An Acoustic Reconstruction of Sistine Chapel in Rome at the Time of Leo X: The Role of Tapestries

Abstract

The Sistine Chapel, originally designed to accommodate papal ceremonies, featured a system for hanging tapestries that ensured they were deployed according to the liturgical calendar. These textiles not only served as temporary decorative elements but also contributed to the acoustical environment. Historical records suggest that Renaissance popes, particularly Leo X, were attuned to the impact of textiles on sound, experimenting with their placement to optimize acoustics for sermons and polyphonic music. Given the lack of direct historical acoustical measurements, this study employs a computational simulation approach to model the chapel’s acoustics with and without the presence of tapestries and human occupancy. A crucial first step involved characterizing the absorption coefficients of surface finishings in order to obtain a reliable model of the space and investigate modifications induced by tapestries. The study revealed that the presence of tapestries reduced reverberation time at mid-frequencies from 7.4 s to 5.1 s in the empty chapel and from 4.1 s to 3.4 s when occupied. The results corroborate historical observations, who noted the effects of tapestries on vocal clarity in papal ceremonies. The findings demonstrate that textiles played a significant role in controlling acoustics within the Sistine Chapel, complementing the liturgical experience.

1. Introduction

The study of past soundscapes and the acoustics of historical spaces has long fascinated researchers across disciplines, from archaeology and architecture to musicology and digital humanities, leading to what is now commonly identified as “archaeoacoustics”. The core aspect of this research is the acknowledgment of the role of acoustics in the definition of the cultural heritage of the building or, more generally, of the site under investigation, despite its “intangible” nature [1,2,3,4].

Archaeoacoustics, as an interdisciplinary domain, integrates principles from acoustical engineering, virtual reconstruction, and heritage studies to revive lost sonic environments. The field has evolved significantly in recent decades, benefiting from advancements in digital modeling and immersive technologies that allow researchers to simulate the acoustic behavior of a number of past or vanished structures [5,6,7,8,9,10]. One of the primary goals of such studies is to evaluate how architecture influenced auditory experiences, whether in the grand halls of ancient palaces, the reverberant interiors of sacred temples, or performance spaces designed for specific acoustic effects, as well as pointing out the interplay of other social and cultural factors.

The literature on acoustic reconstruction of spaces reveals a variety of methodologies employed to approximate the sound properties, the major concern being represented by the significant uncertainties that may characterize the spaces that are being reconstructed. The use of textual descriptions, artistic representations, and analogies with existing structures to infer possible acoustic behaviors are among the most frequent strategies. However, given the difficulties often associated with creating and calibrating acoustic models of existing spaces [11], it has been proposed by several researchers to adopt rigorous practices in order to create and validate virtual acoustic models of non-existing spaces [12], while collecting the greatest possible amount of acoustic data about materials and objects that could contribute to improved accuracy in simulations [13,14,15].

Thus, even though advances in immersive technologies as virtual and augmented reality hold the potential to further enhance the accuracy and experiential quality of reconstructions [4], when dealing with archaeoacoustic reconstructions, it is important to take results “cum grano salis”, remembering the limitations related to the absence of precise architectural data, use of different materials (or in a different conservation state), etc. Additionally, the subjective nature of human auditory perception means that reconstructed soundscapes, though grounded in scientific principles, still involve interpretive elements. The role of cultural context further complicates the analysis, as the meaning and perception of sound vary across societies and historical periods.

In recent years, research on the relationship between sound, architecture, and ephemeral installations in the Renaissance has gained increasing attention. Studies such as Sound and Space in Renaissance Venice by Howard and Moretti [16] paved the way to other studies on the relationship between art, architecture and sound [17,18] that highlighted the significance of sound in the perception of artistic and architectural spaces as well as the positive effect of tapestries in large reverberant churches [19]. Within this context, our study focuses on the role of tapestries in the Sistine Chapel, not only in terms of their visual and decorative impact but also in their acoustic influence, starting from a calibrated acoustic model of the Chapel in its current state and simulating the effects of modifications by means of geometrical acoustic modeling. The aim of this paper is to understand whether such modifications induced audible differences that might have stimulated a discussion on their acoustic role in the space.

2. Historical Outline

The Sistine Chapel is the result of the restoration of the old Cappella Magna (or “Maggiore”) that Pope Sixtus IV (from whom it derived its name) promoted between 1473 and 1481 [20]. The Chapel was conceived from the outset as a highly decorative space intended for major liturgical celebrations and functionary activities of the Pope and his household (the Papal Chapel or Papal Court), in addition to the Papal Conclave to elect the new Pope. At the time of Sixtus IV the Papal Court comprised about 200 people, including clerics, officials of the Vatican and distinguished laity. The official gatherings of the Court were distributed throughout the year according to the Papal Calendar, and among the 50 occasions prescribed, about half of them were masses that could be held in a smaller, less public space, for which the Cappella Maggiore was used before it was rebuilt as the Sistine Chapel.

