Author Topic: A missing shadow ?  (Read 53516 times)

0 Members and 1 Guest are viewing this topic.

Offline Nemo492

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 769
  • Karma: +30/-1
    • Ovnis-USA
A missing shadow ?
« on: August 04, 2008, 06:11:16 pm »
I have been following this research very loosely,
due to the quite sophisticated technical approach...
But I don't even see why the Raj drone would have to cast a shadow
since the sun must be quite low when the time is around 6pm...

As MidusTouch wrote :

Really no need to carry on arguing about the missing shadows
to prove it's cgi when 3D software do not make mistake on
shadows casting calculation at all since a decade ago.
« Last Edit: August 04, 2008, 06:27:11 pm by Nemo492 »
http://ovnis-usa.com
The only motivation for the DRT is finding the truth.

Offline HPO

  • Administrator
  • Hero Dronie #1
  • **********
  • Posts: 317
  • Karma: +27/-0
Re: A missing shadow ?
« Reply #1 on: August 11, 2008, 11:30:06 pm »
Hi,...these are my first findings.

Kris gave me the telephone pole model he used in his render, but didn't gave me the height of the pole or the units he works with, as you may know I work in metric and kris may work in imperial, when I first loaded the pole it was very small.

I asked him in the forum, about the height of the pole and what mm camera lens he used but didn't got a reply.
So I went for a 40 feet pole (about 12 meters), positioned the sun in the 278 Azimuth and the right angle so it produces the shadow from the crossbars on top of the pole.
Then I put a camera in the scene, now here it gets interesting, I couldn't get a lens that recreates the render or the photo, so I tried to get as close as possible, and then put the drone in to match the photo, I got the same result as the others, the drone is in a very odd position, and of course the shadows don't match.



And the final render

But then I thought maybe Kris didn't use a camera view but the perpective view to recreate the pole, so I turned to perspective view, now I could match the picture way better than before but of course the position of the drone was changed also, so I corrected that.
I was surprised how much I needed to turn and tilt the drone to get it's original position back, you can see the difference in the screenshot below (I did a viewport render this time, so I didn't have to put an extra render in this post).  I'm surprised how much the camera lens distortion influenced the position of the drone in relation to the pole. But still the shadows are not the same as in PICT016.



So there are still a few options that needs to get looked into.

1- recreating the pole in camera view, but the question is how long is the pole and what mm lens to use.

2- tilt the pole in the direction of the cables, and try to recreate the shadow of the crossbars on top of the pole in lower angle of the sun and then position the drone to match the photo.

3- a combination of the two options above, try to recreate a pole in camera view when the pole is in a tilted position.

-HPO-





« Last Edit: December 02, 2008, 07:11:25 pm by HPO »

Offline tomi

  • Hero Dronie #2
  • ********
  • Posts: 668
  • Karma: +37/-26
Re: A missing shadow ?
« Reply #2 on: August 12, 2008, 01:19:38 am »
Hi HPO,
You have done so much marvelous work, I really hate coming in and saying anything, but I guess I have to throw my thoughts into the pile for what they might be worth..

I studied the pic 16 and your last rendering here side by side.  I have the same issue with Kris, and others.  I do not see the same angle in the rendered work that the drone has in the original picture.

I know everyone has made a lot of very complicated analysis about this, but I still think that there are just simply 3 elements here to worry about.

The camera's eye, what it saw, what lense maybe as well, but it is the cameras viewpoint that is important here.

The position of the pole and drone relative to each other exactly and exactly at the angle the camera saw.

The exact position of sun or light source.

You must have 2 of these 3 elements to get the third right.  The main ones are the first two I mentioned. 

I have yet to see any rendering that shows the position of that drone exactly as it is in the photo, and until that happens, there can be no way to tell if the shadows are wrong.  It seems very simple to me here and I wish someone would produce a cgi rendering that has that drone correctly positioned, only then can I believe anything like shadow inconsistencies on it.
Cheers, and thanks for listening :)

PS:  Please ignore any admin status I have, please explain to me like I am your willing student.. what am I missing here in my estimation of the problem.. let me know where I might be wrong please :)  Cheers..
« Last Edit: August 12, 2008, 01:39:56 am by tomi »

Offline HPO

  • Administrator
  • Hero Dronie #1
  • **********
  • Posts: 317
  • Karma: +27/-0
Re: A missing shadow ?
« Reply #3 on: August 12, 2008, 07:17:12 am »
Hi Tomi,...yes you are right, maybe you didn't get what I was trying to say, but I also found that it is very important what the camera eye saw, only a slight difference from camera view to perspective view allready made a great difference in the position of the drone.

