How Do The Mic Arrays Work?

[JBuck]

So the venues were captured with a specific number of mic arrays. Teldex was captured with only two but Karl Böhm Hall was captured with four. My question is:

If I move a mic array 2 metres to the left in e.g. Teldex, how does MIR produce a representation of Teldex from that position when no recordings of sine sweeps were created from that position? And how accurate is this representation?

If the sine sweeps were played through a loudspeaker, how were all the different instrument profiles like French Horns created when French Horns face the opposite direction to a loudspeaker?

Cheers,

[Dietz]

A bunch of very valid questions! 👍

MIR's graphical representation of its microphones is fully based on the Ambisonics decoding process used for the selected (or custom designed) Output Format. If you open the Output Format Editor you can look directly into MIR's brain. 😉

Ambisonics can be roughly described as three-dimensional M/S, which means that we can define pretty much every sonic aspect in post production, like the number of virtual capsules used, their polar patterns, their angle and tilting, and so on. The possibilities of this format are virtually endless: If we move the Mic Icon two meters to the left, the decoding geometry of _all_ Instrument Icons adopts to this change. Of course, this only sounds convincing within a reasonable range of, say, two or three metres in any direction, but It is sufficient to compensate for architectural oddities to achieve a refined acoustic symmetry, or to increase the perceived width of a Venue.

@JBuck said:
how were all the different instrument profiles like French Horns created when French Horns face the opposite direction to a loudspeaker?

MIR's unique Instrument Directivity Profiles are the result of a meticulous research project that we have been conducting in the early years of MIR development. I tried to cover the basics in a little add-on to the manual called "Think MIR!". This primer covers other aspects of your questions, too:

-> https://www.vsl.info/en/manuals/mir-pro/think-mir

-> https://www.vsl.info/en/manuals/mir-pro/think-mir#instrument-directivity-profiles

PS: Teldex Studio Berlin was actually recorded from three positions - two in "portrait" orientation like suggested by Teldex staff members, and one in "wide" landscape orientation from the side wall.

[JBuck]

@Dietz said:
If we move the Mic Icon two meters to the left, the decoding geometry of _all_ Instrument Icons adopts to this change

Are you saying here that by moving the Mic icon 2 metres to the left, MIR is actually processing that as if the instrument icons have been moved 2 metres to the right?

@Dietz said:
but It is sufficient to compensate for architectural oddities to achieve a refined acoustic symmetry, or to increase the perceived width of a Venue.

Are you saying here that by moving the Mic icon we are altering the natural acoustics of the venue by increasing the perceived width of the venue?

[Dietz]

@JBuck said:
Are you saying here that by moving the Mic icon 2 metres to the left, MIR is actually processing that as if the instrument icons have been moved 2 metres to the right?

No, the geometry of the Ambisonics decoding gets changed. The IRs triggered by the positioning of the Instrument Icons remain the same.

@JBuck said:
Are you saying here that by moving the Mic icon we are altering the natural acoustics of the venue by increasing the perceived width of the venue?

In a way, yes - but it's not the "natural acoustics" that change. The chosen IRs remain the same (unless you've actually chosen a different position for the Main Mic), it's just the geometry of the decoding that alters.

It's the same logic you would apply as the recording engineer in that hall. If the perceived width of the source(s) is too narrow, you have two options: Either you move the individual instruments further to the far left and right, or you bring your Main Microphone closer to the orchestra / sound sources.

****

As a rule of thumb:

• The selection of the Main Mic position determines the complete set of IRs.
• The position of an Instrument Icon will determine the actual IRs used, also depending on its width and rotation.
• The Instrument Profile determines the weighting and spectral pre-processing of these IRs.
• Any further change to the Main Mic will alter the Ambisonics decoding (i.e. position, rotation, tilting, form and number of the virtual capsules, etc.).
• Any further change of the Instrument Icon will change the IR-selection and the positional pre-rendering of these IRs. (There might be an exception to this last "rule" when you apply only minuscule changes to the Icon: Typically, small changes will again only be reflected by the Ambisonics decoding, as there are no discrete IRs for each and every square centimeter, of course. :-))

HTH!

[JBuck]

Really appreciate the time you have taken to explain this to me. Though the term 'Ambisonics decoding' doesn't mean a great deal to me. I gather this is just the way the algorithm construct the appropriate IR's with position, rotation, tilting etc. taken into account.

Last question, could you please point me to some literature that explains which of the mic arrays are used for e.g. back of the hall and front of the hall type applications. I am predominantly working in stereo and I tend to use the same HOA DnMix capsule every time and cannot really hear a lot of difference between the different presets you have made.

Cheers,

[Dietz]

The ‘literature’ used to create the MIR output formats is mostly my ears 😄

Nothing wrong with using MIR 3D's Default Mic, as I've spent a considerable amount of time making it work as meaningful starting-point in most circumstances.

Most presets also provide a little info text describing the underlying concepts and techniques:

However, there is a "mental map" available on the website of the University of California in Santa Barbara, which offers a structured collection of links to all kind of primers and research papers that cover most relevant aspects of Ambisonics (... quite a few of them published by our ingenious HOA development partners at IEM Graz):

-> https://w2.mat.ucsb.edu/240/D/notes/Ambisonics.html

Maybe this helps you in your quest for more detailed information.