Vienna Symphonic Library Forum
Forum Statistics

201,681 users have contributed to 43,255 threads and 259,320 posts.

In the past 24 hours, we have 2 new thread(s), 11 new post(s) and 54 new user(s).

  1. Home
  2. Sample Libraries
  3. Vienna Instruments Libraries (legacy)
  4. Notes drop out (funnily for some patches earlier than on others)

  • Notes drop out (funnily for some patches earlier than on others)

    L
    L

    I have quite often issues with notes dropping out for the string ensemble sustain patches (using the VSL Special Edition) while playing more than one note with one VI-Player.

    I use the newest version of the player and already installed the VSL SE Library Update.

    I just did some tests to find out, where the issues come from. I used for these tests the standalone version of the player to eliminate the factor of the host (but after all plugin version and standalone seem to act the same for this issue).

    My Buffer-Size of the ASIO-Driver of my M-Audio Delta Audiophile 2496 is 256, but the problem seems to be exactly the same with 512.

    Funnily I found out, that the notes from the String Ensemble Sustain patches (tried Celli and Violins) drop out earlier than any other patches I tried (Solo Cello Sustain, Solo Trumpet Sustain, Trumpet Ensemble Sustain ... all have 3 Layers and Release Samples btw.). The notes don't drop out so early if I disable "Play release".

    The notes also rather drop out, if I hold for example one note and play very fastly one or more other notes. But it even sometimes occurs by just playing two or more slow sustained notes.

    When I load a project for example, which has parts, where more string-sections play together and I play it back or render it directly, I almost always have heavy dropouts! After playing it back a few times, it works fine for these notes.

    I really don't understand, what's happening here. Can't be a simple voice-limitation-problem. My system should neither be the problem.

    2x1GB 667Mhz RAM with CL3 timings in Dual Channel Mode, an E6600 processor and a SATA-II Samsung Spinpoint SP2504C with 7200RPM should be fine I guess ...
  • H
    H

    I think it's a voice limitation (64 voices per instance)

    The string patches will use more voices then dedictaed single instrument patches, because most keys of the strings patches do have mapped two different instruments performing on the same key.

    If you switch off Release Samples you get defenitely more voices, because the Release Samples do also have these double instrument layering.

    best

    Herb
  • L
    L
    last edited
    last edited

    But if I just hold one tone with one finger, and then fastly hit repetative another key, it's already enough to let the holded note drop out.

    I'd wonder if I really reach 64 voices by doing that.

    @Another User said:

    The string patches will use more voices then dedictaed single instrument patches, because most keys of the strings patches do have mapped two different instruments performing on the same key.
    Sorry, if it wasn't clear. But I wasn't talking of the actual "Strings" patches, but for example the VC-8 sustain patch from the "Orchestral Strings", which earlier drops out here than a Solo String Sustain patch for example (or a trumpet one).

    64 voices should be really enough, and like I said, I don't think it's the problem.

    Almost always when I play back a loaded piece, which has multiple voices on one point, some notes drop out. This even happens on other patches as well, but it appears earlier on the patches from the "Orchestral Strings" folder.

    After some playbacks, it works fine.

    For example I have only two VIs, which each play two sustained notes. In one VI there are VC-8 + VC sustain patches loaded, in the other one DB-6 + DB (I like to layer Solo and Ensemble Strings sometimes).

    While first playing I have here always dropouts, sometimes even before any note ends or new notes appear (and just tested it: even without using release samples). So it should be really only 12 voices (2 tones played by 2 patches with 3 layer) per VI.

    It's really very annoying ... any idea, where this could come from ?
  • H
    H

    Fast repetitins with sustain patches with activated release samples can also cause an overflow very fast.

    The Release Samples do have an average length of 1.5 sec to 2 sec.

    If you have velocity crossfade active you trigger with a 4 layer instruments 4 voices plus 4 release samples =8 voices with one key.

    Release samples alway run through their whole length.

    So if you trigger a key 10 times in one second, you will trigger more than 64 voices only with one key, because 40 or more release samples are ringing.

    The release samples of the ensemble patches do have a longer duration than solo instrument release samples.

    I suggest to use performance repetition patches for fast repetitions passages.

