Jump to content

gdenby

Members
  • Content count

    1,844
  • Joined

  • Last visited

Everything posted by gdenby

  1. Hi, rudyp16 Techniques. The image you posted was fairly "dirty." Obvious compression artifacts, and pixellated. So I used a denoise adjustment to get rid of the artifacts. Then went over the pink areas w. the flood selection tool w. a low setting, 8% tolerance. Where I couldn't get a clean outline line, I used the selection brush at 4 px to add or subtract. Then Used the HSL adjustment to change to shift and deepen the selected pinks. There were some problems, so I used a blur brush here and there to smooth the pixelization. I probably should have resampled the image to a larger size to reduce the pixelization. Later, I looked up images of the car, and saw that the red was darker, more like "fire truck red." So I went back, made selections using the pen tool, turned to layers, then flooded w. reds sampled from photos. Adjusted opacity and merged down. Better results, but I've been doing vector pen drawing for quite awhile now. Not "auto-magic" and a little too saturated to fit the rest of the image.
  2. A beer in one hand, a pulled pork sandwich in the other, about 5 min.
  3. Hi, jefferis, I've never had any buggy problem w. any of the built in shape tools. Can you post what the tool context bar looks like?
  4. Hi, American, I'm just a hobbyist when it comes to image processing, and after years of puttering around, I still have only the vaguest notions of how these things work. I think your perception of redundancy is because there can be a certain amount of overlap among filters and blend modes that are similar. But there are typically situations where one or another will give slightly better results. An example. Up until recently, if I wanted a grey scale image, I simply used the HSL adjustment, and dropped the saturation to 0. (Note, in GIMP, desaturation can work on the color average, or its luminance. Slightly different implementation of the same thing.) But then I noticed the the B&W adjustment could produce a wider range of grays depending on the color component. W. some settings, it would be roughly the same as plain desaturation, in others, far more grey levels within specific colors were possible. So if I want to get a grey scale image from something like an old yellowed photo, desaturation works well enough. But if I have something w. lots of colors in it, B&W is the choice. When I 1st started using Affinity, I noticed that it had more blend modes than photoshop. My daughter who uses PS extensively caught that right away. Affinity being relatively new at the time, had less documentation on the modes. I read a bunch of stuff about the PS ones, but then noticed a user here wondering what the exact numeric values for the modes were. That is, what range of hues or luminances, and what percentage of enhancement. From further reading, I found that individual developers implement the same notion is slightly different ways. FWIW, there was a nice explanation about Affinity blend modes here. My take on it is that there are certain filter, adjustments, etc that are most useful for specif images, or specific portions of images. Picking the right one is a matter of finesse.
  5. Tremendous progress. I'll bet it was mostly fun. I think that when centering the focal point, the problem may be a simple manual one. The mouse button has not been released completely, and so the point shifts slightly as one moves ones hand away.
  6. Hi, 828 Design, I've had some result like yours, but nothing quite as extreme. Try this. Select the zig-zag section of nodes. Make them smooth, not sharp, and then use the smoothing widget. My experience is that most of the excess nodes are gathered into a smooth curve, tho' slightly distorted. See the attached showing what happened when most of the teeth were corrected AFAIK, the stroke is created by a series of "dabs", and the expand routine tries to follow the edges of the dabs. When there is a tight curve in a stroke, the dabs stutter. When the stroke is scaled up, the stutters are big enough that the expand routine can follow them. In this instance, not scaling up works better. The same shape shown in the attached when expanded at original size did not make the saw teeth.
  7. Hi, Gitmesteak, I think the problem is based on terminology. A pixel brush can be applied to a vector shape, but does not conform to the vector outline. It can be limited by the vector, but is applied manually. A vector brush is a pixel brush stretched and perhaps repeated along a vector. An outline of the shape, such as you presented, can have a vector brush traced over it. It can not have a simple pixel brush applied. While the stroke dialog shows an icon for a brush, that only applies to vector brushes. See the attached for examples of pixel brushes used within vector shapes, and a vector brush tracing an object. Smear.afdesign
  8. Erm, a duplicate thrice fold. Actually, the original, and 3 others, so 4 copies available, 1 at 90 min interval, 1 noon and night, 1 each mid-morning, which is what I used to review the data entry integrity of the previous days work.
  9. I managed a department's digital resources. I had triple duplicates made every 90 minutes to different machines in different fire protected zones. The fellow that took over from me contracted for off site encrypted in addition. For myself, storage is now so ridiculously cheap, and my cloud connection so hideously slow and until recently unreliable, that a handful of SD cards squirreled away in various places does the job. As a BTW, I'm going to wipe out an old snowball iMac before recycling, and the HD on that is stores less than the SD card plugged into my current machine.
  10. Try the cog tool, and amazing shape maker. The cogs can be formed to thin lines, and made in different depths. They can then be converted to curves, broken at the nodes, and superfluous lines deleted. Attached, a quick and sloppy version which was started by making 1 48 toothed cog, and altering it to ones of 24, 12, & 4 teeth w. different hole radii. Duplicated, and reduced proportionately for the second ring.
  11. gdenby

