culling of very tall cylinders (objects?)

as1 · February 18, 2020, 12:53am

1. A concise explanation of the problem you’re experiencing.

when having very tall objects, they are culled on zooming in - it seems there is a visible sphere around the globe that cuts off…

also it seems the positioning of very tall objects does not align properly - please see attached test file:

2. A minimal code example. If you’ve found a bug, this helps us reproduce and repair it.

see the attached cesium.html file for the complete testing code

3. Context. Why do you need to do this? We might know a better way to accomplish your goal.

please check the example here:

[https://api.iidj.net/coronavirus/](https://api.iidj.net/coronavirus/)

4. The Cesium version you’re using, your operating system and browser.

1.66, Chrome, Safari, Firefox

cesium.html (1.6 KB)

omar · February 18, 2020, 3:48pm

Does turning off logarithmic depth buffer help? If so, it may be this issue:

https://github.com/AnalyticalGraphicsInc/cesium/issues/6573

Otherwise, you may need to move the camera’s near and far planes a little closer. Instead of the default near/far of 1/5e8, try dividing both by an order of magnitude (like 0.1 and 5e7). See the camera doc here:

https://cesium.com/docs/cesiumjs-ref-doc/Camera.html?classFilter=Camera

as1 · February 19, 2020, 3:10am

hi OMAR

thanks for prompt suggestion -

omar · February 19, 2020, 3:47pm

Hey Andreas,

You can turn off logarithmic depth buffer like this:

viewer.scene.logarithmicDepthBuffer = false;

``

No need to modify the source code.

as1 · February 20, 2020, 2:27am

hi OMAR,

You can turn off logarithmic depth buffer like this:

viewer.scene.logarithmicDepthBuffer = false;

great ! this helps certainly.

very much appreciate your effective follow-up!

what still remains is that heights of above ca 1000000 the positioning seems to get more and more displaced.

is it possible that entities(cylinders) are not positioned referenced to their center?

in the example, red one has height of 1000000, red and green have exactly the same lat/lon values.

cheers,

andreas

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IIDj
Institute for Information Design Japan

Andreas M. Schneider

T +81-3-3666 5600
M +81-80-5538 9767

Tokyo 103-0012
Chuo Ku, Nihonbashi
Horidome Cho 1.2.9

[map.iidj.net](http://map.iidj.net)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SPAM FILTER notice:
Please note that your mail may be intercepted halfway through SPAM filtering.
Using a unique SUBJECT line helps to separate the good from the bad

as1 · March 15, 2020, 7:53am

hi OMAR

after installing the most recent version of SAFARI on MacOSX 10.13 (High Sierra) it is no more possible to instantiate a cesium.viewer...

the simple message is BAD REQUEST.

this happens also when trying to instantiate a view in its most basic form...

the site runs problem free on CHROME/FIREFOX

https://api.iidj.net/cv/

i have no idea what could be the culprit - nor any idea on how to deal with this.

cheers,

andreas

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IIDj
Institute for Information Design Japan

Andreas M. Schneider

T +81-3-3666 5600
M +81-80-5538 9767

    Tokyo 103-0012
    Chuo Ku, Nihonbashi
    Horidome Cho 1.2.9

map.iidj.net

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SPAM FILTER notice:
Please note that your mail may be intercepted halfway through SPAM filtering.
Using a unique SUBJECT line helps to separate the good from the bad

omar · March 16, 2020, 7:42pm

Are you seeing this issue just in your app or on any Cesium app? Are you able to run the Sandcastle here:

https://sandcastle.cesium.com/index.html?

Can you provide a screenshot of the error(s) you’re seeing? Seeing the error in the developer console or network tab would help.

as1 · March 17, 2020, 3:26am

hi OMAR

thanks for your prompt reply.

Are you seeing this issue just in your app or on any Cesium app? Are you able to run the Sandcastle here:

https://sandcastle.cesium.com/index.html?

yes, same behavior.

the page loads, reloads and then shows the message (see attached screenshot)

here i made a bare-bones test page:

https://api.iidj.net/tst/

tested also with cesium 1.66 - to no avail.

not able to log what is going on @ SAFARI, as the page reloads - and then only provides such basic statement '-(

cesiumOnSafari13.05.mov (538 KB)

omar · March 17, 2020, 2:34pm

We tested with Safari on a 2015 macbook with discrete graphics (AMD) using OSX 10.15.3 and wasn’t able to reproduce.

Can you show the full browser window when this constant reload is happening? Your video only shows the console.

