Vienna Symphonic Library Forum
Forum Statistics

180,813 users have contributed to 42,142 threads and 254,365 posts.

In the past 24 hours, we have 2 new thread(s), 6 new post(s) and 73 new user(s).

  • Instrument dynamics - real instruments vs libraries

    Good evening,

    I have a few questions, which I have researched for tens of hours without getting the big picture I am after -

    I write music as a hobby;  In the old days and up to not-so-long ago, I used sequencers (so just adjusted little colored bars here and there and pop goes the music), now I work with music a notation software.  I am very self-insistent at following orchestration rules, and to align usage of libraries (it's 2015 after all ...) with how this would be played by an actual orchestra.  For me it's about learning the Craft of orchestration through spending about 40% of my music-writing time researching, learning about instruments (who has had the pleasure to go to a music shop and to have a look at a timpani WOW), score-reading, etc.  It's very rewarding, but there is a good reason why it is indeed a Craft.

    This brings me on the topic of instrument dynamics, referring here to annotated p (piano), f (forte) ff (double forte) etc in a score.  I would say in many respects the most taken-for-granted and nebulous topic.  First, nowadays we equate dynamic with MIDI velocities, but at its core, dynamics has to do with power, as in "physics", so a felt sensation, also twinned with a change of how the instrument sounds at various levels of "loudness" (think of a French horn), and there you go, you have a texture with a felt sensation.  Let's keep textures and timbre out of the equation.

    My questions:

    1- Does anyone have a full map of instruments, not only with their low-to-high range, but which also shows the POSSIBLE dynamics throughout all their full range ?  As an example, a high range piccolo or flute will typically not be playable softly, so let's say "loud", but how loud is loud?  I have searches the web, consulted countless books on orchestration etc, and found few solid references which cover all instruments.

    2- I use Notion 5, I can see pppp and ffff (and the entire range of dynamics in between) appearing in the picklist for dynamics, it's a software tool available in a box, OK.  I get softer or louder sounds when introducing dynamics in the score, so I get the gig, but still wondering - Are dynamics individual to each instruments? i.e.  a trombone playing its loudest will be louder then an english horn playing its loudest.  Is it a fair assumption to say that dynamics are individual to each instrument?

    3- When you set your instruments for each part using a library (I have VSL Special Edition) in the mixer of the notation software, do you place each instrument at the same "volume" setting on each slider of the mixer, and adjust dynamics through what you put in the score?  I came across a few people in forums/sites which recommend to do soundchecks for each instrument (like picking something like mf mezzo forte as a baseline) and to adjust each mixer slider individually so that a given selected dynamic sounds with a similar "volume" for each instrument when playing one note from each, one after the other for the whole orchestra.

    4- A bad comparison on seried limits with libraries, the possible NOTE range of each instrument is often mapped with no-more-sound limits in libraries, so you don't end up playing notes which the instrument can't physically play.  As a point of comparison, my observation on libraries is that dynamics are typically available without "dynamic imposed limits" from very soft to very loud, but I am not sure if this actually reflects reality, simply because of post-production when the library is compiled, a sound can be made louder at the design stage.  Is it fair to say that it is for the composer to have the knowledge of what is dynamically possible for each instrument ?

    Many thanks in advance, folks !

    Regards

    Yvan


  • last edited
    last edited

    Hello Yvan

    I will not directly answer your questions but I'll try to give some information here which hopefully can help to see the matter not in a too theoretical way.

    Technically seen we always have the 127 steps of the Midi-Volume. More, we are able to adjust how strong this effect shall take influence on our signal. But in addition to a midi sound, sampled instruments often have two or more layers for different sounds. You mentioned that louder played instruments do have another sound - often a richer sound in the overtone range in addition to more volume over all. The different "layers" take this fact into account. So playing a sampled instrument louder and louder will also trigger the layer which was recorded for the certain volume. Unfortunately we don't get for all 127 midi-steps a corresponding real sample. Actually we have two, three, four sometimes five such layers. They need to cover a certain velocity range.