    Designer Pocket Watch Illustration

    Hi, doar, Exquisitely simple. Everything just right. Hat is off to you, and thanks for the example. I opened it w/o problem in the 1.6 consumer version, and the 1.7.0.10 beta w/o problems.
  12. gdenby

    My First digital painting on affinity designer

    Yes, good. Obviously a 1st step, but it feels good to look at it.
  13. If the bottling company needs all vectors, try this. Add the stripes that will be the etched glass together. Duplicate the text layer, convert it to curves, ungroup, and add the letter shapes into 1 shape. Then subtract from the stripes. Repeat w. the star array. I don't know how the bottle making process works, but I suppose they 1st etch the glass w. a stencil that protects the areas to remain clear, and then apply a stencil that will hold the white enamel for the stars. If they can provide info about their process, it might help setting the vector file up.
  14. Just read your reply. I'm not certain why you made them paint layers, but if they are vectors, they can be rasterized after the text is subtracted. Most work can be done as all vectors, and rasterized as needed. That way, the shapes can be moved, resized, recolored, etc w/o generating pixel artifacts till the end.
  15. Sorry if I was not quite clear, I was just about to go out the door to a legal meeting. I should have said add them together, not add them in. That will create a mesh of stars that can be subtracted as one object from the stripes, which would also need to be added into one object.
  16. Hi, LukeJS, Looking at the layers panel, I infer the groups are rows of stars. Objects cannot be subtracted from groups. Ungroup the starts, all of them, and add them in. Add all the diagonals. Subtract the star curves from the diagonals. The text can then be subtracted from the diagonals w. star holes. If the document has a white background color, and is not set up to be transparent, that white remains under the subtracted remains.
  17. Did not know the text could be edited, tho' in the few things I tried, adding letters could cause theletters to over-run one already there. So, yes, dealing w. spacing would not be easy.
  18. It appears all the text was converted to vector objects. Its been years since I used either photoshop or illustrator, but once text was converted to curves back then, it was no longer able to be changed as text. It was just a vector shape that could be used by a postscript capable printer. It was something that was done because not everyone had access to specific fonts.
  19. Lucio, If you edit an image in the tone mapping personna, you can save the settings, such as exposure, contrast, etc as a new preset. AFAIK, the image needs to be in RGB 8 color space.
  20. Hi, Lucio, In my experience, styles are generated from the strokes and fills of a vector object. They don't come from pictures. Pictures can have tone mapping presets added to them. Is that perhaps what you want?
  21. A result from running the sample image thru A-Photo to clean it up some. Then a trace in Inkscape. Then some color work back in AD. The resulting vector file was quite large, 4+ Mb, so offering a visual only.
  22. Hi, chocomel, Do you mean you want everything to be a vector representation of the odd slice thingy? BTW, what is that?
  23. Hi, dhb, My complicated solution. I tend to work that way. Lets hear it for fuss and bother. I noticed a couple of things in your file, and wonder why. The two rectangles have different gradients, and at the adjoining overlapping edges, they have different grey values. Seems like there is no way they could not have a noticeable edge. The texture images has different values as a result. The banding seems to be worse w. the shadow fx turned on. Probably not the thing to do. Also, the rectangles, having straight edges, create a sharp line wherever there is a difference in the texture image. Attached is a file where I tried something different. In stead of having a gradient fill, w. a child image layer, I saved off the concrete texture, and used that as a bitmap fill. That way it could be reshaped as needed for variety, and the rectangle could have transparency applied to it, causing more variation if you need bands. I made the rectangles curves so I could roughen edges. Also added another layer of curves to obscure the seams more. SHARP_EDGE.afdesign
  24. Hi, Mark, I'm pretty much a novice when it comes to A Photo, but here are 2 things to try. Make the selection brush smaller, and turn on snap to edges. That should grab contiguous areas of the flat colors. Alternatively, use the flood selection tool, set it to a small tolerance, maybe 2%, and turn off contiguous. That should pick out any area in the image w. the sampled colors.
  25. Hi, PLSNDL, I'm assuming you have snapping turned on. It might help to start the construction w. a grid in place, and have snapping to that also turned on. No need to draw a complete triangle, just the bottom section, and the diagonal. Use the pen tool in polygon mode, and adjust the lines to arcs as you like. Duplicate the shape, flip, and slide to the side till you get snapping indicators. With both portions selected, switch to the node tool, and use the join curves widget. There will remain a red node dot, and if you zoom in on that, you will see that there actually 2 nodes on top of each other. You can slide the top on to the side. Slide it back over the lower node till it snaps, and marquee select both. Then use the close curve widget. Assuming you didn't happen to nudge the node after snapping, the two should fuse. Note, the 1.7 beta works a little better w. node positioning, and other shape fine tuning, but I tried the above in both 1.6 & 1.7, and it worked OK.
×