Once upon a time it was only necessary to put the word digital in front of a product to make it sound modern, cutting edge. It was as if the very title ensured a quality and performance previously unattainable, and just as often unquestioned. Still watching that old pair of hands to tell the time - get real, get a digital watch to be part of the new world. It may have started with a few gadgets until everything was digital and the description became meaningless.

But of course the blanket description hid as many tragedies as triumphs as we all know now with the benefit of hindsight. Now the buzz phrase is AI, and so many people seem to think that with the guiding hand of another intelligence, everything will be perfect. Most of them, of course, don’t work with computers on a regular basis.

Those of us who do will already know, and agree with the general theme of this column, that a human hand on the mouse is needed in order to avoid the sort of mistakes that electronic logic is rightly famous for. It was always important, and will be increasingly so as the relentless introduction of AI reaches all points of contact with the real world.

Ironically, as one of the big giants in the field, Meta, announced a massive investment in AI, its shares on the US Stock Market took a big slump, so not everyone is quite so excited about some of the possibilities the future may hold. Much like investments, the health of IT companies can equally go down as well as up.

To be fair to Adobe, underneath the hype of recent promotions, their adoption of AI is very much measured, and as cautious as you can be in a fast moving, competitive business world, while still retaining creative integrity. Some big updates have come in April, too soon to absorb before penning this month’s column, but the direction is very clear, and despite what some claim, is much more likely to put imaginative design into human control than you might expect. But that’s a longer tale.
In the last couple of issues I’ve been explaining the benefits of using Adobe Bridge - an often ignored resource compared to the instant resource of the ubiquitous Windows explorer which everyone turns to automatically. Bridge needs a bit of patience to learn, and to organise, but the rewards are well worth it. It’s interesting that Windows 11 now includes a much improved version of the preview pane, which makes identifying individual files that much easier when searching through packed folders of anonymous icons.

The advantage of Bridge is that it talks directly to other Adobe applications so that information, edits and other adjustments are seamlessly shared, and displayed, without having to save and refresh, and you can rename, rearrange, colour code -in fact pretty well everything you need to do to organise your digital sock drawer. The more you use it, the easier it gets, and you can quickly find things, and things you need to know, at your fingertips.

The other feature Bridge enables is a portal into the magical world of Camera RAW. Initially this was a filter that allowed the editing of certain high resolution photography images, hence the name. More recently it can be used to manipulate any picture file. It appears as a standalone app outside of Photoshop, and can be opened directly from Bridge as well as from the Filter action in the main app itself. Once edited, the file can be saved into PS. Multiple files can actually be edited and the settings applied to all or none, including Adobe’s Image Enhancement tool for upscaling pixels.

Apart from those tricks, RAW enables a much more intimate analysis of an image, whether colour, texture or shape, for example, as well as a selection of very accurate masking tools.

All of these are also available in Adobe Lightroom, which as a photographer, I personally prefer to use, but as most operators are more familiar with PS I am concentrating on that. The reason we need to know and use these tools is simply the quality, or lack of it, of the majority of customer input over the counter or emailed in. You can take the attitude that what they give is what they get, but that won’t win you any regular customers unless you can do at least a little something to balance what will print against unrealistic expectations from phone screens.

You don’t need to have an old grey head like me, or be a boffin and confuse the customer with science, Adobe is here to help. You have the tools, you are paying for them, so you might as well learn to use them properly. And because of AI it is getting a lot easier because the machine will not only anticipate predictable actions, but provide options so you can prompt it in the right direction.

A very simple example is the matter of colour correction. Most phone and instant camera images provided would benefit from some adjustment to compensate for the light conditions they were taken in. Years ago most people would not dream of taking their camera out after dark, or take a shot without using a flash. Now they just point and squirt and let the software - mostly programmed on a sunny day in Seoul or wherever - work out the rest. But it probably doesn’t have much of an idea what a wet winter evening in Worthing looks like. Photoshop as always had colour correction options, but in its early forms that was tedious and manual or automated and more often wrong. This is why AI that restricts you to just pressing a button is so random.

