Delete a segment from a circle

[Gaz]

If I have a circle, and I want to delete half of it - to create an arc (i.e. the circle is no longer closed), how can I do that ? :-)

 

I've tried convert to curves, thinking this may stop it from closing. However every time I delete a point, a new curve is generated between the two open points.

Any ideas ? Is there a check box somewhere to not close end points ?

 

Thanks

Gary

 

 

[MEB]

Hi Gary,

Welcome to Affinity Forums :)

You have to convert to curves to be able to edit its nodes, then select the Node tool, click on one of the nodes, go to the context toolbar and press Break Curve from the Action section.

Do the same for the node on the opposite side of the circle, then select one of the segments with the Move Tool and delete it.

[Gaz]

Excellent, a slightly long process, but once you know where everything is, it's easy. Thank You :-)

[jkehoe]

Hi Everyone,

 

I am stumped by this and I can't delete a segment of a curve, or at least it doesn't work as spelled out above. I opened a jpeg of an insect and set that as the background layer. I then added a new layer, chose the Pen tool, and outlined the main body. I repeated this by adding more layers for wings, legs, head, etc. Now I'm struggling with Designer while trying to delete segments of these shapes that are pen tool lines. I select the body layer, click "Convert to Curves," select the Node tool, click on one of the nodes and it becomes solid blue, I then click "Break Curve" and it become outlined in red (white center, like all the rest). All other nodes are still showing, as if selected. Next, I click another node that's four stops away from the first one--it turns blue. When I click "Break Curve," that selected node turns white, just like all the others, aside from the first one which is still outlined in red. If I hit the backwards Delete key (the main delete key), the entire shape gets deleted (all nodes appear to be selected, so that makes sense). Please help =( I know the answer is in front of my nose, but I'm just not seeing it.

 

Thank you so much. I really like both Designer and Photo (bought both) and want to migrate to them.

John Kehoe

[Oval]

 I know the answer is in front of my nose, but I'm just not seeing it.

Hi John, just deselect everything (a click besides), then click on the segment and BACKSPACE. Or: type V, click on the segment and BACKSPACE.

[jkehoe]

Hello Oval,

 

Thank you very much! I got it to work much like you recommended, but I wasn't aware of the use of the V key. I really appreciate your input.

[Gaz]

To delete a segment, it looks like you need to select the nodes either side of the segment you want to delete, and then break their connection. However for it to work you need to do the process to each node individually, you can't SHIFT select and disconnect two nodes at the same time, it'll just ignore you.

 

Once that's done, click off the shape, then using the object selection tool (move tool) select the segment you've just isolated and it should be separate from the rest of the shape / curve.

 

They are supposed to be improving this process in later versions.

 

 

 

I really like both Designer and Photo (bought both) and want to migrate to them.

 

So have I, and have already migrated. Apart from the odd quirky process like this, I'm not really missing Illustrator or Photoshop at all. In many ways I think they are both faster and better.

[uncle808us]

Good tips thanks.

[Oval]

Yes, but not good enough to make it quicker than in other apps or to replace a better knife tool.

 

Quicker way without the need to change to the move tool: click (to add a new node in the segment), break curve, marquee (the new nodes or more), delete

 

or: just add a new node, break curve, select, delete, select, delete

 

 

@ jkehoe: already using ESC (key) 4 deselecting?

 

@ Gary: with the Crescent Tool it is possible to create an arc very quickly by dragging one curve over the other (yes it is closed)

[SKT7]

Is there any update on this workflow?
The current workflow is too tedious, esp, if you have to it many times? 

[MEB]

Hi saikat,
In the latest version you can select and break multiple nodes at the same time. There's no other method yet, sorry.

[PixelPest]

I still don't understand why Serif is so reluctant to implement a parametric circle tool. 

Cheers

[SKT7]

10 minutes ago, MEB said:

Hi saikat,
In the latest version you can select and break multiple nodes at the same time. There's no other method yet, sorry.

Ok. That at least reduces 1 step. Thank you!

[MEB]

4 minutes ago, PixelPest said:

I still don't understand why Serif is so reluctant to implement a parametric circle tool. 

Cheers

Hi PixelPest,
Have you filled a request for this in the Feature Requests section (in case there's none) detailing how you expect/want it to work?

[Alfred]

7 minutes ago, MEB said:

Hi PixelPest,
Have you filled a request for this in the Feature Requests section (in case there's none) detailing how you expect/want it to work?

I’m sure this has been discussed before, Miguel. In Inkscape if you drag outside the ellipse you get a segment (i.e. an arc plus two radii) like the result of using the Pie Tool in the Affinity apps, but if you drag inside the ellipse you get an arc on its own.

[PixelPest]

Sorry but "No" - you will end with a closed shape not an arc.

 

Cheers

[MEB]

I will pass this feedback to the dev team. Thanks.

[PixelPest]

Thank you!

