Vienna Symphonic Library Forum
Forum Statistics

194,006 users have contributed to 42,905 threads and 257,892 posts.

In the past 24 hours, we have 3 new thread(s), 16 new post(s) and 96 new user(s).

  1. Home
  2. Sample Libraries
  3. Vienna Instruments Libraries
  4. Are the VI's streaming

  • Are the VI's streaming

    J
    J
    "Extremely powerful proprietary streaming audio engine, developed by Vienna Symphonic Library’s engineering mastermind Christian Teuscher."

    When I'm loading the patches into Vienna Instruments there is a size in MB's for each patch. Is this just the amount of RAM required to run each patch (i.e.the size of sample headers that need to be loaded) rather than the size of the samples related to that patch?

    I ralise that the architecture of the samples is different so you might get say 100 sample headers created and loaded into RAM that all come from and reference back to the same .dat file.

    If the Vienna Instruments are streaming from the .dat files instead of it all being in RAM can there be some software configurations available in the instrument (like in the EXS24 preference) to optomize the Vienna Instruments to each computer's RAM and Disc speed configuration.

    Therefore if you have a very fast hard disc you could have less sample RAM loaded to allow for a greater number of articulations.

    Christian, could you give a little bit of info on how the VI's play out samples and if they can be configured computer by computer?

    Thanks,

    Julian
  • T
    T
    Julian,

    I'm sure VSL will answer, but I recall reading somewhere that the preload is 64kb. So for a 1000 samples you're looking at (64x1000)/1024 or about 62.5MB.

    AFAIK there is no way to change this but I would hope in future releases VSL might at least give us the option. I know people like paynterr have requested a 0s preload for both VI and HALion (different subject that though) [;)]

    Best

    Tim
  • G
    G
    Julian,

    we won't make the cache size configurable - even Herb has no access to it [:)]

    this has many reasons: the engine has been tweaked for quite some time
    and we found this being the optimal balance.

    and there is this display telling you how much RAM the pre-cache will occupy, this is not a realtime calculated value rather an text attachment to the file.

    how it works? for every wave mapped there is a tiny snip of wave data in RAM for immediate access. when hold time exceeds this value data will be reloaded from disk.
    RAM access is reliable and constant, whereas disk access is not. there are quite some variations on how quickly the OS provides the data and we do _no_ kernel streaming.

    I would rather investigate the issues you raised in another thread (5.4 GB ...)
    Let's see what we might come up with.

    I am very thankful to power users like you who try to push the limits. This way we are able to constantly improve our product. Thanks.

    christian teuscher
    development
  • J
    J
    Christian,

    Thanks for the info.... and I'm looking forward to those reduced load times!


    Julian [:D]