Vienna Symphonic Library Forum
Forum Statistics

180,804 users have contributed to 42,141 threads and 254,364 posts.

In the past 24 hours, we have 1 new thread(s), 7 new post(s) and 73 new user(s).

  • Samples inconsistencies regarding different articlulations

    I stopped using Vienna instruments years ago due to having various issues.

    I recently came back to it for a project and am still very frustrated, as instrument levels and tone is changing depending on articluations that should not involve different tone or loudness.

    Here are a couple of examples using instrument 54 Horn Vienna, but I'm coming across similar issues with many instruments

    1) Say I have a sustained note using sus_Vib, then switch to staccato. Suddenly the tone on the staccato is much edgier, as if it is incredibly accented. My understanding is that staccato merely means the note is shortened, not viciously attacked.

    2) I switch from sus_Vib to perf_rep to perf_rep_por or perf_rep_sta_fa and suddenly the notes are louder, so for the same dynamic I need to lower the velocity of some notes. Again, niether portato nor staccato should be louder.


  • I don't know.

    Hearing about these kinds of issues with a VSL library strikes me as a bit odd, because In my personal experience, they produce the most consistent, clean and accurately edited samples on the market. I I always considered this one of the strong points of their libraries.

    I would personally expect a staccato sample to have a clearly more pronounced attack than a sustain patch. If I wanted less edge, I would go to a portato patch. Or just play a short note with sus.

    I also don't think that it makes sense to expect all kinds of different articulations and patches to be normalized to absolutely identical amplitude levels. It kind of defeats the purpose of having articulations, IMO. There's a reason why we have CC2, CC11 and CC7 controllers. [8-)]


  • last edited
    last edited

    @Another User said:

    I also don't think that it makes sense to expect all kinds of different articulations and patches to be normalized to absolutely identical amplitude levels.

    I do. Articulations are not dynamics. Why should a repeated note at mf be louder just because it's repeated? Say I decided after some composing Ii want to change an articulation, why would I also have to change a creaully programmed velocity?


  • For this, I typically adjust the velocity curve. The wider, the quieter it will sound. I have had many cases where the velocity curve is different for two patches in order to achieve comparable loudness. Adjusting that parameter is the first thing I do before I start anything.


  •  You don't need to adjust individual velocity because a patch sounds louder. Just do as I said above.


  • last edited
    last edited

    @nektarios said:

    For this, I typically adjust the velocity curve. The wider, the quieter it will sound. I have had many cases where the velocity curve is different for two patches in order to achieve comparable loudness. Adjusting that parameter is the first thing I do before I start anything.

    Widening the velocity curve results in exaggerated loudness-differences compared with timbre differences. E.g., the loudness difference between a 'p' and an 'f' will be unnaturally large.

    What I'm trying to describe is easier to describe in the case where you narrow the velocity curve all the way. Then you get all the different timbres (from ppp to fff) playing at the same loudness, which is clearly unnatural.

    Only one setting for velocity-curve can yield the natural correspondence between timbre and loudness, so when I'm trying for a natural sound, I don't move the velocity-curve from its default setting.


  • I see. But when you are dealing with repetition patches you can have 3 velocity layers, and performance legato gives you 2 velocity layers, whereas a staccato patch can have 4. My point is that because some patches have a different number of velocity layers, you may need to adjust the velocity curve to achieve comparable loudness. How would you deal with this? I do understand your point on realism.


  • last edited
    last edited

    @petethomas said:

    stopped using Vienna instruments years ago due to having various issues.

    So after all those years may I introduce you this function of VI?


  • last edited
    last edited

    @nektarios said:

    My point is that because some patches have a different number of velocity layers, you may need to adjust the velocity curve to achieve comparable loudness. How would you deal with this? I do understand your point on realism.

    I don't have a solution, besides adjusting velocity and volume on a case-by-case basis, which is what OP was wanting to avoid.


  • last edited
    last edited

    @Another User said:

     Articulations are not dynamics. Why should a repeated note at mf be louder just because it's repeated? Say I decided after some composing Ii want to change an articulation, why would I also have to change a creaully programmed velocity?

    Because the alternative is a "creaully programmed" machine gun effect.  If you're not already doing so, you're going to have to use the Velocity Crossfade together with the Expression fader in realtime as you are recording to smooth out the volume differences. 

    I think you can turn off the round robbin playing where a different sample is used for identical notes played one after the other if the machine gun effect is what you want but I might be wrong on that.


  • Concerning the attack of a staccato, you could also stretch part of the sample using VIPRO. For example, you can stretch the start of a staccato sample but not the end part. Finally, adjust the attack slider for further adjustments.


  • i also find some different levels in different articulations here and there... i.e. often the portamento legato is louder than the normal legato in most of the strings. the easiest thing for me is to lower the level of the articulation cell in the mixer of VI(pro) and save the patch, so it is adjusted for all future needs. inside groups of round robins in one single articulation i never had inconsistencies in any vsl library (in opposite to many libraries of other companies).

  • I wish attack in samples would be independent of sound volume, period. In this way, not only improvement in attack would be achieved, but better legato would also be achieved. This is available in some of the much less expensive libraries still in circulation. Notes in the MIDI editor can be selected, singly or in bunches, and numerical values for whatever designated CC that samples respond to can be entered. I am here especially referring to tweaking string samples, but also winds.

