Microphone placement?
-
- KVRist
- Topic Starter
- 134 posts since 13 Apr, 2016
Does anyone know of any information about microphone placement/position algorithms? I'm not even sure what to google, I always end up finding stuff about modeling spaces in the reverb world instead. What I'm looking for is something along the lines of how Pianoteq and others allow users to position microphones in relation to the source and how that's modeled.
Thoughts anyone? Thanks.
Thoughts anyone? Thanks.
-
- KVRist
- Topic Starter
- 134 posts since 13 Apr, 2016
Thanks, but that's just using impulse responses from specific positions in a room. I'm talking about an algorithm that takes into account microphone in arbitrary positions.
- KVRAF
- 15205 posts since 8 Mar, 2005 from Utrecht, Holland
Think about it, mic positioning IS acoustics, thus can be simulated with an IR.
We are the KVR collective. Resistance is futile. You will be assimilated.
My MusicCalc is served over https!!
My MusicCalc is served over https!!
-
- KVRian
- 617 posts since 30 Aug, 2012
I haven't worked with this but I suspect it is typically done using multi-point IR samples and an interpolation/blending means.
Even though it's not directly what you are looking for the convolution methods used for room simulation likely apply. For example, the length of a convolution reverb can be made continuously variable by simply trimming and/or fading the length of the IR sample. By using multiple IR samples of various microphone positions (IR "maps") you could interpolate/blend between those samples to generate continuously variable positions.
That said, some of the "cheapy" guitar amp simulators I have used seem to be simply applying various EQ curves which you could model using cookbook IIR filters.
- KVRAF
- 15205 posts since 8 Mar, 2005 from Utrecht, Holland
Can you think of a simple mathenatical model?
Foir example an anechoic chamber with one reflective panel. Would that be computable, would that work?
Foir example an anechoic chamber with one reflective panel. Would that be computable, would that work?
We are the KVR collective. Resistance is futile. You will be assimilated.
My MusicCalc is served over https!!
My MusicCalc is served over https!!
-
- KVRist
- Topic Starter
- 134 posts since 13 Apr, 2016
Maybe I am overthinking it. Maybe it's just as simple as the wet/dry ratio is related to the distance of the instrument to the source instrument...and of course, using some room model for the wet signal.
-
- KVRian
- 617 posts since 30 Aug, 2012
It could get very complicated depending on how realistic you want it - but you could start with some basic filtering as I mentioned above. For example, if you are trying to emulate microphone placement in front of some object (like a speaker cabinet) you could measure or get some plots of on-axis vs. off-axis frequency response of various speaker cabinets and microphones and look at how the high end rolls off, bass bumps, etc. You could then model those characteristics using simple IIR filters having frequency and gain mapped to a "mic axis" placement control on your plugin.
That's just one idea for something very basic.
A common, more advanced approach seems to be a combination of impulse responses and algorithms. For example, the guitar cabinet example above could be emulated with a couple of basic "mic positions" (e.g., 3 select-able impulse files) and then various flavors of continuously adjustable EQ on top for tone adjustment. If you want to include "room tone" in that mix then you could have additional impulses and/or algorithms to generate reverb, delay, etc.
BTW - be careful when using "wet/dry" controls, especially with short delays and filters. The region in between full wet and full dry (the blended region) may not give you the ratio of mix or sound you expect due to phase interactions (comb filtering).
- KVRian
- 871 posts since 6 Aug, 2005 from England
I’ve heard interpolating between different IRs gives acceptable results. I’ve not tried it myself. So if you have a 10 foot recording and a 3 foot recording, your IR can be a mix of the two.
Dave Hoskins. http://www.quikquak.com