All the files in in this comparative set are 'raw': no EQ or any other processing or effects, other than identical normalisation for each file, and then identical gentle gain envelopes to balance background and foreground sound levels. Essentially, I wanted to let the microphones and the arrays 'speak for themselves'.

Although in one sense, this a kind of 'technical' field recording study or experiment, in another sense, it is also a kind of aesthetic work in itself. You can listen to it out of 'technical' and practical interest; or, you can listen to it aesthetically as a pure sound experience, or as a deep meditative listening experience. Or all of the above, simultaneously.

The motivation for these comparative parallels was that, rather belatedly, I found out that it was possible, after all, to record 32 bit float isolated tracks with a MixPre. I had assumed it wasn't, because I'd tried splitting 32 bit files using Sound Device's (free) application Wave Agent, which doesn't support 32 bit files: the result was just 'static'. I'd also tried using iZotope's RX and Steinberg's WaveLab, both of which can open Poly files, but can't split or pair the isolated tracks. So, I'd been doing all my multitrack recording at 24 bit, and splitting and pairing the Poly files with Wave Agent.

But it just seemed far too implausible that 32 bit float ISO files weren't possible... Eventually, I dug a bit deeper: first, I came across AudioMove, a (free) application that can split 32 bit Poly files into separate 32 bit single tracks; but the split single tracks then have to be manually assembled into stereo pairs with another program (I used WaveLab). Then, I found out that Audacity (also a free program) can open Poly files; and it is very quick and easy to pair individual tracks into stereo pairs and export them as 32 bit stereo wav files. It takes less than a minute.

Consequently, one of the first tasks for which I wanted to use 32 bit float multitrack files was (at last) to make some comparative parallel recordings of trains (specifically, the Indian Pacific transcontinental passenger train, and freight trains).

The parallel recordings in each track are presented in this order: AB 63 cm (omnidirectional microphones in parallel), ORTF (cardioid microphones with capsules 17 cm apart at a 110 degree angle), Binaural (omnidirectional microphones at a 180 degree angle, i.e., pointing in opposite directions, set into silicon human-type ears, about 15 cm apart, with a partial baffle between them). More exactly, I should say 'semi-binaural', as the baffle is not a representation of a human head, either in size or in shape. I sometimes insert a small oval-shaped cushion as a baffle between the microphones, which is more similar in size and shape to a head, but I wanted to avoid doing this for the present recordings.

Microphone and Microphone Array Parallel Comparisons: Trains

The Indian Pacific is an Australian passenger rail service that operates between Sydney, on the Pacific Ocean, and Perth, on the Indian Ocean, with Adelaide, on the Southern Ocean, as a midway station: a journey of 4,352 kilometres (2,704 miles) and four to five days. This recording was made at night, at around 23:30, in the countryside, one to two hours out of Adelaide: here, the Indian Pacific is en route to Perth. Traversing a long stretch of straight track, the train was travelling at a fairly brisk pace.

The recordings were made with microphones that I've been using for years: a pair of MKH 8020 omni microphones in AB 63 cm (i.e., parallel to each other and spaced 63 cm apart) and a pair of MKH 8040 cardioid microphones in ORTF (i.e., divergent at a 110° angle and 17 cm apart), all mounted on the same mic support bar, at a distance of 9 steps (approx. 5.5 - 6 m) from the railway tracks, simultaneously recorded the passing of the train into a MixPre-6 II. The preamp gain for the 8020s was 35 dB, and for the 8040s it was 40 dB. Even so, the signal level of the 8040s was far below that of the 8020s. Not surprising, given the omni polar pattern of the 8020s, and their greater sensitivity.

The comparative sensitivity of these microphones is as follows (according to the manufacturer's specifications): MKH 8020s -30 dBV/Pa, MKH 8040s -34 dBV/Pa. Although fairly close, the very different polar patterns (omni vs cardioid) of the two microphones make this small difference in sensitivity quite significant. ['Sensitivity' refers to the amount of output that a microphone provides for a given input. Microphone sensitivity is typically measured with a 1 kHz sine wave at a 94 dB sound pressure level (SPL), or 1 pascal (Pa) pressure. Thus, a -30 dBV/Pa output is a more 'sensitive' (and 'louder') response than a -34 dBV/Pa output.]

No processing has been used on the samples presented, other than the following simple adjustments: both samples were normalised to -1.00 dB, and then precisely the same gain envelope adjustments were made to both samples to rebalance the overall temporal gain levels after normalisation. The original files are 96 kHz/32-bit float; the resampled version presented here is 44.1 kHz/16 bit.

The most evident and most significant point of difference between the two recordings, from the recordist's perspective, is that between the stereo imaging/spatial staging of the same event as recorded from the same position. Of course, it would be quite meaningless to suggest that one image is 'more true' or 'more faithful' than the other. There are also some more subtle differences in the timbral detail and texture of the two recordings, due to the subtly different frequency response contours and polar patterns of the two types of microphone. It would be very interesting to make further comparisons with reference to a parallel binaural recording rig.

Track image: Comparative spectrograms of both clips at 96/32. Drone aerial photograph of Indian Pacific from https://www.journeybeyondrail.com.au/indian-pacific/.

2025.01.30 Indian Pacific Transcontinental Train: MKH 8020s and 8040s Parallel Samples.

2025.02.10 Mic Array Parallels: Freight Train (2).

2025.02.10 Mic Array Parallels: Freight Train (1).

2025.02.11 Mic Array Parallels: Indian Pacific Transcontinental Passenger Train.

𝔈𝔵𝔭𝔩𝔬𝔯𝔦𝔫𝔤 ℌ𝔢𝔞𝔯𝔦𝔫𝔤, 𝔱𝔥𝔢 ℌ𝔢𝔞𝔯𝔞𝔟𝔩𝔢, 𝔞𝔫𝔡 𝔱𝔥𝔢 ℌ𝔢𝔞𝔯𝔢𝔯, 𝔞𝔰 𝔞𝔫 𝔈𝔳𝔢𝔫𝔱, 𝔞 𝔐𝔬𝔡𝔢, 𝔞𝔫𝔡 𝔞 𝔇𝔦𝔪𝔢𝔫𝔰𝔦𝔬𝔫 𝔬𝔣 𝔅𝔢𝔦𝔫𝔤.

2025.02.10 Mic Array Parallels: Freight Train (2).

𝕳𝖊𝖆𝖗𝖎𝖓𝖌 𝕭𝖊𝖎𝖓𝖌𝖘