Prepp3d Views: Peripheral Distortion Correction

[denali 74]

So, I believe an apology is appropriate in this case because we are not using any bit of your algorithms because they are just not correct.

  I don't owe anyone an apology.  I have been a developer for over 20 years, and I have written patent applications.  I know for certain that the exact math, readily available to anyone willing to work it out, was not worked out anywhere prior to my doing it.  It was APPROXIMATED.   I don't know why it was the case, but it was, even in the photography field.  I happened to have asked many experts, and everything I was directed to was nothing more than a "fisheye" correction, which was not a true mathematical solution, it just looked much much better.  

 

However, I say "stole my work" tongue in check because, as I said, it was readily available to anyone willing to do the geometry needed - able to wrap their head around it.  That it being common math I really cannot "claim" it.  I think all I did was be the first to run the 4 minute mile, prove it could be done.   But I am very very certain that until I published what I did in this very forum out in the open as a Post Process shader for P3D, that no such lense existed in the wild.   

 

It may be the case that others had worked it out, as is very often the case behind firewalls, but It most definitely did not exist for PC  Part of completing patent applications is performing due diligence, looking for prior art.  I am very thorough.  Mostly cause I don't like to do work someone else did already.

 

It would not be the first time that ideas are taken and ran with, it's the way we all make progress together.  I struggled too long to do anything with it, have to pay (many medical) bills with real work that pays real money.  That's why I even posted that LM can do whatever they want with it at will.   There's only a year I'd have any claim to it anyway (laughable not even close to patent worthy), and it's been what?  2 , 3 years?  How old is my kid now?   

 

What is with everybody playing smear the queer on the Internets all the time anyways.  When are all you n00bz going to grow up?   I wish AOL had never connected to the internet.  And I want my execpc and gophernet back.

 

And You're Welcome!!1

" If you can spend the money [iDP with LCD thingy] is probably a worthy addition "

 

You don't need IDP.  Why don't you just publish your work as a direct Post Process Shader into P3D, instead of using injection.  Then you don't need to worry about the menu system.   I think you just want to sell IDP.  The only thing I was missing was grabbing enough off screen bitmap, clipping out only what I needed, and doing some kind of AA process on that.  It was readily available FOR FREE using Reshade.  Why are you charging $170 to do the same thing?  Gerard has my Reshade shader.   I'll just polish that up and cough it up here.  

 

BTW, did I mention I got hit by a Semi yesterday??  What a world .. what a worl ...

 

 

there are single-pass multi-projection techniques that work in DirectX11 on current gen hardware

 Notice I said earlier "  It's actually what it was originally intended for in the current processors, the core structure exists in them."  One of the delays I had was I'm a perfectionist and I was trying to get my head around this.  Didn't quite get there.

Call it Synchronicity.  But I had the core math up first.  Here.  I'm not saying you stole it, actually stole it.  But I'm not going apologize.  I really do think you've got a great product, for people that can pay half a grand for a card that can run 3 screens and have a few hundred left over, They Should Really Buy It.  Cause I can't stand the distortion so much, I wrote an algorithm all by myself ... in about 3 weeks total work days wrapping my head around the geometry (maybe about a half a year complaining to myself out loud on here about it seeding my mind for the work), not 5 years, because you didn't have it available 2 years ago.   I don't expect to get paid for 5 years of work.  It's a hobby.  

 

You have a beautiful product and you should get paid for it.  Often once a product is released the job is like being a fireman;  Just man your post.  Hopefully the field you set on fire keeps you busy, or you at least have fun.  Peace.

[ngidalov 0]

Nikola,

 

One wide view gives better framerates than 3 seperate views.

 

I now use 3 seperate outside views for 1 wide view , so no distortion.

P3Dv3.2 has improved multi monitor support and 3 views now have the same performance as 2 views prior v3.2.

 

Does you programm connect 3 views to 1 and then corrects the distortion through ID Pro ?

( so about the same performance as 1 wide NVSurround view )

 

Gerald,

 

It is not mathematically possible to correct for the distortion when using only one view. So, for each LCD a view is created and this view is corrected for correct geometry given the LCD size and position as well as eye-point.

 

Regards,

Nikola

[denali 74]

It is not mathematically possible to correct for the distortion when using only one view

   Well.  I did it.  Like I said: 4 minute mile.

[GSalden 4,708]

Well.  I did it.  Like I said: 4 minute mile.

That's is true.

I have tested it and the NVSurround Distortion is completely corrected..

[ngidalov 0]

That's is true.

I have tested it and the NVSurround Distortion is completely corrected..

 

OK, then enjoy the "4 minute mile" correction, but your view is not geometrically correct.

[denali 74]

your view is not geometrically correct.

