Introduction

Hi, my name is Eyal. I’m the Waves developer of presets for the IR-Series Parametric Convolution Reverb project. Sit back and let me tell you a bit about our process for capturing Impulse Responses.

Concept

The basic idea behind the measuring procedure is to capture the Impulse Responses (IRs) of renowned and acoustically interesting venues. The results can be used for archival purposes for posterity, and also for use in developing a state-of-the-art reverb tool. The capturing has been done in some of the most beautiful sounding spaces that exist, selected from around the world. Our list of venues continues to grow. My experience of capturing the Wembley Arena is a great example of a large venue that I measured recently.

Setup

I arrived on a quintessentially English, rainy Monday morning at the Wembley Arena. I was concerned, realizing that the pouring rain would affect the noise-floor, but intuition told me to proceed as planned.

After a gracious welcome from the local staff, we considered the right spot for placement of the mobile control panel. The process of set-up up began with unpacking the 4 large equipment cases.

I always begin with my computer. It is a strong PC without a ventilation system, which makes it very quiet. I connect a professional I/O device with A/D converters to my computer. This allows me to record up to ten channels directly to my hard disc, without the use of a mixer.

The next procedure is the microphone setup. It is critical to consider the best spot in the venue for mic placement. In the case of the Wembley Arena, the microphone is at the front of the stage, where the sound engineer is usually positioned during live shows -- the best seats in the house!

Three different microphone techniques are employed for the measurements:

The first one, the ORTF technique, is comprised of 2 cardioid microphones, set 17 cm apart at a 120° angle. This technique is well suited for recording in large rooms where there is considerable reverberation.

The second technique is the AB technique. I use 2 omni directional microphones, set up parallel to each other, usually at a distance of 1.5 meters from each other. This technique usually produces a good, wide stereo image. This image is created by the level, time and phase differences due to mic placement.

The third technique is the B-format technique, which utilizes 4 different capsules residing on the same microphone (called a Sound Field microphone). This enables the reproduction of the true surround audio perspective. The ORTF and the sound field are placed on a computer controlled rotating table that is used to measure over a full 360°, taking measurements every 15 degrees. There are 24 measurements in all.

Next, I connect the speaker, which is placed where the live performers would be located. I connect the send of the I/O device to the speaker and then loop it back to an input of the I/O device. During a sampling session the speaker will be moved from center-stage to a position on the left, and then to a position on the right side of the stage.

Fortunately, during set-up in the Wembley Arena, the rain stopped. My intuition had proved correct! I could then proceed with the measurements without concern for the noise-floor.

Process

After the equipment is physically in place, the preparation of the electronics for the recording begins. Having checked all the parameters on the I/O device, I launch my Digital Audio Workstation (DAW) recorder. Prior to recording, care has to be taken so that there are no interference noises. These can include electrical devices, air conditioning and disturbances from the outside. In this case, I had to ask the stage manager to turn off their lighting system, which was large enough for an entire town and produced noise accordingly!

A sound check is done to ensure that all microphones are functioning properly and have proper input levels to the DAW. Then the recording begins!

A sweep signal (sometimes referred to as "a pure tone") which changes continuously from low to high frequency is used to excite the room for capturing the Impulse Response. This is one of three common types of signal sources used for capturing IRs. The others are MLS (Maximum Length Sequence) and Chirps. We have selected the sine wave’s log sweep because of its key advantage with non-linear distortion elements. These elements are pushed in front of the linear responses after processing. Therefore they will appear before the direct arrival of the IR, making them very easy to cut later in the processing stage. WAVES uses a 15 second sweep that results in a noise floor of around -100 dB. By using a longer sweep we could theoretically get a better signal to noise ratio, but in fact, for longer sweep durations the IR starts to smear due to slight air pressure variations in the acoustic space. As a consequence, not only the IR is distorted, but also due to this smearing, the actual improvement of signal to noise is only by 2-3 dB for a 4 times longer sweep tone. Therefore, we found that a 15 second sweep is the optimal duration. The frequency range we use for the sweep is 22Hz to 32 kHz. Each session is composed of 24 sweeps, where after every sweep the rotating table turns 15 degrees. I always conduct 3 sessions for each point in the room, moving the speaker (the sound source) to center, left and right of the stage in order to give the recording the realistic character of the place.

At the end of my day I have recorded 72 sweeps, covering the venue from all directions of reflection expressed in different angles of arrival. Combine this with the 2 stereo channels of the B-format and 2 types of stereo placements, and the final number is 288 individual samples of the room from one microphone position. This is an unprecedented degree of information on the venues that have been sampled for this project.

In some venues, I find more than one spot to place the microphones (aisle, balconies, front rows, etc.) and therefore, conduct more sessions. When measuring live rooms in recording studios, I usually capture more than one position in the room because I find that the recoding sounds different at different angles and heights in the room. The factors taken into consideration are the distance between the microphones and the sound source, height of the room, wall reflections, etc.

But, it's not over yet. What about the physical characteristics of the space itself? The dimensions of each venue, length, width and height, are carefully measured using a precision laser device. I also take measurements of the distances between the microphones and speaker. All of this information will be useful for the processing that will be done to derive Impulse Responses from these recordings.

Now its just a simple matter of packing it all up again...Aaaah, but the work has just begun...

Processing

After the recording session, the audio files undergo processing in the Waves laboratory. Here the sweep signal is removed from the recording, and the reverberation of the room is reconstructed as an impulse response. This is a considerable process in and of itself using more "Waves science".

Now the captured IR can be loaded onto the IR-1 Parametric Convolution Reverb engine. It can be used to process a studio recording session to reproduce sound as if the performer was actually standing in the Wembley Arena. Having been in these venues, it's an amazing and very satisfying experience to listen to this uncanny effect!

Enjoy!!!

Eyal

Behind The Scenes - The creation of the IR Library (wmv movie, 8MB)