Original or post-war paintings? The fixed wooden scenery of the Teatro Olimpico in Vicenza: a guided multidisciplinary approach based on scientific analyses and HBIM

Abstract

Archival resources, a photogrammetric survey, analysis of materials and techniques, and evaluation of the state of conservation were interconnected phases in the interdisciplinary research on the perspective scenery of the Teatro Olimpico in Vicenza. To fill existing gaps related to the history of this unique artefact and to the undocumented retouching of paint layers on the scenery during post-war reconstruction work, a broad analytical approach (optical microscopy, XRPD, SEM–EDX, μFTIR and μRaman) was applied to micro-samples, the selection of which was guided by consolidated geometrical and archival data. The aims were to obtain a characterisation of the pigments and binders and to evaluate the microstratigraphic sequence, the state of conservation of the scenery flats, and the deterioration processes involved. The findings unveiled a discernible variability in the production techniques of the finishes, frequently lacking the classic microstratigraphic sequence associated with the traditional method for painting on wooden panels from the sixteenth century. Moreover, by identifying many pigments that could be used as temporal markers we were able unequivocally to establish, for the first time, that a significant part of the scenery designed by the architect Vincenzo Scamozzi was heavily repainted in the post-WWII period. These analytical results and all the data collected on the artefacts over time were related through an HBIM model, enabling the historical and technical-analytical information to be linked to the geometrical survey and thus to provide guidance for actions based on current and future knowledge for the maintenance and monitoring of the Teatro Olimpico.

Introduction

The Teatro Olimpico in Vicenza is the oldest, permanent indoor theatre of the modern era, designed by Andrea Palladio in 1580. Vincenzo Scamozzi designed the fixed wooden flats for the stage scenery in 1585. Today, these represent an unicum of the Renaissance era (Fig. 1). After they reassembled the scenes in 1945, restoration work was carried out on the pictorial layers and supports [1, 2], operations of which there is little documentation remains in the archives of the Soprintendenza of Verona. The first survey campaigns began in 1996 [1, 2] aimed at ascertaining the state of conservation of the theatre and the wooden scenery [3,4,5,6,7]. However, original and restoration elements were not clearly and analytically distinguished at the time. This is fundamental since the complex (chemical and mineralogical) composition of the paint layers of artworks can be linked to many factors, such as the painting technique, the formation of decay products, including those associated with the environment, and the effect of past conservation operations. This characterisation is extremely useful for detecting historical events and improved understanding of the artistic heritage [8,9,10,11].

Fig. 1

Evaluation of iconographic-archival data: relationship between the condition of the scenery in its current state via laser scanner (lower portion) and comparison with the same portion of the items as photographed in 1944, before they were dismantled, (left photo) and the same viewpoint repurposed in 2023 (right photo). The lower image shows five of the seven streets of Thebes (17 single flats in total)

In this context, in order to develop a conservation project for Scamozzi’s scenery, it was necessary to adopt a multidisciplinary approach which combined archival documentation and further data acquisitions, enabling a global understanding of the materials and their history, together with controlled and accessible storage of the documents produced. For this reason, this research proposes a dual approach to knowledge, both technical-analytical and historical-documentary, as advised by the international Burra Charter [12].

This dual approach involved a detailed scientific investigation of materials, guided by archival information, in order to enhance what had been partially investigated and to support the documentary study. Paintings on wooden supports have been widely studied in cultural heritage [13,14,15,16,17]. The preciousness of the materials and their small dimensions have often led to the use of non-destructive investigation techniques of various kinds [18,19,20,21,22,23,24]. In the specific case of the Teatro Olimpico, the areal extent of the artwork and the presumption of a very complex pictorial history resulted in a preference for a classic study of materials carried out through micro-sampling and laboratory analysis.

In addition, a method was developed to include, in a single organised work space, the data resulting from: (i) the study of documentary and iconographic sources conserved in the archives of the Soprintendenza of Verona; (ii) the investigation of previous cognitive campaigns conducted on the building; (iii) the recent photogrammetric survey of the scenery, which allows the simultaneous acquisition of geometric and radiometric data [25]; and (iv) the current analytical study of selected micro samples. The project found direct application within the HBIM (Heritage Building Information Modelling) environment [26], generating an interoperable and updatable model that links information and documents of a multidisciplinary nature to the geometric element. The HBIM makes it possible, for example, to date the dismantling and reassembly of the wooden scenery in order to understand and date restoration and conservation operations and to identify the materials studied through specific analytical investigations, linking them to the final reports. Moreover, the HBIM environment offers the opportunity to plan future actions and to monitor or define necessary investigations in a regime of controlled maintenance. The multidisciplinary approach adopted proved to be fundamental for correct planning of the survey of materials and the maximisation and validation of the information obtained through the coherent interpretation of the scientific data obtained. The integrated analysis of interdisciplinary and multidisciplinary skills (historical-archival survey, scientific laboratory investigation and HBIM) allowed the results to be read in an organic and interconnected way, and interpretation of the data to be raised to a higher level.

Materials and methods

The object of the study

Vincenzo Scamozzi designed the scenery for the Teatro Olimpico in Vicenza in 1585 for the first staging of Oedipus Rex [3, 27,28,29,30,31].

The installation consists of seven sets of flats made of wood with gypsum-based finishes, representing the seven streets of Thebes, arranged radially from the centre of the stage and constructed using the perspective technique. They therefore provide a section both in plan and elevation that narrows toward the perspective vanishing point. This peculiar construction typology meets the demands of perspective by adopting a ground plan in which the flats on either side of the street scene draw closer to each other as the street appears to lengthen, and also by reducing the height of the flats and the buildings they depict as the eye is drawn towards the end of the street.

According to archival documentation in the Soprintendenza of Verona,¹ the flats were dismantled during World War II, and the theatre’s mobile decorative apparatus was moved to a safe location to preserve it from possible damage caused by the bombing of 1944.² After the end of the war, reassembly operations began in November 1945³,⁴ and the restoration begun then has continued, off and on, to the present day. Bibliographic research revealed that restoration during the post-war reassembly operations involved the removal of 19th-century pictorial film [1]. In addition, the 1979 interventions on the wooden set focused on fixing the film, cleaning, plastering, and supplementing the pictorial pigments [2]. No information regarding the materials used in the various restoration operations is to be found in the archive documents.

From 1996 until 2009, thanks to the collaboration of several institutions (Istituto Centrale per il Restauro-ICR, Centro Internazionale di Studi di Architettura Andrea Palladio-CISA, Municipality of Vicenza and Università Iuav di Venezia/Laboratory for Analysing Materials of Ancient origin-LAMA), several campaigns of investigations were conducted on the theatre mainly to verify the state of conservation of wood components [3,4,5,6,7].