    But then, at the same libraries, high strings are completely inferior sound quality wise. This is another matter, though, different from our problem. Separation of attack from volume would also solve your (and my :) problem because by lowering attack for all of the used samples they become more even. And, by the way, piano players work a lifetime on evening attack. Just imagine playing Scarlatti or Mozart, or any piece that has copious amounts of Baroque, Rococo or Viennese "perlage", with all kinds of bumps caused by non-uniform attack. This would be unthinkable.

    Just listen to renditions by the likes of Horowitz or Glen Gould or Schiff and you'll hear what I mean. I also have a question: Is there ANY possibility to give vibrato to the Solo Trumpet in Special Edition Vol. 1? The Trumpet overall is not very good, and without vibrato it is completely lifeless. Thank you all, John.


  • This will not do the trick. I have the same problem with the staccato part of a Trombone patch: TTB 01_basic. From velocity 0 - 99 there is a well banlanced cres. from 100 - 127 the sound gets very edgy and remarkable lower (softer). This means that the trombone dissapears in the mix if you get "louder" than 99. Thats not logical and I to my ears its an error. The same thing happens with the Trumpet in C patch.

    Right now I can't find any compettetors to VSL regarding solo brass, but they will be there in time. I think VSL should consider correcting all these minor (I admit) problems in their libs instead of creating new ones.


  • Agreed. However, I don't know of any library that has all their samples equally well calibrated. Unfortunately this is reality. In general, demos sound quite well, but they're written by very highly skilled top notch programmers who have tons of plugins that add in the quest of correcting sample deficiencies. And, of course, very expensive plugins that a poor guy like me can't afford. So I spend tons of time trying to live with what I have with not much result. I am a melodist and would like it that high strings were much smoother and without extremely bothersome white noise that accompanies their higher registers and strident harmonics that are louder than their fundamentals.

    I have 4 libraries, all with the same deficiencies. No, it's not the sound card. It's a good, pro card, so the sound channel's problems are in general solved. No matter what, I can't render my sweeping and tidy pop melodic lines with the aid of violins as sweepingly and tidily as I want. And yes, I also write 12 tone and world music. Maybe in 10-20 years time we'll have better samples. As far as Trombones, to my taste the best I've heard to date are those by Sample Technology. They probably use Physical Modeling. Well, at least the ***demos*** sound great. Wallender's (Physical Modeling as well) demos also sound well, but when I tried their software the Trumpet had such strident stray harmonics that I decided not to have the package because I couldn't take them out. I bought VSL Special Edition Vol1 because that's all I could afford.

    I only tried the Trumpet, which doesn't have native vibrato and the patches don't react to cc17 + After touch. Far less expensive packages' patches DO. VSL, CAN WE HAVE VIBRATO WITHOUT SPENDING TONS OF MONEY? I think you'd only gain in popularity and, therefore, more customers :) Anyway, my 2 cents. John.


  • To be fair:

    There are no magical, wickedly expensive audio plug-ins available only to the big players in the business that by themselves make sample libraries sound great. IMO, the smallest part of getting your levels and articulations balanced is a matter of inserting the right plug-in - it comes down to editing and tweaking your parameters to no end. And I doubt it will be much different even with the "next generation" of sampling technology.

    Pudik, I'm curious as to what exactly you mean with "white noise" in the higher register of string samples. Generally, all the popular string libraries today are pretty low on noise. VSL strings in particular are super-clean. And even in the most ambient, super-wet libraries that are out there, I can't imagine that you're able to find any seriously detrimental noise in the samples - least of all white noise!

    Perhaps you're confusing that natural "scraping" bowing sound with white noise?


  • last edited
    last edited

    @Stig Christensen said:

    This will not do the trick. I have the same problem with the staccato part of a Trombone patch: TTB 01_basic. From velocity 0 - 99 there is a well banlanced cres. from 100 - 127 the sound gets very edgy and remarkable lower (softer). This means that the trombone dissapears in the mix if you get "louder" than 99. Thats not logical and I to my ears its an error. The same thing happens with the Trumpet in C patch.

    How are you measuring loudness? It is very unlikely that it is getting softer.

    DG


  • last edited
    last edited

    @DG said:

    How are you measuring loudness? It is very unlikely that it is getting softer.

    The first 4 harmonics get quiter as you increase velocity here.


  • last edited
    last edited

    @pudik said:

    VSL, CAN WE HAVE VIBRATO WITHOUT SPENDING TONS OF MONEY? I think you'd only gain in popularity and, therefore, more customers 😊 Anyway, my 2 cents. John.

    John, I answered your question in your other thread, but in case you didn't see that, I will answer again.

    There are vibrato samples available, but it sounds like you didn't buy them. In any case, orchestral Brass players playing in a Western style don't usually use vibrato, so it's debatable whether or not VSL is losing any money by not having them in your particular Collection.

    Regarding using LFO and such devices, VSL produces orchestral sample Collections, not synths, so an LFO vibrato wouldn't be appropriate and furthermore would sound terrible. The best Brass product I've heard for constructing a fake vibrato is Sample Modeling, but that's recorded in an anechoic chamber and all the dynamics are fake as well. I doubt that this could be achieved using an instrument recorded on a scoring stage.

    DG


  • last edited
    last edited

    @DG said:

    ... Regarding using LFO and such devices, VSL produces orchestral sample Collections, not synths, so an LFO vibrato wouldn't be appropriate....

    I don't see how LFO vibrato would be so different from any of the other sample-manipulating functions in VI. The Humanize function alters the pitch already, and I don't think of that as inappropriate.