Vienna Symphonic Library Forum
Forum Statistics

181,765 users have contributed to 42,181 threads and 254,574 posts.

In the past 24 hours, we have 1 new thread(s), 18 new post(s) and 58 new user(s).

  • VSL strings: somehow frustrating

    after some years with the VSL, i'm getting more and more frustrated with the sound of the strings, the big groups as well as the chamber strings. The vibrato never sounds like musicians playing music, but rather than musicians who weren't allowed to go on the toilet for some days. It just never sounds relaxed or airy... I know that vibrato on sampled instruments is a problem, since in reality it is depending on the (musical) context, but i ask myself, why for example the old Vitous samples are simply more "musical"? The legato/portamento samples are really a great thing, but all in all, it's frustrating. The velocity-switches simply are not enough; the articulations not either. Are there any plans to create more variations on the string sounds, specially the violins?

  • I know what you mean.  It sounds great but missing that magic of a real performance.  Maybe the only answer will be to one day record the world's best string players, in their best mood, perfectly rehearsed (not too much!), playing a beautiful concert for their beloved conductor's last performance before he retires to die of cancer, and then use advanced algorithms to turn every heart-rending nuance into playable parameters.

    Come on VSL!  We demand it!


  • I would be satisfied having the sound-culture of the Vitous Library with the technical possibilities of the VSL... Somehow, the VSL's articulations are exagerated and very incomplete. Something like a sfz for example is very context-depending, also a fp and so on. ...but the biggest problem is that the vibrato never sounds relaxed...

  • That's the problem when you have so many players and only one silent toilet.

  • last edited
    last edited

    @Tripit said:

    That's the problem when you have so many players and only one [b]silent toilet[/b].
    ROFL!!! Thanks for that ... you've made my day! :-D

    /Dietz - Vienna Symphonic Library
  •  Recently, I listend to Nevill Marriner recording of Respighis Antiche Aries et Danzes, which are orchestrations of old lute music. Simple and easy string parts, a bit like beginners pieces for piano etc...

    It was still an amazing experience to listen to all the endless variations in colour  and especially vibrato in the string section. With this simple music they did some magic. I guess, to make a VSL version of such a piece would  expose some of the above mentioned shortcomings. 

    What I miss are the evolving vibrato violin sounds. But it would be difficult  with samples because at what point the strings are altering their sound with vibrato depends on the music (tempo, chord changes, etc)

    But maybe it would be possible to programm a violin sustain sound in such a way that the user can decide when the vibrato sneaks in. Garritan violin can do this, but the overall sound is not really that great. 

    So why not split the sample into two parts: 1: Attack+ sustain (non evolving)  , 2: (evolving) vibrato section.

    The two would be joined BUT the length of the sustain portion would be variable followed by a seamless connection to the last bit.

    The result would be an espressivo sound that is very versatile in many situations.


  • last edited
    last edited

    @veetguitar_273 said:

    But maybe it would be possible to programm a violin sustain sound in such a way that the user can decide when the vibrato sneaks in.
    As you say, this is difficult. At the user-determined 'join point' two different samples would have to cross-fade, producing a sound change which would almost certainly be noticeable to the listener. Garritan did it on a solo violin by using modelled (not sampled) vibrato, I don't think the same trick would work with a section where every player's vibrato is different.


  •  Of course, the vibrato would be the performance of the strings ensemble. But just think of the legato mode. There, you join a multitude of samples to one organic flow. That made me conclude, it could be possible in this case also. But for sure, only the experts can tell, if they can get this to work. Maybe it is just naive of me to think that VSL didnt consider this already....But on the other hand... It would really add to the pallette of possibilities. 

    Meanwhile, we can just try to come close to guys loke Guy Bacos and Rob Elliot who really make the samples sing. There is still so much to learn, at last for me.


  •  Hey veetguitar - thanks for putting me in the same league as Guy [:)].   Samples are for sure - frustrating at times. The 'snapshot' of the sample works in many cases but not all.   The BIG brains at VSL - have to be looking at combining current sample technology with modeling to achieve a more 'live' and variable performance.  My gut just says - they have sat around many a conference table exploring this 'future' development.

    Rob


  • I don't agree that they are frustrating, but  think they are incredibly exciting. 

