HSR of a Solo Guitar
The goal of this session was to further test different approaches for creating Holographic Sound Recordings of any sound source. This session put it up against a guitar. Two different Arrays were attempted simultaneously.
In this experiment, two arrays were used around the perimeter of the guitar, a hexagon and a square. Two different miking systems were used directly in front of the guitar, XY and AB. I also want to say that I did this session before I had realized the idea of Faces or Radiation. I was still thinking about building a symmetrical shape with the mics in the place of the vertices. I’ve since discovered (and I’ll explain more in this post) that this is not the best way to think about an instrument’s directional radiation characteristics. So, I’ll go through this session and how I did it and then explain further into the concept of Faces of Radiation.
The hexagon consisted of five hyper-cardioid MKH 800s with the sixth point being the XY and AB systems. Each mic was equidistant (2 feet) to a center point that was placed essentially in the center of the guitarist’s chest. All the mics were placed 3 feet off the ground, except for the one in the rear which was placed 3 foot, 8 inches off the ground to get a better quality of the sound. Maybe that shouldn’t have been done in hindsight. I think one must be realistic with the fact that not every angle has to sound good. But one can go as realist as they want with these recordings. In this craft, there’s always a fine line between making recordings for the point of entertainment or documentation. I’m trying to do a bit of both in these experiments.
The sides of the square was made up of 2 Schoeps CMC-6 multi-pattern condensers set to cardioid. Both of the systems shared the same back mic. This honestly didn’t pose much of a problem in the renderer I think because of the low level of importance that the back mic carried. As long as there’s some muffled sound coming from the performer’s back in Unity, the radiation seems real.
The issue with these diagrams is that the microphones are envisioned to be placed in the vertices of the geometry. This made me think that there was a single point for each microphone to be placed, a perfect symmetrical spot. In reality, this is not a helpful way to envision and capture instrument radiation. If you instead think of the shapes tilted a number of degrees (45º in the case of the square and 90º for the hexagon) the microphones are put in front of the shape’s faces like in the images below.
Envisioning the radiation like this frees the engineer from the idea of perfect symmetry and opens to mind to the idea of, “How many faces are there to this instrument, and how should I capture each face in an interesting mono or stereo way.” Because one of the most important things to understand is that, even though this is a holographic recording, if you’re going to experience it in a game engine, the recording will be experienced binaurally. This is to say that stereo/3D miking systems are work the effort on more complex and important faces.
Even though my thought process was kind of off for this session, the front of the guitar was miked in a way that follows this ideology. Like in the piano session, the mic-line array worked well at capturing the open-lid side of the piano. That side is a larger, more dynamic, and overall more important face. The addition of extra mics helped capture its size and dynamic range. In this case, the guitar doesn’t have six sides so the Unity renders ended up leaning towards the square being favorable. More on this below in the Unity section.
I think the arrays used for capturing multi-timbral sound sources (like a drum set) should still utilize the idea of single points on a shape. Although, a point could include an entire array. The Faces of Radiation method of miking is more useful when miking instruments that radiating a single harmonic timbre, with directionality. Whereas the inverse-tetra system works best on instruments that radiate more purely omnidirectional, like a violin or viola. When creating a sonic hologram of an instrument, it’s important to understand that these systems are just tools for the audio engineers HSR toolbox. No one is going to always be better, it all depends on the scenario one is in. Always use your head and ears before you start relying on math.
The Dolby Atmos Stereo Re-renders are available to listen to on this page. Keep in mind that these experiences put the listener in the position of the performer. This is due to the intended experience to be the inverse of this, with the listener walking around the sound sources instead of sitting in the middle of them. Due to the nature of Holographic Sound Recording, there is no accurate way to listen to these recordings on any stereo, surround, or 3D surround system that puts the listener inside of the ring of speakers. The only way to truly hear the recording is to be on the radiating side of the sphere, which would be outside the ring of speakers. This makes these Atmos render’s nothing more than a novelty item for easy listening on my website and gives you little to no real insight on how the arrays actually sound.
Below is a video that explains the array in detail.
Unity Scene
In Unity, the two surround and two frontal arrays were reconstructed using Google Resonance. I coded in the ability to switch between the square and hexagon surround arrays with the “Z” key and the AB/XY arrays with the “X” key. This gives the listener the chance to switch between all combinations of the arrays.
I made my own notes and then sent the recording to my colleague Pari Songmuang to hear her thoughts on it as well.
The biggest thing learned with this session is that the amount of mics should equal the size and dynamic range of the instrument. More mics does equal more spatial fidelity, but it also results in a larger sonic image in the renderer. In our case with the hexagon vs the square system, the hexagon proved just that. The Hexagon array had seemly less gaps in the image, but doing so created an image of a guitar that was bigger than real life. This could be important to know if you are trying to go for a more surreal size of each instrument. But Pari and I both stated that the square was more “true to the size of the guitar.” Also, the square sounded more frequency balanced than the hexagon, I think to capture such a small instrument from so many angles creates unwanted phasing between the microphones.
As far as the front systems. Pari said that there wasn’t much of a notable difference between them, and I’d have to agree. They definitely sound different in timbre but this is probably more due to mic choice than array. To me, the XY sounded a little wider when standing directly in front of the guitar, but the AB was still wide, only with slightly different characteristics.