Real VR inside Second Life with Google Cardboard and Trinus Gyre

[Coby Foden]

---

Phil Deakins wrote:

 

Parallel views when taking the pics are bound to produce an effect when viewed in different eyes, of course, but proper 3D requires convergent views when taking the pics.

---

What do you mean by "parallel views" and "convergent views"? I guess that you perhaps mean how the camera lenses are arranged.

 

I take it to mean:

 

• 1. in your "parallel views":  the camera lenses axis are parallel

• 2. in your "convergent view":  the camera lenses axis converge

You claim that picture pair produced by method 1 produce an effect, but it's not proper 3D.

You claim that picture pair produced by method 2 will produce proper 3D.

Unfortunately, you are wrong, in spite of in spite of your "experience from years ago".

The truth is:

• Cameras setup so that the axis of the lenses are parallel produce proper 3D image pair.

• Cameras setup so that axis of the lenses converge produce proper 3D image, but it has distortions.

All the three methods I have mentioned earlier to make an image pair for 3D viewing do produce proper 3D view. The converged (Toed-in) method is the worst as it causes distortions in the image (because of the reasons explained earlier).

 

One link: :smileywink:

http://www.binocularity.org/page3.php

"Ideally all stereoscopic images for human viewing should be captured using a parallel camera arrangement. In this approach cameras are setup to have parallel optical axes. Images captured with parallel cameras do not have the geometric errors inherent in the toed-in (i.e. converged) approach."

[Coby Foden]

---

Phil Deakins wrote:

Coby. I'm not going to spend time going through yout copy and paste posts, because you're not the author. I prefer to rely on my own experience from years ago.

---

Naturally you're not going to debate with those "copy paste" texts and pictures because you know that they explain clearly and accurately how things really work. Sometimes you ask for "official documents" to verify something. In this case the copied texts and pictures from various sources appears not be be a good thing. Anyway, the majority of the text in my posts I typed with my own fingers, there is actually very few copy paste texts in them. (Well, I "copied" many texts into my memory and "pasted" from there by typing.) :smileyvery-happy: The images are copied, I could have drawn them myself but why bother as they were freely available on the net.

I prefer to search various articles to learn how things work. Then I can do my own experiments to see and verify is it really so what has been said in the articles. In this 3D matter I can say that it indeed does work as has been explained in the articles. Your ideas how it works have aren't quite correct.

 

Isn't it better to find out the real facts for a discussion, than just to rely solely on one's own (possibly misconceived) ideas about the matter?

[Kelli May]

---

Phil Deakins wrote:

Parallel lines don't converge at infinity

Parallel views when taking the pics are bound to produce an effect when viewed in different eyes, of course, but proper 3D requires convergent views when taking the pics.

---

"No true Scotsman" argument. Foul and forfeiture.

[Coby Foden]

---

Phil Deakins wrote:

Parallel views when taking the pics are bound to produce

an effect

when viewed in different eyes, of course,

but proper 3D requires convergent views

when taking the pics.

---

---

Phil Deakins wrote:

Coby. I'm not going to spend time going through yout copy and paste posts, because you're not the author. I prefer to rely on my own experience from years ago.

---

Ok, to not to rely what various sources say about the subject, I did this experiment myself.

I'm going to show here what really happens with the following conditions.

    • an image pair is produced by camera setup which have the axis of the lenses parallel

    • the image pair is viewed with divergent (aka parallel) eyes

We will see does it only produce "an effect", or does it produce a proper 3D view.

I produced the image pairs in Blender and also in Second Life (the same principle applies to real life photos).

1. Blender

There is "Stereoscopic Camera" add on for Blender, available here: http://www.noeol.de/s3d/

Three different camera set ups are available:

The python script creates nodes to produce image pair rendering.

So I created a scene in Blender with some monkey heads and cylinders; setup lighting, materials, and the camera.

Then selected the "Preset" to side by side image creation. Finally just command "Render" and Blender will create the image pair.

Image pair for parallel eye viewing (left lens image on the left, right lens image on the right):

(Note: If, when looking at the image above, you see the text "Parallel viewing" floating somewhere towards the front then you're not looking at it with parallel eyes, you're looking at it with crossed-eyes. With parallel viewing the text will appear somewhere at the same distance as the farthest two cylinders are located.)

