Losing focus for selection when zooming

• Oct 11, 2015 - 06:48
Type
Graphical (UI)
Severity
S4 - Minor
Status
needs info
Project

GIT commit: f51dc11

Hi

ZOOM function has few bugs/annoyances

1. Ctrl+ (zoom in) key does not work (Crtl- and Ctrl0 hot keys work, but see 2 below)

2. When zooming in or out, Musescore IGNORES User-selected elements and/or cursor placement in the score. Navigating large complex scores is really annoying, because after every zoom in or out operation User needs to scroll through the entire score to find the measure that he/she tried to zoom into. User selection should be *visible on screen* after each and every zoom in or zoom out operation.


Comments

Loss of focus is far more annoying than lack of Ctrl+=

Navigating a complex score is really annoying when you need to scroll the whole score after each zoom just to find where you were...

Tom

I don't understand what you mean; there should be no need to scroll through the entire score. I would suggest you ask for help in the Support forum, attaching the specific score you are having problems with a D precise step by step instruction to reproduce the problem.

take ANY score that is few pages long and set 100% zoom.
Select some measure on the 4th page and zoom in.

The segment that you just selected disappears from screen.

To find this segment in the newly set zoom view you need to scroll through the score.

When you find the chosen segment - zoom in again.
Again your selection disappears from screen and to find it you need to scroll through.

Am I clear enough?

Status (old) active needs info

Sorry, I still don't understand. When you zoom in, the view stays focused on the same page you were viewing - some portion of what you were seeing before will be what you are seeing after. Of course, zooming in also means that some portion of the page that was previously visible will now be just barely off screen. But even if the measure you were interested in happens to be now offscreen, you don't have to scroll through the "whole score" to find it.

FWIW, "selection" is irrelevant - you might have nothing selected, or the selection might be on another page entirely, or it might span multiple pages, etc. So MuseScore cannot really use selection as its way of guessing wehat part of the page to center on. No matter what it guesses, it's going to be right sometimes, wrong sometimes, but the measure you care about will always be nearby.

It seems to me that if the selection is on-screen now and I click zoom then it should stay on-screen. If the selection is not currently on-screen then it should be ignored when zooming.

What if the selection is bigger than the screen? Like the whole score is selected? Or, what if *part* of the selection is on screen, part not? Or, what if the selection was made by ctrl+clicking one element from the top left corner of the page and another from the bottom right corner? I really don't think using selection as a key here is a great idea; I think it will just make the behavior look unpredictable to the avergae user who doesn't know the secret.

They don't strike me as insurmountable problems. LibreOffice zooms on a selection if it is currently visible, and it zooms in on the end of the selection if it is only partly visible.

Note also that if you do want more control over where the zoom is focuses, simply use Ctr l + mouse wheel; then the cursor position becomes the focal point. I guess we could consider using that same technique for Ctrl++, but I am not sure that would really seem very intutive - if you are using the keybaord shortcut, you probably aren't paying much attention to the mouse position.

FWIW, it seems currently we try to keep the top left corner anchored. I could see shooting for the center instead. But having gone back and forth on these things a few times before over the past couple of years, I know it's trickier than you might think to come up with a behavior that makes sense all the time.

Fair enough. I haven't tried implementing it myself. GIMP seems to ignore the current selection entirely, so there is no standardised behaviour here. To me it would seem intuitive that Ctrl + mouse wheel would follow the cursor (since the mouse controls the cursor) and Ctrl++ (or View -> Zoom) would follow the selection (since the mouse is irrelevant here), but I haven't actually tried them.

I use touchpad, no mouse with the wheel.

My ONLY option to zoom in is via the zoom menu, which totally loses the focus each time.

Musescore comes alive with a mouse with a wheel, but Ctrl+= keyboard shortcut is still needed for those who do not use mouse.

Just because you don't have a mouse doesn't mean you cannot use a mouse wheel gesture. Depending on your specific computer, it might be configured as two fonger swipe, or maybe a swipe along a special zone to the side of the mousepad, maybe in conjunction with holding the Ctrl key, or perhaps something else. See the settings for your touchpad to find out if this feature exists - almost all modern touchpad drivers provide this.

Anyhow, again, as explained above, zooming by shortcut does not "lose" focus. MuseScore simply does not know know what part of the screen you might happen to be interested, so it makes a choice - currently, the top left corner is normally retained. No matter what it picked, though, it would be right sometimes, wrong other times - it cannot read your mind.

Zooming in using "zoom menu" ALWAYS loses focus.

When I select a measure (or other element) Musescore is precisely informed what my focus is.

When I use "zoom menu" this focus should not be lost, but it is. Always.

Keyboard shortcuts for zoom-out (Ctrl-) and for zoom-reset (Ctrl0) both work fine,
but there is NO KEYBOARD shortcut for zoom-in. WHY?
It looks like a bug in Ctrl+ code.

Keyboard shortcuts have some advantages over mouse or touchpad actions: For example they are reproducible, reversible and in many cases faster to implement.

Tom

Not sure what you mean by "lose focus", but by definition, zooming *always* loses sopme part of what you are viewing. It is impossible to zoom otherwise. the only question is which part of the view is retained. As I have said, in MuseScore, it keeps the upper left. Presumably you would prefer some other part of the view, but again, MuseScore cannot read minds. No matter which part of the view it retained, it would be right sometimes, wrong sometimes.

The shortcut for zoom in is Ctrl++. If your keybaord happens to require Shift to access "+", then of course you need to press Shift as part of that. You are also welcome to redefine the shortcut via Edit / Preferences / Shortcuts.

For example: here is a typical view:

zoom-1.png

And here is the what happens after zooming in via the menu or via the shortcut:

zoom-2.png

As you can see, the top left corner remained in view.

