Vienna Symphonic Library Forum
Forum Statistics

196,038 users have contributed to 43,014 threads and 258,388 posts.

In the past 24 hours, we have 2 new thread(s), 18 new post(s) and 136 new user(s).

  1. Home
  2. Sample Libraries
  3. Vienna Instruments Libraries
  4. Level Inconsistencies Within Instruments

  • Level Inconsistencies Within Instruments

    S
    S

    I started this thread because it seemed a bit OT to discuss it in another thread on levels, the MIR thread. So here goes...

    Having serious problems using VSL on occasion because of level inconsistencies from one note to the next within a given instrument. I'm using Vienna Special Edition with Logic 8.0.2. Case in point: Orchestral Strings VI-14, sustain. With middle C = C3 as my reference, and with Velocity XF turned on and the Vel XF fader all the way up for all notes I mention in this post, the note B4 is noticeably lower in level than A4 and D5.

    Another problem with B4: it sounds like it's got an artificial envelope on it, where shortly after the attack of the sample there's a brief and sudden decay down to an 85% (approx, by ear) sustain level. And this isn't the only note that exhibits this quality. And it's very easy to see the effect of this "envelope" just by looking at Logic's meters. The ballistics are such that the peak is held for a while after the sound falls to that sustain level, making it very easy to see the level difference. It's pretty drastic.

    Another example of level inconsistency: compare playing a slowly broken octave of C4 and C5 and then D4 and D5. The D5 is noticeably louder than the C5.

    For the arrangement I'm writing at the moment, there is literally no way to compensate for this difference in level when the soft notes (or weirdly-enveloped notes) occur. I've given up trying to "chase" the level with volume automation, something I do all the time to excruciating degrees. So here's a case where the sound of these violins is exactly what I need for this nosebleed string part I've written, but because of the inconsistencies in level it's just easier to find a similar sound in a different library. Sound's not as good, but I'll live with it.

    Wondering what can be done to fix these problems.

    And while I'm at it, how difficult would it be to offer users some form of "getting under the hood" so that we can make adjustments as needed to various parameters such as individual sample levels, or sample start trim?
  • Z
    Z
    Hi Ski,

    Some interesting aspects to your post.
    One of the things I've noticed is that dynamics of many instruments appear to be based on the EFFORT required to play them rather than the sonic outcome. In other words, a fortissimo level (say Vel value of 110) means that the resulting sound is what is heard when a player attempts to play at that level. This is most noticeable when switching between low and high ranges of instruments where the formant curve of that instrument dictates dramatic changes in intensity. Its also very apparent when playing with or without mutes. In order to get a violin to appear to the listener as the same volume with and without mutes, the former will often need to be 40 or 50% greater Note Velocity to the latter. This is interesting as it doesn't always tally with how a real world violinist might interpret a con sord. dynamic marking. I've made a note of this to musicians playing scores written with VSL as it can be confusing at times ("But Mr. Composer, you have a fff with mutes here... that's impossible").

    I've not checked the Orch Strings problem you mention but wil have a look soon. My guess is that this might be one of the many "characteristics" of the VSL library which we just learn to live with (such as strings looping or not looping depending on the dynamic they are played at). I don't think getting under the hood is the answer to be honest as I imagine this is just too messy to implement. I DO think the soloution is for VSL to produce updates and revisions to libraries in which these many 'characteristics' are ironed out. However I don't think this is even technically possible as every time I've mentioned it there has been no response.
  • J
    J

    I've noticed these inconsistancies here and there and they seem to be more profound in the SE then in the larger collections.  I fixed it by upgrading but if that's not an option here are some things you can try:

    First of all, if you are playing your midi notes instead of programming them make sure that your playing is consistant and that you're not all over the velocity scale.  This is pretty much a no-brainer  

    When it comes to orchestral strings, I've come to the conclusion that if you are not layering the strings with at least solo strings then you're wrong because your strings will sound too thin or synthy.  Now you might think that if I layer my OS with solo strings then when I play B4 won't the solo string be more pronounced in the mix?  Yeah! Exactly! That's the point.  Sometimes you want the solo strings to come out because it makes the strings sound more natural.  In a real professional string orchestra there's always somebody playing too loud or too long or not on time or whatever.  I can't imagine what VSL went through to get the string ensemble libraries to sound as prestine as they do; "OK Take number 523 because Hans sitting in the back can't stop playing too loud!  If he gets it right this time we'll let you all go to lunch." 

