Re: Transparent Object Markers
Posted: 16 Nov 2017, 19:42
I've been looking at Skyrim editor and it has the constant screen-size, rendered over environment style manipulators, basically same as Blender.
Those are good points, and further reason to test it.AnyOldName3 wrote: ↑16 Nov 2017, 20:31 you can see all parts of the tool at all times, always have it appropriately sized for the current view
I found this to be inaccurate, at least in regards to blender. In blender, the marker is not the center of rotation for the camera. Also, depth confusion with the markers can easily occur. As a test, I set up a scene with two boxes. After selecting the two, I manipulated the camera until the marker was behind one of the boxes. How did I know the marker was behind one of the boxes? Certainly not from looking at it. The markers always have a screen space size, as you mentioned, and they are also always drawn on over everything else. I knew it was behind the box because I was mentally keeping track of its spatial position. Had I first selected the two boxes from that orientation, I would not have known the position of the markers.AnyOldName3 wrote: ↑16 Nov 2017, 20:31 but also serving as the centre of rotation for the camera, ... don't get confused about its depth
I think in Blender I mostly get the depth clue by rotating the scene all the time. If I have something selected, I can hardly tell how far away it is, but I keep rotating and translating the camera slightly, which gives me that 3D perception (like when you move your head IRL to get a sense of distance). And it works. I think we should consider if this is how we want the editor to work though, i.e. if the users are supposed to keep their hand at the mouse all the time and do these small movements, as they're also filling in text data, for which they need both hands, etc. The best way would simply be to make two implementations and then test which one feels better.
Neat, I didn't think of that. Still, that is with rotation around a center point. That is only one of the camera modes that the editor supports. It will be less effective with linear movement.drummyfish wrote: ↑17 Nov 2017, 20:37 I keep rotating and translating the camera slightly, which gives me that 3D perception (like when you move your head IRL to get a sense of distance)
Blender does have a have an indicator in the bottom left corner to help with that. We still need to do some testing, but I'm starting to wonder if the depth really matters. Well, I guess we'll find out.AnyOldName3 wrote: ↑17 Nov 2017, 18:48 I guess keeping track of the marker position doesn't really matter.
If the camera is perspective, not orthographic, then I'm pretty sure comparison as it moves or pans or whatever will still lead to it looking fine.That is only one of the camera modes that the editor supports. It will be less effective with linear movement.
It was a quick idea where the solution used for translation and scale could be used for rotation, though I'm not convinced by it either.
In my workflow I have them disabled 99% of the time. For regular and demanding work, the keys are much faster. When I'm explaning Blender in person, I suggest turning markers off, while also explaining the shortcuts. However with no prior knowledge of Blender or OpenMW-CS, not seeing any markers by default is a sure way to confuse the user.aesylwinn wrote: ↑17 Nov 2017, 17:57 When I first went through some blender tutorials a couple years ago, one of the first things mentioned was that the user should hide the markers and use the keyboard shortcuts instead. That was several years ago, so perhaps that is no longer suggested, but still, how many people actually use the markers in blender for anything serious?
In light of these problems the shape alone won't help much. My vote is on having a screenspace-sized marker drawn on top everything else.Mojo wrote: ↑15 Nov 2017, 12:27 1) Doesn't it already get confusing right now when you have several markers? Mabe we should also use the pivot point or only draw the marker of the last selection.
2) How are you supposed to move an object, when one axis of the marker is hidden in other objects like in this case? That's happening quite often when you work with wall pieces.
In Blender, you can specify the pivot point before a transformation and each is useful for something.
A single transparent marker definitely increases usability compared to the current implementation. This should be the goal, except for case A there would multiple rotation markers also make sense. (see last answer)
Using multiple markers just for "not being forced to look at a single location to click on the markers" would only help for moving the objects, but it does not work for scaling and rotation (except case A). A single marker would be a better choice there. Also multiple marker make the scene more confusing, which is a larger drawback than being forced to "look for" the single marker. Also if the marker is drawn above the object, you don't have to look for it.
Yes we need a way to choose between shared and individual centers. Explanation below.
Unfortuantely the transformations with pivot points get more complex for multiple objects than single objects. For single objects you only change the place (=translation) or the rotation or the scale. This is also the case for translating multiple objects or rotating multiple objects around individual origins, but the transformations can also get combined for multiple objects.Lamoot wrote: ↑19 Nov 2017, 14:58 In Blender, you can specify the pivot point before a transformation and each is useful for something.
* Bounding box center
* Active (last selected) element
* Individual origins
* Median point
* 3d Cursor
OpenMW-CS would benefit having at least a pivot for all objects together and a pivot for individual objects (the latter is what's currently happening).
I just want to point out that D is the case where the "pivot point" is at the center, while E is the case where the "pivot point" is at the center of the selection.