Vienna Symphonic Library Forum
Forum Statistics

203,186 users have contributed to 43,321 threads and 259,575 posts.

In the past 24 hours, we have 3 new thread(s), 6 new post(s) and 35 new user(s).

  • MIR impulse format - open or closed?

    I am looking forward to the MIR product, but I have one very important question that I have not seen addressed yet.

    Will the MIR convolution library be open (like Emagic Sound Designer, Pristine Space etc) or closed and proprietary (like Sony and Altiverb)?

    It is of critical importance to me that I can use convolution libraries supplied with one reverb with the convolution engine from a different reverb. E.g. Today I can load the Sound Designer libraries into Altiverb and IR-1. Unfortunately, I cannot load Altiverb libraries into anything else, so we are phasing out our use of Altiverb. Although Altiverb has a nice library, it has limited features and Audio Ease have been slow to update it. If they had made their library open we would have kept the product as part of our suite, but Audio Ease seem determined to shoot themselves in the foot.

    Likewise I will not spend $900 for the Sony DRES777 libraries because they cannot be used on any other platform. In fact, the closed nature of the Sony system was one of the reasons we will drop that over time.

    I use a wide range of convolution reverbs in my work and I can't afford to invest in closed libraries such as Altiverb because it limits my choice of platforms and locks me in to one reverb engine.

    I know VSL supports open standards for sample libraries such as Giga, EXS24 etc so I am optimistic that MIR will follow the same philosophy. I have utmost respect for the technical skills of VSL but please do not lock us in to just the VSL developed reverb engine but allow us to choose the best possible convolution reverb engine from all the alternatives out there on every platform, and ensure MIR works with them.

    If MIR is a closed system, it's very unlikely that I would purchase it, just like I wouldn't purchase Opus 1 if it only came in one sample format and VSL didn't allow us to use it with the sampler engine of our choice.

    I'd like to know this so I can either budget for MIR for our mixing rooms later in the year - or look at alternatives if it is going to be a closed system.

  • So we will loose you as potential customer, sorry to say so. :-/

    The MIR.E(ngine) will rely on a completely proprietary format (allthough we will of course include the option to _import_ third party IRs). The reason for this is system-immanent: The MIR.E will have to "know" the input signals to work properly regarding directivity, while the IRs themselves will have a set of very demanding requirements to be fulfilled.

    I don't want to spoil too much info about an unfinished product, but don't consider the MIR just another "reverb". Actually, it will be a complete mixing-frontend for our products. The integration will be _much_ closer than in an ordinary "sample library / reverb / mixer"-combination.

    Thanks for your interest!

    /Dietz - Vienna Symphonic Library

    /Dietz - Vienna Symphonic Library
  • Well, that is very disappointing. We don't need another mixing front end as we already have that in our studios. You have to remember we have a lot more invested in SSL and Pro Tools than we do in VSL, so we are unlikely to change.

    I will point out that there are many companies out there with excellent convolution reverbs, and by limiting the MIR format in this way you limit the ability of users to choose another reverb engine apart from the one you supply with MIR.

    Although I acknowledge VSL's expertise in sampling, unless I am mistaken MIR is your first foray into this level of sophisticated audio programming, so it is unrealistic for us as users to expect the MIR reverb engine to be better than companies such as Waves who spent two years and hired some of the worlds top experts in creating their convolution products.

    I am sure the problem of the reverb engine knowing the directivity of the input signals could be solved. Small companies like Propellerheads put themselves on the map by proposing standards like Rewire. Likewise, VSL could propose a standard for convolution impulses and make MIR an open format library.

    Until then, it's unlikely users like myself will consider it. In fact it is very worrying - is VSL planning to take this same closed approach to future sampling products as well?

  • last edited
    last edited

    @Another User said:


    Small companies like Propellerheads put themselves on the map by proposing standards like Rewire. Likewise, VSL could propose a standard for convolution impulses and make MIR an open format library.


    But - that's what we do! Everybody is invited to develop MIR-sets once the product reaches the market, according to our requirements and proprietary standards. What's the problem with that?

    /Dietz - Vienna Symphonic Library

    /Dietz - Vienna Symphonic Library
  • I am really that this got your dander up Dietz. Going into MIR is really getting me excited about the possibilities. When do you think this will be ready for market?

    Rob

  • Nonono, I'm not angry at all, more kinda amused. The whole history of the VSL is full of disbelieve and bewildered reactions, especially from the "no-one did it this way until now, so it _can't be done like that"-variety.

