Vienna Symphonic Library Forum
Forum Statistics

191,957 users have contributed to 42,821 threads and 257,504 posts.

In the past 24 hours, we have 10 new thread(s), 59 new post(s) and 232 new user(s).

  1. Home
  2. Sample Libraries
  3. Vienna Instruments Libraries
  4. VI dream set-up questions

  • VI dream set-up questions

    P
    P
    Hi Forum People,

    So I've decided to take the VI plunge, and I have some questions before I do. I never cease to be amazed at how useful forums like this can be — answers to any of these questions would be greatly appreciated!

    1. I use Digital Performer under DAE (my hardware is Pro Tools HD Accel 3) on a 2 x 2.5 GB G5, latest OS X version. The ONLY plug-ins DP will run in this mode are RTAS, or VST if you have an RTAS wrapper, which I've never tried. Since (unless I'm mistaken) VI is not yet out in RTAS format, are there any options for me currently if I want to use VI hosted by DP itself? If not, can I access the VI stand-alone instrument on the same Mac on which I'm running DP in DAE mode, and will that be a reasonably CPU-efficient way to work?

    2. Re the "stress tests" results listed in one of the other threads: If memory serves (no pun intended) the tests were conducted with 4GB of RAM. Let's say I install the full complement of 8GB of RAM in my G5. Is it reasonable to expect that I'd double the number of samples I can load?

    3. Will more RAM also increase my maximum polyphony, or is that purely a function of processor speed?

    4. If I get a new Mac and use my current G5 as a straight VI host machine, will I get more polyphony out of the G5 as a stand-alone VI host than I would using the G5 as a combination sequencer/VI host?

    5. Let's talk polyphony — Will a complicated "universal" matrix need to reserve a bunch of voices for each note because of the need to accommodate the ability to switch dimensions quickly?

    6. Right now, with the computers that are now available: if I want to be able to sequence and play a full-blown, CPU-straining full orchestra action cue "live" from the sequencer (i.e., without first "freezing tracks" or otherwise mixing down part of the audio), how many computers would I need? How would the library best be divvied up among the computers? (Let's pretend for the sake of this question that money is no object — just thought it would be a useful point of reference to know how the "big boys" are doing it.)

    Vielen Dank,
    Peter
  • J
    J
    last edited
    last edited
    Hey Peter,

    1. Not sure about the VSP/RTAS wrapper.

    2. No, because in the 32-bit world, you won't be able to load more than about 3 GB of samples. Even if you installed 8 GB in your compuer. The only other option at this point would be to use the "Stand-Alone Multiple Instances Trick", but it creates specific issues (ie. audio outputs).

    3. Nope, the real issue here is CPU power.

    4. Yes, because your G5 will be dedicated to running VI, and thus its processing power will not be used by anything else.

    5. I don't think so.

    6. It depends how many instruments / articulations you want loaded at all time in the background. This article is a good starting point to give you an idea.

    Personally, If I had an unlimited amount of money and space and a perfectly sound-proofed technical room, I'd get 6 Mac Pros with 4GB and run the whole thing.

    Jerome
  • P
    P
    Hi Jerome,

    Thanks so much for your prompt response, and equally for your highly informative website. The Mac-mini farm approach may be just what I'm looking for - very creative! Here are a couple of follow-up thoughts/questions, if I may beg your indulgence once more:

    1. How are you getting MIDI into/out of your Mac-minis? Couldn't seem to find any info about that in any of the three related articles on your site.

    2. Just to give you background on my "meaning of polyphony" question: When I first fired up my Gigastudio, I figured "400 voices -- no problem!" Then I tried a legato 4-way switched Garritan String patch w/release samples -- one quick string run was quickly getting me up above 100 voices (obviously there are workarounds, but still...). Just wondered if VI had similar issues re polyphony, or whether the dynamic allocation of voices was more flexible than in Giga.

    3. When you buy VI, do you have to make a format choice between AU and VST? Or can you make that choice after that fact? I ask because it might be possible for me to kludge together the ability to use the VST version of VI in DP under DAE using an RTAS wrapper, but obviously I'd need the AU universal binary version for the Mac-mini farm system.

    Thanks again,
    Peter
  • J
    J
    1 We use MidiOverLan.

    2. Well, there is a difference between voice count and polyphony. VI will allow you up to 64 voices per instance. However, the number of instances is limited by your CPU.

    You can only use one patch at a time in a VI instance (except if you load two patches on the same cell). If you use the "Release Samples ON" feature, then every note played uses twice the number of voices. If you use the "Velocity Crossfade" feature, then every note played use 4 times the number of voices.

    So, let's say you have both features on, you will be able to play a chord of eight notes. The ninth note will make the first note disappear.

    However, if you don't use the Release Sample feature (we never do, because we found the difference to be minimal on most patches), you can play twice as many note.

