Adobe Photoshop CS3 Extended

It has been exactly two years since I reviewed the last release of Photoshop, version CS2, Adobe's powerful imaging tool that enjoys near ubiquity when it comes to image editing and compositing in architectural firms. That release included several new features that were especially useful to architects in their imaging tasks, including the ability to define perspective grids in an image and subsequently apply editing actions based on that perspective; a Lens Correction filter for easily correcting many common camera flaws including the wide-angle lens distortion and perspective of an image; the ability to warp an image, which could be used to simulate wrapping a texture on the surface of an object; a Spot Healing Brush for fast and efficient retouching of dust, scratches and other image flaws; and the ability to work with line drawings by importing them through Illustrator as "smart" vector objects.

And now, the new version of Photoshop that has just been released goes even further by having a dedicated product specifically designed for professionals in architecture and engineering as well as those requiring advanced digital imagery in science, healthcare, film, and video. While the standard version of Photoshop CS3 continues to cater to creative professionals who rely extensively on it for their image editing needs, the introduction of Photoshop CS3 Extended marks an important step in the development of the application to make it more focused towards the specific needs and workflows of certain industries, one of them being AEC. Let's explore this new product to find out how it better serves the needs of architects and engineers. Photoshop CS3 Extended includes all the features and enhancements in the standard version of Photoshop CS3 as well, so this review can also be used to learn about the new features of the base application.

Working with 3D Models and Textures

The most important new capability in Photoshop CS3 Extended for AEC professionals is 3D compositing and texture editing. You can now import 3D models into Photoshop that are in any of these popular formats: 3DS and OBJ, which can be exported from most 3D modeling applications; COLLADA, an open XML-based file format for interactive 3D applications that can be directly exported from SketchUp; KMZ, the file format used by Google Earth; and U3D, an open specification which can be directly exported from BIM applications such as ArchiCAD and Bentley's solutions or created by Adobe Acrobat 3D. You can directly open the 3D file, specify the size of the image within which it should appear, and the entire model will be placed on a separate 3D layer. Alternately, if you already have an image file, say a site photograph, you can also choose to create a new 3D layer from the 3D file; this will allow you to see the model against the backdrop of the 2D image that is to be used to provide the context for the model. Figure 1 shows the example of a SketchUp model that was saved in the OBJ format and then directly opened in Photoshop. I also tested this capability with the 3DS and COLLADA file formats, and found that they worked just as well. The size of the Photoshop file containing the 3D model is the smallest for OBJ (6 MB for the SketchUp model shown), followed by COLLADA (7.8 MB), and then by 3DS (8.6 MB); the size of the original SketchUp file was 1 MB. Needless to say, the more complex the model, the larger the file size in Photoshop, and the higher the system requirements needed to work with it effectively.

Once you have the 3D model in Photoshop, there are several aspects of it that you can manipulate as you fine-tune the visual composition of the image. One of the first things that you will probably want to adjust is the position of the model and the camera view in which it is seen. This requires activating the 3D layer containing the model in the Layers Palette by double-clicking on it. This action takes you to a 3D editing mode in which all the regular Photoshop tools and commands become deactivated and cannot be used. The 3D tools appear in an Options bar at the top of the image window, as shown in Figure 1-b. You can toggle between editing the camera or editing the 3D model. The Camera set includes standard tools such as Orbit, Zoom, Pan, and Walk. Using some of these tools to manipulate an architectural model is a little tricky as the tools are designed to work with generic 3D models and do not have constraints such as a fixed Z axis, which is common in tools geared towards architectural modeling. Although holding down the Shift key does constrain these tools to a single direction of movement, it still takes some more effort to adjust the view while maintaining the vertical orientation of the model. A 3D Camera Settings dialog is available where you can view or edit the numeric values for the x, y, and z position, rotation or field of view of the camera. There is also an option to display the model in an Orthographic view at a specified scale. You can create and save views for later use in the Views menu, which also contains preset views such as Front, Left, Top, etc., as well as the Default camera view of the model when it was first imported. The view shown in Figure1-b was saved as a custom view.

