Hello –

We hope we find you well and enjoying a Happy Holloween! We have been busy here this past month. First of all, we went to Photoshop World in Miami to show off the Asiva Plug-ins and there we debuted the Shift+Gain plug-in. You can see the pictures from the event posted on the forum. We are now currently working on another plug-in that will use the Asiva technology and mapping techniques to create selections in Photoshop. This is for all those people who just can’t move away from masking. More importantly, it can give you selections directly in a 16-bit per component file. This is without making a duplicate 8-bit file and transferring a selection!

Asiva.com has also gotten a face-lift. Well, it has changed looks completely and we are hoping it is a little more user friendly and easier to load or move around in. Check it out and let us know what you think.

In this newsletter, we go deeper into the value of good pre-production and in honor of the Shift+Gain plug-in being released; we have an introductory article on the differences between shift and gain. Enjoy reading and we’ll be back next month.

Shapiro Consulting Group, Inc.

Our latest plug-in was released and announced at Photoshop World in Miami at the beginning of October, with great success. The director of photography for a major U.S. corporation said it was the best software he saw at the show besides Photoshop CS itself. The shift and gain filters in this plug-in were taken directly from Asiva Photo's Shift and Gain Operations. The following discussion about shift and gain applies to the Shift+Gain plug-in as well as Asiva Photo’s Shift and Gain Operation.

Lets get past the basic terms that seem to confuse users: shift and gain. In the context of digital image processing, shift means to add or subtract an amount to or from a given component of a digital image. For example, shift the red component (or channel) by 6,000 would mean to add 6,000 to the red component for every pixel. We are assuming a 16-bit representation, of course, where the red, green, and blue channels for each pixel are each represented by a value ranging from 0 to 65,535. Shift red -6,000 would conversely subtract 6,000.

What does this do though? If we were to look at the red component on a histogram, it might look something like this:

Shifting red by +6,000 would result in the histogram looking like:

But what does the image look like? In our example, the image would not only get redder, but would increase in brightness as well, since working in RGB changes color as well as light level. Everything would get redder, especially those pixels that were already primarily red in hue.

Shifting the value or luminance component is like the brightness adjustment on a TV - everything in the image will get brighter or darker.

Gain behaves quite differently from shift. Gaining the value component with Asiva gain is basically similar to a TV's contrast adjustment. Gain is really no more than multiplying values or dividing them by some number. If brighter values are represented by higher numbers and darker values are represented by low numbers, then multiplying by the same amount would impact the HIGHER number more than the lower numbers. That is why it appears to increase the contrast - because the whites get whiter but the shadows stay almost the same.

Taking the same histogram example above, lets say we will gain the red channel by 1.5. This means multiply all red pixel components by 1.5. For pixels that have relatively high red values, the red will get redder, but for pixels whose red values are small the red increase is minimal. On a histogram we would see:

Basically we have stretched the histogram to the right, anchored from the very left side (the 0 value of red). Notice that the ‘hump’ has reduced in size. That is because many of the pixels that made up the ‘hump’ are now towards the right side of the histogram.

The left side has not moved like a shift would produce. That is because a sift will modify all component values the same amount, regardless or their original values. A gain, on the other hand, will always affect high values more than low values. This is true whether you are gaining up (multiplying) or gaining down (dividing).

In the most general sense, if you want to impact high values of some component or channel but not so too low values, use gain. If you need to alter low values, typically to increase them, use shift.

In next month's newsletter, we will give very specific examples for the appropriate use of the Asiva Shift and Gain Operations.

The Value Of Good Pre-Production II

Two common misconceptions about proper pre-production planning is that most problems can be fixed in post with technology and, at the other extreme, everything has to be done during the production process.

The result of these two disparate thought processes is that the first misconception results in problems that cannot be fixed cost effectively. This is due to a of a lack of understanding of the technology available to correct the problems. The best case is that it will be costly in terms of post-production time and the worse case is a re-shoot. The second misconception results in an inordinate amount of time and expense spent during the production process to ensure that the shot is set up to perfection before the shutter is triggered. This second option, although safer, really limits your flexibility in terms of the final result. Very often what you shot is all that you end up with.

A classic example of this is the studio photographers trick of changing the background scenery depending on the requirements of the client. It used to be a simple changing the backdrop during the shoot, but often once the backdrop is chosen, it is practically impossible to change after the shot is taken. To solve this problem and to allow for more flexibility, technologically savvy shooters have been using blue or green screen backdrops so that they can be easily replaced with a desired background during the post-production process. This technique which is old in the video world and is still very valid today, is at best still relatively new in photography. Of late, with photographers switching to high quality digital cameras, this technique of 'fixing it in post' has been given new validity, as you no longer have to go through the convoluted and lossy process of developing and scanning the image before you can make use of today's advanced post-production software technologies.

In the video world, special effects personnel and post-production supervisors are costly to get involved during the pre-production process. This is because cutting out backgrounds and compositing at 25 or 30 frames a second, is a very complicated process. The simple making a mask or matte, which is a common in a still frame world, is just not viable when you're dealing with moving pictures. This is due to our human visual physiology, making us more sensitive to movement than to changes in color. As a result, a mask that appears perfect on a still frame would be often be unacceptable when applied to 30 frames per second.

The Asiva technology was designed to deal with this problem precisely. No masks are created when defining an area that any given correction is to occur. Traditional methods required a physical selection (mask) to be made based on a pixel's physical location in a frame. This results immediately in associated boundary and edge artifacts. The Asiva technology makes selections based on a pixel's color properties, so no masks are generated. By adjusting the HSV curves that define the properties of the pixels that are to be affected, very smooth transitions are possible between pixels that are being affected and pixels that are not. No masks means no associated boundary artifacts.

To be successful today photographers need to be abreast of not only the advancements made in acquisition technology but also with the technologies available to them after the production process is complete. In the next article we will guide you through proper set-up for a background replacement shot and give you an brief overview of the technologies available to make full use of this process.