The earliest decorative campaign, executed between 1481 and 1482, involved a team of leading Florentine and Umbrian painters of the time, including Sandro Botticelli, Domenico Ghirlandaio, Pietro Perugino, and Cosimo Rosselli. These artists, working in a coordinated yet stylistically diverse manner, created the lateral wall frescoes depicting scenes from the lives of Moses and Christ, thereby establishing a visual parallel between the Old and New Testaments as a legitimizing framework for papal authority.

The chapel’s decorative scheme underwent its most transformative phase in the early 16th century under Pope Julius II, who, seeking to assert the spiritual and cultural supremacy of the papacy, commissioned Michelangelo Buonarroti to repaint the ceiling. From 1508 to 1512, Michelangelo worked largely alone, replacing the original star-studded ceiling (Figure 1a) with a monumental fresco cycle that redefined the scale, complexity, and theological depth of Renaissance art.

A subsequent commission in the 1530s, initiated by Pope Clement VII and completed under Pope Paul III, saw Michelangelo return to the chapel to paint The Last Judgment on the altar wall. Executed between 1536 and 1541, this colossal composition reflected the tumultuous spiritual climate of the Counter-Reformation and Michelangelo’s own evolving religious sensibility.

The Sistine Chapel also included sculpted architectural elements. The screen or cancellata made of marble by Mino da Fiesole, Andrea Bregno, and Giovanni Dalmata was intended, just like the jubé typically found in late medieval churches, and like the iconostasis of the orthodox churches, to divide the chapel into two parts, one for the members of the Papal Chapel close to the presbytery area and the other for the worshippers. The position of the cancellata, originally dividing the space into two halves, was later moved by 5 m to accommodate the increasing number of people belonging to the Papal Court. The sculptors of the cancellata also provided the cantoria or projecting choir gallery (Figure 1b).

In fact, the history of this space is strictly interwoven with that of the Sistine Chapel Choir (Cappella Musicale Pontificia), which has historically played a central role in its liturgical function. Established long before the chapel’s construction, the choir gained particular prominence under Sixtus IV, who reorganized it as a permanent papal institution (making it its personal choir). The choir became a prominent musical institution and some of the most influential composers, like Josquin des Prez (from 1486 to 1494), Giovanni Pierluigi da Palestrina (around 1555), and Gregorio Allegri (from 1629 to 1652), worked and composed for the choir, contributing to the sacred and ceremonial identity of the space. As a consequence, it is likely that special attention was given to the acoustics of the Chapel.

An aspect that has only recently been considered in depth is the use of tapestries, which were hung according to the liturgical calendar to adapt the space to different celebrations. Pope Leo X was one of the principal proponents of this practice, commissioning Raphael to design a cycle of tapestries depicting scenes from the lives of Saint Peter and Saint Paul [21]. These were woven in Brussels by Pieter van Aelst, one of the most esteemed tapestry makers of the period and first displayed in 1519. Tapestries in the Sistine Chapel served as an ephemeral “skin” for the building, altering not only its visual appearance but also the spatial perception of those within it. The installations were carefully scheduled for specific occasions. During Holy Week, for instance, the chapel was stripped of all textile decorations, emphasizing its stark solemnity. In contrast, on significant feast days such as Christmas Eve in 1513, golden tapestries from Saint Peter’s Basilica were brought in to enhance the space’s magnificence.

Beyond their decorative purpose, tapestries played an essential role in the chapel’s acoustic environment. The architectural features of the Sistine Chapel, characterized by its elongated shape and high vaulted ceiling, caused a long reverberation time that can obscure speech intelligibility and affect polyphonic music performances. The issue was well known in Leo X’s time, and various configurations were explored to improve sound quality within the sacred space, also in consideration of the introduction of polyphonic music in an increasing number of sacred celebrations, as decreed by Leo X soon after his election [22].

One of the members of the Pope’s court, Evangelista Tarascono, suggested removing the tapestries to allow sound to reflect more clearly off the bare walls, making it “sharper and more pleasing” [18]. However, as a demonstration that excessively long reverberation negatively affects clarity of singing, tapestries continued to be used, often indulging in richer versions including golden and silver threads.