A similar set of tools is available to move, rotate, or scale the model itself instead of the view, with options to adjust the position, orientation, and scale numerically. There are also other options for adjusting how the model appears. You can choose from a variety of light settings as well as interesting render modes such as Line Illustration, Solid Outline, etc., in addition to standard modes like Solid, Wireframe, and so on. Two of these are illustrated in Figure 2. Another capability that affords a lot of potential for creative model display is cross-sectioning. This allows you to intersect a model with an invisible plane that slices through the model at a desired angle and displays content only on one side of the plane, as shown in Figure 2-a. This can be useful for displaying cross-sectional plans and sections of buildings if the interior has been modeled. You can also use the cross-sectioning capability to combine two rendering modes on a model, as shown in Figure 2-b. This effect was achieved by first duplicating the 3D layer, setting the cross-section to the same position on each layer, then flipping the cross-section on one of the layers. The two layers can now be set to two different render modes to achieve the dual-rendering effect. The ones used in Figure 2-b were Line Illustration for the left half of the building and Shaded Illustration for the right half.

Photoshop CS3 Extended also lets you access the textures that were applied to the model surfaces in the original modeling application and edit them with the wide range of Photoshop painting and image adjustment tools. As shown in Figures 1 and 2, all the textures in the model appear listed under the 3D layer containing the model in the Layers palette. A texture can be opened for editing by double-clicking on it. Figure 3-a shows the dark grout brick texture applied to the entrance wall opened for editing, and Figure 3-b shows the edited texture and how it is applied to the model after being saved. This capability can be useful for fine-tuning the display of a texture in Photoshop rather than taking it back to the original modeling application. It should be noted that you cannot apply any textures, or even plain colors, to those surfaces of the model that were not textured to begin with. Thus, in the case of the sample model shown in Figure 3, the roof surface did not have a texture applied to it in SketchUp, so it cannot be changed to a different color in Photoshop. Even the Paint Bucket tool cannot be used on a 3D layer until it is rasterized and converted to a 2D image. Thus, Photoshop cannot really be used to texture-map a model—you would have to do that in the original modeling application.

Other AEC-Relevant Enhancements

Recall that Photoshop CS2 had introduced a new Vanishing Point feature that was available as a filter and let you define perspective grids in an image. These could subsequently be used for applying editing actions relative to the grid, and were useful for editing architectural renderings where it is important to be consistent with the perspective in a scene. Photoshop CS3 Extended extends the power of the Vanishing Point feature with some new capabilities. You are no longer restricted to adding connecting grid planes at 90 degree angles—they can be at any angle. A new Measure Tool is now available in the Vanishing Point interface to extract measurements from images. You would use it by first drawing a measurement line over an object on a perspective grid that you know the size of, and specify a length for the measurement. All subsequent measurements on that plane as well as on connecting planes would scale to the initial measurement. You also have the option of re-setting the grid spacing to match with the initial measurement. The measurement line displays the length and the angle that the line was drawn relative to the perspective plane. Both the grids as well as the measurements can be rendered back into the image, if required.

The new measurement feature in the Vanishing Point interface is illustrated in Figure 4-a, where the height of the grid over the planter enclosures was first specified as 5, and then the length and width measurements were placed using the Measure tool. As you can see from the illustration, the measurement results are close but not fully accurate. As far as I recall when I took this photograph, the planter enclosures were square, which is not reflected in either of the enclosures. Also, both enclosures should have had similar dimensions, but even though the height was specified as 5 in both cases, the base dimensions calculated by the Measure tool are quite different.

Another interesting new capability with respect to Vanishing Point is the ability to create a 3D model from a connected set of perspective grids that have been defined for the image. The 3D model will also capture the texture for each surface corresponding to the grid. You can export the 3D model in DXF or 3DS format, or bring it back into Photoshop as a 3D object on a 3D layer. Figure 4-b shows the 3D object derived from the perspective grid over the planter enclosure that was nearer to the camera in Figure 4-a. It should be noted that this capability works properly only if there is one set of grids; therefore, I had to delete the grid over the second planter before generating the 3D object for the first planter grid.

