dmstraker

Not subtitles, but perhaps this will help. I wrote notes on all the Affinity Photo video tutorials. The post is here: https://forum.affinity.serif.com/index.php?/topic/37281-notes-on-in-house-affinity-photo-video-tutorials/ I don't use Designer (I'm a photographer). Regarding sub-titles, as there are ways to add these to videos, such as http://www.3playmedia.com/, I wondered if there is somebody in the hearing-loss support community who would type up the words.

Does Photoshop do more display smoothing of pixelated images?

I've got it too and works fine. When exiting Particleshop it offers returning a whole image or just changes. Accepting changes results in these appearing on a layer with white background. I use filters/colours/erase white paper to leave only the particle adjustment as opaque.

When dragging layers, where your mouse is horizontally changes where the dragged layer goes. There are three places it can go, which are signified by transparent blue bars as you are dragging. Keep the mouse to the left of the layer icons to drop the layer at the top level. A horizontal blue bar the width of the layer panel will appear. This is typically done with pixel layers (or adjustments that will affect everything below them). Drag onto the layer icon for masking or nesting. A small vertical blue bar will appear to the right of the icon. This is typically done with masks or adjustment layers. Drag it onto the name of the layer for clipping or nesting. A horizontal blue bar will appear that start under the name and goes across to the right This is typically done with shapes or adjustment layers. A classic trap when dragging a pixel images is to drop it into a nested position, so it seems to disappear. Remember to keep the mouse to the left of layer icons when doing this kind of move. For further help, I'd recommend watching the excellent official training videos. I've put notes about them here: https://forum.affinity.serif.com/index.php?/topic/37281-notes-on-in-house-affinity-photo-video-tutorials/

Update added in first post - version 3, 22-Jun-17

Having said all this, in another test I got some coloured pixels - red, yellow, magenta, green, blue. So I did a further test - Duplicate layer. Blend Ranges (gear icon) Pull left node on left graph to bottom right corner. Add empty layer below, filled with black. This also selects white. Comparing the two methods, the coloured pixels are black on the second method. Then a lightbulb went on somewhere. The pixels which appear non-white are saturated in at least one RGB channel but not all of them. So partial burn. I guess that's useful to know too, but as masks are not per channel (without fudging) not so helpful in this.

A way to identify pixels which are burned out (white), which can be handy for example in creating mask to clean them up: Duplicate layer (Ctrl-J) Invert (Ctrl-I) Blend Mode: Colour Burn Any burned pixels will appear white. Everything else will be black. This works because (across RGB channels): The blend calculation formula is: Result = 1 - (1 - Base)/Blend For white in bottom layer, Base = 1. For black (inverted white) in top layer, Blend = 0. In this case, Result = 1 - (1 - 1)/0 = 1 (white) For all other colours, say 0.4, Result = 1 - (1 - 0.4)/0.6 = 1 - 1 = 0 (black) To see what is in the image below, get a white brush and hover over any area - it acts like a torch.

When you select a part of an image, you are choosing less pixels so by definition you have lower resolution overall - could it be something to do with this?

You can also change opacity of a pixel image with Blend Ranges. Click on the gear icon in the Layers panel. In the left hand graph, drag down the dot on the right side. Play with shaping this for further effect. You can add dots by just grabbing a part of the line. This graph is luminosity on the x-axis (black on left, white on right) and opacity on the y-axis (transparent at bottom, opaque at top).

Edit/Fill is nice. It would be even nicer if you had a '50% grey' option (for use with contrast blend modes, of course). Thanks!

AP adds blend modes Average, Negation, Reflect, Glow, Contrast Negate and Erase. Is there a detailed technical description of how these work, including the blend logic, please? Rationale for inclusion and ideas for use also would be appreciated. Also, is there any reason why Divide blend mode is not included? Thanks!

Main thing I do is to use live adjustment/filter layers as much as possible. Fortunately, AP is very good at this. You can even do things like add an adjustment and then use the blend mode in this rather than duplicating the background, adjusting this, and doing blend mode at the pixel layer level.

Should work. As toltec indicates, it's so easy to try and apply an effect and nothing happens because you don't have a pixel layer selected. I still fall for this all the time. If in doubt, click on it and check it turns blue immediately before selecting the tool to use on it.