    Now, the polyphony is more of a CPU issue. Let's say you have 10 instances of VI running on your slave. If you computer is not powerful enough, you're going to get crackles when accessing too many samples at the same time, especially when you have fast runs. That's because the CPU can't keep up. The only solution is either to have less instances, or to increase the buffer.

    3. VI comes as a VST and AU, and works on Mac & PC. You don't have to select a version before buying.

    Jerome
  • P
    P
    Interesting. I'd say if you're using chords with more than 8 notes in a single instance... time to crack open that orchestration book again! [:)] (Unless you're channeling Ligeti.) I.e., sounds like voice allocation is very well done.

    SO... that's 64 STEREO voices, correct? And presumably, if I created a matrix that had say, a velocity crossfade legato violin in one layer switchable via the mod wheel to a staccato violin, voice count would be further stressed, yes? (Or would does VI just allocate the voices for whichever layer is currently active?)

    BTW, for those of us with no separate "machine room" for the computers: You've commented on the low heat-low power usage of the Mac minis. How brutal would the noise problem be with 8 of those in the room with you?

    Thanks again for your answers -- I'm also enjoying reading the other threads in this forum that deal with the Mac mini slave concept.

    PL
  • cmC
    VSL Team
    cmC
    last edited
    last edited

    @Another User said:

    you're going to get crackles when accessing too many samples at the same time, especially when you have fast runs. That's because the CPU can't keep up
    i can't provide the proof right here and now, but i'd say the harddrives and the *road to the CPU* is the problem (file access needs hight priority kernel time, if data feed is too slow, too much CPU is blocked to do anything else).
    everything between the CPU and the drive increases overhead, so to have samples on sATA disks connected to an ICH (integrated controller hub, eg. version 7) would reduce this load to a minimum - therefore firewire disks are the second-best choice only.
    too bad the mac-minis don't have their 2nd sATA port (from the chipset) accessible from outside ...
    christian
    and remember: only a CRAY can run an endless loop in just three seconds.
  • V
    V
    Plurye, I'm HD3 user as well and using Mac Minis as VI slaves based on Jeromes setup - it works great fyi. [:)]

    About RTAS, I'm obviously awaiting this feature with huge anticipation like urself, but I fear that the VSL chaps have dropped the idea of integrating this type of support as no info has been posted or communicated about it for a long time. As for RTAS wrapping, VSL reported unsatisfactory results doing this, so not a viable solution it would seem. Ill do some testing one of these days to check how bad the problem really is.
  • J
    J
    last edited
    last edited

    @cm said:

    i can't provide the proof right here and now, but i'd say the harddrives and the *road to the CPU* is the problem (file access needs hight priority kernel time, if data feed is too slow, too much CPU is blocked to do anything else).
    everything between the CPU and the drive increases overhead, so to have samples on sATA disks connected to an ICH (integrated controller hub, eg. version 7) would reduce this load to a minimum - therefore firewire disks are the second-best choice only.
    too bad the mac-minis don't have their 2nd sATA port (from the chipset) accessible from outside ...
    christian


    Our samples are on the internal SATA drive. They have been upgraded to 7200rpm drives.

    I've been wondering if the problem could also come from the fact that the samples are on the same drive as the system drive...

    By the way cm... are you saying that there's a free SATA connector inside the Mac Mini???

    Jerome
  • P
    P
    Vagn L.,

    RTAS seems to have more trouble "playing well with others" than any other format. The REAL answer would be for Digi to dump RTAS and support AU, but... peace on earth would be nice too. So I hope VI supports RTAS, but I believe them when they say the format ain't cutting it. Anyway, thanks for the tip (however sad) about the ineffectiveness of the VST RTAS Wrapper.

    BTW are you using Logic? If so I'd love to ask you some questions about Logic and DAE. Much as I prefer the design of DP (I used Logic for a few years in the mid to late 90's), I'm increasingly frustrated by the half-baked way DP interacts with the Digi hardware.

    Peter
  • cmC
    VSL Team
    cmC
    jerome, i didn't look into a mini mac by myself so far if there might be even a *header* - but see by yourself http://www.intel.com/products/chipsets/945gm/945gm_diagram.htm
    christian
    and remember: only a CRAY can run an endless loop in just three seconds.
  • J
    J
    last edited
    last edited

    @cm said:

    jerome, i didn't look into a mini mac by myself so far if there might be even a *header* - but see by yourself http://www.intel.com/products/chipsets/945gm/945gm_diagram.htm
    christian


    hmmmm... I gotta pop open our backup Mac Mini today...

    J.
  • V
    V
    Plurye I wouldnt touch a Logic-Digidesign combo with a 10 feet pole. I'm in Protools and enjoying every minute of it, and see no reason why Logic would be a better solution. Afaik, the whole TDM bridge thing is the devils brew, asking for plugin and DAE problems of all kinds. [[;)]]

    As you say, Digidesign should just get in support of AU and be done with it, but as company they have always gone their own ways - often to great dismay of us users.
  • P
    P
    Ah, I WISH I could just use Pro Tools. Mixing implementation is still the best, but the MIDI implementation is still so far behind DP that I only use PT for songs. Without a notation page (and many, many other features), I just find it too hard to use for serious MIDI. So I put up with the problems and curse a lot (first and foremost: no separate outputs from virtual instruments when working in DP in RTAS mode. So if you have 16 tracks in one Mach Five instantiation, you have to send them all out the same output pair. Also problems sending timing sync to Stylus RMX, no ReWire capability in RTAS mode, etc.).

