About semi-dry recording

Introduction

Sonic documentation of pipe organs for posterity, scientific analysis, or virtual instrument creation, exists since the 1980s using different approaches and equipment. One of the most important challenges of recording a pipe organ’s individual sounds is that the instrument - usually - cannot be taken into a recording space with preferable environmental conditions; therefore measurements are conducted on-site, often under undesired sonic conditions.

Dry, wet, semi-dry

Different approaches can be employed to record pipe organ sounds for various purposes, including musical recordings, creation of digital organs or virtual instruments, parameter analysis, etc. These approaches involve a number of subjective and intuitive properties (such as in terms of equipment selection or placement), but they also differ objectively, most importantly in the amount of desired reverberant sound in the recoded data.

Commercial recordings (CD/DVD) usually aim at an acoustic impression having a large apparent source width (ASW) and great listener envelopment (LE), which we can call the ‘wet approach’. These recordings often use spaced microphone techniques, and yield a larger stereo or surround image than what can be experienced in the real space. The result is a sound coming from subjectively everywhere, has little to do with a reproduction approach or what humans actually hear, and they often produce a sound which is sometimes called 'muddy'; yet we are very much used to this kind of sound and such an exaggeration of spatial properties can work well in home environments creating an 'impression'. Recordings for digital organs, however, are often aiming at the other end, preferring to minimize the reverberant sound in order to be compatible with reverberant on-site installations in halls or worship spaces, and also to allow a high signal-noise ratio even if a single quiet pipe is speaking; this can be called the ‘dry approach’. The dry approach may also require entering the organ case and recording with a probe placed close to the speaking pipe. The majority of sample-based or sample-driven digital organs do not aim at reproducing a particular instrument's sound as a whole, with all its ranks, but they offer a mix of various pipe sounds originating from various recording sources and instruments. The real pipes, however, are adjusted or 'tuned' to the room where the organ is located, and without major adjustments, combining pipes originating from various instruments may be difficult or unpleasing, if subjective quality is of major importance. Tuning to the room is alone a form of art, and involves, among other processes, adjusting the physical characteristics of a pipe to modify is loudness, frequency characteristics and other properties matched to the acoustic conditions.

Influence of the organ case

Additional difficulties of the dry approach include over-simplification issues, such as neglecting some important radiation characteristics of the pipe and the pipe organ case itself, which is preferable to be simulated afterwards, if possible. Each pipe is located in a physically different position inside or at the facade of the instrument. The pipes usually have very different acoustic parameters such as directivity, spectrum, transient response, etc. and it is not rare to see loose coupling between them, too. The radiation pattern of a pipe and the pipe organ case and its inner obstacles is complex and depends on small and large-scale structural properties as well, and is highly frequency and distance-dependent. To show the importance of the organ case and its acoustic complexities, one can try placing a small loudspeaker deep into the pipe organ case and play back some organ sound from it and listen from outside the case. The resulting sound can sound very similar to as if coming from a real organ itself, given the sound pressure levels and organ type in the recording closely matches the instrument. In an acoustic source, directivity is very much influenced by the physical shape of the sound source and its neighborhood, and in a pipe organ, the resulting directivity of each pipe is affected by the placement relative to the other pipes (as obstacles), too. These effects are quite complicated to model and take into account adequately limiting the applicability of the dry recording technique as a reproduction-type approach. Even with per-pipe convolution, the complex effects of the organ case cannot be easily reproduced, since a sound source with similar directivity as the pipe would be required to measure the impulse response for each pipe. A dual-channel or multi-channel recording technique with the pipe being the sound source is maybe possible to realize, although far from being straightforward, again due to directivity issues.

The semi-dry idea

The main idea of the semi-dry approach introduced by IA is to embed everything in the sound that is difficult or unfeasible to be modeled, and exclude or minimize the influence of those factors that are relatively easy to model and control. Additional reverberation, for example, can be relatively easily modeled using multi-source stereo or surround convolution reverberation; therefore, the reverberation in the sound is preferable to be at a controlled, smaller overall level in the recorded sound. The organ case, however, is too complex to model or to measure, because dismantling the organ is sometimes impossible and the lack of appropriate sound source is a major limiting factor. Therefore, the semi-dry approach aims at recording exterior of the organ case, which guarantees that all its effects will be embedded in the recording, while recording at a position where the direct sound level is still dominating the reverberant energy; yet at a distance that is large enough to be able to model the organ as a concentrated sound source by the sensing array (of microphones). This optimization criterion is sometimes difficult to achieve, and may require excessive prior acoustical measurements and calculations, but it allows for later convolution reverberation with a realistic and subjectively preferable result. Inspired Acoustics believes that semi-dry recording is more future-proof than other techniques in a sense that is it forward compatible with emering reverberation and simulation techniques.