MIR signal insertion concept scrutinized from an audio engineers perspective

[GLM]

We purchased and used multiple VSL products including MIR already a little while ago and MIR especially for projects we had to cooperate with other studios which already were using MIR in their instrument channels (DAW) in these projects. But before starting MIR for own newly started projects, I really have to deeply scrutinize the concept of MIR again for me, from a audio signal processing point of view and audio mastering issues and audio physics. I mainly worry about phasing and signal flow strategies here with MIR.

From my understanding it has been developed for being placed on each channel to position the channel's signal in the MIR instance (virtual room). Which is a great idea so far. But what I wonder is: how is this signal being routed back into the channel without "doubling" the room on each channels insertion point? Doesn't the great idea of placing signals in a virtual room lives from the "overlapping" of the room effects? I perfectly go with the idea to "place" the channel in a room, but from there I have problems to understand the further signal route MIR goes as audio engineer. Shouldn't the whole "room" go out on a summary aux then, instead of getting out on each channel again, wich is connected the channel volume of each signal, etc.?

I am sure it is just something in my head I am confused with the concept of MIR and I would kindly like to ask other audio enginieers and product developers to chime in here to help me getting my worries sorted. Thanks in advance and HAPPY NEW YEAR!

[Ben]

MIR is not just a reverb, but much more a spatialization plugin. The natural room reverb is part of that, but so are also positioning and mic pattern simulation.

It's like a room-simulation (but based on a lot of real recorded IR signals), where you can place your dry recorded audio / libraries, after which you get a new wet signal that replaces your original. The Dry/Wet ratio inside MIR controls the amount of reverb, but the "Dry" signal inside MIR is already altered with spatialization information and no longer the same as the dry signal in your DAW.

[Dietz]

@GLM said:
I am sure it is just something in my head I am confused with the concept of MIR and I would kindly like to ask other audio enginieers and product developers to chime in here to help me getting my worries sorted.

Hi and welcome,

All very valid questions! Thanks for your interest in MIR Pro 3D.

First of all: If you approach MIR from a sound engineer's POV, it should be more or less self-explanatory: You will find yourself working inside a recording space where you set up a mic array and ask the musicians to take place at chosen positions on a stage. Signal flow is always top-to-bottom, and what you see is what you get. That's about it. 😊

MIR is indeed meant to be used as an inline/insert effect. The plug-in routes the signal from your host to the MIR Venue, where it is represented by a MIR Icon, and after processing back to your host with both(!) the readily positioned(!!) dry signal component and the individual room information created for this signal.

MIR's processing is based on Ambisonics throughout its complete signal path. This means that the dry input signal gets EN-coded to Ambisonics first, which allows us to put it _exactly_ in the position of the impulse used to excite and record the room response, in all three dimensions. Think of it as the ideally positioned close-mic of your input signal, while the wet part is the output of the Main Mic array, but without(!) any direct / dry component (see below). The ratio between this dry and the wet signal can be chosen manually for each Icon, and by default it will also depend on the distance between Icon and Main Mic, much like in Real Life (... this behaviour can be switched off, though, and you could also decide to "turn away" from the aforementioned virtual spot mic).

The final decoding of both the dry wet signal is determined by the Output Format. It defines all aspects of the Main and Secondary Mic. IMPORTANT: Each Venue (organized in Tabs on top of the Main GUI) can have its own settings, but there will always be the first instance of MIR that sets the maximum number of output channels for all following instances (stereo, surround, 3D ...).

To avoid any phasing issues, we have also cut the immanent direct signal component from all eight directional IRs that make up for each individual position on a MIR Venue's stage. Consequently, “100% wet” is NOT the sound you would hear in the Venue! MIR Pro 3D expects some dry signal component mixed in to sound “natural”.

Even when there are "overlapping" sources, as you called them (and there will be many of them in a classical orchestra! 😊 ...), the mixture of them will not cause any phasing - quite contrary: This is where the sought-after "glue" between the instruments happens. All of them will interact differently with the hall, according to their assigned Directivity Profile, their position, their rotation, their width and so on - much like in nature.

So, like I mentioned before, MIR is meant to be used as an inline/insert effect. However, there are two hosts (namely Vienna Ensemble Pro and Pro Tools) that allow for splitting the wet signal from the actual track and route it to an auxiliary target, either for sub-mixes or stems, or simply for any other special treatment. Again there's no reason to fear any kind of phasing.

Finally, there is _one_ scenario where phasing might become an issue if not implemented carefully: We call it MIR's AUX-send mode, which means that is used as a send effect rather than as an insert. While it is not true to the original concept, this still might come handy every now and then to use MIR as some kind of "glorified reverb machine", e.g. with two or three super-wide Icons in different distances. Make sure to use MIR full wet in these scenario, otherwise you will mix its positioned direct dry signal with your unaltered original send track.

... please don't hesitate to ask if further questions arise!

Kind regards,