I to have problems with the way the drone looks in the last render, the arm is bend towards the end, but this "odd" position of the drone is what you get when you try to match the photo with that angle of the pole, but.... nobody knows for sure how the pole is positioned in relation to the earth in real life, and the exact position of the camera.

I agree with kris and others that the shadows do have a match on the pole when the pole is in this straight upwards position, the shadow on top of the pole from the crossbars is important in this case. But who is to say that the pole is indeed in a straight upwards position, (if only somebody could find that pole), I think when the pole is tilted towards the cables on the right side, the drone would be more in a level position in the render and could get a better light, also the sun can be in a lower position to match the shadows on the pole. (but that has to be tested)

There is also a positive point in the last render, altough you see the angle of the shadow of the main arm on one of the crown "wings" the ambient light from below "flashes out" (I don't know a good english word for that) the shadow on the main torus.


Offline tomi

  • Hero Dronie #2
  • ********
  • Posts: 668
  • Karma: +37/-26
Re: A missing shadow ?
« Reply #4 on: August 12, 2008, 08:58:11 am »
Thanks HPO, I've spent a lot of time thinking about this.  It is like a brain puzzle.  I've tried moving things, in the sun, putting little match sticks here and there while I observe the effects of slight tilt to shadows and time over a few minutes.  It's just my opinion, but what the camera saw is all that is important.  And until someone can get that drone to be exactly the way it is in the picture, the question of shadows is irrelevent. 

The slightest tilt of an object from the observer (camera) point of view makes a huge difference.  That's why for all of Salads demonstrations, he and others can't seem to bring that pic 16 to a shadow consistency.

If you use the eliptical center as your guide, I think that is the only way..
Then lock those objects together once you have gotten as close as humanly possible to their relationship to each other perfectly positioned.  Then the lighting effect issue would be relevent.  Until then I can't see how recreating anything in cgi would prove what is missing or not.

Also, we do have the issue of the lighting effects not being accurate in the cgi programs.  One of Kris's main interests in this situation is to what degree and range they are lacking authentic reproduction.  He accepts this limitation as being the possible killer of all these "shadow recreation must be spot on in cgi" theories to end with the final decrees of what the cgi engineers have to say.  But that will take time for people to come to accept and the degree of what everyone is asking of these cgi programs will have to be considered.

If someone wants to declare that the shadow inconsistencies can not be matched by a variance of less than a degree, then I think right now, based on what we are learning already from the companies, there is no proof because recreating lighting effects to that degree of accuracy has been almost negated already by the companies who provide these programs. We must find out as much as we can though.  That's what it's all about.

Unbiased judgement means when the proof is in your face, undeniably making itself obvious and the means of showing this abnormality is verified and accurate, then hey we got a hoax.  But that so far is not what has been presented.. sigh ..  but bring it on :)  Cheers and Thank you!!

Offline tomi

  • Hero Dronie #2
  • ********
  • Posts: 668
  • Karma: +37/-26
Re: A missing shadow ?
« Reply #5 on: August 12, 2008, 09:22:01 am »
I just watched an old sightings clip on u tube.  A Vernon Miller was being interviewed about these sightings caught on camera in Canada.
http://www.youtube.com/watch?v=mg7Bfd9c9Ps

He is from the Brooks Institute of Photography.  This video of him is ages old, so I imagine he may not still be there.  But Brooks is the place I'm thinking of bringing these photos to for more expert analysis.  This place is the established leading institution of the highest regard for photography and fine arts.  It is also luckily in Santa Barbra, Ca.  Very liberal I'm thinking and possibly open to giving us some consultation.  VERY ESTABLISHED in their field.
http://en.wikipedia.org/wiki/Brooks_Institute_of_Photography

I'm not thinking of spending money here, I'm thinking of building a relationship of inquiry with some friendly base within, anyone want to give it a try with me or take this idea and run with it?  I've got kinda a full plate atm..

HPO you may have heard about this institute or people in your field may have a contact there?

Their website is here:
http://www.brooks.edu/
« Last Edit: August 12, 2008, 09:57:00 am by tomi »

Offline nekitamo

  • Administrator
  • Major Dronie
  • **********
  • Posts: 224
  • Karma: +28/-0
Re: A missing shadow ?
« Reply #6 on: August 12, 2008, 10:59:25 am »
Well, here are some things I considered while trying to perform 3D analysis...