The preliminary investigation conducted in 2006 on the paint finishes of the fronts of flats 3B, 3D, and 5B, which is unpublished and not open access, was carried out through optical microscopy (in reflected, transmitted and UV light), X-Ray powder diffraction and Scanning Electron Microscopy and energy dispersive X-ray spectroscopy analyses. It indicates a paint palette that includes ochre (red, yellow, brown, and also raw umber), carbon black, white lead, lithopone, titanium white, barium white, rare green earth, and chromium oxide. The observed stratigraphy is straightforward, with few (maximum 4) paint layers applied on gypsum-based preparations that sometimes contain pigments such as carbon black and ochre. The presence of binders of a generic organic nature is occasionally noted, as is the presence of homogenization layers without aggregate and always of an organic nature. Additionally, very thin layers based on titanium oxide and/or barium sulphate are frequently found. The investigations conducted present the reported data without providing an interpretation at a global or stratigraphic level (other than the indication of some of the more recent pigments).

Information management and reading resource

In 2021, both topographic and laser scanning surveys of all portions of the theatre were carried out to generate the geometric consistency of the existing proscenium and wooden scenes [25]. The resulting data, optimised by the multiple techniques applied, have been processed to obtain an HBIM model of portions of the Theatre. BIM is the digital information system for the construction of a building consisting of a 3D model integrated with physical data from multiple disciplines (architectural, physical, installation, structural, etc.). Today, BIM models applied to Built Cultural Heritage (HBIM) [32] allow the inclusion of abundant geometric-type information [33] as well as data referring to the construction techniques and history of the building, its state of preservation and the materials employed [26, 34,35,36,37]. As is necessary when working within the HBIM environment, the 3D model was progressively simplified [38] and the scenery was modelled by sector without detailing the development of the decorative apparatus. The theatre was subdivided into macro-elements (i.e., perspective scenery, proscenium, orchestra, cavea, roofing, odeum, antiodeum, elevations, garden) to which information of both a historical–critical and scientific nature could be attached, as well as future possible operational guidelines. In this way, the efficacy of the restoration work can be simply managed and monitored.

The archival, bibliographic and iconographic information acquired during the document search was subdivided by type and stored in a digital repository, as illustrated (Fig. 2). In addition to the source type, another element of file subdivision was the date of data acquisition or document creation. In the case of archival sources, these were sorted by date of document creation, thus year by year until 2020. Case study analyses were sorted by the date of information acquisition.

Fig. 2

Structure of the digital repository for cataloguing information related to the Teatro Olimpico

The architecture of the information repository allows for correspondence between the sources and the three-dimensional element within the HBIM environment. The data collected, whether historical, technical, analytical or geometric, were all linked within the 3D model. The HBIM models can support different types of analysis in a single digital model [39] and can be continuously updated [40]. By interrogating a three-dimensional element in the HBIM model, it is possible to retrieve differentiated types of information. Some can be experienced by the user immediately, such as scenery construction date, architect or ongoing analysis. Others, such as the final reports of investigations performed in previous data acquisition or photographic campaigns, are connected through dynamic links directly available in the model. The following information has been linked to each element in the HBIM: dating, author, transformations, materials, construction characteristics, state of preservation, investigations carried out, studies to be carried out, historical iconography, photographs, critical issues, action to be taken.

In particular, the model used made it possible, through the interpolation of historical and analytical information, to guide the new investigations that are the subject of the current research, orienting the sampling campaign towards the areas that had been less analysed. Thanks to the HBIM model, it was possible to make a direct comparison between the previous survey campaigns (mineralogical-petrographic and physical–mechanical), and it emerged that the most significant ones were carried out between 2006 and 2009 and concerned flat 1 (front 1B), flat 3 (fronts 3B, 3C and 3E) and flat 5 (front 5A), the upper part of the scenery representing the sky, the cave roof, the odeon, the sub-scene and proscenium areas [7]. To evaluate the state of conservation of the fronts, the survey orthophotos were compared with historical photographs conserved in the Archives of the Soprintendenza of Verona and taken during the disassembly phases in 1944, which provide an accurate picture of the state of conservation of the paint surfaces in the post-war period. Critical evaluation of the historical images extended our knowledge of the historical state of preservation of the building through mapping the ancient pictorial layers and their state of conservation. This methodology gave specific indication of the areas needing to be sampled during the present research [41]. The samples taken in the previous campaigns were mapped (for the first time) in the elevations of the wooden flats and linked to the HBIM model to obtain a complete framework of preliminary knowledge of the materials of the stage flats and their state of preservation. This made it possible to orient the new samples to be taken from flat 1 (frontal areas 1A and 1B), flat 2 (frontal areas 2B and 2C) and flat 3 (frontal areas 3A, 3B and 3C) (Fig. 3).

Fig. 3

Static visualization of the dynamic path of research: the first column shows archival sources; the second column shows past analytical campaigns; the third column shows the location of all points where photographs and samples of pictorial materials were taken on the 3A front during the current research project. The image highlights comparison with the new process of analytical survey methods through the use of HBIM

The HBIM model also allowed for the inclusion of other investigations carried out in previous campaigns, such as structural diagnostics in the sub-scene spaces; endoscopic investigations in the walls and sub-scene; magnetometric and thermographic investigations in the proscenium, in the ceiling above the stage scenery, painted to depict the sky, and the ceiling over the cavea and the Odeum, highlighting the fact that the theatre's internal microclimatic parameters had never previously been monitored. These investigations commenced recently in order to gather data on the fluctuations in internal temperature and relative humidity, as well as to establish possible correlations between these factors and the development of deterioration phenomena or pathologies on surfaces or structures. In addition to microclimatic investigations, acoustic monitoring was carried out to verify the response to acoustic stresses of the interior spaces of the Theatre, which serves as a sounding board. It is currently being determined whether the structural alterations made during the second half of the twentieth century may have impacted the acoustic and microclimatic behaviour of the interiors of the Theatre.

The HBIM model makes it possible to have a diachronic and overall reading of the events and analyses of each of the theatre's architectural elements organised in a multi-scalar manner (from the flat to the individual sample). Moreover, by being able to integrate past and current investigations, it is a useful way of monitoring the effectiveness of protection and conservation actions, both those already conducted and those in progress ("Three-dimensional model data visualisation and cataloguing for preservation" section explains what information is available in the model when an item is selected).

