Innholdsfortegnelse for July 2023 i Practical Photoshop

01 MAKE A SELECTION Just as with Photoshop’s older fill commands, like Content-Aware Fill, we need to make a selection of the area we want to fill in before we can use Generative Fill. So grab any selection tool and make a rough selection over part of the image. 02 START FILLING Once done, go to the new Context-Sensitive Taskbar (Window > Context-Sensitive Taskbar) and click the Generative Fill button. Here you can choose to add a text prompt to guide the process, or simply leave it empty. Click the Generate button. After a few seconds, content will appear within your selection. 03 VARIATIONS There are three variations of the generated content that you can choose between. Use the arrow keys in the taskbar to scroll through them, or choose…

If you want to generate a specific object, the shape of the selection you make can have a huge difference on the look of what is generated. So rather than simply using the Marquee Tool to drag a rectangular box, try using the Lasso Tool to draw a freehand selection that roughly fits the shape of the object you want to create. Be specific about what you want to include in the fill. In this example, the first attempt to create an aeroplane was made with a simple box selection; while in the second attempt, we used the Lasso Tool to draw a rough plane shape at the angle we wanted, which gave better results.…

After making a complex selection – say, with the Select Subject command – you may get better results by eating into the edge slightly, so that there’s a bit of overlap. Try using Select > Modify Expand or Contract to add a few pixels either side of the edge. Similarly, while expanding the bounds of an image, after cropping outwards, make a selection that overlaps the existing pixels slightly: this helps make a gentler transition between existing and generated pixels.…

Today’s generation of AI image generators struggle with hands. Extra fingers, wonky wrists, smudgy edges and missing extremities can all occur, and are an instant giveaway. If the first attempt at a hand looks bad, make a rough selection over the hand with the Lasso Tool, then type ‘hand’ into the prompt box and generate again to see if things improve. If the subject’s original hands are available on a layer below, the best thing to do is exclude them from the initial selection. If this doesn’t work, try selecting and copying the original hands to a new layer and blending them with the final image.…

Generative Fill is an amazing rendering tool, but it doesn’t render your old Photoshop skills obsolete. There are likely to be messy areas in the fill, like dodgy details or mangled hands that need to be fixed. This is where your cloning skills are essential. Make a new layer, grab the Clone Tool and set Sample to All Layers, then hold Alt to click a clean source area, and paint to clone out messy details. You might also want to paint on the Layer Masks attached to each fill, to hide any parts you don’t want to see.…

AI image generators are trained by analysis of millions of photos, so it’s important to know which photos it uses for training. Firefly is based on the Adobe Stock library, so it naturally works to a different set of parameters than other generators. While it’s probably impossible to form an understanding of how it works, it is possible to gain experience in what works and what doesn’t, and how to get the most out of the tool. A great place to explore is the Adobe Firefly Gallery (https://firefly.adobe.com/gallery), which shows incredible AI images, along with the prompts used to create them. Try out the prompts in your own work, and see which aspects are successful. Of course, Photoshop’s Generative Fill is a different beast to Firefly, as it also takes…

A burst of flash lets us freeze motion, while a long exposure lets us create motion blur. Both are great skills to learn in their own right – but if you’re looking for a creative approach to action shots, the slow-sync flash technique lets us have it all in a single frame. Sometimes called shutter dragging, this is where we combine a brief flash with a longer-than-normal exposure to introduce motion blur. It’s a great trick to try for action or sports subjects – like our taekwondo expert Mitch, where we’ve created streaks of blur to convey the movement of the subject, and used flash to freeze the motion. Whenever we use flash, we are effectively making two exposures in one. First there’s the flash exposure, where the subject is…

01 ENHANCE THE COLORS Open the image into Camera Raw or Lightroom and press K to activate the Brush Tool (or select it in the Tools Panel). Paint roughly over the blurry parts of the scene, then try increasing Clarity using the settings shown. This will crisp up the flowing lines for more defined blur. You can also lighten or darken the blur using the tool. 02 TIDY THE EDGES There are likely to be parts of the frame that look messy, either because you’ve included areas of backdrop or because of lens flare caused by the various lights. Pressing K again will create another brush mask. Paint around the messy areas, then decrease Exposure to darken them. Repeat, if necessary, until they’re fully black. 03 CHECK THE SHADOWS Use…