    Try tapping the high pass filter down by about 10 to 15.  Sometimes that'll suck some of the shriek out of the louder notes to bring them in line with the quieter ones.

    Manually edit the expression fader MIDI data of the offending note to match the volume of the other notes.

    Again, since you're talking about strings, when you layer them these inconsistancies tend to negate each other for an overall even sound.

    "Good luck.  We're all counting on you." 

         
  • S
    S

    Thanks for your replies so far, Zentrum and Jasen.

    Zentrum, it's nice to get a reply from someone who understands formants and how fff mutes can't play fff! That is, unless you want that airy/hairy/cloudy yet strident sound then of course, mark the part fff but expect the dynamic to be mf or so.

    Jasen, what you said about inconsistencies being more noticeable in SE than the larger collections is not something I would have expected, but that's good info to know. Upgrading isn't really an option at the moment, but something I might consider in the future.

    As far as technique, I don't mean to sound ungrateful regarding your suggestions but they're all no brainers (except for the HPF suggestion, because there's no HPF in the Vienna player I'm looking at here). Perhaps the settings I posted (brute force "full on") suggested that I don't know what I'm doing and need to learn some mockup rudiments, but that's not the case. Anyway, the problem is, I think, very basic and totally unrelated to technique or approach, layering or not: the volume of certain notes is inconsistent in the range I wrote about.

    I'd love to see a fix, even if the problem is most prevalent in SE. Or perhaps Vienna can offer an alternative mapping (lose the B4 and stretch C5 down to it, or something along those lines). The VI-14's are well-suited for nosebleed strings and it would be nice to use them more often for that purpose.

    "It's that easy, and it's that hard." -- Space Ghost
  • Z
    Z
    I agree...

    ...what we need from VSL is some sort of open discussion as to the degree to which changes or fixes can be made to already released libraries. I would really love if VSL could turn around and tell us what options are available to develop already released libraries and implement changes which are clearly causing (some) users some problems. After all, VSL is a serious investment for most of us and certainly for my own work, represents the core of my daily work as a composer. I update all the programmes I use (even if these are not free updates) and would willingly invest in the purchase of say a yearly DVD containing updated elements of existing libraries. For instance Solo Strings library package with looping implemented consistently, with harmonics which don't stop before the end of a bar, with harmonic tremolos fixed so that they are playable in the same range as normal harmonics... (Solo Strings is a good example as it was one of the first libraries released and much has been improved in more recent libraries).
    Basically it would be great if VSL could just turn around and say categorically whether they are willing or even able to consider such revisions. VSL is a professional tool, and whilst most of the users are happy to accept that not everything is perfect surely the ability to make improvements would reflect this further.

    By the way do you mean LPF or am I misunderstanding the problem?
  • Beat KaufmannB
    Beat KaufmannB
    last edited
    last edited

    @ski said:

    ...I'd love to see a fix, even if the problem is most prevalent in SE. Or perhaps Vienna can offer an alternative mapping (lose the B4 and stretch C5 down to it, or something along those lines). The VI-14's are well-suited for nosebleed strings and it would be nice to use them more often for that purpose.

    Hello sky

    Even if VSL-Libraries got just a very (very, very,...) little amount of "bad samples" it is a tricky situation to find a solution with such a "troublemaker".

    If you have the full library you often will find another articulation which can solve the situation.

    For a sustain sample you can choose a long dim articulation, sustain without vib, sustain progressive, legato, perf-trill etc.

    Most time you are able to replace an articulation for this one situation. 

    Maybe you can play it on another level (layer) and decrease the volume on the audio side.

    If this doesn't help: Use  the pitch function. Transpose the midi note +2 and turn down it with pitch -2. You use another sample this way. If you got the full Library (DVD) you only need to transpose the note +1 / -1 because the samples are available in halftone steps. 