 

Okay I'm game. You've been at this for longer than I have. I'm open to criticism. Maybe there is something I haven't wrapped my head around yet. Please explain why it's not geometrically correct.

I can actually think of reasons why you would say that. Considering monitor angles is one and I get that. I have not factor that in yet and you have. But the main point is that going from a spherical projection on a plane back to a sphere is entirely possible mathematically.

[OmniAtlas 93]

Not too sure why this got into thrash talking one another; there is a commercial option out in the market now for those who would like to pay for it; if there is an alternative or more affordable solution I would gladly use it. 

[denali 74]

 

 

Not too sure why this got into thrash talking one another; there is a commercial option out in the market now for those who would like to pay for it; if there is an alternative or more affordable solution I would gladly use it. 

 Done

[Mike_CFII_MEL 815]

Hi,

 

Am I to understand there is an professional item out there that can be purchased and used with P3D to correct for the distortion?. If so, I have two questions. 1. Do you need a PHD to figure out the software and 2. Where can one find this software?

 

What is this 4 minute mile for NVSurround?, I'm running three 75" 4ks via NVSurround and cannot stand the distortion, it ruins the flight deck view.

 

Thanks

[denali 74]

For my software, a PhD is required, and it cannot be found anywhere. 

 

Anyone wanting to have a go at this feel free to msg me.  

[ngidalov 0]

Hi All,

 

No PhD required to make or use LCD Panel Designer software.

 

Just to let you knwo that we just released an new version of LCD Panel Designer.

Thanks to everyone who used LCD Panel Designer v1.0r1 and provided us with useful feedback.

We integrated some of the feedback and suggestions in this new release.  

Give it a try.....

 

http://fly.elise-ng.net/index.php/news/78-news-immersivedisplaypro10r2

 

 

Regards,

Nikola

[denali 74]

NOT "100% Geometrically Correct".  Stretching is still visible in the peripheries of each view.  You are only minimizing it's presence with individual views.

[ngidalov 0]

NOT "100% Geometrically Correct".  Stretching is still visible in the peripheries of each view.  You are only minimizing it's presence with individual views.

 

It is, when you put yourself at the designed eyepoint no more stretch. That's how it works.

 

N.

[denali 74]

It is, when you put yourself at the designed eyepoint no more stretch. That's how it works.

 

Nikola,  You have a great product.  I even recommended it to someone just last night that asked me if my thing would help them with a projected image on a dome issue.  I do believe your software is exactly what people need in that situation.

 

And I will not argue with you that much that you have a solution often taken by programs like assetto corsa for flat LC Displays and it is effective.  You have worked hard in your software to present a solution.  A very pricey one, but also a functional one.   Where for projected displays your product is very worthy, but as for an LCD solution I do not think it is fair for you to charge it as a bundled price, as you do, e.i. you must purchase your flagship product at it's premium price.  However, that is not up to me, but the markets.  

 

And again, if you did not use the math that I have published, you are not getting a 100% geometrically correct solution.  Not even "when you put yourself at the designed eyepoint".    Your claim is false.

[ngidalov 0]

denali,

 

I do not know your actual name. Sorry. So I address you as denali.

There is one simple way to proof you are not correct about the stretching LCD Panel Designer algorithm. 

If you are good in Direct3D or OpenGL or any other scene graph software you can do the following:

 

Create a scene in this way:  Draw a cylindrical surface with a radius 1 m (or whatever you like) and high enough to be completely seen from the eye-point (usually 0,0,0). Then on each 10 deg along the radius of the cylinder draw spheres. So you should end up with 36 spheres along the perimeter and you can then stak them along the cylinder height. This is a very nice scene to evaluate any stretching or distortion of the perspective projection used by DX or OpenGL "camera"

 

Now, lets imagine a LCD setup with  3 LCD (the left and the right rotated by about 40 deg) and an eyepoint in front of the second LCD put in such a way that the total horizontal FOV is  around 180 deg . For the simplicity, ignore the bezels.

 

You can position camera in this scene , use whatever FOV you need and try to capture images with this camera.

 

Now, you can try both your algorithm (single view) and LCD Designer algorithm (multi-view) and try to produce bitmaps for all LCD panels. You can again use LCD Designer to paste the bitmaps on the LCD displays, move your position to the eyepoint and rotate your head to observe the spheres pon all LCD panels.

 

If the correction is OK in both algorithms, then on the LCD panels you will be able to see about 18 spheres (180 deg view) and the spheres along the cylinder perimeter will be seen as circles and not as elongated ellipses.

 

Will you give this suggestion a try?

In the next version of LCD Designer I will add this "spheres" scene as a standard scene in the software to evaluate the distortion.

 

I propose to take this discussion offline. You can always contact me directly using my email box in this forum or contact Fly Elise-ng directly.

 

 

Regards,

Nikola