11 Replies Latest reply on Feb 24, 2009 7:12 AM by Neil Larsen

    Camera view: thoughts for coders

    Neil Larsen
      Mr Sweeney is this your domain?..
      ..continuing to explore 09sp0 coming from 05 I have got to camera view..yes.

      now this is quite a useful tool even without PW..
      but here I found an issue - the view wont auto-fit to the 3d window properly-
      if you have the taskpane open and pinned the camera view when displayed auto-fits on the right (and when the edge is dragged) but with the FM open and docked the view wont auto-fit on the left (or when the edge is dragged)..it gets hidden behind it (and this behaviour is true with a left side tiled viewport as well) -bug? or is it designed/neglected this way?

      also if the display field of view box is selected wouldnt it be handy to auto crop the image to that size if you 'save as' or screen capture? - just a thought..and on that track how about showing the actual pending capture size in pixels somewhere.. (if the auto-fit worked as I anticipate it should you could have some control over the capture size by dragging the FM, TM or viewport divider to cramp the auto-fit..)

      also - couldnt the display pane allow you to toggle on and between any camera views using the hide/show column ?

      further - there seems to be weirdness when using F9,F10,F11 to get a full screen view and back from re what bits you expect would come and go and what does - requiring more presses to get it right again...

      keep up the good work folks Concord wasnt built in a day - there's always next release..

      PS how about an option for a grid overlay (or rulers) for screen captures - A,B,C.. by 1,2,3.. just for reference purposes when making a presentation or such...maybe you could fit a nice wee title in there too at the bottom...if you are not tooo busy ATM

      Happy holidays etc

      Edit: moved to UI forum instead of general

      Bonus offer for good behaviour:
      here is a paper of interest re ambient occlusion that might help Realview display quality - specifically reflected colour
      http://www.mpi-inf.mpg.de/~ritschel/Papers/SSDO.pdf


      Edit2: here's another oddity (well I think so)
      re new reference triad functionality: in perspective mode this will not align the views properly even if the F,T, R planes are in the middle of the object..nor even for the non perspective case, necessarily centre the object with each view change..
      and it you have selected 'zoom to fit when changing to std views' option for either perspective or not it doesnt do it...hmmm... oh well...
        • Camera view: thoughts for coders
          John Sweeney
          Hi Neil,
          I'm not an expert with the Camera view stuff. This is handled by our Graphics guys. I've talked to Ron, and he'll get back to you.
          Best Regards,
          John
            • Camera view: thoughts for coders
              Neil Larsen
              thankyou Mr Sweeney what a fine fellow you are,

              as GWBush observed:"I know the human being and fish can coexist peacefully."

              regards
              Neil

              BTW
              he also apparently remarked:"There's no question that the minute I got SW, the storm clouds on the horizon were getting nearly directly overhead."

              and perhaps a little more loosely: "And truth of the matter is, a lot of reports in Concord are never read by anybody. To show you how important this one is, I read it, and Jeff Ray read it."

              but the truth is that in the coders cubicals they know: "I'll be long gone before some smart person ever figures out what happened inside this program."

              ...what a vintage year for SW this is going to be..







                • Camera view: thoughts for coders
                  Ron Bates
                  Hi Neil,
                  Big apologies for the long delay. You covered quite a few points so let me see if I can capture it all.

                  View won't fit into 3D window - I assume you mean that the left hand side of the split view, which is displaying your standard view along with the camera manipulator? The feature manager is now a floating window. In other words, it doesn't disect the graphics area like the left hand edge of the taskpane does. In other words, the graphics area goes behind the feature manager. The advantage being that when the feature manager is resized, we don't need to do any repaints. e.g. if you have a massive assembly that repaints at just a few frames per second, dragging the feature manager used to be painfully slow. Now it's real-time. In most cases, we've tried to account correctly for the fact that zoom-to-fit should ignore the area behind the feature manager. It's clearly not always behaving correctly with camera views. We'll take a look at this.

                  Honor aspect ratio and size for Save As .jpg - This has been discussed as part of a future project. In fact, more overall control of the output size/resolution of .jpg/.tif for parts and assemblies in general is the real topic. Honoring the current camera settings is part of that.