3. Materials and Methods

3.1. Room Characteristics

The room is a simple rectangular space about 40 m long and 13.5 m wide, with an overall height of 20 m. The floor, finished in Cosmatesque style (using geometric decorative stonework), is flat and is only raised in the presbytery area to which it is connected by four continuous steps. The space is subdivided into six bays, each one having windows on both sides that open about 10 m from the floor. A flattened barrel vault covers the ceiling. The cantoria is 3.6 m above the ground floor and located in the third bay from the west. Its marble balustrade projects 1 m from the wall. All the surfaces are, as described before, frescoed with architectural elements like corniches and pilasters interrupting the large flat surfaces. A careful analysis of the surfaces was possible thanks to a high-definition 3D virtual tour available on the website of the Vatican [23].

Considering that, along the time, the only significant changes in the room’s configuration were represented by the displacement of the cancellata and by maintenance of the frescoes, in order to reconstruct the acoustical conditions at the time of Leo X, the interior space was first modeled using the current configuration under empty conditions and then modified moving the cancellata back to its original position, halfway into the room, and adding occupancy and tapestries. To this purpose, a famous engraving by Étienne Dupérac (Figure 2) was taken as a reference for the way the interior space was used during ceremonies. The Chapel is represented during a solemn Papal service. The attendees were seated or standing on the foreground of the figure, (i.e., the side of the cancellata closer to the entrance), while on the side closer to the altar, the Pope was seated on his throne, the cardinals and other members of the curia on benches, Roman nobles and officials on the steps, and the choir standing at the cantoria. It is known from the records that the choreography of the ceremony dictated each person’s movements and position in the Chapel, and that Leo X requested “stationing officials” to hush those who talked amongst themselves in order to reduce unwanted ambient noise during the Mass.

3.2. Acoustical Modeling

As mentioned earlier, the acoustic model of the Sistine Chapel was developed in two steps. First the current state under empty conditions was used to fine tune the acoustical parameters characterizing materials and surfaces, and then, starting from the above descriptions and other historical records, the Leo X configuration was modeled. For the current conditions, drawings and photographs, including the virtual tour available on the Vatican website [23], provided most of the supporting information. Acoustic modeling based on geometrical acoustics (GA) typically involves a trade-off between geometrical detail and simulation accuracy. Thus, in the creation of acoustic models, small details were neglected as well as small differences in the dimensions of the different sides of the chapel that made it non-perfectly rectangular. The resulting geometrical model (Figure 3) was made of 380 surfaces developing an area of 3236 m² and a volume of about 9500 m³.

The subsequent step was the characterization of the surfaces in terms of sound absorption and scattering coefficients. The largest surface was that of walls and vaults covered by frescoes, followed by the floor that was finished in marble in Cosmatesque style. Windows and wooden doors completed the relatively short list of materials. In the first stages, for each of these materials’ sound absorption coefficients were derived from the relevant literature [24,25] (Table 1). However, these values, based on “typical” material properties, were then slightly adjusted to reflect the acoustic influence of decorations and other site-specific features, ensuring a reliable model.

In order to perform such adjustments, onsite measurements of at least reverberation time are a necessary complement. In this case, given the impossibility to directly access the Chapel, the publication [26] of a set of measurements of reverberation time carried out in 1976 by Prof. Fricke was taken as a guide to calibrate the model under unoccupied conditions (Table 2). Unfortunately, at the time of the measurements, no environmental data were collected, and consequently temperature and relative humidity were estimated as 25 °C and 60%, values which represent the upper limits recommended for frescoes conservation and were in better agreement with the measured reverberation time. In fact, as shown in Table 1, the high-frequency absorption coefficients were reduced in order to match measured reverberation, which suggested that air absorption also had to be kept at a minimum, which, according to ISO 9613-1 standard [27], takes place at higher temperatures and higher relative humidity. However, considering that frescoes conservation requires strict control of environmental conditions and that the difference in air absorption coefficients at 25 °C and 30 °C is small, it seems more realistic to use the above-mentioned values as a reference. Anyway, as a consequence of this approximation, a slightly larger tolerance in the high frequency range was adopted. It is important to point out that apart from frescoes restoration, no acoustically relevant modification has been made from 1976 on, as most of the subsequent works dealt with air conditioning and lighting systems.

All the simulations were carried out using the commercial software CATT-Acoustic 9.1g, with the TUCT v.2.0g engine, which relies on geometrical acoustic simulation. Simulations were run using the “Algorithm 1”, which uses a randomized cone-tracing that switches to randomized ray-tracing when the expanding receiver sphere touches a surface and is suitable for closed and proportionate (mixing) spaces. The number of rays to send was chosen following the recommendations of the software, according to the purpose of the simulation. Thus, in order to obtain high-quality impulse response to be used also for auralizations (e.g., the aural renderings of the room arrangement), the number of rays varied between 7 × 10⁶ and 15 × 10⁶, with the highest number of rays used for the occupied configurations.