    Sustain patches with activated Release Samples are defenitely the wrong patch for this.

    best

    Herb
  • L
    L

    Yes, that's totally clear of course ... I only mentioned it as one part of the test I just did.

    But if you read further, I described, that it also happens on long sustained notes and in situations, which should definetly stay under 64 voices per VI.

    2 tones played by 2 patches, which each has 3 layers ... makes 12 if I didn't miss something.

    And like I said, it happens the first time played back, sometimes also the second time ... but after a few times playing it back, the notes don't drop out anymore for those slow sustained notes.

    And then I still find it strange, that the patch "VC-8 sustain" seems to drop out earlier than the "VC sustain" patch for example.

    PS.:

    Would it maybe be better to continue this topic per eMail in german ?
  • M
    M
    LazyPeon, I do not know if the following could help? http://community.vsl.co.at/forums/t/18179.aspx
  • L
    L
    No, didn't solved the problem (tried 128 voices).
  • T
    T

    Hi LazyPeon,

    What OS are you running on the PC - XP or Vista?

    Tim.
  • L
    L
    XP Professional (32bit) with SP2.
  • T
    T

    The reason I ask is that I had similar problems but I was using Vista 32. The solution I found was to use Vienna Ensemble instead of VI. I was losing notes with only one patch loaded in one instance of VI in Cubase but using VE as a plugin to Cubase I do not loose any notes across16 (very busy) instruments. It is hard to explain why this should be so but it has completely solved my latency problem.

    BTW, I think the reason you find less notes lost when you play them repeatedly is that data loaded from disk (the samples in this case) remains in the the Windows file cache (an area of RAM managed by the kernel) for a period of time so that frequently accessed data on disk does not require repeated disk access. In this case, for these samples, no disk access is required and so there is less chance of audio not being reading in time. Of course, with disk streaming we don't wish to rely on the file cache but I suspect this is the effect you are seeing.

    I hope VE works for you. Tim.
  • L
    L
    Thx for your reply Tim! I'll definetly try out the VE in near future.
  • L
    L

    In my first test with the VE it actually seems to be better then the VI. Where I had always dropouts with VI, I didn't got any using VE.

    I activated "Synchronous Processing" since it doesn't work without that option in FLStudio. The manual states, that this will cause a higher CPU load. Is this drastic or not so much an issue ?
    What do you think, how much difference it would be, if I'd move my 16 VI-instances template into the VE using that option ?

    By the way, I have a question regarding VE and RAM:

    Since it seems to be a standalone-application, which is connected to the host through some kind of bridge even in plugin-mode (isn't it ?), would it be possible to use more than 2 or 3 GB of RAM on a 64bit operating-system in the VE, while the host still works in 32bit mode ?

    I ask because FLStudio doesn't support 64bit yet.

    Regards

    Christian
  • D
    D
    last edited
    last edited

    @LazyPeon said:

    Since it seems to be a standalone-application, which is connected to the host through some kind of bridge even in plugin-mode (isn't it ?), would it be possible to use more than 2 or 3 GB of RAM on a 64bit operating-system in the VE, while the host still works in 32bit mode ?

    I ask because FLStudio doesn't support 64bit yet.

    Regards

    Christian

    Yes, you can use as much RAM as you have, on either XP64 or Vista64. However, you must make sure that all your drivers have 64bit versions before you take this step.

    DG
  • L
    L

    Yeehaaa!

    This Software rulez soooooo hard !!! [:D]

    Thx for the answer DG.[8-|]
  • T
    T
    last edited
    last edited

    @Another User said:

    I activated "Synchronous Processing" since it doesn't work without that option in FLStudio. The manual states, that this will cause a higher CPU load. Is this drastic or not so much an issue ?
    What do you think, how much difference it would be, if I'd move my 16 VI-instances template into the VE using that option ?

    I found that switching to Synchronous Processing increased my Cubase VSTi performance meter from about 5-10% to about 15-20% when running 16 instruments in VE i.e. w/o synchronous processing my CPU was generally responding in 0.25-0.5 ms (256byte buffer) and with it switched on in about 0.7-1.0ms. So, it seemed to double my processor workload but since it was so far away from causing latency problems it did not seem to matter.

    Tim.