                  Toggle cameras from display pane - neat idea. I'll pass this to Prd. Def. for consideration in the future.

                  Grid/Ruler overlay - I'm curious how you'd use the overlays as it pertains to screen capture.

                  And thanks for the realtime Global Illum reference!

                  Ron Bates
                  Manager, Graphical Applications
                  SolidWorks Corporation
                    • Camera view: thoughts for coders
                      Neil Larsen
                      thanks for taking the time to reply Ron

                      Ok I understand what you mean about the FM but this isnt the way I was expecting it to work for camera view.

                      Same with clicking on the ref triad to change the view - I would have thought it would zoom to fit with that set in the system options as per a change of view from the 'heads up view toolbar' icons.

                      As I said there seem to be issues with how it centres and aligns with perspective enabled as well. I would have thought the main planes would be square/end on to the view. I think it is part of the same issue cos the vanishing point seems to remain the middle of the screen regardless of the FM width?... dunno it just seems to give unusual results...I have a 27" lcd so if I pan an object across/updown the 3d window and it changes the perspective view slightly, fine- but clicking on the triad and going through view 90 deg changes away from the vp but in the view centre produces kind of puzzling attitudes/shifts sometimes...maybe its just me

                      Attach is an overlay idea hacked out in MS Paint..- no particular thought gone into it or the aspect ratio- its just to make referencing easier when the images are used in presentations. I guess I would most likely want to use this for LCD projector size stuff

                      Looking forward to future developments
                      Neil
                      (understudy Product Manager for Old Product Problems)

                        • Camera view: thoughts for coders
                          Ron Bates
                          Hi Neil,

                          You are correct on the behavior as it pertains to the VP in perspective. It is off center when the FM is open. This is because we are trying to shift (pan) the view to account for the graphics area behind the feature manager, since the FM no longer splits the graphics area. So a real/true zoom-to-fit would actually result in the model being rather close to the feature manager, and appearing as though it is panned to the left. We try to correct this by panning it back to the right based on the width of the feature manager. And this results in the VP error you're noting in perspective. Try this, collapse the feature manager, and then try your front view while in perspective. You should see the VP behave as you expect. See also the images. We realize this is a rather inconsistent behavior. My recomendation to frame a particular view for a presentation or rendering is to use a real camera, and to FLOAT the camera, and collapse the FM, then use the standard viewing controls to modify the camera properties in real-time. Then RMB on background and toggle camera back to locked if you want to be sure to avoid any further edits.

                          As far as clicking the triad not doing a zoom to fit like the other standard view's do for front/right/top, I'm afraid I'll need to go back to John on that one since it was implemented by the usability group.
                            • Camera view: thoughts for coders
                              Neil Larsen
                              hmm....well ok Ron I'm not going to obsess about it but I find this accomodation/implementation of the floating FM feature to be slightly strange in how it affects the views...

                              It just doesnt quite work the way I expected and I noticed it, thats all, really - there is some small nuisance value in it and there is inconsistancy but I guess you guys did the best you could to make it all work...possibly it just needs fine tuning...

                              BTW I notice too (sp0 anyway)- if the Task pane is open but not pinned the view pan and rotate is slow and lagging, pinned or hidden the speed is ok. I guess this is the redraw effect you are trying to avoid with the FM and especially large assemblies...or maybe its a bug..

                              There would seem to be good technical reasons why the whole window needs to be drawn in this fashion for the sake of the FM but from a user point of view I think a docked FM should be similar to a pinned Task pane in its effect..
                              Perhaps you can add a 'pin' to the FM or a 'view frame lock' feature for the occasions when a user wants it that way? - like setting up camera views, and for actual size screen captures in the future...

