MK
Marius Kintel
Mon, Feb 9, 2015 12:03 AM
I don't the perspective is wrong as such. It is correct when viewed from the much shorter distance needed to make it fill the view. It is just that humans don't normally view objects so close to the eye, only cameras do when taking close ups.
Right, so basically, we set up a camera and render the object using that camera, but we don’t take into account that people look at the object on a screen, which doesn’t fill their entire field of view.
When refactoring the camera code, we set the camera field of view to 90 degress, which is pretty standard. I’ll look into finding a better value.
I’m not sure what it was earlier - the code was a bit convoluted.
-Marius
On Feb 8, 2015, at 18:37 PM, nop head <nop.head@gmail.com> wrote:
> I don't the perspective is wrong as such. It is correct when viewed from the much shorter distance needed to make it fill the view. It is just that humans don't normally view objects so close to the eye, only cameras do when taking close ups.
>
Right, so basically, we set up a camera and render the object using that camera, but we don’t take into account that people look at the object on a screen, which doesn’t fill their entire field of view.
When refactoring the camera code, we set the camera field of view to 90 degress, which is pretty standard. I’ll look into finding a better value.
I’m not sure what it was earlier - the code was a bit convoluted.
-Marius
YS
Yvette S. Hirth, CCP, CDP
Mon, Feb 9, 2015 4:11 PM
On 02/08/2015 10:02 AM, nop head wrote:
This is what otho looks like. It looks bigger at the back because it is
the same as the front but you brain expects perspective.
that one looks "warped", like an old LP left in the sun...
yvette
On 02/08/2015 10:02 AM, nop head wrote:
> This is what otho looks like. It looks bigger at the back because it is
> the same as the front but you brain expects perspective.
that one looks "warped", like an old LP left in the sun...
yvette
NH
nop head
Fri, Feb 20, 2015 7:30 PM
Another aspect of the camera needing to be closer is you are far more
likely to have negative objects hit the front clipping plane and being
ignored.
On 9 February 2015 at 16:11, Yvette S. Hirth, CCP, CDP yvette@dbtgroup.com
wrote:
On 02/08/2015 10:02 AM, nop head wrote:
This is what otho looks like. It looks bigger at the back because it is
the same as the front but you brain expects perspective.
Another aspect of the camera needing to be closer is you are far more
likely to have negative objects hit the front clipping plane and being
ignored.
On 9 February 2015 at 16:11, Yvette S. Hirth, CCP, CDP <yvette@dbtgroup.com>
wrote:
> On 02/08/2015 10:02 AM, nop head wrote:
>
> This is what otho looks like. It looks bigger at the back because it is
>> the same as the front but you brain expects perspective.
>>
>
> that one looks "warped", like an old LP left in the sun...
>
> yvette
>
>
> _______________________________________________
> OpenSCAD mailing list
> Discuss@lists.openscad.org
> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org
>
MK
Marius Kintel
Thu, Mar 5, 2015 3:13 PM
Will fix, see https://github.com/openscad/openscad/issues/1236
NH
nop head
Thu, Mar 5, 2015 4:05 PM
Thanks.
On 5 March 2015 at 15:13, Marius Kintel <marius@kintel.net> wrote:
> Will fix, see https://github.com/openscad/openscad/issues/1236
>
>
> _______________________________________________
> OpenSCAD mailing list
> Discuss@lists.openscad.org
> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org
>
MK
Marius Kintel
Thu, Mar 5, 2015 10:03 PM
Fixed in master.
Note: This will still break images generated with custom camera settings.
However, we do now have —viewall and —autocenter, which may help generate better views.
-Marius
Fixed in master.
Note: This will still break images generated with custom camera settings.
However, we do now have —viewall and —autocenter, which may help generate better views.
-Marius
NH
nop head
Thu, Mar 5, 2015 10:24 PM
Does it still produce the same image with the same command line parameters?
On 5 March 2015 at 22:03, Marius Kintel marius@kintel.net wrote:
Does it still produce the same image with the same command line parameters?
On 5 March 2015 at 22:03, Marius Kintel <marius@kintel.net> wrote:
> Fixed in master.
> Note: This will still break images generated with custom camera settings.
> However, we do now have --viewall and --autocenter, which may help generate
> better views.
>
> -Marius
>
>
> _______________________________________________
> OpenSCAD mailing list
> Discuss@lists.openscad.org
> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org
>
MK
Marius Kintel
Thu, Mar 5, 2015 10:47 PM
Does it still produce the same image with the same command line parameters?
Most probably not - the camera settings changed slightly.
..but they also changed slightly from 2013.06 to 2014.03, so you’d have to regenerate images anyway, as discussed earlier.
-Marius
On Mar 5, 2015, at 17:24 PM, nop head <nop.head@gmail.com> wrote:
> Does it still produce the same image with the same command line parameters?
>
Most probably not - the camera settings changed slightly.
..but they also changed slightly from 2013.06 to 2014.03, so you’d have to regenerate images anyway, as discussed earlier.
-Marius
NH
nop head
Thu, Mar 5, 2015 10:56 PM
IIRC the previous change was just add 90 or something similar. Is there
programmatic way to get the same view?
On 5 March 2015 at 22:47, Marius Kintel marius@kintel.net wrote:
Does it still produce the same image with the same command line
IIRC the previous change was just add 90 or something similar. Is there
programmatic way to get the same view?
On 5 March 2015 at 22:47, Marius Kintel <marius@kintel.net> wrote:
> On Mar 5, 2015, at 17:24 PM, nop head <nop.head@gmail.com> wrote:
>
> > Does it still produce the same image with the same command line
> parameters?
> >
> Most probably not - the camera settings changed slightly.
> ..but they also changed slightly from 2013.06 to 2014.03, so you'd have to
> regenerate images anyway, as discussed earlier.
>
> -Marius
>
>
> _______________________________________________
> OpenSCAD mailing list
> Discuss@lists.openscad.org
> http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org
>
MK
Marius Kintel
Thu, Mar 5, 2015 11:27 PM
IIRC the previous change was just add 90 or something similar. Is there programmatic way to get the same view?
No, we changed the camera parameters so that gimbal and vector cameras now have the same field of view.
Camera handling is likely to change yet again when we get around to using a real OpenGL framework.
..so using vector cameras, perhaps combined with —viewall and —autocenter would be recommended.
-Marius
On Mar 5, 2015, at 17:56 PM, nop head <nop.head@gmail.com> wrote:
> IIRC the previous change was just add 90 or something similar. Is there programmatic way to get the same view?
>
No, we changed the camera parameters so that gimbal and vector cameras now have the same field of view.
Camera handling is likely to change yet again when we get around to using a real OpenGL framework.
..so using vector cameras, perhaps combined with —viewall and —autocenter would be recommended.
-Marius
