Tom Arah explores the vector and bitmap capabilities in Adobe Illustrator and Adobe Photoshop and how to make the most of them.

Photoshop and Illustrator are the twin graphical pillars of Adobe’s Creative Suite. As standalones they each offer their own unique mix of state-of-the-art vector and bitmap-based handling but, for maximum creative power, their respective strengths can also be pooled.

In fact the Creative Suite only makes full sense if the end user makes full use of both Photoshop and Illustrator both separately and together. To understand just what power they offer and how best to put it to use, it’s necessary to see where it came from in the first place.
Moreover it’s important to realise that things were once very different. Back in 1992, when I first got my hands on Photoshop and Illustrator after their recent port to the Windows platform, these were entirely different beasts. To begin with, meaningful integration was not an issue – in fact the two programs were presented as opposite poles that were best kept apart. Generally users worked with either Illustrator’s vectors or with Photoshop’s bitmaps and never the twain should meet.
Each application was also unrecognisable from the creative powerhouse that it has become today. This was particularly evident with Illustrator, which despite its name and the use of the famous Botticelli Venus logo, was in no way an application for artists. It’s difficult to believe now but, even in its fourth release, Illustrator’s creative formatting options were beyond dismal. You could apply flat CMYK or Pantone-specified colours and uniform width lines that were either solid or dashed and that was about it.
The reason usually given for this poverty of creative options was that Illustrator was a vector application and that, if you wanted creativity flexibility, you should be using the bitmap-based Photoshop where the underlying unit wasn’t the mathematically-defined open path and filled shape but the individual pixel. But that didn’t wash. In particular the same limits simply didn’t apply to Illustrator’s biggest rival on the PC, CorelDRAW, which happily mixed imported bitmaps and vector objects complete with bitmap fills and bitmap effects including advanced graduated transparency. Features that other programs, most notably Xara and Macromedia Fireworks, took further to produce creative, even naturalistic, bitmap-based images built on vector-based handling.
The bottom line is that, while vectors and bitmaps are very different, if you want to be seriously and artistically creative they are best combined. After all, all graphical output is eventually rasterised for bitmap-based output whether that’s to a screen or printer. Moreover all hands-on input originally starts off as vectors – after all that’s how the computer monitors your movement of the mouse turning it into a vector path along which a brush’s pixel-based effects are applied. Seen properly, vectors are the means to a bitmap-based end, so why did early versions of Illustrator so carefully avoid all bitmap-based handling and the creative power that went with it?

Early versions of Illustrator were focused on high resolution, CMYK, PostScript-based output