                              thanks for your replies
                              I'll leave you in peace to do your work now.
                              I'm sure if you think of a better way to do this you will follow it through

                              regards
                              Neil

                                • Camera view: thoughts for coders
                                  Neil Larsen
                                  BTW I think you should be able to use the camera view for honouring video recording size as well
                                    • Camera view: thoughts for coders
                                      Neil Larsen
                                      Ron,
                                      Sorry to drag this topic up again but I have noticed another issue with the graphics probably related to the same screen drawing issue at the root of the above mentioned misdemeanours-

                                      In an assembly with perspective turned on if I select a part in the FM and Ctrl drag it into the 3d view to make another instance the existing assembly appears to 'bend/flex/banana' in the middle about the vertical axis as I move the new part back and forward across the window - weird effect....the effect is most pronounced just as the cursor crosses the FM edge and again where the display pane edge would be if it was showing

                                      I think you guys really need to do this floating FM/viewpoint stuff differently...

                                      How about incorporating an option in the large assy mode settings (where the size for applying the option is user defined..) to re/draw the screen as you have it presently if it is better for large assy, and revert to the old steady state but slower re/draw for std mode?
                                      I may be alone but I find these vp errors/orientation shifts, and odd windowing to be disconcerting, 'incorrect' or at least counter-intuitive .
                                      I compare this to how 05 handles the same issues and there is no side effects there, the perspective views align properly or as expected, while the redraw delay for the FM edge drag is present it is not disruptive, and the views resize logically...

                                      HTH
                                • Camera view: thoughts for coders
                                  John Sweeney

                                  Neil Larsen wrote:

                                   

                                  ...
                                  Same with clicking on the ref triad to change the view - I would have thought it would zoom to fit with that set in the system options as per a change of view from the 'heads up view toolbar' icons.
                                  ...

                                  Hi Neil,
                                  Clicking the axes on the reference triad was meant to behave more like the left/down/up/right arrow buttons. For example, using the shift key along with the arrow buttons rotates the view 90 degrees but maintains the view scale. Likewise, using the shift key and selecting a ref triad axis rotates 90 degrees about the selected axis and maintains the view scale. Similar behavior for incremental rotate using the Alt key and Alt-Ctrl key combinations. For these cases, I think it would be a bit disorienting to perform a zoom-to-fit for each incremental rotate.

                                  I do agree that clicking an axis without any control keys (which performs a "normal to" operation) would be OK if it performed a zoom-to-fit at the same time, but it is currently implemented to be consistent with the other ref triad clicking commands where we avoid the zoom-to-fit.

                                  I hope this helps explain our reasoning.

                                  Best Regards,
                                  John
                                    • Camera view: thoughts for coders
                                      Neil Larsen
                                      John,
                                      thanks for getting back with an answer I feel a little spoiled for attention - unfortunately the reasoning sounds a little like a Jim Wilkinson near encounter with a magnifying glass handle.. he wasnt sitting in on the meeting was he?

                                      ok so in regard to the said ' view orientation enhancements' in 09 there is no tie in with clicking on the axes (only) to get a std view, with the system option' zoom to fit when changing to std views' if it is selected - and.. the whole axes clicky thingy just relates to doing the same as rotate view using the <-- keys ,with or without Shift, already allows, gotcha... (roll eyes)

                                      I have a handle on it now.. hmmm.. now I am sure Jim is involved..

                                      and we are going to have to live with the strange case of the misbehaved VP and the cloaked camera view and other tales of the barely explained, I take it...


                                      off topic..its just a passing thought you know.. but I think there is something slighty uncomfortable about SW coding ATM in that it seems to have a dual personality arising from the serial SW as we have always known it and the new multi access one, that is leading to some identity confusion both for coders and users.
                                      It just seems the realities of each dont coexist and understand each other very well and even the UI is a little warped by it.
                                      I sense some folk are expecting and seeing the 'logical serial' SW as they have known it and using it that way and leaving aside or oblivious of the new stuff, and the new stuff is a little hard to grasp and use with the same mindset because it is somewhat random and 'illogical', parallel and simultaneous.
                                      We now have so many ways of looking at and doing things it gets disorientating.
                                      A bit like loft and boundary coexisting and being the same but different..
                                      I wonder if there should be a little more emphasis given to identifying in the users and coders mind not a 'tool' to use but focusing on a process that leads to the result - the purpose.
                                      agh well I know what I am trying to say, I think.. even if no one else can follow it

                                      thanks again for the reply
                                      I appreciate your time and effort
                                      keep away from Jim he's a little difficult