AP - Curve thickness question

[DarkClown]

I wonder if there is a way to get a 1pt vector curve in Affinity Photo that is black (and not 50% opaque) - See attached Video.

Cheers, Timo

[Pšenda]

3 minutes ago, DarkClown said:

I wonder if there is a way to get a 1pt vector curve in Affinity Photo that is black (and not 50% opaque) - See attached Video.

Line position (for horizontal line Y) must be 0,5 px.

[DarkClown]

Hmmm, it "somehow" solves the problem .... but comes with interesting additional effects (See video)

[Pšenda]

You're welcome.
I do not see any problem.

[R C-R]

5 hours ago, DarkClown said:

Hmmm, it "somehow" solves the problem ....

but comes with interesting additional effects  

This is related to antialiasing -- curves of odd thicknesses fit perfectly into rendered pixel boundaries if positioned on ½ pixel boundaries but ones with even thicknesses won't, resulting in ½ pixel thicknesses on either side being antialiased. 

You can use the Blend range Options to force off antialiasing, but this will result in the effect sometimes known as the jaggies if the curve is not perfectly horizontal or vertical.

This antialias example.afdesign may make it clearer what I mean. Note that you will have to set AD to the Pixel View Mode to see the effect, or if you are using Apub, switch to the Photo Persona.

[NotMyFault]

For lines with stroke alignment to "center":

• for lines with odd width (1,3,5,7 ...) it must be positioned to .5 position
• for lines with even width (2,4,6,8 ...) it must be positioned to .0 position.

When you change the alignment from middle to inner/outer, you can leave the line at .0 position (only for closed curves, not for open curves).

 

[DarkClown]

On 12/24/2021 at 2:40 PM, NotMyFault said:

When you change the alignment from middle to inner/outer, you can leave the line at .0 position (only for closed curves, not for open curves).

I played a bit around with the settings and it left me even more confused. Regrettfully I could not find any "proper" information on the detailed curves settings (Maybe the Affinity help files could do with some improvements). The alignment buttons have no effect on an open curve. Maybe because there is no defined "Out" an "In" in an open curve. Despite that the setting would make sense - it probably would need an adittional "invert" button. I'd just like to ask the programmers, why a setting is offered that makes no sense? Would be more intuitive not to offer the buttons at all or at least to deactivate them as long the curve is not closed.

Also the "join" settings leave me a bit confused. The join mitre setting seems to switch from what I understand as "mitre" to "beveled" - depending on some parameter I cound not find an explanation for. And it surprisingly also changes the way a curve is shaped outside the joint point. Certainly on sharp angles.

See this example: Mitre Joint vs Bevel Joint on a sharp angel of a closed curve (NOT a joint)

 

From my understanding this is a setting for "Joint" behaviour and not the curve behaviour ...
As I said ... confused ... maybe a feature.

Cheers, Timo

[NotMyFault]

26 minutes ago, DarkClown said:

Also the "join" settings leave me a bit confused. The join mitre setting seems to switch from what I understand as "mitre" to "beveled" - depending on some parameter I cound not find an explanation for. And it surprisingly also changes the way a curve is shaped outside the joint point. Certainly on sharp angles.

https://developer.mozilla.org/en-US/docs/Web/SVG/Attribute/stroke-miterlimit

stroke-miterlimit

The stroke-miterlimit attribute is a presentation attribute defining a limit on the ratio of the miter length to the stroke-width used to draw a miter join. When the limit is exceeded, the join is converted from a miter to a bevel.

[Pšenda]

46 minutes ago, DarkClown said:

The alignment buttons have no effect on an open curve. Maybe because there is no defined "Out" an "In" in an open curve. Despite that the setting would make sense - it probably would need an adittional "invert" button. I'd just like to ask the programmers, why a setting is offered that makes no sense? Would be more intuitive not to offer the buttons at all or at least to deactivate them as long the curve is not closed.

For interrest.

 

[NotMyFault]

The video shows how miter limit works.

 

Stroke width: 10px, centered. Miter set to 2 (factor)

For miter, 1/2 of 10 = 5px is relevant when centered. As soon as the miter exceeds 2*5px (larger circle), it switches to bevel. Bevel is cut within 5px circle around edge node.

2022-01-03_15-17-36.mkv

[DarkClown]

Thanks for your explanation and the link to the discussion.
I don't see the usage of the "Mitre factor" .... seems like a random limiter - but I guess there might be some usage for it I did not reveal until now 🙂

Still there I have no explanation why the joint setting changes the curve behaviour far away from any joints ...

[NotMyFault]

14 minutes ago, DarkClown said:

Still there I have no explanation why the joint setting changes the curve behaviour far away from any joints

Unsure if a got what you mean by "joint" and "far away from joints". Can you elaborate?

Join (without t) in the sense of the stroke UI is every node (Knoten) (= join, Verbinder) of the curve, except the start / end node in case of an open curve. And the "Join" settings only impact the node/join within the limits of stroke width * Miter factor.

[R C-R]

4 hours ago, DarkClown said:

The alignment buttons have no effect on an open curve. {...}  I'd just like to ask the programmers, why a setting is offered that makes no sense? Would be more intuitive not to offer the buttons at all or at least to deactivate them as long the curve is not closed.

The stroke alignment setting becomes a persistent property of the curve, whether open or closed, but the inner & outer options have no effect unless/until the curve is closed. So for instance, this can be useful as in this 3 alignments.afphoto example if it is desired to close two or all three of these open curves without changing the stroke alignment previously applied to them.