In the present work, micro-invasive sampling (all samples taken measure a millimetre at most) was carried out with a view to characterising the pictorial finishes of the wooden flats, taking into account the chromatic variability of the decorations. The most characteristic and repetitive colours were therefore sampled in significant numbers from each of the most important flats, as indicated by the study of archival resources. Sampling was performed on portions that were already damaged and/or hidden so as to avoid affecting the aesthetics of the scenery in any way or damaging the paintings further; in all, fourteen micro samples were collected from the flats. Table 1 contains a complete list of the collected samples, while Fig. 3 shows the path followed by the research team in building the HBIM model so as to understand where the new samples have to be taken from. The starting point of the process was the point cloud obtained in the laser scanning campaign, which was necessary for construction of the model (column 1), but more important was the comparison between the survey and archival documents related to the previous campaigns and historical photos of the reassembly of the scenes. These three elements visualized simultaneously in the HBIM made it possible to locate where new sampling should take place, avoiding repetitions, overlaps or disadvantageous locations. Figure 3 is also a static representation of the dynamic path of the research, from which it is clear how the model can be used for research purposes: the first column shows the archival sources used to gather information about the historical development of the building and the restorations to which the Teatro Olimpico was subjected in the past; the second column shows the recent history of the scenery, specifically the past analytical campaigns; the third column displays the location of all the sampling points of surface film and pictorial materials carried out on the 3A front during the ongoing research project. For the purpose of the representations in Fig. 3, the flats have been virtually dismantled to pinpoint previous survey locations and to highlight comparison with the new process of analytical survey methods using HBIM.

Table 1 List of samples taken from the fixed wooden scenery of the Teatro Olimpico in Vicenza

Analytical methods

The samples were investigated using a multi-analytical approach to define their stratigraphy accurately and identify their composition in terms of binders and pigments. A polished cross-section was prepared for each sample by embedding it in epoxy resin. Polishing was performed dry and/or using petroleum as a fluidifying agent to avoid any hydrosoluble material (e.g., gypsum) being in contact with water. Where necessary, certain analyses were conducted directly on the materials as they were. The analytical methods adopted are summarised below.

Reflected light optical microscopy (RLOM)

Reflected light optical microscopy was adopted on polished cross-sections to investigate, describe and measure the microstratigraphy of the layers and identify their various components (pigments, binder). This study also used ultraviolet light (UV) observation to estimate the fluorescence induced by any organic substances. For this purpose, use was made of a Leica DM2500 optical microscope equipped with a mercury vapour lamp and a set of filters suitable for visualising the fluorescence emission response (green emission: excitation filter BP 450–490 nm, dichroic mirror FT 510 nm, emission filter LP 515 nm; blue emission: BP 320–380 nm, dichroic mirror FT 400 nm, emission filter LP 425 nm).

X-Ray powder diffraction (XRPD)

Mineralogical analyses were performed by X-ray powder diffraction on fine-grained powder samples (obtained with an agate mortar and pestle) using a Panalytical Empyrean diffractometer, operating at 40 kV and 40 mA, in Bragg–Brentano reflection geometry and equipped with CuKα radiation and an X’Celerator detector. Qualitative diffraction data analysis was carried out with X'Pert HighScore Plus® software (PANalytical) and the PDF-2 database. The reference intensity ratio (RIR) method, rather than only the intensity of the peak channel, was used to semi-quantify the amount of the recognised materials in the multiphase XRPD patterns. This analysis was carried out, where possible, on single layers or several together as long as they could be separated manually with the support of the stereomicroscope and scalpel.

Scanning Electron Microscopy and energy dispersive X-ray spectroscopy (SEM–EDX)

Micro-morphological, textural and elemental compositional investigations were carried out using a ZEISS EVO15 scanning electron microscope (SEM) equipped with a LaB6 cathode coupled with an X Flash 6160 BRUKER energy-dispersive X-ray spectroscopy (EDX) system, on platinum-coated cross-sections. A series of analyses was performed on each recognised layer to characterise the different materials and measure the thickness of the layer where possible. Elemental maps were also generated to enhance comprehension of the samples’ multilayer structure.

Micro-Raman spectroscopy (μRaman)

Micro-Raman spectroscopy was performed with a DXR2 Thermo Scientific™ Raman microscope, equipped with two frequency-stabilised single-mode diode depolarised lasers at 785 nm and 532 nm. Spectra were acquired on the cross-sections, using laser power varying between 1 and 30 mW for the 785 nm laser and between 1 and 10 mW for the 532 nm laser. Spectrograph aperture of 25 μm pinhole and 10X and 50X objectives were also adopted, therefore with a spectral resolution in the range 2.7–4.2 cm⁻¹ and a spatial resolution of 2.1 μm. Here too, a series of pinpoint analyses was performed on recognised layers and coloured particles to characterise the different materials and pigments where possible.

Fourier-transform infrared spectroscopy (μFTIR)

The identification of the organic compound was performed using a Thermo Scientific™ Nicolet™ iN10 MX Infrared Imaging Microscope for pinpoint analysis of both the cross-sections and the untreated samples. In the latter case, a small amount of powder sampled from every layer identified by stereomicroscope was taken using a needle, then deposited and pressed on a standard KBr pellet for each micro-fragment. The analyses were performed in transmittance mode, with the minimum spectral resolution equal to 1 cm⁻¹ and 60 consecutive scans. In addition, fingerprint peaks were mapped on cross-sections using germanium crystal micro-tip ATR with 0.35 mm Ø for tapping-mode investigation of the component distribution in the microstratigraphy.

Field acquisition

A photogrammetric survey provided enhanced knowledge of the wooden flats and the theatre proscenium. The use of Structure from Motion software in processing photograms yields various outputs, including high-resolution textures and orthophotos. These outputs represent two-dimensional renditions of three-dimensional space, providing geometrically scaled and measurable representations [25, 42].

In the frame acquisition campaign, 964 shots of the central street scene were acquired. Due to difficulties encountered inside the theatre, such as poor lighting and insufficient space, the data were obtained partly with an action camera and partly with a Canon 5D Mark II, as described in Ballarin et al. [43]. In particular, 24 of the 964 frames were taken with the DJI Osmo Mobile 3 action camera at maximum resolution (4000 × 3000 px), and 940 with a Canon 5D Mark II coupled with Canon EF 24-70 mm F.4 IS II lens at full resolution (5616 × 3744 px). The frame capture was manual and supported by a soft LED box with variable colour temperature and illumination intensity to correct interior lighting conditions proactively. This expedient was adopted because the existing illumination caused strong contrasts of light and shadow, resulting in an almost complete loss of information in the darker portions.

The shots were processed using the commercial software Agisoft Metashape Pro. The photogrammetric model was oriented on the same reference system previously used for processing the topographic and laser scanning survey. The decision to use the same reference system was made for two main reasons: (i) the previous survey was already available to the research team, so it was not necessary to perform a new topographic survey; (ii) the use of a shared reference system opens up the possibility of including the results of the photogrammetric survey within the HBIM environment. Thus, it became feasible to link the photogrammetric point cloud to the pre-existing information model, employing the same reference system.

