PREVIOUS   NEXT   CONTENTS   HOME 

3.5 Data formats for display and analysis

After any needed colour correction, a high-resolution OBJ file was exported. It is expected that this format will be the primary one for scholarly analysis purposes as it preserves the maximum amount of 3-D and texture detail. However, these OBJ files are quite large, often 50-100 MB or larger. As a result it was felt that additional formats were needed. One was a low-resolution OBJ. The low-resolution OBJ preserves most of the visual properties of the object and can be readily used in other animation and 3-D software, as described in more detail below. To develop the low-resolution OBJ the models were decimated or reduced to a standard polygon count of 25,000 triangles for full colour models and 100,000 triangles for gray scale models (without significantly altering the shape of the mesh surface) and a low resolution OBJ file for the object was exported. These values were chosen as a good compromise between reasonable performance in most software and quality of the visual product, remembering that the high-resolution OBJ was always available if needed. Because of the ubiquity of VRML/X3D in many web applications this was selected as a third format. The low-resolution OBJ was then imported into Vivaty (originally Flux) Studio and used to generate the VRML/X3D file.

As we discussed above, the Virtual Hampson and Amarna Museums are intended for the general public as well as for scholars interested in studying the collection. After reviewing a range of options we decided that the primary mode for the general public to access the collections would be through the use of Adobe 3D-PDFs, though the other formats are also available. Adobe 3D PDFs were chosen because they have several advantages over the more traditional 3-D viewers. 3D PDFs are easily embedded into a webpage and are cross-browser and cross-platform compatible. Most online 3-D data viewers require the end user to download a plugin that will often only work in a desired platform/browser configuration (i.e. Internet Explorer on Windows). Instead Acrobat 3D is cross-platform (Mac and PC) and cross-browser compatible (Internet Explorer, Firefox, Chrome, Safari, and others). 3D PDFs only require the user to have Adobe Reader 8.0 or higher. The online viewer might require the user to download the latest version of Adobe Reader. But, because it is a software with which they are likely to be familiar, it was assumed that the user would be more comfortable with this option as opposed to downloading a viewer from a third party. In addition, there are a number of advanced tools in Adobe Reader, including measurement and cross-sectioning. Finally, it retains the scale of the artefacts carried over from the digitisation process allowing end users to actively measure and perform cross-sections on the artefacts; a functionality that is currently unparalleled by any other free, 3-D viewing application. Going forward it is possible that emerging HTML5 based applications may replace the Acrobat 3-D viewer, but that is not true at this time.

To achieve reasonable access times for the PDF download it was necessary to reduce the polygon count from the high-resolution OBJ mesh versions. Typically, PDF files were reduced to 5 MB, a size that streams reasonably quickly over many connections. In these and the other various decimation steps an essential part of the process is to preserve the original image detail of the artefact. When the object is scanned by the VIVID 9i, the colour or RGB information is tied to the individual vertices of the model (XYZRGB). When the model is decimated at this stage, reducing the triangle and vertices count, the colour information is also decimated (see Fig. 3) leaving an unrealistic appearance.

Figure 3

Figure 3: (Top) Decimated object with colour information not preserved, (bottom) decimated object with original colour image preserved and re-applied. [View object record in tDAR]

In the workflow initially developed for the Hampson Museum and then also applied to Amarna, prior to decimation, Rapidform was used to extract the RGB information from the original model and assign it to a separate JPG file. The model can then be decimated and the original image texture preserved and reapplied to the reduced model.

In addition to the various 3-D file formats a short video illustrating the full 3-D nature of each of the objects was created to be part of the representation on the web. The video animation for the artefact was prepared using the high-resolution OBJ file in Polyworks IMInspect. All videos were rendered as AVIs in 800X600 resolution and later converted in a batch operation to Flash FLV format with 400X300 resolution.

Figure 4 Figure 5

Figure 4: Final design for viewing artefacts in the Virtual Hampson Museum.
Figure 5: Final design for viewing artefacts in the Amarna Virtual Museum.

In total, each artefact from the Virtual Hampson and Amarna Museums is available for download in four different formats (high-resolution OBJ; low-resolution OBJ, VRML/X3D, and 3D-PDF). A variety of download options are provided to the user in an effort to streamline access to the digital Hampson and Amarna collections (see Figures 4 and 5).

The comments facility has now been turned off.


 PREVIOUS   NEXT   CONTENTS   HOME 

© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions | File last updated: Tue Jun 28 2011