    HOWEVER... since you're using Mac Minis with PT HD 3: How many Mac Minis are you using? If more than four, how are you getting their audio outputs into PT? Do you have more than one Digi I/O unit? I have one 96 I/O and would love to find a cheaper way of bringing in additional digital audio streams than buying a second and maybe third one.

    PL
  • J
    J
    last edited
    last edited

    @cm said:

    jerome, i didn't look into a mini mac by myself so far if there might be even a *header* - but see by yourself http://www.intel.com/products/chipsets/945gm/945gm_diagram.htm
    christian



    Hey cm... check this out : http://flickr.com/photos/jsnell/106961279/in/pool-macintosh/

    I completely forgot that there's no such thing as a standard SATA connector in the Mac Mini... there a big connector (above the Airport card) that goes to this daughterboard which is designed to plug directly to a SATA drive.

    No luck 😕

    Jerome
  • D
    D
    Vagn Luv / plurye: I am using Logic TDM (ie using Logic as a front end for ProTools HD) and it's 100% stable and an excellent platform, ESB bridge works fine too.

    I'm churning out one soundtrack after the other on it - I use it for programming, recording, mixing (stereo and surround), also orchestral recording when I just patch it into an other studio's 192s. Having to do the sequencing and recording/mixing on different platforms (eg Logic and ProTools) would just cost far too much time.

    Plurye: In Logic TDM you can use AU and TDM plugs (not RTAS) but they are essentially in different environments and connected by the ESB System Bridge, an internal; 8 channel connection. Some people also use a separate soundcard for native audio and then connect it to a 192 via ADAT.

    I haven't yet used VI with this rig - I hear there are problems with Vi and Logic TDM in terms of screen re-draws or something like that. I am just using an external PC for samples - coming in via ADAT.

    If you're interested I can give you the details (versions etc). I think nicks@aubergine from the forum here is also on Logic TDM (How are you Nick?)

    Dom
  • cmC
    VSL Team
    cmC
    yep - if i'm right the interconnect slot holds all connections, but too sad apple has not donated an external sATA connector to us. possibly OSX does not recognize sATA disks as plug&play devices? XP does not always ...
    christian
    and remember: only a CRAY can run an endless loop in just three seconds.
  • P
    P
    Hi Dom (and Nick?),

    I would LOVE to know the details re Logic TDM use, as you (Dom) so generously offer! My big issues using DP in DAE mode:

    1. Can't send the individual instruments from a multitrack virtual inst (e.g., Mach Five or Stylus RMX) to separate outputs, so they can be processed by my lovely, expensive Digi cards. Does the ESB give you this capability, and is it straightforward or hair-raising?

    2. Timing sync to RMX (presumably this is no problem, since you'd use the AU version).

    3. ReWire inactive (also presumably no problem, since this is outside the RTAS realm).

    Could you elaborate on the "separate soundcard for native audio/lightpipe to HD" idea as well? Don't quite get that.

    Thanks,
    Peter
  • D
    D
    last edited
    last edited

    @plurye said:

    I have one 96 I/O and would love to find a cheaper way of bringing in additional digital audio streams than buying a second and maybe third one.


    If you're on 44k or 48k you can connect a 2nd hand Digidesign ADAT Bridge (24bit) (around £250-£300 on ebay) to the legacy port of your 96. That'll give you an extra 16 ADAT inputs.
  • P
    P
    Dom,

    Excellent tip re the ADAT bridge! Thanks.

    PL
  • D
    D
    1. It does, but you have only 8 channels in total that connect your native side with TDM. Unless you use the "core audio" on separate soundcard idea. (see 4 below). You can however run EXS24 - Logic's built-in sampler - as a TDM plugin, and then you don't have the 8 channel restriction.

    2. Fine. RMX just runs on the native side.

    3. I don't know about rewire - I've never used it.

    4. You could have for example and RME card with 16 ADAT outputs that you use for Core audio. You then stream these 16 channels of ADAT into a Digi 192 Digital for example. Again both environments - native (core-audio) and TDM are both controlled by Logic as part of the same session. Some people find thus combo more stable (for me they're both stable). One problem with this is though the timing between the 2 audio engines can be erratic when you have Logic's plug-in delay compensation switched on.


    There are some overall disadvantages of the Logic/TDm combo: surround TDM plugins are not supported. Multi-mono plugs are not supported. Bussing and routing is not as flexible as Protools. It can take longer to instantiate TDM plugins. (up to 3 seconds or so).

    Dom