Results and discussion

Paint microstratigraphy

The microstratigraphic study of the pictorial finishes of the wooden scenery proved to be varied and sometimes very complex (Fig. 4). The analyses and observations conducted enabled the pigments, fillers and binders used to be identified and recurring sequences of layers exhibiting similar characteristics to be recognised and categorised. In the most successful cases, also thanks to the identification of specific pigments, it was possible to attribute a relative temporal collocation to the paint layers concerned in the series of restoration and/or retouching operations to which the flats have been subjected over the centuries.

Fig. 4

Cross sections in reflected-light photomicrographs of samples TOV_02 and TOV_17 (a, d); elemental EDS maps in false colours of samples TOV_02, TOV_04 and TOV_17 (b, c, e respectively). Reflected-light (f) and UV green emission (g) photomicrographs of sample TOV_04 compared with μFTIR map (h) of the characteristic peak of acrylic polymers (ν C = O, carbonyl, 1730 cm.⁻¹)

In order to describe the microstratigraphy more efficiently, the layers identified were named alphabetically starting from the deepest layer (i.e., layer A) to the top (i.e., layer C for a sequence of 3 layers) (Table 2).

Table 2 Description, composition and thickness of each layer of the samples analysed

Wooden supports

The use of HBIM and the introduction of all prior investigations into its archival system prevented unnecessary repetition of characterizations, and enabled assessments to be based on these findings as regards the condition of finishes and supports. The study of the wooden supports, therefore, relies on outcomes from earlier investigations as revealed by archival research.

According to this research, Zurli was the first Superintendent to order the scenery to be studied during the 1979 restoration campaign [2]. His predecessor, Forlati, mentions the removal of "the nineteenth-century paintwork" to restore "the original paintwork" [1] during post-war restoration, but he makes no reference to investigations guiding the restoration efforts. The first evaluation of the condition of the theatre's wooden elements dates from 1996 when the ICR photographed various signs of aggression attributable to xylophagous insects and fungi (woodworm holes, wood decay) during an exploratory survey⁵ supporting the development of a suitable fire-fighting system [6]. Between 2006 and 2009, CISA, in collaboration with the LAMA Laboratory (Università Iuav di Venezia) and the Municipality of Vicenza, conducted an investigative campaign to develop a conservation project for the Teatro Olimpico [7]. This campaign involved dendrochronological analyses to classify and date the wooden elements. Results achieved demonstrated that the wooden perspective flats were crafted from Silver Fir (Abies Alba) and Spruce (Picea Abies) and probably date between 1520 and 1539.⁶ Lastly, in the report on the restoration work on the ceiling of flat 3, it is noted: “Inspection of the front and back of the wooden structure does not, in fact, reveal the presence of rodent teeth marks…except for just two points, but the phenomenon as a whole appears to be of previous origin"⁷ [6, 44].

In the survey campaign described in this article, there was no repetition of investigations to characterize the wood and its state of preservation. Nonetheless, microstratigraphic analyses conducted on the three wood samples taken from the back of flats TOV_13, TOV_14 and TOV_16 confirm the presence of some woodworm holes and deposits of soft, brown rot [45].

In addition, all the analyses performed on the grey deposits recognised on the surface of samples TOV_13 and TOV_14 revealed the presence of gypsum residues. Moreover, a discontinuous yet cohesive layer consisting of sub-microcrystalline silica was also identified through XRPD, μFTIR, and EDS analysis, exhibiting abundant Si and O content (Fig. 5a). Perhaps this latter detail can be linked to the use of silicate-based products adopted to treat the backs of the flats, although their application on wood is not widely attested in the literature [46].

Fig. 5

Backscattered SEM–EDS images of a microcrystalline silica deposits on wood sample TOV_13; b and c two areas of sample TOV_17 showing different calcium sulphate micro-textures

Ground layers

Since at least the twelfth century, the recipe historically used for ground layers [47] involved gypsum or anhydrite as the inert material and animal glue, usually rabbit glue, as an organic binder. This mixture was laid on top of wood previously treated with strong glue and covered with a canvas soaked in glue, a procedure that offered excellent chemical stability. The ground layers might then be numerous, starting with coarse plaster mixtures and gradually decreasing in grain size. Finally, each ground layer underwent the 'rasatura' procedure, i.e., its surface was levelled with a special tool to remove protrusions and to smooth the surface. Sometimes, this procedure was replaced for the final layers by applying a layer of glue on the last ground layer [47]. In the case of the Teatro Olimpico, although in some places there is a correspondence with the chemical and mineralogical composition of the preparation, no evidence of canvas has been found in the samples analysed. The basic composition of most of the ground layers (Fig. 4) found is calcium sulphate (CaSO₄), but in different degrees of hydration (often coexisting in the same sample) and thus in different mineralogical phases, from anhydrite (CaSO₄) to bassanite (CaSO₄⦁½H₂O) to gypsum (CaSO₄⦁2H₂O). Anhydrite is a mineral (naturally available or synthetically produced by calcinating gypsum) which, in the presence of water, hydrates and transforms into gypsum. Thanks to this characteristic, it also serves as an inorganic binder, and is typically used in mixtures that require a particular lightness (e.g., fine plasters or stuccoes). The presence of bassanite, an intermediate hydrated phase between anhydrite and gypsum, can be linked to an incomplete hydration process of anhydrite. The microstructural study of sulphate phases conducted via SEM–EDX has shown the simultaneous presence in many samples of calcium sulphate crystals with both the typical morphological characteristics of ground granules – a compact, crystalline structure with an average size of about 50 µm (Fig. 5b)—and as an agglomerate consisting of thin acicular crystals (10–20 µm in size) forming a felt-like structure (Fig. 5c), the latter being peculiar to the hydration phenomenon of a calcium sulphate-based binder. Therefore, it can be assumed that even though theoretically, calcium sulphates in the mixture should act solely as an inert filler, in this case, they have been intentionally added, probably in the form of fine powder, to perform an additional function as an inorganic binding agent, likely supporting a more typical organic binder.

In detail, at least three types of ground layers have been recognised:

• i. The first is from pure white to greyish white, depending on the number of impurities (i.e., carbon black, ochre) or carbonate fillers (calcite) included (Fig. 4a, b). Its mineralogical composition is based on calcium sulphate (gypsum and anhydrite), both used as inert and binding agents. μFTIR demonstrated the presence of an organic binder based on animal glue. Where this was not detected, weddellite was identified both through μFTIR—by identifying the typical peaks at 1621 cm⁻¹ (νas C = O), 1322 cm⁻¹ (νs C = O), 1136–1048 cm⁻¹ (νas C-O), 882 cm⁻¹ (νs C–C)—and through XRPD (in samples TOV_04, TOV_06, TOV_12, and TOV_17). This demonstrates the probable oxidative alteration of organic matter (i.e., the organic binder in the ground layer mixture) over time or possible biological mechanisms, which induced calcium oxalate development [48,49,50,51]. The layers attributable to this type tend to be the deeper and therefore older layers of the more complex stratigraphic samples (TOV_02/A, TOV_02/B, TOV_04/A, TOV_09/A) (Fig. 4).