This may make more sense to you if you keep in mind that previously created closed curves can be opened, divided into two or more open curves, & some or all of those open curves can be closed again.

The short version is it can be useful to set the stroke alignment even for open curves because they can be converted to closed curves at any time, one at a time or in groups.

[DarkClown]

18 hours ago, NotMyFault said:

Unsure if a got what you mean by "joint" and "far away from joints". Can you elaborate?

Sorry - my spelling mistake (Start and Endpoint of a curve look like "Joints" (Gelenke) - so I got mixed up here .... Correct wording is "Join" (Knoten)
And what I mean is that the join setting does not only affect the end and startpoint of a curve but also a thight angle of such a curve (far away from the joins). See attached example:

We are looking at a single curve (1 start and 1 end point). See the behaviuor at the bending angle of the curve (Arrow).

Mitre Join:

 

and bevel Join:

 

[DarkClown]

14 hours ago, R C-R said:

The short version is it can be useful to set the stroke alignment even for open curves because they can be converted to closed curves at any time, one at a time or in groups.

I get you.
But I wonder, why the alignment is not made available for open curves as well ... (despite the fact there is no inside and outside). The definition of inside and outside could either be based on what would be inside if the curve would be closed - or you call it "rigth" and "left" looking at the curve from the starting point - or the "one side" and the "other side"

[NotMyFault]

Ah now I see. Interesting edge case.

Moonshot explanation : 

You (Affinity) created a kind of virtual node / joint, and the mitre limit applies to all nodes, self created and virtual where the curve makes sharp turns.

 

[Pšenda]

3 hours ago, DarkClown said:

The definition of inside and outside could either be based on what would be inside if the curve would be closed

Just out of curiosity - what would be used to determine what an open curve/line would look like if it were closed? If the curve is open - for example a straight line, it can be closed on both sides.

[DarkClown]

3 minutes ago, Pšenda said:

Just out of curiosity - what would be used to determine what an open curve/line would look like if it were closed? If the curve is open - for example a straight line, it can be closed on both sides.

That probably has been answered by the designers already. Elsewise R C-R s comment would not make sense:
 

Quote

The stroke alignment setting becomes a persistent property of the curve, whether open or closed, but the inner & outer options have no effect unless/until the curve is closed. So for instance, this can be useful as in this 3 alignments.afphoto example if it is desired to close two or all three of these open curves without changing the stroke alignment previously applied to them.

 

[R C-R]

6 hours ago, DarkClown said:

But I wonder, why the alignment is not made available for open curves as well ... (despite the fact there is no inside and outside). The definition of inside and outside could either be based on what would be inside if the curve would be closed - or you call it "rigth" and "left" looking at the curve from the starting point - or the "one side" and the "other side"

Consider for example an S-shaped open curve or an open curve that looks like a series of ocean waves. Those curves are among the many that have no obvious inside/outside or left/right sides. A few examples: inside or outside?.afdesign

Regardless, like I said the stroke alignment property is always available, even when it has no obvious effect.

[DarkClown]

42 minutes ago, R C-R said:

Those curves are among the many that have no obvious inside/outside or left/right sides.

I'd disagree ... there might not be an "inside" "outside" - but for sure there is always a "left" and "right". Seeing that every curve has a defined startpoint (first click) and defined endpopint - standing at the startpoint following the curve there's always a left and right as long you stick to 2D.

[R C-R]

7 minutes ago, DarkClown said:

I'd disagree ... there might not be an "inside" "outside" - but for sure there is always a "left" and "right".

What if you use the Reverse Curve option on the Node Tool's context toolbar? Wouldn't that flip the left & right sides? 

Regardless, there is no "Align Left" or "Align Right" stroke option in the Affinity apps. If you think there should be, the place for that is in the feature request section.

[Old Bruce]

1 hour ago, DarkClown said:

I'd disagree ... there might not be an "inside" "outside" - but for sure there is always a "left" and "right". Seeing that every curve has a defined startpoint (first click) and defined endpopint - standing at the startpoint following the curve there's always a left and right as long you stick to 2D

Falls apart quite easily. I duplicated the lower item and then used the node tool to move the C and D nodes to their new locations.

[Pšenda]

1 hour ago, DarkClown said:

but for sure there is always a "left" and "right"

The discussion of about 4 years old thread (see my previous link) is repeated again - some users just do not understand, that each curve has a beginning and end (just follow the red node), and therefore always has a left and a right side.

 

15 minutes ago, Old Bruce said:

Falls apart quite easily.

Unfortunately, you haven't displayed the curve nodes (just select the Node tool), so it's not clear where the curves begin and end in your examples. So there's nothing you can say about your claim that something fell apart.

[Old Bruce]

2 minutes ago, Pšenda said:

Unfortunately, you haven't displayed the curve nodes (just select the Node tool), so it's not clear where the curves begin and end in your examples. So there's nothing you can say about your claim that something fell apart.

Take a few minutes and try it your self. The nodes are at the start (A in both cases) and end (D in both cases) then there are two more at B and C.

[Pšenda]

1 minute ago, Old Bruce said:

Take a few minutes and try it your self. The nodes are at the start (A in both cases) and end (D in both cases) then there are two more at B and C.

It would be enough to make a screenshot with the Node Tool selected. I have nothing to try because I don't understand your picture - I don't know what points A - D are, I don't know what the left / right texts are. Is it one curve or several independent segments?