As the minimum difference perceivable by human subjects (known as just noticeable difference or JND), for reverberation time is estimated to be 5%, calibration procedures typically consist of changing the absorption and scattering coefficients until the spatially averaged values of measured and predicted reverberation times differ by less than that amount [11,12,13,14]. The adjusted absorption coefficients were given in Table 1, showing that minimal variations were needed, ensuring the coefficients did not deviate substantially from the literature values while matching the averaged reverberation time obtained from measurements. Given the relative lack of sculptural decorations, scattering coefficients, which are another important (although often underestimated in their acoustical importance), acoustic parameter that accounts for the diffusion of sound due to surface irregularities, were assumed to have a constant value of 0.05 for frescoes and other large flat surfaces. For the cancellata, the balustrade of the cantoria, and the pillars separating the successive bays, as well as other small protruding or recessed elements like side walls of windows, doors, and niches, the “auto-edge” option available in CATT, was selected. This option aims at taking into account that in GA, when a surface is small in relation to the wavelength, or is close to another one with different absorption, it cannot give a valid specular reflection, so the energy can instead be transferred to diffuse. Thus, according to the algorithm, such surfaces receive a scattering coefficient that is proportionally increased to account for the edge area (which is frequency dependent, as it is considered to extend up to a quarter of wavelength from the edge), which is assumed to be fully scattering. The upper half of the cancellata and the cantoria balustrade were also given a transparency coefficient of 0.8 to account for large perforations.

With the above assumptions, as shown in Table 2, the differences between measured and predicted average reverberation time were well below 5%, with the exception of the 4 kHz octave band where, as anticipated, the uncertainty about environmental conditions required an increased tolerance.

Table 1. Sound absorption coefficients used during the different stages of the simulation. Values in parentheses are those obtained before calibration.

	Area m2	125 Hz	250 Hz	500 Hz	1000 Hz	2000 Hz	4000 Hz
Frescoes 1	2380	(0.02) 0.042	(0.02) 0.048	(0.03) 0.05	(0.04) 0.06	(0.05) 0.045	(0.05) 0.03
Stone floor 1	620	(0.02) 0.04	(0.02) 0.04	(0.03) 0.04	(0.04) 0.05	0.05	(0.05) 0.04
Windows 1	100	0.30	0.20	0.14	0.10	0.05	0.05
Cancellata and balustrade	70	0.10	0.05	0.05	0.05	0.05	0.05
Wooden doors 1	30	0.14	0.10	0.06	0.08	0.10	0.10
Seated audience (1 p/m2) 2	198	0.22	0.32	0.58	0.89	0.91	0.91
Standing aud. (0.5 p/m2) 2	230	0.08	0.12	0.20	0.38	0.42	0.45
Tapestries 3	322	0.05	0.10	0.25	0.50	0.75	0.80

¹ After Meyer [24]; ² measured by Martellotta et al. [28]; ³ measured by Martellotta and Pon [14].

Table 1. Sound absorption coefficients used during the different stages of the simulation. Values in parentheses are those obtained before calibration.

	Area m2	125 Hz	250 Hz	500 Hz	1000 Hz	2000 Hz	4000 Hz
Frescoes 1	2380	(0.02) 0.042	(0.02) 0.048	(0.03) 0.05	(0.04) 0.06	(0.05) 0.045	(0.05) 0.03
Stone floor 1	620	(0.02) 0.04	(0.02) 0.04	(0.03) 0.04	(0.04) 0.05	0.05	(0.05) 0.04
Windows 1	100	0.30	0.20	0.14	0.10	0.05	0.05
Cancellata and balustrade	70	0.10	0.05	0.05	0.05	0.05	0.05
Wooden doors 1	30	0.14	0.10	0.06	0.08	0.10	0.10
Seated audience (1 p/m2) 2	198	0.22	0.32	0.58	0.89	0.91	0.91
Standing aud. (0.5 p/m2) 2	230	0.08	0.12	0.20	0.38	0.42	0.45
Tapestries 3	322	0.05	0.10	0.25	0.50	0.75	0.80

¹ After Meyer [24]; ² measured by Martellotta et al. [28]; ³ measured by Martellotta and Pon [14].

Table 2. Measured and simulated reverberation time as a function of octave band frequencies.

	125 Hz	250 Hz	500 Hz	1000 Hz	2000 Hz	4000 Hz
Measured	9.0	8.6	7.9	7.0	6.0	4.7
Simulated	9.1	8.6	7.9	6.9	5.9	4.5
Error (%)	1.1%	0%	0%	−1.5%	−1.7%	−4.3%

Table 2. Measured and simulated reverberation time as a function of octave band frequencies.