• ii. The second typology (Fig. 4c, d, e) regards yellow ground layers. The colour is due to the intentional addition of yellow ochre and carbon black particles (both with a considerable grain size) to a calcium sulphate matrix that presents similar characteristics to those described above. The layers showing this composition are listed in Table 2. In samples TOV_04 and TOV_09, these layers overlap the pre-existing ones, indicating a subsequent retouching intervention. Otherwise, samples TOV_12 and 17 are in direct contact with the wooden support, suggesting the absence of the older ground layer (lacunae) and the execution of pictorial integration directly on the wooden panels. It was curious and unexpected to discover that in sample TOV_10, the ground layer was applied on a polystyrene panel used as an additional element to modify the morphology of the lower part of the flat. FTIR analysis (Table 2) proved polystyrene (peaks 3600-3400, 3162-3081-3060-3025-3002, 2924-2848, 1601, 1492-1451, 1028, 758, 698 cm⁻¹) [52] and Polymethyl Methacrylate composition (peaks 3600-3400, 2924-2848, 1943-1871-1802, 1730, 1380-1328, 1276, 1197-1155, 1073, 908, 842 cm⁻¹), whereas macroscopic study of the sampling area reveals a white substrate compatible with the typical aspect of polystyrene.

• iii. The last category consists of layers composed of calcium carbonate (lime) (Fig. 4b, c), an unusual feature in the panorama of paintings on wood and also in those of the Theatre. The layers with this composition are TOV_02/D and TOV_04/C (Fig. 4b, c). In sample TOV_04, this mixture occurs, stratigraphically speaking, below layer TOV_04/D, which belongs to the second type of preparation described above. This indicates that this third type belongs to an earlier intervention. The use of calcium carbonate as the principal inorganic binding agent only in these two samples leads to the assumption that retouching and/or restoration work was limited to a few flats.

The sequences of ground layers testified here present a maximum of two successive layers, sometimes with increasing grain size towards the outside instead of decreasing as required. The final layer, traditionally made of rabbit glue and applied on the preparatory strata, was matched in the C layers of TOV_02 (Fig. 4a, b) and TOV_09. SEM–EDX images clearly showed these layers to be composed of a low atomic number element (probably Carbon); furthermore, both the shape and related peaks of the FTIR spectra acquired in a sequence of layers that may also include layer C of sample TOV_02 confirmed the supposed chemical composition—3283 cm⁻¹ (ν N–H), 3070, 2956–2919-2874–2851 cm⁻¹ (ν C-H), 1736–1646 cm⁻¹ (ν C = O), 1541 cm⁻¹ (δ C-N–H), 1445–1340 cm⁻¹ (δ C-H), 1243 cm⁻¹ (ν C = O)—and thus correspond to the typical spectrum of animal glue [53]. In addition, as usually happens with animal glues [54,55,56], a bright yellow fluorescence characterises the layer when observed with RLOM in cross-section under UV light (particularly in green and blue emission).

Two exceptional cases in terms of the composition of the ground layers are identified in samples TOV_05 and TOV_06, which appear macroscopically to be fillings executed at a later date. The two impastos revealed a carbonate composition characterised by abundant traces of acrylic polymers as indicated by μFTIR analysis. It should be noted, moreover, that TOV_05 and TOV_06 do not show a true pictorial microstratigraphy: the former shows a gypsum basal layer (layer A) surmounted by a sort of thick calcium carbonate-based plastering (layer B). The second sample (TOV_06) also presents a grey-yellow, gypsum-based basal layer (layer A), followed by a thick, white, predominantly calcite layer (layer B), which is unusually filled with wood splinters. The presence of chromium oxide in layer A of sample TOV_06 also places it historically in a period certainly later than the early 1800s.

Paint layers

Pigments

Table 3 summarises the most significant data from the multi-analytical study performed on the identified pigments observed in the stratigraphy of the samples. Specifications for each pigment are reported in this section.

Table 3 Most representative results for pigments

Black

Black pigment particles were identified in almost all the paint layers and numerous ground layers (Table 2). In the RLOM study, they appear in the form of opaque black particles with a mixed grain size ranging from a few microns to a maximum of approximately 100 microns even in the same sample (Fig. 6a). When analysed by μRaman these particles revealed diagnostic first-order bands in the range 1300–1600 cm⁻¹ in all of the samples. In particular, two broad, overlapping bands with intensity maxima at around 1580 cm⁻¹and 1350 cm⁻¹ (G and D bands, respectively, in Fig. 7). For a more accurate identification of the pigment, full width at half maximum (FWHM) of the D and G bands was taken into account. Indeed, Tomasini, Siracusano, and Maier in 2012 [62] report a direct connection between the increase in these values and the increase in the degree of disorder of historic pigments based on carbon. In our case, these values exceed 200 cm⁻¹ and 100 cm⁻¹, respectively, placing the materials in the range of carbon-based pigments of amorphous nature. It should be noted that the peak at 961 cm⁻¹ is missing, due to the PO₄³⁻ anion linked to the bone or ivory black. Moreover, the EDX investigation never detects the presence of phosphorus. The particle sizes also exclude the use of lampblack, which typically presents an extremely fine grain not detected here. This information allows us to classify this pigment as carbon-based black of the charcoal type [63].

Fig. 6

Reflected-light photomicrographs of a cross-section of the sample (a, b); backscattered SEM–EDS images of the same samples (c, d)

Fig. 7

Pigments. Raman spectra of the most significant pigments

Green

Two different green pigments have been identified in the paint layer (and sometimes as occasional particles in other layers), indicating, at the very least, their presence in the artists' palettes.