[R C-R]

3 hours ago, PixelPest said:

I still don't understand why Serif is so reluctant to implement a parametric circle tool. 

As I understand it, it is because it would require defining curved vector paths (including but not limited to circles & ellipses) using polar coordinate parameters instead of Cartesian ones like is done now for all vector paths, whether straight or curved.

Among other things, since computer screens must display everything mapped to cartesian coordinate values, this would incur a huge amount of computational overhead, making it impossible to maintain the fast pan & zoom refresh rates that make those operations so smooth in the Affinity apps.

[PixelPest]

You are kidding aren't you? It has been available since my early days as a Freehand.alpha tester more than 30 years ago and no other vector app I have in my portfolio is missing this feature.

[R C-R]

4 hours ago, PixelPest said:

You are kidding aren't you? It has been available since my early days as a Freehand.alpha tester more than 30 years ago and no other vector app I have in my portfolio is missing this feature.

No, I am not kidding. For one thing, it is impossible to do more than to approximate a perfect circle with bezier curves -- see for example this passage in the Converting a circular arc into Bezier arc section of this scholarly paper & the discussion that follows:

Quote

It is impossible to draw an absolutely exact circle with one Bezier curve. But we can approximate a unit quarter of a circle (900 arc) by a cubic Bezier curve with an error 1.96×10-4 in the radius [3], what is acceptable for most practical cases.

That small an error may be acceptable in many apps but consider that the Affinity ones are capable of zooming over a million to one zoom range, so at extremely high zoom levels on large radius circles the error would be quite obvious, making it totally unacceptable for precision work.

Put in another more general way, mathematically an ellipse is a curve for which the sum of the two distances to its focal points is a constant for the set of all points along its path, collectively known as its locus. A circle is just special type of ellipse for which the two foci are the same. For either type, to achieve a truly resolution independent vector representation, the locus must include an infinite (!!!) number of points. AFAIK, even in the last Freehand version released before Adobe killed it, circles were created in the same way they are in Affinity, using the Ellipse Tool & the Shift key to constrain an ellipse to an approximation of a true circle. Both apps limit the locus to a manageable set, thus neither one can create truly resolution independent circles or ellipses.

Of course, it is possible to achieve an approximation to any required or desired limit by increasing the number of points but computationally it is very expensive to do that, particularly in an environment using a mix of polar & cartesian coordinates.

[PixelPest]

I already know that neither quadratic/cubic-Bézier-Curves/splines nor NURBS can describe perfect circles with 4 anchor points for instance.

But all I´m hearing is that you´re saying what works since its invention in the 1960ties and used in computer graphics and for bazillions of TrueType/Postscript/Type1/2/3/OpenType fonts, in CAD/CAM, vector-graphic editors etc will not work in an Affinity.app.

If that is true I´m waisting my time here.

I often need partial circles for animation purpose as splines:

and all I want for partial circles now.

(Aside 9 other things AD can´t do for me but Inkscape will; Inkscape user since 2004. Responsive and listening developer; bugs are fixed within a day if not an hour in my experience. Adjustable/rotatable/lockable ruler/guides, real tangents, interpolation between splines for guilloché-like pattern, path offsets, Spiro curves, Envelope/Lattice-Deformation, Cloning by formula, Pattern along Path, Mirror Symmetry, Dimensions+measure Paths, Bend, quick+reliable auto-tracer, (for complex selection-to-SVG I use GIMP) Access to XML structure.)

All the best.

P.

6 Examples from Intaglio, Inkscape, Gravit Designer, LibreOffice/Draw, EazyDraw, FreeHand

[R C-R]

25 minutes ago, PixelPest said:

6 Examples from Intaglio, Inkscape, Gravit Designer, LibreOffice/Draw, EazyDraw, FreeHand

Do any of those actually create geometrically perfect circular segments?

EDIT: off topic but I noticed from your Freehand example it appears that your are running Macromedia Freehand on what appears to be a recent version of the Mac OS. If so, how did you manage that? For me, just to run Adobe's Freehand MX I have to fire up an old iMac that can run it on an OS X version earlier than 10.6.

[Alfred]

12 hours ago, R C-R said:

No, I am not kidding. For one thing, it is impossible to do more than to approximate a perfect circle with bezier curves -- see for example this passage in the Converting a circular arc into Bezier arc section of this scholarly paper & the discussion that follows:

Quote

It is impossible to draw an absolutely exact circle with one Bezier curve. But we can approximate a unit quarter of a circle (900 arc) by a cubic Bezier curve with an error 1.96×10-4 in the radius [3], what is acceptable for most practical cases.

 That small an error may be acceptable in many apps

Speaking of small errors, I’m a little surprised that in the published version of a “scholarly paper” a superscript zero was allowed to slip through in place of a degree symbol. It should also say “which is acceptable”, not “what is acceptable”, but the (Lithuanian) author clearly didn’t have the article proofread by a native English speaker: even the abstract contains errors (e.g. “problem for” instead of “the problem of”).