As mentioned before, judging by the utility pole, some barrel distortion seems to be present in Raj's PICT0016 - or perhaps not, if the pole is actually bent in just the "proper" place to make it appear that way. However, I found it easier to align my 3D scene with slightly barrel-corrected version of Raj's image, so there may be something to it. Note that there are significant differences between the original and corrected version, not just a few pixels missplaced:



Also, there's a great angle assessment tool present on the drone itself that I also used for better alignment of my 3D model. If we assume that the three small fins on the drone are all of equal length and the angle between them is exactly 45° (and hope they aren't tilted up or down wrt drone torus), we could actually calculate drone tilt in relation to the camera by measuring perspective-transformed angles and lengths of the fins in Raj's image (barrel-corrected version):



I didn't do the math (yet), only paid special attention to match the exact angles and length ratios while adjusting the tilt of my 3D model. This actually makes it easier and is more precise than matching tilted toruses, especially when your 3D model is completely hiding the background image like in 3D software I used.
« Last Edit: August 12, 2008, 11:23:30 am by nekitamo »

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: A missing shadow ?
« Reply #7 on: August 13, 2008, 09:09:26 pm »
positioned the sun in the 278 Azimuth

How can you tell where that is?


Offline elevenaugust

  • Administrator
  • Hero Dronie #3
  • **********
  • Posts: 1231
  • Karma: +34/-1
  • א ב ג ד ה ו ז ח ט י כ ך ל מ ם נ ן ס ע פ ף צ ץ ק ר
Re: A missing shadow ?
« Reply #8 on: August 13, 2008, 09:41:03 pm »
positioned the sun in the 278 Azimuth

How can you tell where that is?
Good question.
Very nice work, HPO; but I'm curious to see how various parameters can modify the shadows rendering, such as:

- Camera lens mm
- Pole Height
- Pole tilt
- Sun azimuth
- Barrel/pincushion distortion
- Main arm of the drone tilt wrt drone torus (as suggested by Arkhangels)
- Reflection/ambiant lightening.

Another question for our 3D specialists: chromatic aberrations seems to be one of the hardest thing to reproduce in 3D rendering; is someone know which one of the rendering software known in 2007 was able to recreate them?
I guess there's probably a short list.
IPACO, the new tool for photo and video analysis is on-line! www.ipaco.fr

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: A missing shadow ?
« Reply #9 on: August 13, 2008, 09:50:50 pm »
Nekitamo,

The reviews of the Dimage X clearly state the camera has a certain amount of barrel effect.

It's also very possible the small arms are not in exact alignment.  Onthefence discovered this when experimenting with a mock up.

I have a feeling the nine tabs that hang below the torus are also not aligned.

Offline HPO

  • Administrator
  • Hero Dronie #1
  • **********
  • Posts: 317
  • Karma: +27/-0
Re: A missing shadow ?
« Reply #10 on: August 14, 2008, 12:48:25 am »
positioned the sun in the 278 Azimuth

How can you tell where that is?



Well yes of course,...that's a good question, I also wanted to write 24 Altitude, the position of the sun on 5-16-2007 at 17h 42m.
With the height of the sun at 24 degrees, I've adjusted the position of the pole so that the shadow on top of the pole from the crossbar/member is the way it shows in pict 16 and 17, but of course, if that is the right position of the 277/278 azimuth can be debated, but that postion of the pole (by a margin of only a few degrees) is the only position that the shadow shows on top of the pole. (1111 had a good animation for this in the OMF forum)


positioned the sun in the 278 Azimuth

How can you tell where that is?
Good question.
Very nice work, HPO; but I'm curious to see how various parameters can modify the shadows rendering, such as:

- Camera lens mm
- Pole Height
- Pole tilt
- Sun azimuth
- Barrel/pincushion distortion
- Main arm of the drone tilt wrt drone torus (as suggested by Arkhangels)
- Reflection/ambiant lightening.

Another question for our 3D specialists: chromatic aberrations seems to be one of the hardest thing to reproduce in 3D rendering; is someone know which one of the rendering software known in 2007 was able to recreate them?
I guess there's probably a short list.