    I wrote it before - I hesitate to give any release-dates, mainly to avoid unnecessary hopes and the impression of uncontrolled delays. Let me just say: There's a lot done already, but still a lot to do ... ;-9 ...

    All the best (and an excuse to the address of "Soundstage", if I made the impression of being disgruntled).

    /Dietz - Vienna Symphonic Library


    /Dietz - Vienna Symphonic Library
  • last edited
    last edited
    BTW:

    @soundstage said:

    We don't need another mixing front end as we already have that in our studios. You have to remember we have a lot more invested in SSL and Pro Tools than we do in VSL, so we are unlikely to change.


    Hey, nobody beats you to the new stuff. Nice gear you have over there! (... allthough personally I'm more the Neve-guy with a preference for floating-point savvy audio-workstations :-] ...). In the end, it's all a matter of taste.

    All the best,

    /Dietz

    /Dietz - Vienna Symphonic Library
  • Dietz, I assume MIR be happy accepting external signals (even though it can't know them the same way it knows VSL programs)? Obviously, most people combine other sounds with VSL sometimes - anything from enemy libraries to synthesizers to live instruments.

  • last edited
    last edited

    @Dietz said:


    Uhm - it's _their_ format that is limited, not ours. We talk about several _dozens_ of impulses for a single instance if the MIR.E, in any arbitrary surround-format. Give me one example of any other convolution reverb doing this. - Of course, you would be able to import conventional mono- or stereo-IR's ... but what for?

    I have no problem with incompatibility if other reverbs aren't as capable as MIR. If MIR only contains special 12 or 24 channel convolution samples then obviously you cannot load those into most reverbs today. Pristine Space supports 8 channel convolution but most of the others are limited to 4 or 6 channels right now.

    What I am more concerned about is what Altiverb and Sony do and that is they deliberately go out of their way and make sure their samples are incompatible with others. Even a basic stereo or surround impulse from the Altiverb library cannot be loaded into IR1 for example.

    Will VSL be doing that? Or is it just a matter of technical incompatibility?

  • last edited
    last edited

    @soundstage said:

    [...]
    Will VSL be doing that? Or is it just a matter of technical incompatibility?


    Just a matter of compatibility. - And we talk about _several_ dozens of IRs here ...

    /Dietz

    /Dietz - Vienna Symphonic Library
  • last edited
    last edited

    @Nick Batzdorf said:

    Dietz, I assume MIR be happy accepting external signals (even though it can't know them the same way it knows VSL programs)? Obviously, most people combine other sounds with VSL sometimes - anything from enemy libraries to synthesizers to live instruments.



    Yes, in one way or the other ... acting more or less as a nice, but conventional reverb.

    /Dietz - Vienna Symphonic Library

    PS: tststs .... "enemy" libraries .... there ain't! 8-]

    /Dietz - Vienna Symphonic Library
  • (I've referred to VSL as an enemy library on other forums too!)

  • Hi Dietz and VSL Team:

    For MIR reverb IRs will you be commissioning folks like Ernest Cholakis, or others, to produce IRs for MIR? Importing existing full stereo Pure Space IRs would be one thing, but a special and full VSL MIR implementation of them sounds to me like a match made in heaven [[[;)]]] .

    The reason I mention Ernest here specifically is that, thanks in very large part to everyone's high praise here for his IR libraries, I licensed both "Pure Space" libraries and find them absolutely extraordinary, in every meaning of the term. And that is not the kind of praise that I toss around lightly, (except for VSL offerings [:D], and a few select others).

    Incorporating them into the more complex spatial and other currently mysterious [[[;)]]] aspects of MIR would be on the critical path challenge if you are considering this sort of approach, however I can't imagine the result being anything other than astounding with the VSL Pro libraries.

    - OR -

    Are you planning on VSL produced IRs surpassing his Pure Space collections in terms of sonic quality? Is that even possible? [*-)]:

    I'll be very eager to hear what you can share...

    thanks, and best regards... ...jim

  • Hi Jim,

    I'm curious what you mean with "existing true stereo IRs" from Ernest? I also have Pure Spaces, but as far as I understand they are only mono (single emitting point) to stereo. For true stereo you need two stereo impulses, one to process left channel info and for the right channel.