The answer is that Illustrator isn’t just a vector application, it’s a PostScript-based vector application. PostScript is the seminal Page Description Language that Adobe devised back in 1984 which enabled typefaces to be described as scalable outlines so enabling resolution-independent output. To describe the shapes that go up to make letterforms you don’t need any formatting beyond solid fills so PostScript’s creative capabilities were rudimentary. However they could still be used for more than just text and the first, Mac-only release of Illustrator, launched in 1987, was Adobe’s way of extending PostScript to handle graphics. The tie-in was such that each Illustrator file was effectively a self-contained, “encapsulated” PostScript file that could be output directly to a PostScript device.
PostScript explains Illustrator’s original creative weakness, but it also provided the program with its greatest strength. By avoiding bitmaps and bitmapped effects in its drawings, Illustrator could completely avoid rasterising its output in the way that more creative rival applications had to. This meant that Illustrator files could be saved to a vector-only version of EPS (Encapsulated PostScript) that could then be incorporated in a layout within a DTP program. Only when output from the DTP application would the embedded EPS file finally be rasterized according to the maximum resolution of the PostScript output device – absolutely crucial for high resolution imagesetting. This also explains the early Illustrator’s limitation to flat CMYK fills as this was necessary to cleanly and reliably separate the artwork into the printing plates necessary to produce four-colour process print. In other words, when producing high resolution, colour-separated commercial print, Illustrator’s creative limitations became its production strength.
This PostScript underpinning explains the apparently fundamental early division made between Illustrator’s vectors and Photoshop’s bitmaps, and it also gives an insight into where early versions of Photoshop were coming from. Again a current user would be bewildered when presented with Photoshop 2.5, the first Windows release. While the core colour correction capabilities would be very familiar, the lack of creative options would be astonishing. In particular the whole concept of creative photo-montaging for which “photoshopping” has today become synonomous would seem near-impossible fifteen years ago for one over-riding reason – before the advent of layers, all images were flat.
In fact with some expert cut-and-paste, montaging was just about possible thanks largely to some early vector path handling that Adobe had grafted on to Photoshop 2 for creating and storing selections. The overriding use of Photoshop though was simply to enhance and output flat photographs. What made it stand out compared to rivals was that it was designed from the ground up to output flat CMYK images with comprehensive control over RGB-to-CMYK conversion, onscreen proofing and gamut warnings. Usually these CMYK images were stored and exchanged in the efficient and flexible TIFF format but for more advanced work, such as handling duotones or spot colours or using those vector selections as clipping paths to produce non-rectangular bitmaps, Photoshop could also output to a dedicated, bitmap-oriented, pre-separated version of EPS.
In other words, Photoshop was designed to handle the production of embedded ink-based bitmaps ready for PostScript-based colour-separated commercial print, just as Illustrator was designed to handle the ink-based vectors. Illustrator and Photoshop were not intended to be creative applications in their own right let alone working together – rather they were devised as separate production tools designed to feed through PostScript-friendly photographs and illustrations for use in the text-oriented, multi-page DTP application that lay at the heart of the publishing workflow. In short, Adobe’s original interest in graphics was based on their publishing role rather than for their own sake.

This central focus on publishing and so on PageMaker and later InDesign still lies at the heart of the Creative Suite but, over time, the balance has shifted and both Photoshop and later Illustrator have burst into creative life. This process started with the ground-breaking release of Photoshop 3 in 1994 which introduced layer-based handling. Rather than handling image elements via complex paths and alpha channels this could now be done with simple layers so unleashing the photo-realistic compositing for which Photoshop is now famous. In addition the introduction of layers also opened up entirely new creative options through the use of opacity levels, masks and blend modes. Suddenly the base photograph had become just the starting point for work within Photoshop.
Bitmap-based layers were only the beginning, soon followed by Adjustment Layers for applying non-destructive colour corrections and Layer Styles for applying non-destructive effects. In another ground-breaking release, 2000’s Photoshop 6 combined the existing path, layer and layer style handling to enable the creation of vector-based Shape Layers and Text Layers. Photoshop’s hands-on pixel-level editing was now augmented with some considerable vector-based non-destructive editability and even a level of scalability. As a direct result the sheer number of creative possibilities exploded requiring yet another layer-based innovation to keep on top of them – the Layer Comp, a snapshot of the current position, size and blend mode of all layers introduced with 2003’s first Creative Suite release, Photoshop CS.

With the introduction of vector-based Shape Layers Photoshop 6’s creative compositing options extended even further

