playable piano in my arsenal is The Grandeur. The sample is not as accurate, yet I feel there is something right in the scripting.