Dazzler

As John said the Affine filter will allow you to move the sides into the middle, so you can create a seamless loop. This is something I used to do a lot in Photoshop (offset filter) when making seamless textures for 3d models. As for how you would print a continuous loop onto a T-shirt, I've no idea, but I imagine the input needed would just be a rectangle that meets at the edges, or supplied in two pieces that are printed with some sort of registration alignment. Best get advice from the people making the T-shirt.

Thinking about the OP, have you tried the in-painting tool rather than the healing tool? It might give a decent result, although I realise the two are quite similar i nature. Also, try to obliterate the entire object you want to get rid of in one hit rather than several steps as these tools will try to blend the edges of your painting back into what the surrounding area is, which may explain the dark graduations coming from the line. If there's not a lot of surrounding area then you may have to work at it a bit to get what you want, but I normally find getting as much as possible in one hit the best start, then work on the things that it may have duplicated into the fill as a second step.

LOL Can't say I've ever done that one, my usual thing for music is to think I pressed the record button but instead I've pressed the play button. This normally happens when your performance is at it's peak and you realise it's gone forever!

Are you using the fill tool to flood fill your drawn shape? If so, you may be experiencing a common problem when using such a tool (and this applies to all packages I’ve ever used). The problem is in the way shapes are represented in the computer. Rather than being a single colour, lines are made up of lots of colours (or to be more precise, different intensities of the same colour, and the reason for this is to make the line look smooth - known as anti-aliasing - zoom right in to see these variations on the pixels, normally ranging from solid to transparent, but it depends on the context). So when you choose to fill a shape the fill works it’s way out until it finds a pixel of a different colour to that you clicked on. Now, if you imagine your outline being surrounded by more subtle pixels to make it look smooth, what happens is as soon as the fill detects the very first of these subtle colours the fill algorithm thinks it’s at the outline and stops there leaving what looks like a gap between the subtle colours and the main part of the outline. I can’t remember if in Photo there is a tolerance setting on the fill (I’m sure there must be, probably in the context bar), but you can normally overcome this issue by making the tolerance a higher value. This tells the fill algorithm to accept more difference in colour before registering as the edge, and closes that gap. Might be worth upping the tolerance if that is your issue. It’s not always possible to get a perfect result, as if you up the tolerance too much it fails to detect your outline properly and fills the outside area. It all depends on the consistency of the outline and whether it has any weak spots.

Have you checked you’ve got the correct layer selected? and you’re not on an adjustment or filter layer applied to it? I’m always doing that by mistake, effectively healing or cloning the mask for an adjustment layer.

I suspect it’s just down to rounding errors in the maths, being that the smaller you go the more pronounced this becomes. I’d be very surprised if it was rasterising this and then tracing it. But I could be wrong.

In Affinity Photo? Yeah, sure, it's up in the context bar on the right - 'Protect Alpha'. Comes up when you select most of the tools that would affect the alpha by painting or whatever.

I do know what guides are, but ultimately, you can make a shape (whether that's a straight line or angle line or whatever you like), put it on a background layer, lock that layer, then use that for objects to snap to. It's not a guide, but it's the next best thing, and helps overcome the limitations of the real guides. The only real difference is that you can't hide and show that layer with a shortcut key, you'd have to click on the eye in the layers panel to hide it. Of course, you'd also need to delete or hide it before outputing too.

AMW, Just out of interest, are these tall screenshots coming from a browser showing a web page? Because Firefox has a screen capture feature built in that grabs the whole page in one go. Saves a lot of time messing about stitching grabs back together. I still have to do this for sites I work on where the only browser available on a specific machine is IE11, and I find Designer very good for piecing these grabs back together (not automated, but quick nonetheless). You can just use the artboard tool to click on the first grab to create a 'to size' art board, then pull the art board taller on the bottom handle, then place the next grab in, pulling the art board down each time, and everything snaps nicely together without much fuss. It's like a dream compared to doing it in Photoshop.

Not sure if this has changed since your post, but you can currently use objects as guides (hovering over gives a nice yellow line along the object you are about to snap to). That includes locked layers, so it would be fairly trivial to add your own angled guides. Also, there's an amazing tool hidden away for some reason called the 'Point Transform Tool'. I'd definitely recommend customising your tool palette and adding that, as it's great for aligning objects to other shapes.

Yeah, I agree, and in many cases you can actually push images way beyond their 100% resolution, but it would depend on the nature of the final output. I used to work for a company that did large format printing and we'd quite often downsample images to much lower resolutions than 300dpi to make the file sizes more manageable (back in the mid 90s when 1Gb was extremely difficult to deal with and would literally take many hours to transfer over a network!). But yeah 300dpi is a good guideline for most outputs. You could probably go with much less for a billboard. I do like the way that readout works though, gives you exactly the information you need to know.

I had an issue initially. I reinstalled designer (think I had 1.7.0 installed not 1.7.1) , then opened up both designer and photo, then launched Publisher and it seemed to click into place.

Up in the top context bar there is a readout for an image that states it's size, effective DPI and percentage scale. As you scale a picture this will adapt to show you the new DPI and scale compared to it's original size. Obviously if the DPI of the image is less than your intended print DPI then you've pushed it too far!

With regards my above solution, it also makes sense to not convert a text frame in the first place, but to use a rectangle object and paste that inside the group also constrained in the same way. That way you can colour at will and apply features such as rounded corners (using the corner tool), and these will also resize in a nice way with the outer parent rather than being stretched etc. It also allows you to choose the amount of margin you want inside the outer frame by simply resizing the text frame to suit, then when you resize the outer group the text frame will maintain that gap. In this way you are sort of constructing little text 'widgets' that can be duplicated and resized to suit - great for speech bubbles for example.