Can you also post the results of opening WebGL Report (https://webglreport.com/) on Safari on that same device?

Are you able to run other WebGL content there, like this: https://threejs.org/examples/#webgl_animation_cloth ?

as1 · March 17, 2020, 2:54pm

hi OMAR,

thanks for the detailed answer and efforts to help.

We tested with Safari on a 2015 macbook with discrete graphics (AMD) using OSX 10.15.3 and wasn't able to reproduce.

i am running SAFARI 13.05 on High Sierra 10.13.6

cesium_1.67_as@iidj.net.mov (643 KB)

WebGl_Report_as@iidj.net.pdf (929 KB)

omar · March 18, 2020, 3:31pm

Thanks for the additional reports. We’re still unable to reproduce this.

Did you check if any other WebGL applications are giving you the same behavior? Does YouTube work? Does this page https://threejs.org/examples/#webgl_animation_cloth work?

Does this happen on Chrome? Can you try updating to the latest Safari/your OSX?

as1 · March 19, 2020, 1:08am

hi OMAR,

thanks for your follow-up.

Did you check if any other WebGL applications are giving you the same behavior? Does YouTube work? Does this page https://threejs.org/examples/#webgl_animation_cloth work?

yes - i thought i had confirmed this in my previous mail...

Does this happen on Chrome?

no - this is a SAFARI only issue.

and - to confirm again - the issue can be observed with all CESIUM examples form the sandbox gallery - obviously as the most basic call already causes the reload (loop)

Can you try updating to the latest Safari/your OSX?

the problem started AFTER updating to the most recent version for OSX 13.06
(i had included screenshots of the SAFARI info with my previous mail)

i tried as much as i can on my side to identify the issue. it must be a very specific thing between SAFARI/CESIUM

i did not get so far yet to see if i can start debugging by adding breakpoints to the CESIUM source code.
will do that i guess...

cheers,

andreas

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IIDj
Institute for Information Design Japan

Andreas M. Schneider

T +81-3-3666 5600
M +81-80-5538 9767

    Tokyo 103-0012
    Chuo Ku, Nihonbashi
    Horidome Cho 1.2.9

map.iidj.net

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SPAM FILTER notice:
Please note that your mail may be intercepted halfway through SPAM filtering.
Using a unique SUBJECT line helps to separate the good from the bad

as1 · March 19, 2020, 3:22am

hi OMAR,

i could at least identify the step when SAFARI starts reloading

CesiumUnminified/Cesium.js'

line 190976

if (!defined(glContext)) {
glContext=canvas.getContext('webgl', webglOptions)||canvas.getContext('experimental-webgl', webglOptions)||undefined;
}

the webglOptions at that point are:

{powerPreference: "high-performance", alpha: false, stencil: true}

hope this helps!

andreas

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IIDj
Institute for Information Design Japan

Andreas M. Schneider

T +81-3-3666 5600
M +81-80-5538 9767

    Tokyo 103-0012
    Chuo Ku, Nihonbashi
    Horidome Cho 1.2.9

map.iidj.net

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SPAM FILTER notice:
Please note that your mail may be intercepted halfway through SPAM filtering.
Using a unique SUBJECT line helps to separate the good from the bad

omar · March 19, 2020, 12:06pm

Thanks for debugging this - I think the issue was the recent change to “high-performance” mode:

https://github.com/CesiumGS/cesium/blob/master/CHANGES.md#additions-tada-2

Can you check if the suggested setting of “default” fixes it? Or this Sandcastle that sets it to low power?

as1 · March 19, 2020, 12:26pm

hi OMAR,

thanks for your fast reply!

I think the issue was the recent change to "high-performance" mode:

that solved it! great!

so obviously that is an issue to note for other OSX HighSierra (10.13)/SAFARI 13.05 users!

very much appreciate your digging in that!

if i may use this opportunity for another question - the representation of cylinders can be done via the cylinder primitive or extruded ellipse - is there a recommendation for performance/stability in both primitive declaration or CZML?

cheers,

andreas

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IIDj
Institute for Information Design Japan

Andreas M. Schneider

T +81-3-3666 5600
M +81-80-5538 9767

    Tokyo 103-0012
    Chuo Ku, Nihonbashi
    Horidome Cho 1.2.9

map.iidj.net

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SPAM FILTER notice:
Please note that your mail may be intercepted halfway through SPAM filtering.
Using a unique SUBJECT line helps to separate the good from the bad