Image pair for crossed-eye viewing (left lens image on the right, right lens image on the left):

2. Second life

Location: http://maps.secondlife.com/secondlife/Glanduin/45/32/24

It is possible to make parallel axis lens image pair with SL's single camera. Take a snapshot, name it "Left lens", then move the camera sideways 65 mm to the right, take another snapshot, name it "Right lens". (Note: In the scene nothing should move because the snapshots cannot be taken simultaneously.)  Put the images side by side, and there you have it: image pair for 3D viewing.

Image pair for parallel eye viewing (left lens image on the left, right lens image on the right):

Image pair for crossed-eye viewing (left lens image on the right, right lens image on the left):

Parallel eye viewing is very difficult for most people without any device. Parallel viewing is limited to rather small images, the separation of the centers of the images should be a bit less than the interocular separation of the eyes.

However image pair for parallel viewing can be viewed properly with devices. I have a 55 inch 3D capable HDTV. So I can make 1920 x 1080 pixels size image pair and view it in full HD resolution in 3D by using the 3D glasses. The objects in the foreground pop out from the screen about one meter and objects far away appear to be on the screen level. Pretty awesome.

If this is not "proper 3D", then nothing is.

 

Finally I would like to quote the following saying:

The proof of the pudding is in the eating.

(Spanish version: Al freír de los huevos lo verá.)

 

:smileyhappy:

[Coby Foden]

PS.

The above post belongs to the series:

:smileyvery-happy:  :smileytongue:

[LlazarusLlong]

It seems that someone here considers dementia as being a valid element of rhetoric argument.

[Fox Blackcinder]

So... Has anyone managed to get either SL OR Viewer or Ctrl Alt Studio (or some other viewer, even) to work properly with Google Cardboard via TrinusVR? I gave both a try today and the "official" viewer, once you enable Advanced HMD Mode and turn on HMD View, forces the window to a resolution way bigger than my screen (which makes TrinusVR unable to render it), and Ctrl Alt Studio outright won't start up unless it detects an Oculus Rift connected to the computer, as far as I can tell.

[AlecDeBoeuil]

I thought I couldn't be less interested in "Real VR",  but I have been proved wrong.

Alec - real and virtual, as two separate things

[Annalee McKeenan]

Any luck on this? I just ordered a cardboard kit and should get it tomorrow. Does it work decently with the regular viewer?

[Coby Foden]

Regular viewer has only one single camera. For stereoscopic view a viewer which has two cameras would be required. The two cameras will provide a slightly different view for each eye - the result is strereoscopic view. So, the regular viewers will not give stereoscopic view because both eyes would see exactly the same image.

The Oculus Rift enabled viewers provide separate slightly different views for each eye. Unfortunately for Google Cardboard those views will not work right for it. The views are specifically made for Oculus Rift - they have a very strange shape which the Oculus Rift corrects to a rectangular shape. Google Cardboard cannot do that.

[AlecDeBoeuil]

Also, whatever VR viewer you use, if you are female you are likely to feel sick. Even if you are playing a male alt.

Alec - to explain what the sellers of these gimmicks would prefer suppressed.

[bwillard]

Hey,

I recently bought one Google Cardboard and I tried FIFA with Trinus. But there is some problem with Trinus settings. I can't see the whole screen. It is not easy to config settings, atleast not yet. 

I was hoping Second Life developers would develop SL viewer that will work in general parallel optical VR gadget. But seems like it's not here yet.

Even though Google Cardboard is cheap and not full-fledged yet. I believe Google cardboard is in experimental phase and they are motivating people to develop apps and software for it. And one day, Google will release awesome VR gadget. Aside from that, people have even come up with their own comfy Google VR.

Anyways, I was hoping Second Life Developer community will make things easier soon for parallel optical VR gadget users other than Oculus Rift.

Thanks.

 

 

[Nalates Urriah]

A new version of the Linden Oculus Viewer has been released 7/2016: Second Life Project OculusRift Viewer version 4.1.0.317313.

Also see my blog. Look in categories to find all the related articles.