A way to be exact and control image sizing is to: Document/Resize your photograph to the size you want it to be. You can choose the resize algorithm here. Create canvas: File/New document. Set the size including the white borders. This will be bigger than the resized document. Unclick 'transparent background' to ensure white border. Copy and paste the photo image into the canvas. Use Move tool (arrow) to drag photo into place. Guide lines will appear when it is centred.

Further tweaks you can do: Make it black/white (as noted above). Recolour, for interesting different coloured outlines. Remove white background with Blend Modes (select top pixel layer and click gear icona above it). This allows a straight overlay without blending the adjusted layer. Mask areas so some has lines and some doesn't. So attached has red outlines applied to everywhere except the car, kind of echoing the red in the car. daves-linedraw-pecos_tx01-DS2.afphoto

My adjustments attached. Nice car, by the way. You may need to drag the adjustment layers around to get them in the right place. Assistant setting may affect this. You can then use this for blending, for example with a Multiply (adjust with opacity) to enhance outlines. You can also mask out areas not needed. Yes, the 'hide the bottom layer' doesn't make sense unless you have already put a blend mode on the duplicate, which I had when I was writing the notes. Oops. Yes, B/W adjustment does help here. I did try it in my original image but that picture didn't need it. Overall, it seems cleaner than doing a Detect Edges and also allows for non-destructive dynamic tweaking of such as blur (I used 1px in this image). daves-linedraw-pecos_tx01-DS.afphoto

Here's a novel way of turning an image into a line drawing: Duplicate Layer (Turn off bottom layer so you can see what’s happening) Denoise, Luminance right up to get rid of background noise (without losing main lines) Gaussian Blur, 0.5px or to taste, Blend Mode: Subtract Invert Levels, black up to about 70% Because this is all dynamic, you can go back and tweak the noise, blur and levels later. You can then blend with the background with Blend Mode: Multiply (or other Darken blend). Turn down Opacity to suit.

So you've just rotated your image and now have four nice transparent triangles in the corners. So what do you do? Here's a way of handling it: Select/Alpha Range/Select Fully Transparent Select/Grow/Shrink... (Radius 1px) Edit/Inpaint This should select the corners and give AP the opportunity to apply its Inpainting algorithm. At best it's a perfect fix. At worst, it's a start and you can clone bits that are not inpainted so well. You do need to do the Grow, by the way, as without it you can get missed pixels and an inpaint that is not fully opaque (I think these are bugs and have reported them - until the fix, you need the Grow).

After resizing the canvas, another way to fill in the transparent border without adding a layer is with Edit/Matte. Set the border colour you like and click OK.

There are now several 40+ inch 4K monitors on the market for £500 or so (cost similar to a cheap lens). I've had a Philips one for a while now. Highly recommended for photo editing (good for spreadsheets too). I place it vertically in front of me on a stand, with a tilted 22" monitor below (angled like a laptop screen).

It would be nice if you could show the aspect ratio when using unconstrained resizing. Even nicer would be the ability to snap to common ratios. Thanks for considering this idea!

Update: Convert edges layer to b/w and add Gaussian Blur (adjust to taste). So sequence is now: Layer/Duplicate layer. Filters/Detect/Detect Edges. Layer/Invert. Layer/New Adjustment Layer/Black & White Layer/New Live Filter Layer/Gaussian Blur (radius 1 or vary to taste) Blend Mode: Colour Burn. Add Curves above it all and pull up midtones. Play with Opacity to vary strength of effect. Worked better in some images. Gaussian Blur reduces cartoon effect and copes with irregular edge detection.

Yes, you don't need filters like you used to. However, graduated ND is good when there's a higher dynamic range than your sensor can cope with (even with RAW). It helps, of course if the horizon is level, though you can compensate for dark-top mountains in AP. The alternative is exposure bracketing and stacking games, though this can get problematic when the subject is moving (good HDR deghosting can help). For movement, also, non-grad NDs turn down the sun very well. After this, it's philosophy. And Aristotle didn't have the benefit of AP, so he would have been biased.

May be possible using macros, depending on what you want to do (AP macros have limitations but are easy to record). File/New Batch Job allows you to run a macro on a set of images.

Some thoughts: Do it in the camera with long exposure, narrow aperture and low ISO. Use a dark neutral density filter on the camera, such as the Lee Big Stopper. to force a longer exposure. Do a number of short-exposure-time images. Blend as a stack in AP. Take one image for ground and mask in the blended sky. Use an AP blur filter on the sky. Either motion for lateral movement or radial if the clouds are moving from a point.