The Context-Sensitive taskbar is a new feature that allows for easy access to the tools Photoshop thinks you’ll need at different stages in your workflow. After you make a selection, the new taskbar will appear and give you the option to launch Generative Fill (if you don’t see the task bar go to Window > Contextual Task Bar to turn it on). The taskbar automatically positions itself to the bottom of the selection, but this can be annoying as it obscures other details. You can click the three-dot icon in the bar to pin it elsewhere. Alternatively, launch Generative Fill by going to Edit > Generative Fill (having first made a selection). Once Generative Fill is open, simply type in a prompt based on the type of subject you want…

The Select Subject and Object Selection tools are brilliant at isolating objects for you, and can speed up your generative filling. If you want to fill in a background, for example, simply use Select > Subject and press Cmd/Ctrl+Shift+I to invert (a useful shortcut to learn for this kind of work). Here we used Select Subject to isolate the car, then inverted the selection to the background and typed in ‘sandy beach’.…

The Generative Fill command is astonishingly good at expanding your canvas, perhaps to create a wider view or to change the aspect ratio of an image. Simply grab the Crop Tool, drag out from the original edges (hold Alt to expand both sides at once, and right-click for different crop proportions). To select the extra areas, grab the Marquee Tool and drag over one side, leaving a slight overlap, then hold Shift and drag to add in more areas. Once done, leave the text prompt box empty and simply click Generate. The results can be seamless, and you can expand out from very tight crops to epic vistas in mere seconds. WATCH THE VIDEO https://bit.ly/3NfOOgk…

Attempts to create content that may be deemed pornographic, racist, offensive, violent, extremist, fraudulent or misleading will trigger a warning. This is to be expected, but perhaps more interesting is Adobe’s approach to copyright and intellectual property, which are contentious issues in this new age of generative AI photography. The terms of use state: • ‘Using Adobe’s generative AI features to create content that violates third-party copyright, trademark, or other rights is prohibited. This may include, but is not limited to, entering text prompts to generate a third-party brand logo or uploading an input image that includes a third party’s copyrighted content.’ It will be interesting to see what approach Adobe takes to licensing generated images, whether they bring in a payment structure to remunerate artists whose work has been…

Sometimes the colors in the original image can bleed into the fill area in unwanted ways. This is easy enough to fix with an Adjustment Layer. Here, the redness from the sky has affected the bushes generated in the left corner. A Color Balance Adjustment Layer reduces the redness, and a Layer Mask restricts the effect to the corner. After making an Adjustment Layer like this, simply press Cmd/Ctrl+I to invert the mask to black, then paint with white to reveal the color change in the affected area.…

GENERATIVE FILL IS VERY GOOD AT: • Replacing clothes• Varying hairstyles• Matching lighting• Generating grass and foliage• Expanding landscapes• Creating new parts of existing objects• Water and reflections• Generating hard objects like cars NEEDS IMPROVEMENT… • Generating full people• Depicting hands• Placing entirely new objects in scenes• Generating animals• Changing backdrops…

1 DARK ROOM We need our backdrop to be as dark as possible, so that the only ambient light from our LED studio lights. Either close all the curtains or shoot outside at night. A plain, dark background will ensure the blur is as clear as possible. We fastened some black material for the background and floor. 2 CONSTANT LIGHTS The studio lights illuminate the subject from behind to the right side as he moves. They’re for the blurry part of the long exposure. We’ve used two LED lights here: one fitted with a red gel, the other fitted with a snoot to control the spread of light and set to a warm color. 3 FLASHES Two flashes are to the left of the subject: a rear flash, with a…