Photoshop CS3 Extended also includes a whole new set of measurement and analysis tools for extracting quantitative data from images that have planar rather than perspective data. While this capability is particularly useful in the medical imaging field, it can also be used in AEC to get measurements from aerial site photographs, plans, elevations, and so on. The process starts with defining the scale of the image using a known value. Figure 5-a shows this being done for an elevation drawing, using the Ruler tool to measure the pixel length of the width of a door in the drawing—which turned out to be 30 pixels—and setting that to a logical length of 3 scale units, which is assumed to be feet in this example. Now that the scale is set, you can use the Ruler tool as well as the Selection tools to measure distances, areas, perimeters, and so on and record these measurements in a log, which can contain data from multiple files. Figure 5-b shows a length measurement of the full height of the building and an area/perimeter measurement of the sliding glass doors at the first level, both of which have been recorded in the Measurement Log. You can customize the log columns, sort data within columns, and export data from the log to a tab-delimited, Unicode text file. A Scale Marker tool is provided for adding a graphic that indicates the scale of the image.

Also included in the Analysis toolset is a new Count tool that allows you to tally features in an image simply by clicking on them with the tool. Each click is marked in the image with a consecutive number, so you can tell which features are already tallied, and you can then send the results to the Measurement Log for further analysis or export.

Other New Features and Improvements

As Photoshop CS3 Extended is targeted towards advanced digital imagery in other fields such as science, healthcare, film, and video as well, it includes various additional features relevant to those fields. Of these, the new video and animation support might also be useful in AEC. Film and video professionals typically use Photoshop to generate content for visual effects or to retouch individual video frames. In the past, they had to export frames as still Photoshop files first, or create a graphic in Photoshop and then animate the image in other software. Photoshop CS3 Extended expands and simplifies this workflow by providing new video-format and layer support which allows a video file to be edited on a frame-by-frame basis, or a layer to be added to the video for creating edits that will appear on every frame. Instead of each frame of video being imported as a separate layer, as in the past, the entire video file becomes one video layer. A redesigned, enhanced Animation palette lets you navigate the video layer and edit based on time, as video and motion graphics professionals typically do. For shorter projects, such as a brief animation file, an additional Frame mode lets you navigate the file on a frame-by-frame basis. Video files can now be imported and exported in a wide range of formats including QuickTime, AVI and MPEG-4 video files, and Adobe Flash video.

Since Photoshop CS3 Extended is built on top of Photoshop CS3, it includes the host of improvements made to the base application as well. These include Smart Filters, which are Photoshop filters that are applied nondestructively to layers, and which remain live and re-editable. Once a layer is converted to a Smart Object layer (the concept of Smart Objects was introduced in Photoshop CS2), any filter that is applied to it automatically becomes a Smart Filter. You can apply multiple Smart Filters to a layer—they appear in the Layers palette below the Smart Object layer to which they are applied, from where they can be selected, edited, or disabled. You can also choose to apply the Smart Filter effects to only selected portions of an image as opposed to the entire image by creating a filter mask. Figure 6 shows a collection of two Smart Filters applied to a selected area of an image. This new capability allows you to freely experiment with different filter combinations and settings, without having to back up and start over with the original image.

Another useful feature in Photoshop CS3 is automatic alignment and blending of image layers that makes advanced image composites faster and easier to create—for example, combining the best parts of multiple images of the same scene into one "best" image. The multiple images would be placed on separate layers in one document; the Auto-Align Layers command would then be used to make a composite photo. It works by analyzing the contents of all the layers and moves and rotates them so that they overlap as precisely as possible. Layer masking tools can be used to reveal areas from different layers and finalize the elements of the composite. Finally, the Auto-Blend Layers command can be used to seamlessly blend the color and shading into the final image. The Photomerge command, which is used to combine several photographs into one continuous panoramic image, also incorporates this new layer alignment and blending capability. Now, in addition to finding the overlapping areas of images and assembling the panoramic composite, Photomerge actually rotates and transforms images to create more seamless results. Improved blending helps smooth the transition between images in the final panorama, especially noticeable with the slight differences in lighting captured from one part of the panorama to another (see Figure 7). The improvements in Photomerge are particularly useful in the AEC field, where it is a common requirement for multiple site photographs to be assembled into one consolidated image capturing the context of a proposed building design.