    The Miraslav Vitous sounds were good - very well played and recorded -  but when the entire violin section is a few dozen megabytes, as opposed to gigabytes, the severe limitations quickly become obvious.  You can use it on block chord music effectively but anything complex is very artifical sounding, with machine gun effects, single-note sample blurred "legato" instead of real legato, stretched samples, etc.  I have the entire MV library but never use it these days.  And that is after buying it for the original price...[:(]

    However the disappointment you are feeling with the strings is coming from useage, not actual limitation of the VSL library.  Not that you are doing anything wrong exactly, but because it is simply very difficult to get good-sounding string performance no matter how good the samples. What you are talking about is the constantly changing flux of live instruments that are extremely espressivo, in fact, are rivaled in expressiveness only by the human voice (which is certainly the hardest of all to sample). So what has to be done is use almost every articulation that exists in VSL, and then you will begin to approach the normal practice of a good string orchestra on an ordinary day.  There is no shortcut someone is going to invent for this.  Of course,  even more variations or additional articulations are always desirable, but I feel that MIDI performances using the VSL instruments as they exist right now have only scratched the surface of what is possible.


  •  @William: You expressed it exremely well. There should be no excuses. It is a learning process that takes its time. I havent even gone through all the tutorials that are available...

    Anyway , besides that, I am extremely open and intrested to pick up any hints that make it easier to get better sounding results. There is an art of painting controller lanes, but I suspect that some guys are very successfull because they are using a breath controller. I even tried one out, couldnt make it work and gave it back.  What a pity!

    @Rob Elliot: Yes I was really very pleased to listen to your music.I found it when I read in an older thread were you presented some film music. How organic the strings were sounding! Maybe it is even still different than a live performance but utterly convincing. It might be easier not to be too much of a slave of "authenticity.  Sure , orchestration has to be studied but then there are tricks like fading solo strings that work nicely.


  • last edited
    last edited

    @veetguitar_273 said:

    So why not split the sample into two parts: 1: Attack+ sustain (non evolving)  , 2: (evolving) vibrato section.

    The two would be joined BUT the length of the sustain portion would be variable followed by a seamless connection to the last bit.

    The result would be an espressivo sound that is very versatile in many situations.

    I think this is a very good idea. I do similar things (crossfading between different articulations, like for example between fp and cresc or between cresc and trill cresc - I mainly use the Pro Edition - although no "evolving vibrato" ) quite often "by hand", means I record to audio and do the crossfade there. The result ist in most cases satisfying, especially if itยดs "hidden" in a large orchestration. I would even do so if there were "evolving vibrato" samples, although doing it with a midi controller inside VI would be highly desirable.

  • last edited
    last edited

    @William said:

    However the disappointment you are feeling with the strings is coming from useage, not actual limitation of the VSL library.  Not that you are doing anything wrong exactly, but because it is simply very difficult to get good-sounding string performance no matter how good the samples. What you are talking about is the constantly changing flux of live instruments that are extremely espressivo, in fact, are rivaled in expressiveness only by the human voice (which is certainly the hardest of all to sample). So what has to be done is use almost every articulation that exists in VSL, and then you will begin to approach the normal practice of a good string orchestra on an ordinary day.
    Hi William, you put me in the wrong box with your statement. ๐Ÿ˜Š (it's maybe also because i made my statement quite drastically) Generally i'm really very glad about the VI-s. I'm used to make a big effort to get good results; as working lot on Volume-Control and small Pitch-Bendings, which incredibly increase the realism and are a big step in the right direction. (the difference between "pure tempered" string octaves and ones, which are reached after a short moment of approaching the right intonation is huge...) And i'm not saying that the Vitous Strings are better. No, the overall result with the VI-s is nearly a lightyear better... ๐Ÿ˜Š What i criticize, is the generel sound of the vibrato, which by the way also hits me in the face in _all_ demos etc. i've heard so far in which it comes to f-vib in the string section. It's simply not a nice vibrato in my opinion/ears. Vibrato in the string section is not just a theoretical term, it's a huge cultural tradition. Different orchestras, even different conductors cultivated different kinds of how a vibrato in general has to be performed. Same with crescendi and decrescendi (Karajan e.g. often let the players start continuously insted of all at the "right" time, etc.) i just mean, that the vitous string section in general play a more "friendly" vibrato. I'm simply missing a certain magic, which in some other samples exist. What's a pity too, is that in opinion the "espressivo" samples mix really, really bad when blended with other samples. It suddenly starts to get that "stringmachine" sound. It's not enough to work on volume control and filtering to create a little bit of dynamic, since when decreasing, the vibrato gets unreal. in piano, nobody treats his finetuning on his violin so fast. On one hand - as it's been said in this thread - the future for sure is in the combination with modeling algoritms. (as far as i know, it's really difficult to write good vibrato algoritms) The other would be to offer a variation of the existing sustained (vibrato) sounds , which is simply performed in a more relaxed way by the musicians. Best, Sigi