	125 Hz	250 Hz	500 Hz	1000 Hz	2000 Hz	4000 Hz
Measured	9.0	8.6	7.9	7.0	6.0	4.7
Simulated	9.1	8.6	7.9	6.9	5.9	4.5
Error (%)	1.1%	0%	0%	−1.5%	−1.7%	−4.3%

3.3. Sound Absorption of Tapestries and Occupancy

Finally, once the model of the Cappella under unoccupied conditions was completed, the effect of tapestries and occupancy as influencing factors could be investigated. With reference to tapestries the obvious choice was to simulate Raphael’s tapestries designed for the Cappella and commissioned by Leo X in the years between 1515 and 1519. The tapestries were hung on the lower part of the side walls beyond the cancellata (i.e., in the area reserved for the Pope), but historical evidence shows that the remaining walls were covered by tapestries as well. The tapestries had an average height of 3.85 m, so that they completely covered the wall from the floor to the first corniche (Figure 4), resulting in an overall area of 322 m². Raphael’s tapestries are well known for their richness and for a significant presence of threads wrapped in gold or silver. From the acoustical point of view, such detail presents interesting challenges as the metal is likely to increase both the surface mass and the flow resistance of the textile (because fibers become less permeable to air), as well as tortuosity, which may consequently result in an increased sound absorption according to theoretical models of porous materials [29]. However, given the difficulty of accessing the original Raphael tapestries, the Barberini tapestries, which the authors had the opportunity to acoustically characterize on site [14], were used as a reference. Although they were woven a century later—likely resulting in a tighter, less air-permeable weave, despite being made entirely of silk—it was assumed that using the measured coefficients would provide a more reliable estimate than relying on literature values or guessing.

It is important to point out that the distance of tapestries from wall is crucial to the determination of the sound absorbing behavior. Absorption coefficients of Barberini tapestries were measured when they were hung at an average distance of 4 cm from a supporting structure, and tapestries in Sistine Chapel were hung to the lower corniche that, contrary to the topmost part that has a remarkable edge projection, is mostly a beading that is well compatible with the above-mentioned distance. Finally, the absorption values that were used are given in Table 2 along with those of the other materials. The wall surface covered by tapestries was simply given the new absorption coefficients, therefore the actual area covered by frescoes reduces to 2058 m² when they are in place.

With reference to the second factor, in order to add the occupants in the Chapel, the engraving by Dupérac (discussed earlier, Figure 2) was taken as a reference. In the picture, the public part appears moderately filled by standing people, while in the private area occupants were seated along the perimetral stone benches, on the stairs, and on an extra row of benches on the right side. The same areas were identified in the model, with standing audience covering a surface of 230 m² in the public area, and seated audience covering a surface of about 190 m² in the private area. To this end, absorption coefficients for standing (assuming a density of 0.5 person/m²) and seated audience (assuming a density of 1 person/m²) were derived from the literature [28] and given in Table 2. For seated occupants that clearly belonged to the Papal court and were likely dressed in the typical renaissance liturgical vestments, richly decorated and characterized by several layers [30], the highest absorption coefficients corresponding to a thermal insulation of 1.3 clo [28] were used. As is customary in acoustical modeling, when no unusual situations are involved (e.g., very high absorption due to very dense audience [31]), they were assigned to the corresponding surfaces that were modeled as blocks (for audience members seated along the benches and stairs), while for standing audiences, additional extruded boxes were added in the public area.

Once the model was completed, and the characteristics of the surfaces defined, the following conditions were investigated: (1) empty chapel; (2) empty chapel with tapestries; (3) occupied chapel; and (4) occupied chapel with tapestries. Six receiver positions were considered for the analysis (Figure 3): one at the seat of the Pope (01), one at the center of the private area towards the cancellata (02), and one at the center of the public area (03), two extra receivers in the private area, at the side of the presbytery area (04, 05), and one extra receiver in the public area, close to the entrance (06). The sound source was located either at the right side of the presbytery area (P), where Duperac located the “Cardinalis celebrans”, and one at the cantoria balcony (C). In all cases, a simple omni-directional sound source was used, in agreement with standard requirements in room acoustics measurements [32]. The acoustical characteristics of the room were analyzed as a function of the influencing factors using well-known descriptors typically used in room acoustics [32,33]. In particular, reverberation time (T30) and early decay time (EDT) representing the sensation of reverberation in the space, strength (G) representing the relative loudness of sounds, clarity (C₈₀) representing the clarity of the musical message, lateral energy fraction (J_LF) representing the spaciousness of the perceived sound, and, finally, the speech transmission index (STI) representing the intelligibility of the spoken message. Although the IEC standard [33] recommends measuring it with a directional sound source, STI was calculated from the same measurements used for the other descriptors, based on an omni-directional source, to allow simpler comparisons among different configurations (as is typically done when characterizing the general room acoustics parameters), with a sound emission corresponding to “normal” voice spectrum, and assuming a background noise level of about 30 dB (A-weighted). For the sake of brevity, definitions of the above descriptors are given in Appendix A, and more extensive background can be easily found in standards and textbooks [34,35]. The reverberation descriptors are resented as spatially averaged values as a function of frequency, whereas the other parameters are presented as averages across the 500 Hz and 1000 Hz octave bands (referred to as “mid-frequency” and denoted as “mid” in the following) from individual positions. Finally, in order to evaluate if the differences between the analyzed configurations are aurally significant, in addition to JND of 5% for T30 and EDT, the subjective limens for the other parameters were taken from standards [32,33], and relevant literature [36,37]. In particular, the following values were used for the relevant parameters: 1 dB for G, 1.5 dB for C₈₀, 0.05 for J_LF, and 0.03 for STI.