Other enhancements in the base application include improved color correction features including better brightness and contrast; one-click conversion of color images to monochrome, and more powerful cloning and healing tools for retouching images. A new Quick Selection tool makes it faster and easier to select parts of an image. The interface has been streamlined to maximize screen space for editing while keeping essential tools accessible. Palettes are now arranged in self-adjusting docks that can be widened to full size or narrowed to icons, or even to a thin, self-revealing strip at the edge of the screen. You can set up and save custom menus, keyboard shortcuts, and workspaces for quick access to frequently used tools, and new preset capabilities allows sophisticated color adjustments to be made in one click. Photoshop CS3 is faster by virtue of compatibility with the newest systems, including native Mac OS X 10.5 (Leopard) on PowerPC and Intel processors, and Microsoft Windows Vista. It also features improved printing controls, better support for Adobe PDF-based content, and improved integration with other Adobe tools including Dreamweaver, Flash, and After Effects.

Analysis and Conclusions

Photoshop CS3 Extended marks the beginning of the evolution of Photoshop as a specialized application for a certain number of fields including AEC that have some specific imaging requirements in addition to being a broad-based product for graphic design and image compositing in general. The new ability to import 3D models and manipulate aspects such as camera views, the size and orientation of the model itself, lighting, render modes, and the textures applied to the model open up a whole new slew of possibilities for the use of Photoshop in architectural design. The final look of the model in a rendered image can now be fine-tuned in Photoshop itself when the model is being composited with other images such as a site photograph. Architects will especially love the cross-sectioning capability which allows them to display not only sectional views inside their buildings but also combine the use of two rendering modes on a model, which can lead to interesting and creative results.

With regard to the new Vanishing Point capability of enabling measurements in 3D views, I found this less compelling as the grids really need to be set up properly to get accurate results. In contrast, the ability to measure distances and areas in planar drawings and views and record these measurements in a log should be more useful, as the results can be very reliable. The other new ability in Vanishing Point to create a 3D model from a connected set of perspective grids is technologically very interesting but its practical value remains doubtful, especially because it is limited only to one set of connected grids. If multiple disconnected grids have been created over the image, the resulting 3D model is not generated correctly.

It is interesting to see the direction that Photoshop is taking from an AEC perspective, given the fact that Autodesk has just entered the "presentation drawings" arena with its new Autodesk Impression product, which was reviewed in AECbytes earlier this month. While Impression has no special capabilities for working with 3D views, it makes the illustration of 2D drawings a snap with a near-infinite variety of artistic styles that can be quickly mapped to objects or layers, along with advanced features such as style mapping and block substitution. Thus, if the objective is to illustrate 2D drawings alone, Impression would be a better choice of product. In a world increasingly going 3D, Photoshop is certainly doing well to find ways of working with 3D models. However, just as with Autodesk Impression, it still doesn't come close to matching what Piranesi can offer for dedicated architectural rendering. Hopefully, future versions of Photoshop can continue to build upon the terrific start it has made with 3D and provide more capabilities such as the ability to adjust the texture maps on the surfaces of the model, apply textures to surfaces for which textures were not defined in the original modeling application, provide more lighting and display options, and improve the 3D navigation capabilities so that they work more intuitively for building models. Photoshop has already established itself as the undisputed leader when it comes to tasks such as painting, compositing, and image editing, and it ranks high on good memory management for working with large files. More powerful 3D capabilities would take it to new heights of mission-criticality in AEC.

From a cost perspective, the $999 price tag of Photoshop CS3 Extended as opposed to $649 for Photoshop CS3 does not make it significantly more expensive. (The corresponding upgrade prices are $349 and $199 respectively.) The numerous enhancements made to the base product itself such as Smart Filters, improved Photomerge with automatic layer alignment and blending, improved color correction features, and faster performance through compatibility with the newest systems already make Photoshop CS3 a compelling upgrade, and I can see architectural firms choosing to upgrade at least a few of their copies to the Extended version in order to take advantage of the new 3D capabilities it has to offer.

About the Author

Lachmi Khemlani is founder and editor of AECbytes. She has a Ph.D. in Architecture from UC Berkeley, specializing in intelligent building modeling, and consults and writes on AEC technology. She can be reached at lachmi@aecbytes.com.

Reviews > Adobe Photoshop CS3 Extended > Printer-friendly format