The modal interface of a plug-in is fine for working with a single source image, but it's not really appropriate when working with multiple source images. The modal interface restricts your ability to switch to another process to rename or manipulate files, or to perform some other operation. If you were to begin a process within a plug-in, open a file or two, then realize that you needed to move or rename another file, you would have to exit the plug-in, do your work, then start from scratch again. A plug-in also has restricted access to memory. By making Maskerade a full blown application, we have no limitations imposed on us by Photoshop. It's also very possible that some Maskerade users may not be using Photoshop for editing. There are many competing products to Photoshop that are much better for certain applications.

To make your job easier, however, Maskerade exports native Photoshop format files so you can simply double-click your images and open them directly in Photoshop, complete with transparency and layers.

Because one is not enough and three would be too much work. Seriously, with only one image, there's just not enough information to determine which pixels represent the foreground object, which pixels represent the background, and which pixels have a contribution from both the foreground and the background. We'll explain further.

In an image, every pixel has color contributions from the foreground object and the background. In the simplest example of a completely opaque object, the pixels in the center of the object have contributions from the foreground object only, while the pixels surrounding the object have contributions from the background only. However, the pixels along the edges of the object, where the object 'touches' the background, will have contributions from both the object and the background because of something called 'anti-aliasing'. This anti-aliasing occurs because the image capture device (or in the case of film, the media) has a limited resolution and cannot reproduce the edges sharply enough to insure that a pixel is either purely representative of the foreground object or purely representative of the background. It is this 'anti-aliasing' that causes problems when masking.

If you think about an example of a red apple photographed on a white background, you have no problem using the magic wand in Photoshop to select the white pixels which represent the background, or the red pixels which represent the foreground. Where you have difficulty getting an acceptable result is along the edges of the apple where it blurs into the background. This is because of the anti-aliasing that we just described. In the image that you're trying to obtain, you want to maintain that 'blurred' edge in the masked object so that it can be composited onto a different color, yet still have the accurate look of the original image.

In the case of transparent objects, things get more complicated. It is very difficult to determine what is part of the foreground object and what is part of the background. Blue screen attempts to use isolated colors to determine this difference, but it is not correct as anyone who has ever used blue screen techniques for high-end work will attest to.

Mathematically, you are trying to generate four channels of output information (red, green, blue and alpha) from only three channels of input data (red, green and blue). This leads to an indeterminate equation that cannot be solved. So, some assumptions are made and a guess is made at a result. In many instances, like television, the results are visually ok since they are masked by a mediocre resolution and format. But in high end photography, this is not acceptable.

Yes and no. Yes it takes a little extra time to take a second shot, but by doing so, you are completely eliminating the hours of editing time that you have to spend in Photoshop removing the background. In most cases, once the shot is setup properly, changing the background only takes a few seconds. Maybe two or three minutes at most. We think a few minutes in the studio with a skilled photographer to get a perfect image far outweighs the hours of editing in Photoshop that usually results in a mediocre image.

Actually, there are no edge problems like other methods that you're familiar with. Maskerade doesn't select pixels and delete them like you do with the magic wand or lasso in Photoshop. And it doesn't create hard-edged cuts like clipping paths do. Rather, Maskerade creates a totally new image with the correct colors and transparencies by extracting information from the two source images. Nothing is deleted then de-fringed.

Maskerade can save an image as a native Photoshop document or as a PICT file with an alpha channel. If you are doing photo-compositing, nothing else needs to be done with the image. If you want to immediately place it into a Quark document for printing, the image needs to be flattened onto a background of your choice, then converted to whatever color mode you need the image in, such as CMYK. These steps are performed outside of Maskerade in an image editing package such as Photoshop.

No, Maskerade does not currently support CMYK images. Why not? Because CMYK images are not easy to produce. Most users think that you just select the CMYK mode in Photoshop and send it to the printer, but most users have little idea what they are doing wrong here (sorry, we're not trying to bash anyone...). Doing CMYK correctly is really difficult and we're not going to put out a half-baked solution when there are already good tools out there to work with.

CMYK is a device-dependent color mode. That is, you have to know the characteristics of the device that you are going to print an image with to know exactly how to convert your RGB image into that device's CMYK color space. This is why a CMYK image looks different on every device you print it on. Each device prints its own distorted version of the CMYK image.

Anyway, there are many tools out there to do correct RGB to CMYK conversions. Many, like Photoshop, are scriptable so they won't slow your workflow down. The new ColorSync 2.5 release is even scriptable, so you don't even need a program like Photoshop to process your images. All you need are the device profiles for your camera, monitor and printer to get excellent results.

Maskerade creates transparent images. Unfortunately, transparency and clipping paths are not the same. Clipping paths are hard edged "knife" paths that knock out selected areas of an image. They cannot be used to simulate the transparent parts of something like a glass object because pixels are either turned on or off by the clipping path. There are no in-between values.

If you mask a completely opaque object that has well defined edges, however, you can generate a clipping path from the resulting mask in Photoshop. Check out the on-line help included with the application for more information on this.

We might add this feature in a future release, so we're not ruling it out completely.

No, Maskerade requires two source images photographed a specific way (well, not that specific). If you have an existing image, you have to resort to old fashioned methods. Sorry. It's just that Maskerade needs info that can only be gathered from multiple source images.

No. The analytical process is completely different from blue/green screen techniques. Maskerade can produce detail that is impossible to capture with blue/green screen techniques. The only thing that they share is the requirement to shoot your source images on a specific background.

Your subject needs to remain perfectly still from one image to the next, and no person we have ever met can remain perfectly still. When we say perfectly still, we mean perfectly still. If the person moves at all, the hairs on their head (and maybe elsewhere) would move out of registration in the second shot. This is why Maskerade is really suited to still objects only.

No, but backgrounds of these colors are commonly found in any photographers supply room and they will yield the optimum results. But if you must, you could shoot on pink and green, yellow and purple, or any other two color combos that have sufficiently different luminosity characteristics.

Not really. The reason is that only a couple of compression codecs support alpha channels which are needed to produce the effect that you see. These codecs use "lossless" compression techniques which maintain much higher image quality, but are not as efficient as lossy compression schemes like JPEG.

It really isn't. Maskerade doesn't mask an image like Mask Pro or Magic Mask. It creates a new image using information from the two original images. However, our target market consists of the individuals that are using traditional masking tools now because these tools are the only solutions available. We want to be associated with these tools since we are an alternative to them for many photographers and designers.