Naturally all this layer-based creative power needed to be stored, so increasingly work was saved to Photoshop’s own layered PSD format, usually in RGB mode. Only when all creative work was finished was the multi-layered PSD converted and “flattened” to produce the single-layered, PostScript-oriented CMYK TIFF or EPS that was to be incorporated into the DTP layout. It was a delayed approach to working with PostScript that Adobe began extending to Illustrator, starting with 1997’s version 7. Crucially this release finally broke with Illustrator’s CMYK and vector-only past to embrace RGB colour handling complete with support for rasterisation and even the application of Photoshop filter effects to embedded bitmaps. It also introduced vector brushes that remained fully editable as stroked paths when stored within Illustrator’s native AI file format but were “flattened” and treated as filled shapes when output to EPS.
Illustrator 7 was no longer a vector-only, PostScript-throughout editor and its creative options began to improve but still, compared to more flexible and bitmap-friendly rivals, its output looked appallingly flat – inevitably so when its formatting options were still restricted to those flat fills. Eventually this limitation began to be eroded with 1998’s version 8 which introduced gradient mesh handling which massively boosted colour shading capabilities while still managing to avoid rasterization and so maintaining full vector scalability and quality when output to PostScript 3 devices.
Illustrator’s limited formatting capabilities were finally blown out of the water with 2000’s version 9 which at last brought in the ability to apply transparency effects. Crucially, whereas rivals such as CorelDRAW tended to rasterize the results of its transparency effects when printing or outputting to EPS, Illustrator again chose the more difficult but rewarding route of flattening– breaking down the artwork into as many discrete shapes as necessary each with their own flat or gradient fills. This approach maintained the vector nature necessary to ensure scalability and maximum quality for imageset output and commercial print.
Just as importantly, Illustrator 9 also introduced the ability to apply bitmap filters non-destructively to both bitmap and vector elements and to combine multiple fills, multiple strokes, transparency, vector brushes and filter effects as non-destructive styles. Whenever creatively necessary Illustrator 9 now provided the bitmap-style and bitmap-based effects that ensured that, from nowhere, the program suddenly seized the creative high ground compared to its rivals. At the same time, wherever possible, Illustrator 9 kept true to its original PostScript-inspired mission and maintained as much vector scalability as possible in its flattened EPS output. It’s a balance that the program maintains today.

With the advent of PostScript-friendly gradient meshes and vector transparency, Illustrator burst into full creative life

By shifting from its original PostScript-only, flat file formats and publishing focus, and moving to the rich AI and PSD formats for ongoing work and then flattening the results when outputting to the PostScript-friendly EPS and TIFF exchange formats, a vast amount of combined vector and bitmap-based creative power was unleashed. However the shift went against Adobe’s original vision of the universal, PostScript-based publishing workflow from beginning to end. In particular having separate native and exchange formats adds greatly to workflow complexity.
In the longer term it’s likely that Adobe will solve the problem by resurrecting its original universal PostScript vision, but this time based on PDF (Portable Document Format). PDF is built on PostScript and can act as replacement to the EPS exchange format but it is superior in two main ways. Firstly unlike EPS, PDF has been designed to be viewed onscreen as well as output to paper. Secondly PDF isn’t limited to storing PostScript information so it can maintain the editability of features such as transparency, brush strokes and RGB handling while enabling the eventual flattening and conversion to CMYK which is necessary to ensure successful high resolution PostScript output. In fact it’s already possible to save PSD and AI files to PDF while maintaining all editability and when InDesign falls into line, a single universal PDF creative file format for ongoing work and fixed exchange should emerge.
In the meantime the focus is simply on closer integration. In particular, by enabling InDesign to directly import both AI and PSD native file formats, file handling is kept much more manageable and, if you need to edit the linked image, it’s just a case of selecting the Edit Original command. In addition such integration means that advanced features such as vector-based transparency remain scalable and only need to be flattened during final output so ensuring maximum quality. It also maximizes flexibility so that, in the case of PSD files, you can quickly swap between layer comps to try out image variations. Finally it enables some unique in situ formatting so that if you apply the drop shadow or feathering effects to an imported AI file, these are applied not to the surrounding bounding box as you might expect, but to the actual objects.
The integration with InDesign is crucial to the publishing production workflow, but for maximum creativity it’s the integration between Illustrator and Photoshop that is most important. Unsurprisingly this starts with the ability to place linked PSD files in Illustrator documents complete with support for all advanced features such as transparency and spot colours and a choice of any embedded layer comps. You can also choose to embed the PSD which breaks the link and provides the option to convert the image to a single flattened layer or to convert each layer to a separate object – text layers even become fully editable. When moving work from Illustrator to Photoshop the usual route is simple cut-and-paste which has long offered two main choices: the occasionally useful options to copy the vector framework to either Photoshop paths or shape layers or the much more regularly useful option to rasterize the artwork complete with anti-aliasing at the size that you specify when you hit the Commit button.
These capabilities are impressive and allow the vector-oriented Illustrator to gain the benefits of Photoshop’s more advanced bitmap controls and vice versa. Inevitably though, as we’ve seen, bringing in the creative flexibility offered by bitmaps immediately loses the benefits of vector scalability. Or that’s what you would think. Brilliantly in the last release of the Creative Suite, Adobe found two very different ways of further merging vector and pixel handling while maintaining scalability and so maximum output quality.