You can sort of achieve this in Designer with a workaround. Create your text frame, add text etc. Set it to the jusification you require (mine was centred both vertically and horizontally). Group that layer using CTRL+G / CMD + G and then add a stroke on the parent group layer. Select your nested text frame and go to the constraints panel and set those so the two resize arrows in the middle square are showing but the arrows around the outer edges aren't. You should be able to resize the outer parent box and the text frame will expand with it. However, it fails with regards to adding text - the text frame won't resize automatically when you add more text, but you can always pull the parent box open more. Hope that helps.

If you want to reduce the amount that the sliders affect the picture when adding an adjustment layer such as shadows/highlights then you can take the opacity down a bit on the adjustment layer (it's just there in the corner, so no need to switch back out of the dialogue). That will make the adjustments more subtle as it's then mixing it with the original. Also, don't forget you can add more than one adjustment layer even if it's the exactly the same as a previous one, so you can always add a second shadows/highlights to use with the shadows and have that a different opacity than the one you use on the highlights, so they each have their own aggressiveness.

+1 for this feature to be added. I'm a little bit surprised that you can't already do this, as it's not exactly a new thing, pretty sure you could do this in some vector packages in the 90s? For some inspiration, take a look at the way the 3d package 'Modo' by The Foundry does it's replicators - they are superb, allowing not only one item to be duplicated, but a group of items can be used and it picks items to duplicate from that group - very nice for more organic ideas. Would love to see this sort of thing in a 2d version.

Just wondering now whether that setting is initially dependant on the document type you choose? That would make a lot of sense.

If I remember correctly, you can just click on the pixel button on the toolbar to unselect it (it’s a toggle button). Pixel based preview is a very useful thing for people who work with web or where the output will eventually be rasterised. It enables you to draw assets that don’t have that fuzzy ‘half a pixel’ thing going on and is one of the great features of Designer, which may be why they have it enabled as default.

Yes, there are three methods of masking, one is the 'Rasterise to mask', which is pretty much equivalent to the Photoshop way of doing it. The second is by nesting layers in other layers. This is more interesting and would be the way to go if you want the mask to be a vector shape rather than a rasterised image. The parent layer becomes the mask and the inner layers are then masked by it. The third is using the blend mode set to erase. Not so keen on that way of thinking, but it's there if you need it. It looks like in your case the mask is a solid flattened image with a background colour containing a greyscale image map, so you'd have to use the 'Rasterise to mask' method.

I wouldn’t worry about it. If you’ve been doing it for years and there’s never been a problem then you wouldn’t realise. Really, the problem lies with the software companies where no strict standard has been maintained. I’m a web dev, and as any web dev will tell you, consistency and strict rules are absolutely essential to being able to provide a consistent experience for the user. If there is no agreement at the start of new technology then it goes off in all directions and then you find yourself putting five css rules into a style sheet instead of just the one, because the browser vendors all decided to implement the same thing in different ways. It’s crazy stupid. But that’s the nature of cutting edge technology, where it’s expensive to bring new technology into places like printers, so workarounds happen to allow new technology to talk to old technology. Ultimately a tiff file should really just be a tiff file, and it should fit the exact specification of a tiff file, not have some extra hidden data stuffed in there. But now it’s too late, it’s already happened, and so if you stop supporting odd files where things have been abused, then suddenly you find yourself with a load of files that no longer work as expected with certain software. Exactly the same with the web, where you can’t suddenly make all browsers behave the same without breaking a ton of old legacy content. So we continue to adapt everything to accept these formats, and so the problem lingers on. Ho hum, fun and games!

Chris, I think the thing that can be learned from this, is it’s not really your fault, file formats sometimes get skewed to suit certain situations, and it looks like that’s s what’s been going on. So we’ve ended up with a file which actually has two methods of storing it’s transparency. That’s kind of dumb, but I guess there are reasons somewhere along the line, probably due to compatibility. It must mean however that the files are larger than they need to be.

Thanks Walt, the bit I was doing different was embedding the images rather than linking ... got it now! I realise there is transparency in the original file, as you say, but is that transparency coming from the extra psd data rather than the true tiff part of the file? Being that AFPhoto can open PSD files it may be? Certainly looking at the files within my explorer window the ones that are working have a transparent background in the thumbnail, whereas the ones that don't look solid white. They ALL open in Photoshop and AF Photo with transparency intact, but the ones with the white backgrounds don't come through with transparency when placed into a document, however a TIFF file that has the transparency checked upon save does work perfectly with the place image tool in publisher. So I maintain that the original file was saved with that option not checked, and it's the extra psd information that is where any transparency is coming from. Would be nice for it to work the same across the board though. But if the place tool is reading it like a true tiff and the open is reading that hidden psd data then it would explain it.

Thanks Patrick. So surely this is the answer then? The original seagull file has been saved out of Photoshop (or whatever) without the transparency being checked? Any ideas about why I can't see the edit image button? I've tried doing all sorts, changing personas etc. but I just can't get it to appear? Should it be there for me? As I say, I have all three apps installed, and can happily switch between the personas. Is there a reset for the workspace somewhere that I can use?

Ok, so upon further investigation, saving out from Photoshop if I don't check the save transparency option (below the compression settings dialogue), then the file comes in with a white background into publisher when placed. However that same exported (supposedly no transparency) file can be opened back up in Photoshop and has the transparency still. Which is a bit weird isn't it?