Changing dialog values via mouse cursor

[DarkClown]

I'm struggelng with a functionality that seems I have not fully understood. When a dialog is open there is this nice feature to change the value of an numeric input field by pressing the left mouseky and dragging the mouse outside the dialog.

1) Is there a way to define which field is selected to change the values? It seems it's always the first value that can be changed by dragging the mouse outside the dialog and even clicking on a differrent field and activating it the mouseclick+dragging will still only change the value of the first entry field

2) dragging the mouse across the screen will reduce the filed value to a random(!) minimum value and than start increasing the value again but it never reaches the minimum value possible in the field (e.g. it starts with 70%, goes down to 40% and than starts increasing up to 100% again when dragging from left to right) This seems to be very strage and I don't know what 'm doing wrong ...

Cheers, Timo

[walt.farrell]

Which dialog is giving you trouble, Timo. And which application?

[DarkClown]

Sorry, Walt, talking Affinity Photo here - productive and beta version.

As you can see in the video the parameter for radius never goes below 20px - despite the fact that I'm continuosly moving the mouse from one side to the other. At a certain point the value does not decrease any more but increases again. As well I can't select a different edit field to get the functionality with dragging the mouse. I can activate a different field but dragging will still only change the values of the radius field. See attached video.

Cheers, Timo

[walt.farrell]

Thanks. I see the same thing, using the beta.

[DarkClown]

So it's not between my ears but in the code?

[DarkClown]

In that case maybe someone should move it to the "Bug" section ....

[Chris B]

Hey DarkClown,

It is looking at the X and Y axis. So if you have say, a 1000 x 1000 px document and your mouse pointer is in the dead center, (500 px on both axis) you should be able to get back to 0 px in the field.

It looks as though the reason you can't get it below the 20 px is because your mouse is not returning to the dead center of the canvas—it is probably just off on one of the axis. It's hard to tell as you don't have the ruler enabled. 

The middle would be much harder to find on a canvas that doesn't have an equal vertical and horizontal pixel ratio. 

As a test, work out where the middle is and place a dot, then drag and make sure your cursor hits the dot and you will be able to go below the 20px (hopefully 😅)

*Note* It looks like this is true when you start the drag from the center... If you start it elsewhere, you might be up the creek. This is how my brain works it out anyway. I'm happy to ask dev.

[DarkClown]

Hi Chris,

sorry, I didn't fully understand what you try to explain. As I showed in tha previous video the problem existst independent from wether I use the X or the Y axis.
I gave it another - more analytical - go in this video. Marked the center for X and Y in a small document. started with a clear radius offset and moved the mouse from the one side to the other. Even at the center the value never reaches something even close to 0px. Please check this other example ...

Cheers, Timo

[Andy Somerfield]

The value is *distance from mouse down* - not simply *horizontal difference*.

[DarkClown]

Meaning?

To be able to adress the full value range I need to precisely find the mid of the document and start dragging from there?? How is a user supposed to find this spot in a 24MP photo? In case this is correct the function is simply not usable ... it seems simply impossible to get down to reasonable values for the unsharp mask radius (of e.g. 0.2px or 0.3px) if you precisely have to hit the mid point of a huge document to reach this value.

Aside this: is there a valid reason for this behaviour? Horizontal difference from any starting point would easily make the function perfect, understandable ... and usable.

Cheers, Timo

[Mark Ingram]

You don't need to find the mid point of anything. You just need to click somewhere, and drag. The value is the radius of the initial click point and the current cursor position. If you can't back to 0 (i.e. get back to exactly where you initially clicked) just press Ctrl+Z to undo the edit.

[Gear maker]

Wow, I had never heard of adjusting a value this way.  It seems to work for higher values but to get it to work for lower values it's quite a task.  It doesn't use the center of the document it was a point off center by a fair amount.  I also had to slow down my mouse to where I could barely move an inch before having to reposition my mouse.  Then with a lot of playing I was able to get the value down it 0.4.  But it took me about a minute with the slow speed and holding my breath to do it.

DarkClown, I also couldn't find any way to use this adjustment on any slider but the first in any adjustment.

This would be handy for large numbers, say numbers over 100 where the slider does not allow.  But otherwise...

[Gear maker]

I take it that this adjustment procedure only works in AP.  So far I can't find any place it seems to work in AD.  It would be handy in FX when trying to adjust "for effect" after reaching the max for the slider.

[DarkClown]

16 minutes ago, Mark Ingram said:

You don't need to find the mid point of anything. You just need to click somewhere, and drag. The value is the radius of the initial click point and the current cursor position. If you can't back to 0 (i.e. get back to exactly where you initially clicked) just press Ctrl+Z to undo the edit.

So instead of just finding the original horizontal (or vertical) starting point you have to adjust the mouse to X AND Y axis (radius) in synchronity to adress lower values of the field. I'm not sure if you ever tried that in practice ... aside the fact that you increase complexity of handling by another dimension this procedure is far away from "intuitive". It seems like a gambling game for underchallenged people who enjoy mikado ... Just my 2 cents ... guess that I'm back to the keyboard. But I really liked the idea and the function! A medal of honour for the initial approach

[Mark Ingram]

1 hour ago, DarkClown said:

So instead of just finding the original horizontal (or vertical) starting point you have to adjust the mouse to X AND Y axis (radius) in synchronity to adress lower values of the field. I'm not sure if you ever tried that in practice ... aside the fact that you increase complexity of handling by another dimension this procedure is far away from "intuitive". It seems like a gambling game for underchallenged people who enjoy mikado ... Just my 2 cents ... guess that I'm back to the keyboard. But I really liked the idea and the function! A medal of honour for the initial approach

If you want to go back to zero just click the document view, it will reset. The idea of dragging on the document is to make the adjustment visually, so you don't really need to worry about the specific values (unless you want to, then they're still available on the dialog).