    With my Pure Space impulses I have to narrow the stereo width of the input, otherwise lateral sounds in the stereo field may sound detached and not in the same space.

    So, I think they could be even better than they are now. [[;)]]

    BTW: here is a nice Windows Media video file with some background on recording impulses for the Waves IR1:

    http://www.waves.com/objects/movies/mov_behind_scenes_ir1_01.wmv

    (and enjoy the professor's English [:P] )

    Cheers,

    Peter
    www.PeterRoos.com

  • last edited
    last edited
    Hi Peter

    @Another User said:

    I'm curious what you mean with "existing true stereo IRs" from Ernest? I also have Pure Spaces, but as far as I understand they are only mono (single emitting point) to stereo. For true stereo you need two stereo impulses, one to process left channel info and for the right channel.


    oops, you're absolutely right about Pure Spaces as stereo, from a single emitting point, [:O]ops: (although I occasionally do have some very interesting fun "doubling up" Pure Spaces IRs in the Pristine Space plug-in in 4 channel mode, and also doing the same in the IR1 plug-in) [:D].

    Interesting that you hear them latterally detached - I've really enjoyed the widths in the stereo fields. If you would be so kind to describe specifically how you are narrowing the stereo field for a particular Pure Spaces IR, I'd like to try replicate it and try to hear what you are hearing.

    I had seen the Waves IR1 wmf, and it is amazing to me what developers are doing with technology in this area - very thoughtful and exciting work. Who'd have thought we could ever put Pink Floyd inside the Great Pyramid Kufu at Giza !!! [H]

    best regards... ...jim

  • Jim,

    One of my problems with the mainstream impulses is that with very lateral sounds (close to left or right, like French Horns and Trumpets), they will "come out of the convolution" with an ambience that is also at their lateral position. Just try this, pan a trumpet to the right, with only a bit signal in the left channel. Now put this through IR-1 or Pristine Space with a single stereo impulse Wave.

    You would expect that the sound of the trumpet will travel through the entire "hall", which should be perceived as a virtual room with its acoustic gravity at the center. But it won't sound like that, because single stereo impulses cary no cross-channel information. The convolution process does not know how to handle position-based information over time... So, the convolution is simply "weighted", with more ambience applied to the right side, where most of the signal is.

    Sad but true. The common fix, is to make the input to the convolution more narrow, for instance with the Waves S1. The original trumpet will then still have its dry component panned to the right, but its sound will be inserted more centrally into the virtual room.

    I think that the applauded quality of AltiVerb probably can be attributed to its support for using two impulses, one for the left side of the stage, and one for the right. Back to my example: we need an impulse for sound emitted from the left part of the stage and one for the right part. This would include the required temporal and positional information into the impulse set.

    I still have not found good stereo-stereo impulses in the public domain, so I can't (subjectively) validate this assumption with my Pristine Space.

    A few years back, Acoustic Mirror was the major convoluter for the PC, and it supported only single stereo impulses... I think this is probably the reason why Ernest is not yet offering true stereo sets, he probably started his projects when Acoustic Mirror was the main thing...

    Peter

  • Very interesting discussion!
    Good point with the mono to stereo thing on convoluters. That might explain some experiences I had.

    I was a bit surprised about the omnidirectional loudspeaker this professor was so proud of. I´m not aware of any instrument which radiates omnidirectional. Even the double bass is quite directional.

    A question to the MIR-project: Will that be a mixing frontend directly attached to Gigastudio, or can it be used also inside an audio editor? The step in the audio editor is very important for my workflow, I couldn´t work directly from Giga to CD.

    Bests,
    - M

  • Mathis -

    as much as I enjoy to talk about the MIR.E-project - please understand that I can't spread too much information here, partly because some aspects are still in developement, while others are simply too "hot" to be told before the actual release. - Stay tuned ... ;-]

    /Dietz - Vienna Symphonic Library

    /Dietz - Vienna Symphonic Library
  • Mathis, I think it's safe to say that VSL's approach to the universe is to roll up their sleeves and - with the patience of Job - apply brute force to a problem, whether they're sampling instruments or spaces!

    What the Prof is doing is less precise, taking more of an "average" response. For that, it probably makes more sense to excite the room in all directions. He's using sine sweeps, not generating individual impulses for each instrument.

  • *apply brute force* - i like that, although if generally speaking this could be understood more as plain power and not so much as the clever system behind
    christian

    and remember: only a CRAY can run an endless loop in just three seconds.