The first green pigment appears in green/blue rounded semi-transparent particles under RLOM (Fig. 6b). Eastaugh et al. [63] The accurate EDX analysis performed on the green crystals in layer B of sample TOV_04 (point 2 in Fig. 6b, e) reveals a typical composition of an iron-based clay mineral with the presence of silicon and aluminium combined with magnesium, potassium, and calcium in addition to iron (Table 3). It is green earth, widely used in its various forms since antiquity, i.e. as glauconite [(K, Na)(Fe³⁺, Al, Mg)₂(Si, Al)₄O₁₀(OH)₂] and celadonite [K(MgFe³⁺)(Si₄O₁₀)(OH)₂] or a mixture of the two [63]. In particular, μRaman investigations confirm the identification of the pigment as a green earth and also help to make a distinction between the two more common green earths (glauconite and celadonite). Indeed, the spectra show (Fig. 7) a doublet between 170 and 200 cm⁻¹, which lacks glauconite, and the wave numbers of the two main peaks between 250 and 400 cm⁻¹ are higher (i.e., around 270 and 394–400 cm⁻¹). Furthermore, in the region between 550 and 600 cm⁻¹, the most intense peak is at around 550 cm⁻¹ and not at 590 cm⁻¹ as required by glauconite [60]. All these Raman features are typically for celadonite (Fig. 7), suggesting the use of the excellent-quality Terra verde di Verona pigment, which was identified by previous studies as green earth precisely based on celadonite [63].

The second green pigment only appears in layer A of sample TOV_06 and layer C of sample TOV_17, showing rounded green particles under RLOM. The μRaman spectrum (Fig. 7), acquired in correspondence with green/yellow particles, shows the diagnostic peaks of the mineral eskolaite (Cr₂O₃), which identify the pigment as chromium oxide, the synthetic form of the mineral (Table 3).

Typically, the pigment is opaque, olive-green, generally known as chrome oxide or Casali's green. It was introduced in the early nineteenth century and first detected in Turner's 1812 paintings [64].

Red

The red on the wooden flats can be linked to a single pigment. It occurs under RLOM in deep red particles, often agglomerated and associated with very fine-grained mineral particles (clay minerals). EDX analysis reveals high iron contents (Fig. 6a, point 3), which can be associated with the presence of the mineral hematite (Fe₂O₃) and lower percentages of silicon, magnesium, and/or aluminium that can be linked to the clay portion. Calcium and sulphur are also often present, possibly due to background noise from the gypsum component surrounding the particles (Table 3). The pinpointed μRaman investigation identifies hematite, which undoubtedly appears with the two very strong intensity bands at around 220 and 290 cm⁻¹ and the medium intensity band at about 400 cm⁻¹ (Table 3, Fig. 7). The coexistence of hematite and clay minerals enables the pigment to be defined as red ochre.

It is well known that ochres have been in constant use as pigments since prehistoric times [63]. Moreover, their provenance is extremely vast and very difficult to discern due to the extensive distribution of deposits from which this pigment can be obtained [65].

Yellow

As in the case of red, the analyses discovered only one type of yellow pigment. It occurs under RLOM in opaque yellow particles often agglomerated and associated with very fine-grained mineral particles (clay minerals), having grain size from a few microns to a few tens of microns, depending on the sample and layer (Figs. 4a,d and 5a). Also, in this case, the analyses carried out by EDX on yellow particles (point 1 in Fig. 6a, d) revealed the abundant presence of iron – most probably referred to goethite [α-Fe³⁺O(OH)] or other forms of iron hydroxide and oxides (i.e., limonite, here intended as unidentified massive hydroxides and oxides of iron, with no visible crystals) – together with the presence of silicon, aluminium, magnesium, attributable to the clay portion, with traces of zircon (Table 3). Once again, sulphur and calcium are more likely correlated with the gypseous components surrounding the analysed particles. The μRaman investigation revealed the diagnostic peaks of goethite. Therefore, the fundamental peaks are intense at about 390 cm⁻¹ and the medium intensity at about 290 cm⁻¹ (Table 3). The mixture of clay minerals with iron hydroxides and oxides, including goethite, allows this pigment to be defined as yellow ochre.

Regarding this pigment's historical use and origin, what has already been said about red ochre also applies to yellow ochre [65].

White

The white pigments identified are by far the most numerous, and most of these are modern (i.e., after the second half of the nineteenth century).

μRaman investigation detected diagnostic peaks at 1099, 298, and 178 cm⁻¹ in two paint layers, TOV_09/E and TOV_17/B, and in the ground layer TOV_06/B, which identifies the mineral dolomite [CaMg(CO₃)₂]. The XRPD analysis of sample TOV_06 confirmed this attribution. However, only in layer TOV_09/E can its use be attributed as a pigment, while in the other two cases, it is a generic carbonate filler.

The use of dolomite as a pigment is certainly known from Hellenistic and Roman times, although it was not widespread [63].

An unusual white pigment based on lead and sulphur was detected through EDX investigations only in TOV_09. The elemental mapping reported in Fig. 8 showed the superimposed distribution of sulphur and lead mainly concentrated in layer B (red arrow in Fig. 8) and traces in layers E and H. In the literature, lead sulphate is the only pigment that seems to fit this composition [63]. However, attributing these particles to this pigment seems highly improbable in the context of the Teatro Olimpico. Moreover, the historical-temporal location of lead sulphate is not clearly defined. Indeed, it only vaguely figures among the pigments used for artistic purposes. It is, for example, identified by Winter [66] in thirteenth–sixteenth -century Chinese paintings. At the same time, Harley [67] suggests its use in watercolours made in Great Britain in the late 18th and early nineteenth century, when it appears to have been marketed under the name "Flemish white" by painters who followed instructions dating back to the late eighteenth century for its preparation. Lead sulphate still appears in many modern industrially produced pigments.

Fig. 8

Elemental EDX maps of sample TOV_09, in false colours

The analyses revealed two additional types of white pigments. The RLOM study did not enable the two pigments to be differentiated but only detected the presence of a very thin superficial layer (a few tens of micrometres thick) appearing as a homogeneous white-to-yellowish hue (Table 2). On the other hand, SEM–EDX showed titanium-rich particles, occasionally accompanied by lower amounts of other elements (Si, Ca, S, Al, Na, K, Mg) due to background noise from surrounding fillers. These particles were uniformly distributed in the paint finish with a very fine grain size (Fig. 8b). These were paired with barium-rich particles, which were larger in size (up to a few tens of micrometres) and less widespread in the paint layers; they exhibited a chemical composition comprising barium and sulphur, with traces of calcium and silicon likely due to contamination from surrounding materials (Table 3). Both pigments were found together in extremely thin paint layers covering all the samples analysed, as shown by the SEM–EDX elemental map (Fig. 8). However, recognising dispersed barium sulphate particles using mapping mode proved challenging due to the unavoidable overlap of energy peaks crucial for the recognition of Ti and Ba (i.e., Kα at 4,5 keV and Kβ at 4,9 keV for titanium and Lα at 4,5 keV and Lβ at 4,8 keV for barium). Nevertheless, point investigation obviated this factor using other energy peaks, and confirmed the simultaneous presence of barium sulphate and titanium dioxide in the superficial layers of all the samples. Furthermore, XRPD recognised barite and a generic titanium oxide (Table 3). And μRaman analysis revealed diagnostic peaks for barium sulphate (Table 3, Fig. 7) and attested two different crystalline forms of titanium dioxide (TiO₂). Diagnostic peaks of the rutile mineral form were found in the TOV_11/C layers, while those of the anatase form were found in the TOV_12/C and TOV_17/C layers, which would seem to indicate the use of two different titanium pigments (Table 3, Fig. 7). Various observations led us to exclude the possibility of a natural origin for TiO₂. Firstly, its colour differs from that of the natural mineral rutile, which never exhibits the bright white hue characteristic of synthetic titanium dioxide. Secondly, no accessory minerals potentially originating from natural deposits were identified. Instead, the coexistence of rutile and anatase could be linked to various methodologies of synthetic pigment production. Historically, the chloride process, involving the vapour-phase oxidation of titanium tetrahalides, utilized mineral rutile as its starting raw material. This method was introduced in Germany in the 1950s [63]. Thus, the white pigments are titanium white (titanium oxide, TiO₂) and barium white or bianco fisso (barium sulphate, BaSO₄), historically used in various mixtures.

