3D windows not rendering negative z-values of geometry

[qgib]

Author Name: Jakob Lanstorp (Jakob Lanstorp)
Original Redmine Issue: 17789
Affected QGIS version: master
Redmine category:3d
Assignee: Martin Dobias

---

Adding a 3D-multipoint through the DB Manager hooked to PostGIS, does not render the points with negative z-values.

@select 1 AS id, ST_GeomFromText('MULTIPOINT(505105 6305783 -200, 505105 6305783 0, 505105 6305783 200)', 25832) AS geom@

The same goes for single point geometries:
@
SELECT
1 AS id,
ST_GeomFromText('POINT(505105 6305783 -200)', 25832) AS geom
UNION ALL
SELECT
2 AS id,
ST_GeomFromText('POINT(505105 6305783 0)', 25832) AS geom
UNION ALL
SELECT
3 AS id,
ST_GeomFromText('POINT(505105 6305783 200)', 25832) AS geom
@

Also the geometry in z = 0 is cut in half.

Version: OSGEO QGIS Master (2.99.0-255) of 4th Jan 2018

---

• udklip.jpg (Jakob Lanstorp)

---

Related issue(s): #25710 (relates)
Redmine related issue(s): 17814

---

[qgib]

Author Name: Martin Dobias (@wonder-sk)

---

This is expected: by default the 3D view also renders "flat" terrain at zero elevation with map image as the terrain's texture - that's why the point at zero elevation is cut in half and the point with negative elevation does not appear (hidden under the terrain).

[qgib]

Author Name: Jakob Lanstorp (Jakob Lanstorp)

---

Is the terrain's texture not just another 3D geometry that you should be able to pan through? Negative z-values is important for data like geological layers and sewer network systems. Normally you would set the surface to 0 meters. Negative values below surface positive above the surface. If I just have the 3D-geometry and no other layers in the map, there is still an imaginary 0 meter surface hiding the negative values.

[qgib]

Author Name: Martin Dobias (@wonder-sk)

---

Is the terrain's texture not just another 3D geometry that you should be able to pan through?

That's correct. There are few things special about it though compared to other 3D geometries, e.g. when dragging map with mouse we find out which point on the terrain user grabbed in order to move camera correctly.

The camera controller as of now also has a lock on maximum tilt angle so that user does not get under the terrain - something that is not desirable when not working with subsurface data. In a future release we should have options to allow users move the camera more freely to give more power to those working with subsurface data.

Shall we close this ticket?

[qgib]

Author Name: Jakob Lanstorp (Jakob Lanstorp)

---

Yes lets close it. Thanks for clarification. For the record: Issue changed to bug in Postgres provider (#25710)

---

• status_id was changed from Open to Closed

[qgib]

Author Name: Jürgen Fischer (@jef-n)

---

• relates was configured as Postgres provider retrieves 4D geometries as 2D #25710

[qgib]

Author Name: Jürgen Fischer (@jef-n)

---

• subject was changed from 3D windows not redering negative z-values of geometry to 3D windows not rendering negative z-values of geometry

[qgib]

Author Name: Giovanni Manghi (@gioman)

---

• resolution was changed from to duplicate