SketchUcation 3D Viewer (formerly the '[ignore]' thread)
-
@unknownuser said:
u can embed with <iframe>
Surfer showing the viewer embeded on a site must be registered on Sketchucation ?
-
I couldn't open the nice wireframe sphere on fullscreen.
1 - It took a lot of time to open the viewer embeded on this page;
2 - It took what seemed to be the same amount of time for it to open the fullscreen page (I wouldn't expect I would have to reload it, only that I had to wait for my punny GPU to display it)
3 - Reload was finished but didn't show.Might be related to Nvidia 425M 1025Mb
EDIT: Also, on the full screen window there's a fullscreen icon to make it fullscreen... I guess that's for people to be sure they are in full screen or that this full screen is as full as it gets
-
@pilou said:
@unknownuser said:
u can embed with <iframe>
Surfer showing the viewer embeded on a site must be registered on Sketchucation ?
To upload a model to the viewer you have to be a member. Actually now only a few testers has the ability to upload.
If the model is set to be public it can be embedded into any HTML page and anybody can see it there. If the model is set to private, only those can see it who are allowed by the author.
-
@jql said:
I couldn't open the nice wireframe sphere on fullscreen.
Did you get any error messages. If you open the javascript console do you see any error messages there?
@jql said:
1 - It took a lot of time to open the viewer embedded on this page;
Performance issues will be addressed later. There are several points in the script that are not performance optimized yet.
@jql said:
2 - It took what seemed to be the same amount of time for it to open the fullscreen page (I wouldn't expect I would have to reload it, only that I had to wait for my punny GPU to display it)
Yes, the fullscreen view at the moment reloads the model.
@jql said:
3 - Reload was finished but didn't show.
[/quit]
What was on the screen? Any error messages?@jql said:
EDIT: Also, on the full screen window there's a fullscreen icon to make it fullscreen... I guess that's for people to be sure they are in full screen or that this full screen is as full as it gets
Will address this as well.
Thank you for the testing.
-
Loooking good.
-
[sw:9dyd83bx]9rvku6oIXacm1K8[/sw:9dyd83bx]
-
I had a similar experience with the wireframe model that JQL did.
At one point the "loading" icon stopped and there was just a black image and my computer was frozen. I kept clicking and finally the model was visible.
That is the only model that has acted like that. The last one that Rich posted loaded just fine.
BTW: When I clicked on the Full Page icon when I was already in a full page view, it just loaded a second full page window and loaded the model again in the 2nd page........I assume that is what it was supposed to do.
From a viewer's perspective, everything seems to be working with the exception of that one wire frame model (which BTW I liked
)
-
We are testing an issue with mirrored components negatively scaled.
Sorry for overburdening your browsers but it's a good means of flagging issues.
-
@gĂĄbor said:
@jql said:
JQL wrote:
I couldn't open the nice wireframe sphere on fullscreen.Did you get any error messages. If you open the javascript console do you see any error messages there?
Now everything worked... no error on java console.
@rich o brien said:
We are testing an issue with mirrored components negatively scaled.
Sorry for overburdening your browsers but it's a good means of flagging issues.
Is that why some wires are black? are those reversed faces? I didn't see backfaces on the cornelbox model.
-
There are 2 icons in the controls.
The first respects SU hard edges. The 2nd is the tri'd mesh.
I left wires on to show mesh density.
-
First of all, wow. Your work here is very impressive, though that seems an understatement.I realize that I'm a newbie round here, but I thought to contribute a little.
I'm viewing these on an HP Elitebook 8440p Core i7 M620 @ 2.67GHz laptop with 4GB RAM and Intel HD graphics, Win 8.1 Pro 64 bit and Chrome Version 39.0.2171.95 m in 32 bit mode. Display of models is very fast, though there is some latency in the rotation of models. I have just read the entire thread and I have noted some characteristics of the zoom function that hadn't been mentioned. Don't know the capabilities of WebGL. But I thought I would throw this out there anyway.
Earlier models in the thread seemed to limit zoom when I ran into a surface. Later, once I zoom out, I could not zoom back in to the initial scale. I would have to restart the image to reset the scale. I could however alternately rotate models off center and 'walk' them past these limitations and close enough to move sideways past the outer surface and see the inner surfaces. I found this ability desirable because it more closely resembled native SU views (that's just me). Another item of note is that zoom is not relative to the cursor position within the frame but seemingly relative to the center of the frame instead. Not criticizing mind you. Just noting a few things that differ from native SU behavior.
Absolutely stunning work so far. I hope these observations are a help to your efforts.
-
Sorry but this might be my GPU or a design feature...
I'm talking about wires but not wireframe rendering. Metal wires...
There are some black metal wires there, and I was expecting them all to be steel. Take a look at image:
They are black painted? Reversed faces? Or rendering bugs related to my shi... hardware?
-
I assumed half the wires were 'silver' and half were 'black'.
There is a consistent pattern ??
The form is very complex with many many facets... -
The black is occuring when you negatively scaled a component to mirror it.
it seems this causes an issue on export with textures and other things.
re:orbit/zooming
we know this. for now the priority is presentation. we need to nail it down.
you'll see some big forward steps this month. with a full public beta to follow.
-
I had a similar issue with my OBJ exporter [http://sketchucation.com/pluginstore?pln=OBJexporter now v3.0, but 'fixed' since ~v2.4 - it was initially developed for Octane-Render, where these issue first cropped up].
Simplistically put... what needs to be done is when a group or component-instance is processed its transformation is checked for negative scaling [taking into account that several negative scalings/rotations cancel out!].
I suggest that GĂĄbor and co look at my
OBJexporter.rb
file at lines ~#337-#405 [and elsewhere, e.g. #426...] for ideas - I know it's not quite the same coding but the principals are the same...
That code finds the truly 'flipped' faces and simply reverses them and swaps the remapped textured-materials from front to back etc, all before they are processed in the exporter [the operation is aborted after the export completes, so any such changes to the SKP are undone automatically] - alternatively the [re]calculated face normals etc, could simply be reversed in code alone...We can continue by PM as needed...
-
Thank you TIG. Working on it.
-
[sw:li44u31m]HFqwrWluo8nSgB0[/sw:li44u31m]
-
Like the lighting when I rotate the chair.
I think you may have addressed this before, but why can I not rotate the model 360 in the vertical direction? I have looked at other viewers and they do not have that restriction?
I know this is not the current emphasis of your testing, but have you considered changing the Full Screen icon to a Close icon when I have opened the model in Full Screen. That is not something that needs to be addressed right now but could be looked at near the end of your development. Just a suggestion...........
-
@gĂĄbor said:
Thank you TIG. Working on it.
@GĂĄbor
I've PM'd you a full-fat way to find the 'real' xyz scaling of an object - my exporter's code only traps for scaling of 1 v. -1 etc... There is a better sure-fired way... involving lots of maths on an Array from the object's Transformation -sx = Math.sqrt(ot[0]**2+ot[1]**2+ot[2]**2)
etc... - this covers all scaling/flipping possibilities... -
Advertisement