It is worth mentioning that titanium white, first produced in 1916, did not establish its presence as a pigment until after the First World War, and it was only unanimously accepted and recorded in art manuals in the 1930s [63]. In comparison, barium sulphate was used as a white pigment from the beginning of the nineteenth century [63].

Blue

Only one blue pigment was detected. Under RLOM, it appears as blue angular particles with a vitreous appearance. Although characterised by an intense blue hue, it was found in two layers, G and H, of sample TOV_09, appearing macroscopically brown/grey and green, respectively. μRaman investigation identified it unequivocally in both sample layers, and the spectra collected showed all peaks referred to the natural dye indigo [61] (Table 3, Fig. 7).

Indigo was known and used in antiquity. It owes its blue hue to a natural dye derived from the leaves of various species of Indigofera. The procedures and materials used to prepare stable lacquers are numerous. For example, the clay mineral palygorskite was used to produce the famous Maya blue [63]. Its presence in superficial paint layers characterized by the presence of acrylic resin suggests a synthetic and recent origin. Additionally, certain spectral features recognized in the μRaman spectrum, such as the resolution of one of the characteristic peaks between 1582 and 1571 cm⁻¹, would confirm this hypothesis [68].

Comparing the palette detected on the Teatro Olimpico flats, excluding the pigments found in the more recent layers, with that recognized in paintings from the same historical period or from the Renaissance, it emerges clearly how the pigments here are very simple and basic (ochre, earth, carbon black). The pigments found elsewhere are generally more refined and precious. There is widespread mention of vermilion, madder lake, lapis lazuli, verdigris, and azurite [69, 70], and this is also the case in nearby Venice [71]. Pigments such as azurite or cinnabar can already be found in some productions of decorative wooden ceiling panels from the previous century (fifteenth century), such as those studied in Cremona (northern Italy), which recall the most widespread technologies in northern Europe [73].

Organic binders and retouching

The complexity of the case study is also reflected in the definition of the original binders.

However, traditional organic materials were identified as present in the original painting in only a few cases. Indeed, in layer E of sample TOV_02, μFTIR revealed the presence of the characteristic peaks of a binder of proteinaceous nature (3282 cm⁻¹ due to ν N–H, 3067 cm⁻¹; 2956–2919-2873–2850 linked to ν C–H, 1731; 1646 cm⁻¹ attributed to ν C = O, 1541 δ C–N–H; 1444–1399 cm⁻¹ corresponding to δ C-H, and 1242 cm⁻¹ linked to ν C = O), whereas in sample TOV_03 the spectrum of the paint layer showed the presence of Arabic gum (3346 cm⁻¹ due to ν O–H; 2900 cm⁻¹ linked to ν C–H; 1639 cm⁻¹ attributed to δ O–H, and 1427–1316 cm⁻¹ corresponding to νas C–H). Arabic gum could possibly have been used as a medium for tempera and watercolour. It may also have been applied as an adhesive in the assembly of wood elements during restoration. Ground layers of samples TOV_04, TOV_05, TOV_06, TOV_12 and TOV_17 showed the presence of weddellite correlated to the alteration of the organic substance as described in "Ground layers" section.

In contrast, all the samples have a modern acrylic component in the superficial layers (2990–2870 cm⁻¹ ν C-H; 1730 cm⁻¹ ν C = O; 1475–1440 cm⁻¹ δas C-H; 1235 cm⁻¹ ν C = O; 1145 cm⁻¹ νas C–C(= O)-O, 1026 cm⁻¹ ν O-C–C). The multi-analytical study revealed a very thin external layer (between 10 and 20 μm thick), generally white, based on acrylic resin rich in titanium and barium white on all the samples. This chemical composition is clearly of modern date since the two white pigments only became widespread in more recent times, as already mentioned in "White" section. Furthermore, it was only in the second half of the twentieth century that the advent of plastics led to the use of acrylic-type paints in fine art as well. It can therefore be assumed that the application of this thin film dates from the relocation of the scenery after World War II or later. A good example of this reasoning is shown in Fig. 4f, g, h, in which the μFTIR map of TOV_04 proved the chemical composition of the acrylic binder in layer TOV_04/C just above the green paint layer TOV_04/B. This factor dates the retouching operation to at least the second half of the twentieth century. This layer, which is macroscopically almost transparent due to its thinness, may have been applied in order to enhance/revive the chromatic effect and homogenise the macroscopic appearance of the reconstructed set of flats. However, the characteristic peaks were also detected in the paint layers below all the samples. Since µFTIR maps did not identify continuous layers based on acrylic nature in the lower strata but only sporadic traces, it may be supposed that acrylic paint was used for localised retouching or that acrylic matter intended only for the paint surface in fact penetrated the porosity of the pictorial stratigraphy [72].

Based on the microstratigraphic study, although the paint layers just beyond the first type of ground layer in samples TOV_02, 04 and 09 are not precisely datable, they can nevertheless be considered the earliest of the pictorial finishes. There is a change of colour from the deeper layers of the stratigraphy to the colour of the most superficial layer currently perceptible on the flat. Indeed, there is a change from pale yellow (layer D) to yellowish brown (layer G) for sample TOV_02 (Fig. 4a), from green (layer B) to orange (layer F) for sample TOV_04 and from red (layer B) to green/grey (layer G/H) for sample TOV_09.

Samples TOV_03, TOV_11 and TOV_15 present a very simple and unique microstratigraphy compared to the others. They show a white sulphate preparation layer from the inside to the outside, with a few dispersed red and yellow ochre particles on top of the wooden support. Laid directly on this is a very thin layer (layer C) containing the usual titanium white and coarser particles of barium white dispersed in a binder of an acrylic nature.

