Extra material for a paper by Steve Marshall in
TIME & MIND 2016, VOL 9, NO. 1

Steve Marshall's paper in T&M (above) includes references to a series of recordings made in the West Kennet long barrow (WKLB) and analysed with Audacity, an open source sound editing programme. The spectragrams used for analysis and quoted in the paper can be accessed below, but some readers may wish to install Audacity and make their own spectrograms. DOWNLOAD AUDACITY

The sound recordings were made with a Zoom H-1 solid-state recorder with built-in microphones. Though it is a high quality instrument, the H-1's frequency response falls within the standard 20 Hz to 20,000 Hz range. This raises several problems, since the WKLB has a resonance of about 9 Hz. Audacity's spectrum analyser works extremely well but, again, is not intended to work below 20 Hz. Since this is the range we are most interested in, can we find a solution?

One possible way around this is to treat the low, infrasonic resonance as a 'missing fundamental' and search for its upper harmonics instead. Audacity's noise reduction may also be used to clean up the lowest frequency range, revealing detail that is otherwise hidden. An account of experimenting with these techniques may be found here: AUDACITY EXPERIMENTS

Victor Reijs was extremely helpful throughout the whole process and we had many discussions by email on how to procede with analysing Audacity's spectrograms: WKLB Spectra VR


31 samples, taken from original recordings made inside the WKLB, may be downloaded or played as streaming audio from here: SOUNDCLOUD WKLB
All are labelled 'WKLB' and numbered from (1) to (34). Note that there is no WKLB (12) (13) or (14). The samples are:

Ten albums of pictures may be found on this Flickr account: SPECTROGRAMS
Each album contains a numbered sequence of screenshots showing how a particular recording has been processed through Audacity.

The sound samples and screenshots above are labelled using the same system of abbreviations.
The WKLB's central passage (Ctre Psge) points east to the forecourt; its five chambers are identified by compass directions:
NE Ch, NW Ch, SE Ch, SW Ch & W Ch. Where locations inside the barrow are identified, they are where the mic was sited.
Two sound sources were used for impulses - a balloon burst that was always done in the forecourt, and a paper banger (PB)
that was generally used next to the microphone in various parts of part of the structure.
Recordings were also made at a resonant spot in the passage, between the NE & SE chambers (Ctre Res)
Audacity's noise reduction was used in making most of the spectrograms, labelled NR.
Two different bull roarers were used, both whirled in the forecourt, and identified as BR1 & BR2.

The Flickr albums contain different numbers of screenshots; each album has the various stages of processing for just one sound sample. Hence, the albums with the least number of images should be examined first, as they are easiest to follow! Spectrograms of the five chambers were made first, using an impulse from a paper banger. The balloon burst was too loud to use next to the mic inside the chambers, but the quieter paper banger was useable.

The Flickr albums that hold most images show how several more stages of processing in Audacity can reveal more detail in the low frequencies. In three of the albums, 21.45 Hz and its harmonics (odd & even) have been removed with the noise reduction tool. The reasoning was that being a 'closed pipe' open at one end and 10m long, the central passage should resonate at approx 8 Hz, with no even harmonics -only odd harmonics. But the huge blocking stone no 45, which was added later, might effectively extend the central passage to 15m, producing a 'pipe closed at both ends'. The passage would now have a resonance of 21.45 Hz, with odd & even harmonics. Audacity can generate waveforms, so a sawtooth wave of 21.45 Hz was sampled, and removed as noise.

The central passage has a resonant frequency determined by measurement to be 8 Hz; there were indications though, that this was not the exact frequency. Audacity was used to generate a 9 Hz square wave, made only of odd harmonics. This was superimposesd on several other spectrograms to see whether peaks coincided. This process is illustrated and described in detail in the downloadable pdf 'AUDACITY EXPERIMENTS' (see above).

For reference:
WKLB Recordings made 19th Oct 2015
01 – BR1 W Ch – silence – wind noise
02 – BR1 W Ch
03 – BR1 W Ch
04 – Balloon W Ch
06 – PB in SE Ch
07 – PB in SE Ch shorter - 77, 86 Hz
08 – Voice
09 – PB in SW Ch – 112 Hz
10 – PB in W Ch
11 – PB in SE Ch
12 – PB in NE Ch – 77, 87 Hz
13 – PB in central passage resonant spot – 111 Hz
14 – Voice in central passage res
15 - PB in NW Ch – 113 Hz
BR = Bull Roarer, PB = Paper Banger

Steve Marshall 10-03-2016