- Camera lens mm -> no shadow changes, you don't change anything on the geometry or the light direction, but it can effect the way you see the total of the geometry with shadows, and the position of the drone in the scene when you want to recreate the picture.
- Pole Height -> almost nothing or nothing with parallel light rays
- Pole tilt -> only tilt I tried sofar was 5 degrees towards the sun, almost no changes visible on the pole shadows. (but there are)
- Sun azimuth -> will have a great affect on the pole shadow, only a few degrees projects the top shadow of the cross members
- Barrel/pincushion distortion -> the same as camera lens
- Main arm of the drone tilt wrt drone torus (as suggested by Arkhangels) -> that has to be tested of course but that will have an effect, I'm sure.
- Reflection/ambiant lightening. -> like I wrote to Tomi, It seems like it does, when you look at the last render in the perspective view.

The list for chromatic aberration render software? well, the only one that I know of is "Violet" an tonemapping GUI for Indigo renderer, but I never used Indigo renderer.

http://www.indigorenderer.com/joomla/forum/viewtopic.php?t=1332&sid=2ac25feecdd54f2ea7a52f63381b1d93


BTW. nice work with the corrected barrel distortion picture and line up nekitamo, one of the things I look at when positiong the drone  is the collinear or parallel position of one of the antenna's with one of the half rings inside the torus, they are exactly in one line.
« Last Edit: December 02, 2008, 07:18:35 pm by HPO »

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: A missing shadow ?
« Reply #11 on: August 14, 2008, 01:26:01 am »
I think this is closest from that gif.  Still probably needs some tweaking.  Remember there is no point of reference in the photos and the exif data may not be accurate.




Offline HPO

  • Administrator
  • Hero Dronie #1
  • **********
  • Posts: 317
  • Karma: +27/-0
Re: A missing shadow ?
« Reply #12 on: August 14, 2008, 06:42:09 am »
Hi Numbers,..it's a little hard for me to respond in this topic because we got the software at the office and mostly I do the postings at home, but I don't think that is the right angle for the pole, it looks like the right angle for the drone in the picture but not in the 3D scene.
But I think we have to look the answer in a different direction, I made a animated gif for myself of pict 16 and 17, what if Raj allready is tilting the camera to the left in pict16 maybe because he couldn't walk any further on that point.
when I tilt the camera to the left, the drone gets tilted downwards on the right in the 3D scene, when I correct the drones position it is pointing more to the sunlight. (if you understand what I mean)

Offline elevenaugust

  • Administrator
  • Hero Dronie #3
  • **********
  • Posts: 1231
  • Karma: +34/-1
  • א ב ג ד ה ו ז ח ט י כ ך ל מ ם נ ן ס ע פ ף צ ץ ק ר
Re: A missing shadow ?
« Reply #13 on: August 14, 2008, 11:17:57 am »
- Pole tilt -> only tilt I tried sofar was 5 degrees towards the sun, almost no changes visible on the pole shadows. (but there are)
- Sun azimuth -> will have a great affect on the pole shadow, only a few degrees projects the top shadow of the cross members
- Main arm of the drone tilt wrt drone torus (as suggested by Arkhangels) -> that has to be tested of course but that will have an effect, I'm sure.
- Reflection/ambiant lightening. -> like I wrote to Tomi, It seems like it does, when you look at the last render in the perspective view.
So, to summarize, there's only these four possibilities (or combination of) who maybe can explain the shadows problems.

The list for chromatic aberration render software? well, the only one that I know of is "Violet" an tonemapping GUI for Indigo renderer, but I never used Indigo renderer.

http://www.indigorenderer.com/joomla/forum/viewtopic.php?t=1332&sid=2ac25feecdd54f2ea7a52f63381b1d93
Thanks, HPO.
If I remember correctly, Kris was talking a while ago of another software who can also produce such chromatic aberrations, but I can't find anymore which one.
The fact is that we have chromatic aberrations on some drones pictures, the Chad's one, if I remember correctly.

It's hard to believe that soemone who have the skills to reproduce perfect chromatic aberrations can make a huge mistake like the main arm shadow missing.....

IPACO, the new tool for photo and video analysis is on-line! www.ipaco.fr

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: A missing shadow ?
« Reply #14 on: August 14, 2008, 05:22:12 pm »
Hi Numbers,..it's a little hard for me to respond in this topic because we got the software at the office and mostly I do the postings at home, but I don't think that is the right angle for the pole, it looks like the right angle for the drone in the picture but not in the 3D scene.

That's basically at the crux of this whole issue.  I don't think these 3D apps can exactly duplicate what the Raj camera sees.  The shadows from those apps are crude and not entirely accurate.  The perspective is hard to emulate.

It is because of this that we should not put too much faith in this endeavor.  There is too much room for error, whether that be from the user or app limitations.