The Maryhill Monument Replica of Stonehenge
Models are often used within acoustics in order to conduct acoustic testing before a building is constructed, or when work needs to be carried out in a laboratory. Scale models can be problematic as they scale down acoustic effects, and other factors also have to be scaled, such as absorption by materials. We became aware of a full size scale model of Stonehenge in Maryhill, Washington State, USA. It was based on archaeological plans, and completed in 1926, built as a war memorial. Although it was made of concrete, it was the most accurate model available, and it was decided to travel to Maryhill to carry out field tests to try to find further evidence for the acoustic effects we found through theoretical analysis.
[Youtube movie clip of Maryhill]
The monument was made of concrete that had been polluted by salt water in a flood, and was quite porous. The shapes were all more regular than the original. It had the large central altar stone lying down rather than standing up. It also had surfaces that were deliberately made to look rough in order to make the site look ‘old’. All of this would minimise acoustic effects. We were happy that this was a better situation than acoustic effects being exaggerated, as any results would be conservative. It was far more accurate than any model we would be able to build ourselves. Concrete is still a hard reflective surface and would act similar to stone, in particular in terms of patterns of reflection, echoes, reverberation and acoustic response. This was later supported by digital modelling results. The level and intensity of effects would be reduced, but their nature should be the same. The replica stone circle itself was reasonably accurate, but the space outside was very different. The Heel Stone was closer than it should have been to the monument because it is sited next to the Columbia River Gorge which drops away dramatically to the side of the monument. We therefore focused on measurements inside the circle. No archaeologists had ever used the site to test theories about Stonehenge in any way, and the Maryhill Museum who manage the site, were generous in allowing us to carry out acoustic field tests.
To see a ground plan of Maryhill click HERE
For these tests we used WinMLS acoustic testing software, a dodecahedronic omnidirectional loudspeaker, a large sub bass speaker, a B&K small diaphragm condenser microphone for critical measurements, a Soundfield ambisonic 3D microphone for immersive and subjective recordings, a decibel meter, and a laptop with digital audio workstation, and professional soundcard for recording. We took measurements using a calibrated loudspeaker system playing slow sine sweeps in various source positions. The main source positions were in the centre of the space, in front of the largest trilithon, and on the right hand side of the space (when facing the largest trilithon). We used a large number of receiver microphone positions for the measurements. We took measurements every metre in a straight line measured on an axis from the centre of the largest trilithon through the middle of the space out of the entrance and towards the Heel Stone. This repeated some of Watson’s measurements, we were hoping to confirm some of his findings. We also made measurements across the circle, from side to side, at a 90 degree angle to the other axis of measurements.
Figure 2: The Maryhill Monument, Maryhill, Washington State, USA (author’s photo).
We also played a number of audio source files into the space and subjectively evaluated how they sounded. We had the opportunity to listen to and record a flautist, dijeridoo player and actor in various positions and while moving around, and to evaluate the effectiveness of speech and musical sound. We experimented with using sine and square waves at single frequencies, slowly swept upwards manually, to investigate resonances within the space. We also used rhythmic pulses and waveforms at various tempi and frequencies, to explore frequencies of 1 – 20Hz. We measured variations in loudness caused by resonance by making recordings and using a handheld decibel meter. We explored echoes by using impulses, making short sharp sounds so that we could hear, record and evaluate echoes.
Our measurements found a reverberation of about 1.5s (T30 = 1.5s). Early Decay Time (EDT) was about 0.8s. There was quite a large amount of background noise, around 30db, and so figures below this level were unreliable. There was a considerable amount of wind noise, of which unfortunately we took little notice, especially considering the information on Hardy and wind noise placed discussed here already, but discovered in reality after the Maryhill trip. The wind noise was very loud in the space however, especially as the day continued, and it made measurements of any sort impossible eventually. This was largely due to the positioning of the monument next to a river gorge, but in retrospect the wind seemed louder inside the monument than outside, and it is possible that Hardy was right, and that the sound of the wind whistling through the stones was (and is) an important feature of Stonehenge.
The reverberation time is shown in figure 3 below. The blue Schroeder curve shows stepping, undulations up and down, which is an indication of the presence of resonances and/or echoes. We were able to discover and map variations in loudness, and to create powerful standing wave resonances using both single low frequencies and musical percussion rhythms between 0.5 and 15Hz. This replicated the results of Watson’s pilot study.