Ctrl+ keyboard shortcut for zoom-in, common for most visual applications does not work for me.

I use Windows 10 and the latest Musescore, that says that no update is available.

Try using Google maps and you will see that the focus is NOT decided by the cursor position (which should be free to serve menus and palettes), but by the center of the screen before zoom. This way the focus during zoom operations is never lost, because what you see after zoom will be the center of the screen before zoom.

Tom

Ctrl++ for zoom in did not work on my computer, mo matter if I used Shift or not.
Changing it to Ctrl+= via Edit>preferences>shortcuts did the trick.
Re-defining the shortcut worked too.

Thank you for your help and patience.

But.... using a keyboard shortcut to zoom-in GUARANTEES to see the top left corner of the file, NOT the cursor position as when you zoom-in with a mouse wheel.

So, the problem of losing focus after zoom-in remains.

If you happen to aim to zoom-in on something on 3rd page of your score using a keyboard shortcut, be prepared for having to scroll through the entire score in the new zoom mode.

Tom

The problem is, there are many different keybaord layouts in the world. The default works on most, but it is next to impossible to get one set of shortcuts that works for every possible keybaord layout in the world.

Again, you *do not* need to scroll the the entire score after a zoom. The top left corner always remains in view, exactly as my screenshot shows. So you only might need to scroll slightly down and to the right.

When you zoom-in using a camera, the center is the reference for all zoom operations, not the corner. This is a universally adopted meaning of zoom.

You already use cursor position as a center when zooming using mouse wheel. Why not become consistent and adopt the same strategy for shortcut-based zoom?

Imagine zooming right out to see a 10-page score, placing a cursor on a place of your interest and Ctr++ few times to see the zoomed details in 1 second. No scrolling...

Navigating the score is quite important...

Tom

When using mouse wheel, it makes sense that mouse position is relevant. It doesn't make sense to use mouse position when not using the mouse.

Maps are not read left to right, top to bottom, and thus are different from music. With music, I find it much more intuitive to keep the anchor at the top left rather than at the center most of the time. But again, whether the anchor point is the top left or center, it is what you want sometimes, not what you want sometimes, and the difference either way is negligible - one second of scrolling afterwards gets you where you want.

Your intuition choice makes it necessary to scroll the score in search for a place after zoom.
And if there are repeated segments in the score, User must highlight the fragment of interest, otherwise scrolling is not guaranteed to find the exact fragment.

Why not be consistent and use pointer as a center of zoom? Instead of adopting a dogmatic decision this would allow User to choose.

Expanding Users freedom of choice always results in a better interface.

Tom

As I keep saying, no matter which spot MuseScore would keep in view, it would be the right answer sometimes, the wrong answer other times. Sometimes you will need to scroll slightly after the zoom - that is always true of zoom. it's the very nature of zoom - after a zoom, less is vislbe than behfore, so sometimes you might have to scroll slightly to find the area you were interested in. The actual distance you might need to zoom is the same no matter which spot is chosen as the anchor, and it's always small. Somehow it seems you might be misunderstanding this, as you use phrases like "scroll through the entire score". No, this is never necessary. The point you are interested in is always at most a small distance away, and this would remain true if the default anchor point were change.

Feel free to start a discussion on the forum to see if a consensus develops that a change would make a significant difference. If a large consensus exists saying they would prefer center to top left, we could consider making that change. But I still don't like the idea of a keyboard operation involving the mouse in any way. This is counterintuitive and will make the behavior seem unpredictable, and in any event, in many cases the point won't even be within the score view.

There is really not much more to say here. Right now, the current behavior is by design and as expected by many users. If enough users express a desire to see this changed, it might happen, but this discussion needs to happn in the forum.

BTW, it also occurs to me that if you haven't figured out how to work the "mouse wheel", you probably aren't navigating as efficiently as you could be. Are you perhaps zooming out, dragging the canvas, then zooming back in? Much better to scroll using the touchpad gesture, whatever that works out to be on your computer. Mouse wheel alone to scroll vertically, with Shift to scroll horizontally. Also the Navigator is available in View menu, plus the page up / down / home / end keys. Also Ctrl+F to navigate to a particular measure, page, or rehearsal mark.

Thanks! Definitely good material for further discussion. That last post of mine (#29) might be a key component. I can see that if one is using extreme zoom out and drag of canvas as a method of navigation, one may well have different expectations for what happens when zooming back in to normal magnification than for the more ordinary use case of zooming in starting from normal magnification. Also, if one is starting from an extreme zoomed out state such that page 1 is in view when begining the processing of zooming back in, I can see how one might develop the misperception that the zoom in always starts from the top left of the *score* when in fact it is actually always starting with the top elft of the *crrent* view. Again, when zooming in from extremely small magnifications to normal magnifications, this is going to seem very different from zooming in from normal magnifications to higher ones. This could perhaps be part of the basis of how we consider tweaking the algorithm.

Hi Marc
Mouse-wheel zooming is equally easy for any place in score.
Only resetting zoom to default is awkward using mouse wheel.

But keyboard-shortcut zoom-in becomes much more awkward and complicated in multi-page document when zoom is needed closer to the end of score..
To achieve "minimum-scroll" and avoid loosing sight of the place you try to zoom-in, you need to scroll the place of your interest to the upper-left section of the screen.

Wouldn't it be be easier (and far more elegant) if we could just put the cursor there, without scrolling the whole score?

Musescore already has a function that computes zoom on the basis of cursor position. Can we just use it for keyboard shortcuts?
This would make all zooming functionality consistent, no matter which interface you use to zoom in.

I like Musescore very much, and this is why I try to communicate ideas how to make it better.