In Illustrator CS2 this took the form of the new Live Trace capability. Now when you place a PSD file (or any other bitmap) in Illustrator a new Live Trace command becomes available on the toolbar. Click on this and you can choose between a range of presets or open up the main Live Trace dialog where you can specify just how the bitmap should be converted to vectors. The control is extraordinary letting you set image blurring, path fitting and so on and even providing an option for centreline tracing. Even more impressive are the end results. Select the Photo High Fidelity preset for example and, although this defaults to just 64 colours, the results usually compare surprisingly well to the original but with an extra creative twist.
And of course with the added advantage that the results are now vector rather than bitmap-based. This means that, if you expand the results, you can manipulate them with all of Illustrator’s many tools – for example applying non-destructive effects or vector brushes. Even better, when you export the results to EPS or PDF, or load the AI file into InDesign, the results are fully scalable and will output at the maximum resolution of the printer or imagesetter. And remember that the tracing is live so that if you edit the original PSD (easily done via the Links palette) all changes automatically update.

Illustrator’s Live Trace capability can be used to convert bitmap photos and line drawings to vectors on-the-fly

By contrast Photoshop CS2’s most impressive new feature is its support for so-called Smart Objects. These can be used for handling Photoshop’s own bitmap layers, but come into their own when handling Illustrator’s vectors. In particular when you paste Illustrator artwork you now have a new option to store the results as a Smart Object or, if you use the new Place command to load an AI file, one is automatically created for you.
The fact that Illustrator’s original vector information is now embedded within your PSD offers two great advantages. Firstly you can reload it into Illustrator and edit it with all changes automatically being updated (though surprisingly the link to the original AI isn’t maintained so updating can only be done from within Photoshop). Secondly you can scale the Smart Object up within Photoshop and its maximum quality is automatically maintained – vector-style scalability within a bitmap environment. Based on the Photoshop CS3 beta it looks as if another benefit is on the way with the ability to directly apply filters non-destructively to the embedded Smart Object AI – though of course that’s already an option within Illustrator.
Live Trace and Smart Objects might look very different at first but they are actually very similar. Live Trace overlays the embedded Photoshop PSD artwork with a vectorised version while Smart Objects overlays the Illustrator AI artwork with a rasterized version – in fact in many ways the Smart Object capability is best understood by thinking of it as a “Live Raster” layer. By automatically converting Photoshop’s bitmaps to vectors and Illustrator’s vectors to bitmaps they each enable their host application to take advantage of its partner’s complementary strengths while maintaining the unique benefits of their own primarily vector or primarily bitmap approach.
Ultimately the fundamental distinction between vectors and bitmaps remains and needs to be recognized, especially when producing work for high-resolution PostScript-based commercial print. However with the ever-increasing integration between Photoshop and Illustrator this underlying distinction is no longer an unbridgeable division or obstacle to creative expression. In fact in many ways the many bridges between the two worlds lead to even more creative possibilities.

PostScript: Despite their increasing integration, Adobe Photoshop remains focussed on bitmaps and Adobe Illustrator on vectors. However there is a third way as Macromedia Fireworks showed with its bitmap end results created on a fully editable vector framework. For those concerned that Fireworks would quietly vanish with Adobe’s takeover of Macromedia, there’s excellent news: a new version of Adobe Fireworks 9 is currently out in private beta. Whether Adobe gives it the development it deserves is yet to be seen, but Fireworks has the potential to become the third graphical flagship in the Creative Suite.