Hello all, Michael,

I am regularly facing the following limitation when dealing with the Image Tool : outside of what is possible at the time of initial creation (creation from the top left corner to the bottom right corner), there seems to be no way to match the corners of a reference image to existing points in the scene later on - regardless of them defining a boundary exactly proportional to the image being placed, or an arbitrary shaped rectangle. As shown here :

https://www.youtube.com/watch?v=Js7Zw4qWvOk

The best I manage to do is to make the image match one dimension of the rectangle ; but I cannot seem to be able to place it precisely where it should be. I've tried the Align feature inside the Image Tool, but I do not understand what it does.

I feel like the issue could be addressed by simply adding a center hotpoint to the image manipulation tool. Non uniform scaling could be good too, as that could be an even faster way to snap to three known points (not to mention that non-uniformly scaled images can be useful at times).

I hope this makes sense !

On a side note, it would also be great for the image elements to have at least their corner points be eligible for Object Snap. That way one could at least draw a rectangle matching an already placed image.

Hi pior, in the video it looks like you need to edit an image "corner to corner" but you are in "scale from center" editing mode.

To edit an image corner to corner like when you initially create it you'll want to be in "scale from opposite corner" grip editing mode, not "scale from center".

You can switch the corner grip scaling mode by clicking (press and release the mouse button) on a corner grip instead of dragging on it. It looks like you may be doing this accidentally and inadvertently switching the mode.

You can tell what mode you're in by observing where the center mark appears while your mouse is hovered over a corner editing grip.


Currently background images are designed to maintain their aspect ratio when sized. If you want to change the aspect ratio you'll need to modify the image in an image editing program and then bring in the new image instead.

- Michael

Hi Pior, example vid here:



- Michael

Oh I see ! I wasn't aware of these two modes triggered by this "secret" click. I actually don't think I ever noticed the center mark either, because it is hidden most of the time except when hovering over the corners (essentially putting it in a blind spot that the gaze is never focused on). Thank you for the explanatory video.

Why not simply put a button in the UI for this mode switch, next to Add/Remove/Align/Hide ? This would make it easily discoverable.

The question of how to snap to the corners of an already placed image object remains though. Is there any know workaround for that ?

Hi pior,

re:
> Why not simply put a button in the UI for this mode switch, next to Add/Remove/Align/Hide ?
> This would make it easily discoverable.

It's because the same behavior exists for the edit frame being used to modify regular geometry. It's not specific to just image editing.


> The question of how to snap to the corners of an already placed image object remains though. Is there any know workaround for that ?

When you are in the image command, points for added or edited images will already snap to other image corners.

If you want to snap geometry to image corners outside of the Image command I'd recommend placing a rectangle down before placing the image.

- Michael

Understood - well, if anything I would say that having the option for images to trigger object snapping would be great in a future release :)

On a somewhat related note : Is there a way to tell the application to always embed images, as opposed to the way it always reverts to linking to them by default ? I must say I am often surprised to see this option being turned off because of the memory of "I swear I set is as embedded last time" (as it is easy to forget that the option is applied on a per-image basis, and not as a global setting).

I do feel like having the images be embedded by default would be the safer option anyways - because it has less potential for bad surprises/work loss for the user. Or at the very least not putting this setting behind the "Properties" button, and having it in plain sight instead.

I hope this makes sense !

Hi pior, image embedding is not on by default because it duplicates data and increases file size.

I'll see about putting in an option to make it be turned on by default though.

- Michael

Hello !

I think there are multiple solutions really, the point being that imho the priority should be put on error prevention - in this case making sure that the default behavior doesn't cause any loss of work. I would say there are two ways to do that (hence 4 permutations total) :

- Making the tickbox unmistakably visible when interacting with the image tool (regardless of its default setting) ;
- Setting the default behavior to embeded, as the argument could be made that file size (especially for just a small set of images) is of much less importance than making sure that the user doesn't lose work when moving files around on or sending a file over the network. I can certainly confirm that I have already lost some image planes that way in the recent past ...

Of course an .ini setting that advanced users could enable could be an acceptable solution ... but with the cavehat of potentially losing a few images first, until the user starts searching for a workaround.

Anyways ! Thanks a lot for the assistance as always. I am obviously influenced by the Blender UX here, in which the settings is set once and for all and never needs to be touched again (it even keeps the initial link to the image, meaning that it is even possible to refresh the embedded pictures as long as they are still present on disc). A great solution really.