4. Results

The results showed (Figure 5) that, as expected, the presence of the tapestries determined a significant variation under unoccupied conditions as mid-frequency reverberation time decreased from 7.4 s to 5.1 s, with a relative decrease of more than 30%. Under fully occupied conditions mid frequency reverberation time dropped from 4.1 s to 3.4 s, with a relative variation of about 20%, after application of the tapestries. Thus, in both cases, considering that the JND for reverberation time is 5%, variations could be potentially perceived by attentive listeners that used to attend ceremonies requiring different wall installations. At low frequencies, given the low sound absorption of the tapestries, variations in reverberation time were not so big (around 10%).

Thus, it can be concluded that the presence of tapestries compared to the unoccupied room introduced a clearly detectable variation in reverberation time and, given the frequency distribution of the added absorption, tapestries made it drier and amplified bass sounds. With a congregation in the room, variations due to tapestries were still audible, but less dramatic than they were in the first case. In terms of reverberation, the presence of the tapestries in the unoccupied room provided acoustical conditions that were more similar to those in the occupied room (as seen in Figure 5, where values were almost overlapped in the high-frequency range), likely making rehearsals and other more private celebrations more acoustically similar to events with increased occupancy.

In order to understand the effects of the tapestries and audience in different listening positions, the values for individual source–receiver positions are discussed. Figure 6 shows that for the selected parameters (EDT, G, C₈₀, J_LF, and STI), whose numerical values largely depend on the intensity and time distribution of the reflections, the variations due to different configurations mostly reflected the change in reverberation time, with a few interesting anomalies. Early decay time (EDT) followed an overall change that reflected the spatially averaged reverberation time variations appearing in the same frequency bands (500 and 1000 Hz), resulting from the different configurations already discussed in Figure 5. Despite EDT being a more position dependent parameter, in the present case, it showed only small point-by-point variations. Sound strength (G), reflecting the contribution that the room gives to the loudness of the sound, was highest in the unoccupied room across the considered conditions, in agreement with the theoretical behavior that assumes a logarithmic dependance on the ratio of the reverberation time to the room volume. Thus, if room volume remains the same, as in this case, G is expected to increase by 3 dB when reverberation time doubles. The effect of tapestries is evident when compared to the empty configuration, with the important effect of producing conditions that are more similar to the occupied space (as shown in Figure 6b, where lines tend to overlap). Occupation plus tapestries brings values closer to the optimal range (which is between 4 and 8 dB [38]) and, differing on average by more than 1 dB from the simply occupied condition, suggests a clearly audible change. Point-by-point variations are observed among receivers, mostly as a consequence of different source–receiver distance (e.g., combinations P_R4 and P_R5) and of screening elements like the cancellata (e.g., combinations P_R3 and P_R4). When the source is at the cantoria, the G_mid is much more uniform across receiver positions and the presence of the tapestries (either in unoccupied or occupied conditions) induces an average 2 dB difference that would be clearly audible in most of the positions.

In terms of clarity (C₈₀), when the source was in the presbytery, receivers 03 and 06 in the “public” area showed the lowest values, possibly because of the cancellata preventing early reflection to contribute. The presence of tapestries and occupancy improved the values a little, but clarity remained low. The major change was observed when tapestries were added to the empty configuration (blue and orange lines in Figure 6c), whereas their addition to the occupied configuration produced only a minor effect (gray and yellow lines in Figure 6c) except at a few specific receiver positions with source C. Conversely, in the “private” area, the values were higher and, when tapestries and occupation were added, were closer to 0 dB, particularly for receivers closer to the source. When the source was at the cantoria, C₈₀ values were more uniformly distributed in space and showed differences as significant as for the other source position when tapestries were added to the empty room, with the only exception of receiver positions 4 and 5. Across the considered listening positions, when the source is located at the presbytery, the Pope’s position (R1) benefits from the greatest gain in clarity due to wall tapestries and audience. When the source is placed at the cantoria, the clarity increase at R1 remains significant, although not markedly higher than at the other listening positions. In all cases the role of the tapestries was significant, allowing us to obtain acoustic conditions less dependent on the occupancy.