Three-dimensional model data visualisation and cataloguing for preservation

Through a targeted analysis of front 3A of flat 3, querying an element within the HBIM model yields the information detailed in Table 4.

Table 4 Example of data obtainable from the HBIM system interrogation

Information about the transformations of the fronts and the images, whether historical or current, has been linked to external folders. The HBIM thus offers a further possibility for managing the model: the dynamic link, which can also be updated like all the other information above, leads to folders whose content can be expanded as the search goes on. In the case of this research, it was possible, for example, to link historical post-war black-and-white photographs and shots taken during the surveys, relating the photographic data by individual front, facilitating the interpolation of information and the identification of differences. This operation made it possible to investigate the limits in the pictorial films and decorations between original portions and 20th-century repainting, a completely new element never associated before. Figure 1 shows the relationship with the same portion of the scenery as photographed in 1944, before they were dismantled, and the same viewpoint repurposed in 2023, making it possible to compare the state of conservation of the surfaces.

Through the critical evaluation of previous analyses and the observation of historical images, it emerged that front 3A had not been subjected to sampling. For this reason, the research team decided to focus the investigation on the less studied areas, as shown in Table 1. Table 4, on the other hand, summarizes the information shown by the HBIM when an element has been selected within the model. The information related to Table 1 are graphically proposed in the three columns of the Fig. 3.

It is also possible to have an indicative quantification of the state of conservation of the selected element with a five-grade scale—good, fairly good, fair, poor, very poor—associated with a degradation assessment scale—very slight, slight, moderate, severe, very severe [73].

The laboratory investigations revealed that the pictorial cycle of the scenery that belonged to Scamozzi’s era might have been different, both from a chromatic and compositional point of view, with particular reference to the replaced portions, the additions, and the different materials used to repaint the wooden supports. The results reported here testify in several cases to the chromatic modification undergone by the flats (samples TOV_2, TOV_4, and TOV_9), probably in response to changes in taste or, otherwise, to highlight certain architectural elements over others (a paper is currently being prepared, to explain in greater detail the correlation between the mapping of the historical paintings as seen in the 1944 photos and the micostratigraphies found in the pictorial finishes of the Teatro Olimpico flats). Figure 1 shows the relationship between the appearance of the wooden flats before they were dismantled and how they appear today.

This information can be visualised by three-dimensional axonometric exploded view processing, as shown in Fig. 9. The three-dimensional visualisation allows a detailed overall view from a single viewpoint, guaranteeing, on the one hand, the placement of the individual sample in the general layout of the facade, and on the other hand, an accurate understanding of the laboratory analyses performed as well as the associated results.

Fig. 9

Isometric view of the 3A front of Scamozzi’s scenery for the Teatro Olimpico in Vicenza. The upper part of the figure shows Flat 3A broken down into its architectural components; the lower part shows the location of samples taken from three different points on the flat, the analyses to which each was subjected and the conclusions drawn with respect to the stratification of the paint and ground layers

In particular, through the photogrammetric survey, it was possible to obtain an orthophoto of the flat—in this case the 3A front—which was digitally drawn. Samples and details of the results of the analyses carried out in the current and previous investigations were then inserted. The subsequent analytical determination of their composition showed which parts are original and which were repainted. Figure 9 shows an example of a system interrogation using samples TOV_4, TOV_9 and TOV_10. The upper part of the image displays the entire 3A front, from which the three portions related to the samples examined were exploded. The lower part of the image, on the other hand, shows all the details related to the laboratory investigations performed on the samples taken from the flat, with a three-dimensional visualization of the thin section of the sample that can help in understanding the complex layering of the element. The image highlights the three-dimensional modelling of the individual sample and makes it possible to visualise both the methods of investigation applied and the stratigraphy of the paint films; each layer has been coded with an identifying letter and linked to a legend, which provides details of its composition as detailed in “Paint microstratigraphy” section. The model, of which Fig. 9 presents a simple example of data extraction, can thus generate an infinite number of search possibilities. Through the selection of a single object (a portion of a wooden flat, a single architectural element, etc.) it is possible to obtain detailed information concerning it. This makes it possible to systematize historical data and results from laboratory analyses, starting from the overall view of the architectural element being interrogated (scene, column, portion) and arriving at the single sample taken. The three-dimensional model data visualisation thus offers an agile synthesis between the overall vision of the selected architectural object backed by studies typically conducted on it by art historians, critics and architects, and a detailed view, at a smaller scale, that characterises the typical contributions of applied mineralogists and petrographers, chemists and restorers working in heritage science. This holistic, multiscale view is one of the innovative results of using HBIM as a tool for investigation and design in the restoration of material and tangible cultural heritage.

Conclusions

The methodological approach adopted for research on the Teatro Olimpico in Vicenza proved the potential of an interdisciplinary study to enhance and maintain a site of significant cultural and historical interest. On one hand, the diagnostic enabled significant observations to be made on the materials and technologies used for the painting of the wooden scenery, identifying the materials applied and distinguishing the original ones from those related to modern restorations of the panels, an important advance evidenced here for the first time. The investigations confirmed complex and varied sequences regarding the microstratigraphic sequence and painting technology, indicating that the wooden perspective flats had undergone numerous retouching and restoration interventions. The laboratory data revealed a rather basic pigment palette, especially concerning the hypothetically older layers (i.e., ochres, green earth and carbon blacks). Indigo is the only exception to this uniformity. The pervasive presence of acrylic polymers, especially in the surface layers, as well as the ubiquitous use of titanium oxide and barium sulphate, are evidence of further modern reworking.

The use of a Heritage Building Information Modelling (HBIM) system facilitated the digitization of all the acquired documents and data related to the Teatro Olimpico in Vicenza, and contributed significantly to identifying and guiding the selection of sampling points through meticulous evaluation of historical information. The analyses performed acquire a peculiar importance by cataloguing all the research carried out, both completed and in progress, for a timely analysis of the historical relevance of the traces of original and modern paint that can guide the conservation process in the operational choices related to them. In addition, the model enables a programme to be developed, detailing new insights and investigations to be carried out to increase the level of knowledge. By keeping the model updated with the restoration work carried out, it will be possible to monitor the response of the materials to interventions on the wooden scenery as a function of the microclimatic indoor conditions of the Teatro Olimpico, to set up monitoring programmes and to plan scheduled conservation. An added value of the methodology described is the high potential for sharing which will avoid future data dispersion (as is the case of the unpublished data on the repainted sections of the scenery) as all those involved in the restoration process will be able to update the information in the digital model as studies and interventions continue. Finally, the HBIM system allows consolidation of all the acquired building data into a unified digital model facilitating the creation of connections and correlations between the information collected.