The Impulse response of the space shows a number of reflections and echoes, although these were particularly localised.
Figure 3: Reverberation time at Maryhill.
We found in fact that one could hear clear echoes when one stood in front of the largest trilithon and could see the Maryhill equivalent of the Slaughter or Heel stone. This lies in a straight line between what would be the ceremonial approach to the site, the ‘entrance’, and the centre of the circle. Of course there are no station stones or other stones outside the circle at Maryhill. I found later that at Stonehenge that there are similar echoes that come from the Station stones, and that the main echo in the site now comes from the Heel Stone. It seemed that the external stones worked as echo stones, catching sound and projecting it back into the circle. The placement of these was vital, and carefully aligned with acoustic and visual sightlines.
Subjectively, there was a powerful sense of envelopment and inclusion within the concrete circle. One felt ‘inside’, separated from the outside world. Sounds from outside seemed distant and disconnected. Within the circle musicians and the actor immediately stood on the ‘altar stone’, in the centre. Perhaps rather than standing, this was a prone ‘stage stone’. They noticed that the centre was the most flattering position to play music. Recordings of music were also pleasantly flattered within the space, rather than being disrupted or blurred by the acoustic of the space. We played recordings of traditional music from various world traditions that featured voice, percussion and clapping, in order to simulate possible group musical activities in the space. A pleasant reverberation was added to them, without any severe muddying or colouring of the sound.
Figure 4: Impulse Response of Maryhill
Speech was very easy to hear, intelligibility seemed (subjectively) to be good. One could talk to another person on the other side of the space quite easily, even if they were behind a stone. When walking behind a stone, speech bounced off other surfaces and was still audible, the stones did not seem to reduce the level of sound. This allowed one to stand behind a stone and make it appear the stone was speaking. Speech was diffuse and intelligible in the space, and clarity was good.
Being exactly in the centre of the space amplified perceived sound, and this was definitely a position with special acoustic properties. When entering the circle, standing underneath the lintels of the circle had a very pronounced effect. It felt as one entered the circle as though one was going through an acoustic door. If one stopped under the ring of lintels, with an upright stone each side, there was a different, enclosed, resonant acoustic. However this acoustic was different enough to either the inside or the outside to enhance the process of crossing the threshold from outside the circle to the inside. It felt as if one was entering another acoustic world. Watson speaks in some detail about this effect at Stonehenge itself. Some other positions had very different acoustic properties. The space between the bluestone circle and outer circle was quieter than other positions, whereas being right next to the outer circle was louder.
Occasionally when walking around the space, one would hear a word seem to leap out of the air, or there would suddenly be a strange acoustic effect. By being in a specific position while someone else was standing in another particular place, talking in a specific direction, the acoustic would suddenly come alive. This proved impossible to predict or repeat, but repeatedly one would suddenly hear a strange effect applied to on one’s own or someone else’s voice. It was high frequencies that were affected in particular. These effects would have been stronger in the original site where the stones were more reflective, shaped to be smoother, and curved in a way that would focus sound.
However we also found remarkable low frequency results. We were able to stimulate standing waves, resonances, so strong that sound was equally loud on each side of the circle, rather than being louder by the loudspeaker at the circle’s edge that was making the sound. We made the space resonate like a wine glass or bottle using individual frequencies generated from the computer. A short repeating drum sound, played at the correct tempo at high volume by loudspeakers, was also able to stimulate a sympathetic vibration in the space, and as one walked around, the short attack of the drum was made longer, changed into a low bass throb, more like a synthesizer than a drum, getting louder and softer as one walked across the space. Low frequency boost, patterns of loud and soft and softening of attack are all typical effects when modal acoustic resonance is present.
This video shows a red dot (me) walking around the space, and the sound changing dramatically. The source sound is a bass drum beat at a tempo equivalent to the resonant frequency of the space. The sound seems to change and shift as one moves around. When we returned to the site for the History Channel documentary we were able to get the site to resonate using only the sounds of small hand drums based on prehistoric European designs, as long as the tempo was right. The correct tempo could be found by listening to echoes in the space, or by maths using a modal predictor built in Matlab.
This video is an audio recording of me walking through the Maryhill site with a microphone. The changes in sound are due to the acoustics of the space. The red dot illustrates where I am standing. When I am on the far left you will hear the simple bass drum sound direct from the speaker, a TR808 sample. This is where the loudspeaker was placed.