    So in the end we have an instrument which can play very (if you want very, very, very) quiet up to very (very, very, very) loud with a solution of 127 steps with up to 5 different sounds. Pianos and drum-libraries often do have (much) more layers.

    Please see and listen here: Setting the Dynamic with Samples

    --------------------------------------------

    In practice you play samples like an instrument. Each library comes with a certain amount of articulations. Each of those articulations have their layers... So in the end you have to produce a sound which should correspond to the score as good as possible. So playing a violin with the articulation sustain you can play notes which should sound pp p mf and f or even ff. Of course we only have 4 layers (sounds) and 127 midi-steps. But if it should sound even louder (fff or fffffff) we can choose another articulation - for example "harsh". With this articulation you cannot play p pp ppp or mf no this one is for ff fff...

    Listen here (00:24-00:27 harsh)

    Some libraries are coming with lyric legatos, (normal) legatos etc.

    Seeing and knowing this it is clear that score programs not will play the music in the highest quality which could be eventually possible. While a score program triggers the articulation sfz (because it is notated) do I choose perhaps "portato long" (highest layer) - just because it sounds more the way it fits to the context at that place.

    So when I play an instrument with all the dynamics etc. I choose "the best sound" I have - removed from any name and indication of the certain articulation. "The best sound" means in this case the one which sounds closest to the reality. This also means that you need to know the sounds of all the layers of your Library very good and also that full libraries are able to come closer to the reality because of their larger content of different articulations.

    Musicians often have difficulties because the offer of articulations is so small (even with full libraries). Where is the sad legato? wher is the glorious and the sparkling legato?...

    Technicians often see the play with samples just as a technical matter. They ask how much dB less volume a flute does need compared to the timpani for making corresponding settings in the mix. But that's not the way for getting good results with samples. Those values can be starting points more not.

    We should see the libraries as instruments with their possibilities and restrictions. Play these libraries in their best way and mix the results in the context of all the other instruments just with your ears. In the mean time you should be always very flexible for making compromises all the time.

    Despite all these restrictions and compromises you will get fantastic results.

    All the best

    Beat Kaufmann


    - Tips & Tricks while using Samples of VSL.. see at: https://www.beat-kaufmann.com/vitutorials/ - Tutorial "Mixing an Orchestra": https://www.beat-kaufmann.com/mixing-an-orchestra/
  • last edited
    last edited

    I do not know a source for a detailed list of dynamic ranges, but I would assume you would be more succesful in finding one if you consult books for recording engineers rather than orchestration books.

     

    In VSL - and probably all competitors - each instrument as a whole seems to be normalized, which means, that the loudest possible note will sound at 0db. Therefore you have to set the general volume of a flute lower than that of a trombone.

     

    Vianna MIR (and also MIRx, its little brother) has a function called "natural volume", which will reduce the volume of each instrument according to an internal table, so that when all instruments have this option turned on, the balance of the instruments is close to that you would expect in a live orchestra.

     

    Fortunately, this table is accessible in the MIRx manual (page 29), so if you download the 30d free demo licence you get the manual al well, or you might even kindly ask to get the manual anywhay.

     

    Some examples:

    according to the Natural Volume Table most of the percussion instruments are the loudest instruments, their level will be kept untouched, while the softest instruments are alto and bass flute, as well as the french oboe, their volume vill be reduced by 19dB(!). The tenor trombone in comparison has an offset of -8dB

     

    You can use the values of this table to set this offsets manually in your notation program, so you do not have to use MIRx.

     

    @YJP_Canada said:

    First, nowadays we equate dynamic with MIDI velocities

     

    Not necessarily. Generally, dynamic should still be considered as the possible range between soft and loud. Midi Velocity is ONE midi control value which CAN be used to control dynamics, and is most useful for short articulations. But the dynamic of a Vienna Instrument can be controlled in multiple ways:

     

    - NoteOn Velocity

    - Expression (CC11)

    - VelXFade

    - Dynamic Samples

    - Channel Volume (CC07)

    - Dynamic Range Setting in VI

    - Per Patch Volume Setting in VI

    - A Combination of all of the above.

     

    Naturally, in a notation programm you do not have direct access to all of these controls.