Much better is the smart bomb approach to nudge it on course with a few manual switches.
The Curve tool has always had preset options, and has always been the preferred choice for separating highlights and shadows at a basic level. Increasingly it has had intelligent scrutiny of the image creeping in by increments.




Choose any image that has a decent amount of difference between light and dark and open the curve tool, or create an adjustment layer. Instead of Auto, if you select the options you will see a choice of algorithms that will compensate in various degrees for the features in the tonal range.

If you work your way down the tick boxes, you can easily see how they affect the overall image and pick a preference based on how it will print. You can save that as your default setting or you can further fine tune the options if you need to.
It’s a good example of man and machine working together, but it’s just the beginning. In April Adobe released a massive update to their AI functions in Photoshop, both regular and Beta versions. I don’t normally mention the Beta option, which is for trailing features yet to come, and can run alongside the normal programme. But it can get confusing as to which version you are using, and sometimes there are bugs yet to be sorted. Some of the bugs may not be of Adobe’s making, being conflicts with operating systems and other computer parts yet to be ironed out.

People tend to assume that IT is all about lots of clever people who cooperate with each other rather than the reality of jealously guarded secrets and a desperate attempt to trump each other with the latest gimmick.

Adobe has been very much pushed into speeding up its AI steamroller by developments from other companies which were in danger of stealing the headlines, as well as the sales. But that has not been at the cost of quality, as unlike other applications it can draw on a vast resource of stock images and examples which it can share legitimately, rather than pirate poor versions from other sources. Much of AI is, by its online nature, low resolution and not suitable for print at anything other than small size. Adobe is very much pushing the boundaries of higher resolution that were previously possible, as well as many options available to select from.

That’s why I suggest it’s worth dabbling with the Beta as there is so much included in this package, it’s worth getting familiar with them before they become an integral part of the mainstream. You can download it from the Creative Cloud App, and it has a blue on white patch rather than blue on black to identify it.

Adobe’s AI programme is called Firefly, and this is now Firefly 3, though number twoaaaa seems to have slipped under the table somewhere. At the heart of it is the ability to select, manipulate and create images beyond the realms of previous picture management, and at a level of quality that can be used for print, not just for sharing fantasy on line. That’s why it’s essential for the modern print shop to get abreast of the latest techniques otherwise customers will be staring at their phones and asking why you can’t reproduce something they have just done with the press of a button.

While the more fantastic creations of AI may grab attention, the more basic bread and butter functions will be of much more use. Image editing has always been a matter of selecting one pixel from another - a colour, a shape or whatever. That has always been a problem until the resolution of original files got bigger, and computer software got smarter. Once you can separate one item from another, everything else gets a lot easier. Or at least that’s the theory. Whether you want to just change the background to a different colour, or eliminate it and replace it with a completely different one, the challenge is the same.

But there is more to it, as the changes have to be in context, and look realistic. Features have to be in perspective, sunlight and shadows in the right direction for example, and people’s faces have to be right. We all recognise a face instinctively, because we read its mood, which is the very essence of a pleasing portrait. You don’t put a smiley face in the middle of a tragic scene. Conversely if a particular cactus wasn’t exactly the right shape in a panorama, you won’t worry too much. Somewhere along the line there has to be a human intervention into the equation to determine the difference, and that’s very much what Firefly is all about.




I’ll have some more useful examples of the new stuff next month when I've had a chance to play with it but this is just one from the existing programme already pretty good. The original is a still from a phone 16 by 9 capture of a bunch of mates in the mountains. Using Generative fill both the peaks and the people have been extended to fill a frame shape, using both the technology and a keen eye and a little bit of manual dodging on any stray glitches, like banana hands.
It may not be perfect, but without sending them back up the snowy slopes for a reshoot, it will do the job