[DarkClown]

Mark, I don't mean to be nagging all the time. The functionality seem to be cool and an easy way to edit parameters. In terms of the unsharp mask dialog (and I'm aware this might be different in other dialogs) reasonable radius levels range from 0.2-0.8px ... This on the other side is an value area for the dialog that is extremly hard to acomplish with the current method. While higer values can me accessed a lot easier due to the nature of the implemented algorythm. Of course I can always reset the edit field value or general setting but the feature only makes sense if it makes things easier. (Workflow!   ) Before I try to luckily hit the right number with the mouse I'll always choose the manual keyboard edit since it's not connected to filigran mouse juggling. Making a long story short: Maybe you should re-think the intuitivity (it now took quite a while to figure out and explain the basic functionality to an not unexperienced user) .. and effectivity (despite the fact I love the idea the way it's implemented will not make me use it in future since it is not as efficient as it could and needs to be - from my personal perspective). Feels like you missed out on a good opportunity ... excellent idea but struggled on the last meter ...  maybe keep my input as a remark for future improvements ...
Cheers, Timo

[kirkt]

@DarkClown - If you are zoomed into 100% to view the image, then mousing on the image area to manipulate the slider gives you finer levels of control.  And, when applying Unsharp Masking at such small radii, you should be making the adjustment at 100% zoom anyway, so this makes sense.  That is, the incremental changes in the slider when you use your mouse to manipulate it are related to the zoom level - i.e., the pixel location on the image.  This is a clever implementation of this interface feature.  Perhaps in the future, a modifier key (like SHIFT) could be employed to divide the current increment by 10, to temporarily override the current zoom level increment.

It sounds like the reference point for "0" is established on the image area when you first click the mouse (and then hold it to start sliding the mouse) - so pick a feature in your image that is easily identifiable and click on it to start your mousing to change the slider value.  This way, you know where zero is on your image if you need to slide your mouse back toward it.

Another feature that would be nice for this interface mode would be a key (something akin to TABbing through web fields) that would permit the user to step through dialog fields while mousing the field values.  This way you could mouse the value for radius, then hit the TAB key, for example, and change to the Factor field, mouse the value of that, tap the TAB key again, change the Threshold field by mousing.  Etc.  Currently, the TAB key hides the interface, but you get the idea.

Kirk

[DarkClown]

@kirkt: True, the precision gets better with higher zoom factor. And of course I do the adjustments in 100%. But this is by far not high enough to be able to steer the slider precise enough. The problem is significantly increased by the fact that BOTH axis are the base to the value (it would ba a lot easier if only the x-axis would be the parameter). Eventually it's a lot easier to just click the slider itself and adjust the value this way.

[R C-R]

6 hours ago, DarkClown said:

But this is by far not high enough to be able to steer the slider precise enough.

I think this feature was intended for quick, coarse adjustments, not for fine, precise ones. Accordingly, I use it to quickly get an approximation of the precise setting I want, & then hover the mouse pointer over the value & use the scroll wheel (optionally with the Shift or Alt modifier keys) to fine tune it.

8 hours ago, kirkt said:

It sounds like the reference point for "0" is established on the image area when you first click the mouse (and then hold it to start sliding the mouse) - so pick a feature in your image that is easily identifiable and click on it to start your mousing to change the slider value.  This way, you know where zero is on your image if you need to slide your mouse back toward it.

It does not matter what "zero" point reference is chosen because only the distance the pointer is dragged sets the radius value. So if you press the mouse button & do not drag the pointer at all before releasing it -- IOW, just a simple click -- the radius value is set to zero (because the distance the pointer is dragged is zero).

It is similar to using the Ellipse tool to drag out a constrained circular shape from the center point. Because the size of the circle is determined by how far the pointer is dragged, it does not matter what direction the tool is dragged, only how far from the center point it is dragged.

[kirkt]

15 hours ago, R C-R said:

It does not matter what "zero" point reference is chosen because only the distance the pointer is dragged sets the radius value. So if you press the mouse button & do not drag the pointer at all before releasing it -- IOW, just a simple click -- the radius value is set to zero (because the distance the pointer is dragged is zero).

It is similar to using the Ellipse tool to drag out a constrained circular shape from the center point. Because the size of the circle is determined by how far the pointer is dragged, it does not matter what direction the tool is dragged, only how far from the center point it is dragged.

I think there was confusion earlier in the thread about 0 being at the center of the image. 
kirk

[R C-R]

1 hour ago, kirkt said:

I think there was confusion earlier in the thread about 0 being at the center of the image.

Clearly there was. I think that is because it is tempting to assume the "radius" of any filter always involves setting some origin or center point on the canvas, from which the effect of the filter spreads (IOW, radiates) outward. While this is true for many of the filters, for others it is not.

For example, the radius of the Unsharp Mask determines the number of pixels affected around light colored pixels in everything it is applied to, so it has no single point of origin on the canvas, if that makes anything clearer.