    How to use the pitch function? Click here.

    I wish you success

    Beat

    - 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/
  • S
    S

    Zentrum, I think you and I are on the same page. The sentiment isn't necessarily a complete overhaul but at least some very basic fixes and improvements -- even perhaps some new and useful features.

    Beat, the reason I posted about this situation is because I found it extraordindarily difficult to get a consistent line out of a simple violin patch playing a very simple part. Having to resort to programming trickery for this kind of thing seems unreasonable. We're talking about footballs here (sustained whole notes for those who don't know what that is), not some complicated, difficult to articulate line. Your suggestion that I substitute other articulations is a good one, but I don't have the full version so I don't have access to the samples you mentioned. During production I did indeed try switching to the legato samples but the sound didn't match.The easiest solution was to just use another library where the sound was consistent. A little cheezy, but consistent. Hey ho.

    I'll apply your kind wishes for success towards this thread and whether or not it will have any influence on getting a few of these little problems looked at, and perhaps fixed. Look, it's not the end of the world. No one's going to die because B4 has a wonky envelope and isn't loud enough. On the flip side, we are talking about professional tools here and I don't think it's unreasonable to ask the devs to look at the situation I posted about and comment on whether or not a tweak can be made.
  • J
    J
    last edited
    last edited

    Hello Ski,

    First of all, if my original post insulted your intelligence or MIDIstration experience in anyway I apologize as it wasn't my intention to be condescending.  I just didn't know a whole lot from your OP and it sounded like you wanted some advice on how to overcome these note volume inconsistencies which, I agree, are present in the SE collection at least.

    I think where we part ways Ski is that,even when I was using the SE exclusively, I didn't think these inconsistancies were so much of an issue that I thought VSL should spend any great deal of time fixing.  But I don't know.  Maybe these inconsistencies are an easy fix.  When you layer your strings you don't notice the volumes smoothingt out?  I do, that's why I didn't think it was that big of an issue. 

    @ski said:

    (except for the HPF suggestion, because there's no HPF in the Vienna player I'm looking at here).  

     

      

    I'm refering to the fader titled, "Filter" that is in the Performance view that I think is next to the "Cell Cross Fade" fader.  Although I might be wrong on whether or not it's an HP filter but that's my understanding. 
  • J
    J

    Oh! I see, you're working with solo strings.  Well, you can't do a whole of layering there can you? LOL.  Beat's sugestion about transposing pitch is a good one that I forgot to mention BTW. 
  • S
    S

    Hi Jasen,

    Thanks for your kind reply. No offense taken. I will say that I've found it difficult in the past to read certain people's replies in this forum where it's assumed before anything else that the musician complaining about a problem lacks technique or skill in MIDIstration. So I apologize if that filtered through in my reply to you or anyone else, for that matter.

    Just to clarify, I was using the VI-14 Orchestral Strings, sustained, and that's where the problems are.

    Layering other sounds on top didn't work. I tried two different solo violins, chamber strings, appasionatas (I should have known better LOL) and two flavors of Sonic Implants. The result was as I said above -- ARP. Cheez. In the end I used (another brand of samples) for the part and I'm done. Client is happy. But for the rest of my life, when I hear this part I'm going to be thinking, "damn, it would have been soooo much better if B4 wasn't wonky in that VI-14 sustained patch" [:D]

    Beat's suggestion is something I'd normally be reluctant to try on any other sampler. Just tried it out now with a solo violin patch. Wow, sounded just fine, actually. So thanks for the heads-up on that, Beat!

    Oh, and regarding the filter, you wrote "high pass filter" but that's a lowpass filter next to the Cell XF control. It's a real steep one too (sounds like 24 dB/oct, and pretty resonant) but despite that I've gotten some really wonderful expression out of various instruments like english horn (for example) by controlling both XF and the filter with a CC played in real time. But because the filter has such a steep response I scale the amount of CC closing the filter using a Transformer in Logic's environment. That way, as I pull the CC down to make a passage dim to PP, the filter only closes a relatively small amount in comparison to the XF amount, taking the edge off the sound more than just using XF. 

    Gotta run. Thanks for your replies so far guys.