Lateral energy fraction (J_LF) showed significant point-by-point variations, mostly due to the relative position of sources and receivers with reference to reflecting walls, with little variations as a function of occupancy, as already observed elsewhere [11,31], being the parameter independent of reverberation time variations. Conversely, with few exceptions, the presence of the tapestries caused a predictable reduction in the parameter, with variations that were more audible (being higher than JND) at certain combinations (e.g., P_R2 or C_R1), where strong lateral reflections were attenuated by the increased wall absorption.

Finally, with reference to STI, the trend was very similar to what was observed for C₈₀, with milder point-by-point variations resulting from the nature of the parameter (being less sensitive than clarity to small variations in the reflections sequence). The addition of tapestries caused a clear and audible improvement (corresponding to a 2JND increase) with reference to the empty room, and a smaller but still audible improvement (corresponding to slightly higher increase than 1 JND) under occupied conditions. The receivers that showed the largest improvements due to tapestries were R6 (particularly under empty conditions) and R5 (in both cases), independent of source position.

In order to better understand the spatial distribution of the parameters and the effect due to tapestries, C₈₀ and STI were mapped with reference to the source located in the presbytery and the source located at the cantoria. The results, shown in Figure 7 and Figure 8, confirmed what was observed previously, providing further elements of discussion. In particular, for C₈₀, when the source was in the private area, a clearly bimodal distribution of values appeared, with the cancellata decreasing the early reflections in the area immediately opposite to the source, despite its partial transparency (of the upper part), and the back wall providing extra reflections that contributed to raise the parameter’s values. This distribution appears similarly in all the occupancy/tapestries combinations, with a comparable span of about 25 dB from maximum to minimum in each individual configuration, while showing a significant upward shift in absolute values depending on the surface arrangements, which demonstrated that the presence of the tapestries contributed to improving the listening conditions in all cases.

When the source moved to the raised cantoria, (above the cancellata), the overall span of the parameters’ values reduced to about 15 dB with no bimodal behavior, although the distribution was obviously skewed towards the lower values as an obvious consequence of the prevalence of diffuse field conditions as the distance from the source grew. Again, increasing the absorption in the room shifted the values upward, with tapestries contributing to increase the mean values by about 2 dB with reference to the empty conditions, and by about 1.3 dB with reference to the occupied conditions, with most of the listeners laying in the range from −5 dB to 0 dB, which is optimal for choral singing and music [24].

When the speech transmission index (STI) was considered (Figure 8), although C₈₀ and STI maps look similar in general, STI maps still present a more complex distribution—particularly with the source in the presbytery, where speech mattered the most—likely due to its multi-frequency nature (being a weighted average of all frequency bands). In fact, the distribution of the values was far from being “normal” and local peaks appeared, as observed in the histograms of Figure 8, representing a generally better intelligibility in the private part of the chapel, while in the public part it was generally poorer and only the listeners close to the back wall were favored by early reflections from the back wall. In absolute terms, increasing the absorption STI moves from a range spanning from 0.22 to 0.44 under empty conditions (with most of the listeners perceiving a “bad” intelligibility according to the descriptors of the subjective perception given in Appendix A), to a range spanning from 0.35 to 0.58 under occupied conditions with tapestries (with all of the listeners in the private area having “fair” intelligibility). The interesting part is that the presence of the tapestries contributed to improving the overall conditions in all the cases, particularly in the empty configuration where an average improvement of 0.07 appeared, suggesting that it was audible (being higher than 2 JNDs for STI), and that, for private ceremonies with a limited audience, it could make a difference. However, a positive variation, although less dramatic, was observed also under occupied conditions, allowing all the receivers in the private area to experience at least a “fair” intelligibility.

The analysis of STI values when the source was moved to the cantoria, although less interesting in practical terms (as the cantoria was not used to deliver speeches or sermons), showed acoustical interest because the combined effect of the raised position (resulting in longer minimum distances) and location at the center of the room (resulting in shorter maximum distances), yielded a narrower distribution of values, although with reduced absolute values. However, the positive contribution of the tapestries was again clearly audible.