Figure 5: Resonance at Maryhill showing the volume level increasing as one gets to the edge of the circle
A graph of volume is superimposed on top of an image of the Maryhill Stonehenge. One can see for example the correlation of the position of the ‘bluestone’ horseshoe shape with an increase in volume. These and other readings were used to show some of the circular modes of vibration in the space. Sound does not resonate in a circular space like a string, in a straight line, but in circles, more like a cylinder and/or drumskin. In figure 6 actual volume levels are shown in red. These seem to be a mixture of modes 4 and 10, (shown by yellow and green lines). There are similarities between the red and green lines, except in the centre where red and yellow become more similar. By drawing blue circles lined up with the peaks of the green line, one can see a clear correlation with the positions of the stones.
Figure 6: Real and theoretical modes at Maryhill.
These circles point out how the positions of the circles and horseshoes of stones at Maryhill (and originally at Stonehenge) act to force particular resonant behaviour. Analysis of results is ongoing, and detailed results from Maryhill will be part of a future paper focusing further and specifically on the work at Maryhill. There were many other interesting findings, such as observing the movement of shadows within the space as the sun rose, and the nature of the relationships to the monument of local people and visitors that had developed. Indeed the Maryhill Monument is interesting to study in itself. It has acted much like an experimental archaeology living model of Stonehenge project that has been running for the past 83 years. If nothing else, the research visit to Maryhill provided an interesting phenomenological, experiential, partial understanding of what it might have been like to travel to and experience Stonehenge.
To summarise results at Maryhill there were a number of acoustic effects present. We measured a reverberation time T30 of about 1.5s and early decay time of about 0.8s. It was possible to stimulate the modes of resonance in the space using single frequencies and rhythms. These caused highly localised differences in volume, and unusual sonic effects. Low frequencies in the space were very interesting and require further study. Wind noise was very loud in the space. The results of Watson’s pilot study were supported, in terms of changes in sound when approaching and entering the space, and variation in volume in the space. The axis of the monument that corresponded to the ceremonial approach to the space (and the break in the bank) at Stonehenge was of particular interest, and sound effects seemed to be focused in that direction. We also discovered effects at 90 degrees to this axis, going across the space between the two pairs of equal sized trilithons. Underneath the lintels of the outer stone circle had a particular effect, creating an acoustic doorway and threshold, and enhancing the effect of entering the space. The acoustic outside the stone circle was different to that inside, and there was a subjective sense of aural envelopment, enclosure, and separation from outside when within the space, that grew as one spent more time there.
The central area, contained by the horseshoes of trilithons and bluestones, from the altar stone to the ‘entrance’, had a different acoustic to the rest of the space, and seemed most sonically active, had the clearest acoustic effects. There was an increase in volume when close to and inside stones of the outer circle, and when close to and in front of the trilithons. There were a number of acoustic dead spots where sound was quieter, in particular halfway between the outer ‘sarsen’ circle and the ‘bluestone’ circle inside, and also behind the middle sized trilithons. There was a clear front and back to the space. The Altar stone if laid flat would have made a good stage, for projecting sound as well as aiding sight lines. Speech intelligibility was subjectively assessed as good across the space, and standing behind stones did not seem to mask the voice. The reverberation provided a pleasant, flattering acoustical quality for both speech and various kinds of musical material. There were echoes in the space, only heard when in front of the largest trilithon, and looking in a line out of the space towards where the Heel Stone would be. The echo was at its strongest at each end of a line connecting these two points.
Above are drum modes: not the same as those in air, but a useful illustration
Above: The Slaughter Stone is a similar distance from the edge of the circle as the circumference of the circle. This would enhance acoustic effects.
Above are illustrations of reflections of sound within Stonehenge and modal vibration on the vertical plane
 Maryhill Museum of Art, Maryhill Museum of Art, available at http://www.maryhillmuseum.org/about.html, accessed 14/08/08.
 This projects sound in all directions at equal levels, whereas most loudspeakers point in a particular direction.
 Measurement omnidirectional loudspeakers usually have limited low frequency characteristics, the sub-bass speaker allowed us to explore low frequencies accurately.
 We used square waves because they were easier to hear at low frequencies than sine waves.
 Aaron Watson, 2006.
 We were there shortly after the solstice. The space is aligned astronomically in as similar a way as possible to Stonehenge, considering the difference of geography. It is aligned to the passage of the sun at the solstices, as well as to various sunrises and sunsets and the rising and setting of the moon.