As well as blurring the motion of the subject, we can also use camera motion to create blur. For this shot we had the subject stay still, and I panned the camera from one side to another to create the blur. There are lots of other ways to get creative with shutter drag too. If you have a zoom lens, you can get interesting results by zooming in or out during the exposure. You could also try including other lights in the frame, like fairy lights or colored LEDs. To take it further, why not experiment by firing the flash at different points, like halfway through the exposure? Or perhaps experiment by having your constant lights pulse so that the blur is staccato. Once you start to experiment, you’ll find…

Before you can use Generative Fill, you need to download Photoshop Beta. This is a separate app to Photoshop, but available to all those with a Creative Cloud account (and it’s also available to others through the online version). Open your Creative Cloud Desktop App, go to Apps and open the Beta tab, and you’ll find Photoshop Beta there for download. It’s almost identical to Photoshop, but with new tools that are being developed, like Generative Fill. The tool is still being refined, so there may be changes to the functionality, performance and even the name before it eventually makes its way into ‘proper’ Photoshop.…

After inputting your text prompt (or leaving it empty) and clicking Generate, you’ll be treated to three variations to choose from in the Properties Panel. If you click Generate again, you can create more variations, then cycle through all of the options to see which one works best. If you change the text prompt, you’ll still see the old variations in the list. You can also offer feedback on the variations, which will help Adobe to develop the tool.…

Generative Fill is very good at adding accessories and clothes to people. Try adding sunglasses, changing clothes, generating jewelry and more, and you’ll find that the results will often look very realistic. Here, by adding the sunglasses, the command has also changed the shape of the nose and mouth. Remember that if you don’t want areas like this altered, you should exclude them from the initial selection. WATCH THE VIDEO https://bit.ly/3PpsqUr…

Currently Generative Fill creates a fairly low-res output that, while OK for posting to Instagram, probably wouldn’t stand up to printing at a large size. Generated areas are set to a maximum size of 1,024 x 1,024 pixels. So there’s little point trying to bolt them on to a high-res photo, as you’ll see an abrupt change in quality where the high-res meets the low-res, as shown in the details in the plant here. In fact, you might want to down-size your photo before starting, to speed up processing times. If you want to get more resolution from your fills, you probably need to wait for the full version of the command. In the meantime, there is a way to game the system: simply create a series of fills in…

If you’re wondering whether artificial intelligence will lead to the death of photography, here’s a good news story: AI is amazing at fixing noisy photos. We tested Adobe’s new AI-powered Denoise function on several noisy, high-ISO images. This bird photo is a proper test – a heavily cropped, underexposed image shot in a dark forest at ISO 25,600. The Denoise Tool worked wonders on this and other grainy photos, giving better results than with Photoshop’s older manual noise tools. The tool is particularly effective at preventing colors from bleeding into one another. In our photo, the original colors were barely perceptible but the tool pulled out the blues and reds in the subject to an astonishing degree. It also cleaned up and revealed details that were almost invisible in the…

01 THE FIRST CURTAIN By default, flashes fire at the start of an exposure, which is referred to as the first or front curtain. But when you use a slow shutter speed, it can look wrong. I took a shot of myself running from one side to the other. The blurry parts overlap the sharp parts, and it looks like I’m moving backwards. 02 SET REAR-CURTAIN SYNC To fire the flash at the end of the exposure instead, use rear-curtain sync. If you have a compatible wireless trigger, set it on the trigger. Or you can use a hotshoe-mounted flash, then set the off-camera flashes to optical trigger mode so they fire on detecting the flash. 03 NATURAL MOTION When the flash is to rear-curtain sync, it will fire just…

Photoshop has many types of layers and adjustment layers available, but there are six that you’ll find you need to use again and again. Learning how they should be used may seem a little daunting for beginners, but once you’ve got to grips with them, you’ll find they play a part in the creative process of almost every image you make. 01 LEVELS This should be the first layer you add to an image, because it fundamentally alters the tonal range of the entire image. Create a Levels Adjustment Layer, drag the Black point slider inwards until it touches the left-hand edge of the histogram, and drag the White point slider inwards to the right-hand edge. This remaps the tones of the image to make more of the available tonal…