5. Discussion

The study of the acoustics of Sistine Chapel at the time of Leo X, with an emphasis on the role of Raphael’s tapestries to modify listening conditions, pointed out that with reference to speech-oriented activities, like gatherings and masses for the Papal Court, their positive effect was certainly appreciated. In fact, the addition of tapestries contributed to reducing reverberation time in the room, significantly narrowing the range of variation as a function of the occupancy (that, at mid-frequencies, dropped from 7.4 s for an unoccupied room to 4.1 s for an occupied room to 5.1 s for an unoccupied room and, finally, to 3.4 s for an occupied room). Given the height of the space, its regular shape, and the fact that most of the absorbing materials were located at floor level, the absorption was slightly less effective than it would have been in ideally diffuse conditions (as predicted by simple formulas like Sabine’s or Eyring’s). Nonetheless, as shown, it was enough to be aurally noticeable by an attentive listener. Thanks to tapestries, in the “private” area reserved for the Papal Court, the listening conditions improved significantly as indicated by the STI analysis. In fact, with the unoccupied chapel, without tapestries, only 18% of the private area benefited of at least a “fair” intelligibility, while with tapestries the percentage increased to 38%. Under fully occupied conditions, the private area fraction with at least “fair” intelligibility raised to 54%, but thanks to the tapestries, 95% of the area was above the 0.45 threshold, and 20% was above 0.6, suggesting that verbal communication could take place, with fair or even good intelligibility, provided that background noise was limited. In fact, taking advantage of simulations, it was demonstrated that an increase by 5 dB in the noise profile induced an average drop of 0.04 in STI, basically nullifying the advantage resulting from tapestries installation unless the speaker raised its voice to compensate. Considering that the natural loudness “gain” that human voice can achieve when moving from a “natural” to a “loud” emission level is approximately 12 dB, Leo X’s request for “stationing officials” to discourage chattering during celebrations, thereby limiting background noise, becomes understandable.

With reference to choral singing, the analysis of results when the source was at the cantoria demonstrates the beneficial (and clearly audible) effect of tapestries. In fact, thanks to tapestries, EDT under fully occupied conditions was uniformly decreased to around 3 s at mid frequencies, which is perfectly acceptable for choral music. Sound strength under occupied conditions was further decreased by the presence of tapestries to an average of 7.7 dB, against 8.9 dB without them, which is positive to reduce the sensation of being overwhelmed by music and singing as it happens in small and reverberant rooms. In terms of clarity, the presence of tapestries played a relevant role both in unoccupied conditions (contributing to making the sound more similar to the occupied conditions, which is useful for rehearsal purposes), and in occupied conditions (contributing to increasing clarity within a range from −6 dB to 0 dB, which is optimal for choral performances in worship spaces [24]). The sense of spaciousness, as revealed by the analysis of lateral fraction, is scarcely affected by tapestries (and by occupancy), confirming what was already observed in other studies [11,31]. Finally, unlike what happened for speech, given the position of the source, the previously described advantages were equally available for both the private and public part of the audience.

In order to allow a practical comparison of the previously described acoustic differences based on auditory stimuli, auralizations of the different configurations as perceived at the Pope seat (Rec. 01) when sources were in the presbytery and at the cantoria were created. For auralization purposes, in order to allow a reproduction of audio excerpts without using special equipment, stereophonic impulse responses (in X-Y configuration) generated by the simulation software were used, in combination with anechoic recordings of an excerpt of Nicene Creed read in Latin and Josquin Des Prez’s Missa Pange Lingua sung by faculty members and students from Southern Methodist University’s (SMU) Division of Music. Josquin’s Missa Pange Lingua was selected after consultation with musicologists as being a typical example of the polyphonic Mass performed in Leo X’s times. The auralized material can be found in the Supplementary Materials.

6. Conclusions

The paper explored the effect of tapestry installation in Sistine Chapel in Rome reconstructing its original state at the time of Pope Leo X. Using architectural data that described the space and taking advantage of acoustic measures carried out in 1976 by Prof. Fricke, an acoustic model of the space was made and calibrated by minimal adjustment on absorption coefficients of materials. Starting from that model, and moving the cancellata to its original position at the center of the room, the effect of tapestry addition with and without occupancy was explored. The acoustic properties of tapestries were derived from actual measurements carried out on the Barberini series, which was actually made about one century after the Raphael’s series that was actually hung in the chapel. However, despite possible differences in weaving techniques, using such measured data was considered to introduce fewer uncertainties than taking more generic data from the literature. The results proved that the effect of tapestries is well evident in terms of variations in acoustical parameters and such variations are, at least for some important subjective aspects like reverberation, strength and clarity, aurally detectable. Finally, such variations provided improved conditions for the intelligibility of spoken messages and singing which, at the time of Leo X, when limited alternate solutions were available, represented a remarkable outcome.