Author Topic: New Raj shadow analysis  (Read 35006 times)

0 Members and 2 Guests are viewing this topic.

Offline onthefence

  • Administrator
  • Hero Dronie #3
  • **********
  • Posts: 1045
  • Karma: +50/-0
Re: New Raj shadow analysis
« Reply #45 on: January 10, 2009, 12:08:30 AM »
Today I received the DimageX Nemo send me, but can it still be of any use after Neki found the real pole?, and I think it's proven now that the camera wasn't a DimageX.

Did we exhaust the ideas of vignetting and pinkish imaging in the DimageX.

Now that we have these cameras, can more tests be done by HPO and numbers?

Also, have all the ideas of PICT16 being faked been exhausted, I realize that no one has reproduced the shadows in a 3D program, but is there still some testing that can prove this fake with your model and the real camera?

Edit to add: that the vignetting issue was, no vignetting in Raj's pictures, but PICT13 does show vignetting, and I am wondering if zooming the camera in would alter that result.
« Last Edit: January 10, 2009, 12:14:49 AM by onthefence »

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: New Raj shadow analysis
« Reply #46 on: January 10, 2009, 12:36:59 AM »
OTF,

I saw on a camera review site this about the DimageX:

Quote
This is something I didn't pick up on in my pre-production review but became very apparent when browsing thumbnails of DiMAGE X images. When an image is reduced in size the vignetting effect (darkened corners) tends to be more visible. This appeared to occur at focal lengths between full wide angle and mid-zoom. In the sample below you can clearly see vignetting in the thumbnail image on the left, to make it more visible I've increased the contrast of the image, now it's really quite visible.

http://www.dpreview.com/reviews/minoltadimagex/page9.asp


Offline onthefence

  • Administrator
  • Hero Dronie #3
  • **********
  • Posts: 1045
  • Karma: +50/-0
Re: New Raj shadow analysis
« Reply #47 on: January 10, 2009, 12:48:37 AM »
OTF,

I saw on a camera review site this about the DimageX:

Quote
This is something I didn't pick up on in my pre-production review but became very apparent when browsing thumbnails of DiMAGE X images. When an image is reduced in size the vignetting effect (darkened corners) tends to be more visible. This appeared to occur at focal lengths between full wide angle and mid-zoom. In the sample below you can clearly see vignetting in the thumbnail image on the left, to make it more visible I've increased the contrast of the image, now it's really quite visible.

http://www.dpreview.com/reviews/minoltadimagex/page9.asp

By that logic, a fully zoomed-in image (narrow FOV) would exhibit less vignetting (like some of the "Raj" images).

Offline nekitamo

  • Administrator
  • Major Dronie
  • **********
  • Posts: 224
  • Karma: +28/-0
Re: New Raj shadow analysis
« Reply #48 on: January 10, 2009, 01:13:43 AM »
IMO, impossible FoV angle is much bigger problem than vignetting, especially after spf33 confirmed he's also using 55° FoV for pic #13/#14/#16 reconstructions in 3D. Like I already explained, this problem can't be defeated by moving the (narow-FoV) camera further away, you simply have to have exact same FoV camera at the exact same place to make the same image.

Perhaps we should demonstrate this with 3D software?

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: New Raj shadow analysis
« Reply #49 on: January 10, 2009, 01:25:57 AM »
Perhaps we should demonstrate this with 3D software?

I would really appreciate it.  I have to admit to not completely understanding the issue.  I know what you mean about FoV but have not noticed the descrepancy between the pics.




Offline onthefence

  • Administrator
  • Hero Dronie #3
  • **********
  • Posts: 1045
  • Karma: +50/-0
Re: New Raj shadow analysis
« Reply #50 on: January 10, 2009, 01:52:31 AM »
IMO, impossible FoV angle is much bigger problem ...

I think I understand that you are saying that to get the desired FoV, there will be massive vignetting.

Because it appears to me that the camera does have the required FoV for PICT16, it is just that numbers was not standing in the right position so I rotated and moved the image, here they are both overlaid and it close enough to me:



Taken from 09010739.JPG

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: New Raj shadow analysis
« Reply #51 on: January 10, 2009, 02:12:21 AM »
Here's my side.  I believe I was standing in the correct positions, this I was able to do with my eyes But I was having problems seeing the view screen on the cameras.  I did not have enough time to set up the tripod and that may not have helped any way.  I took multiple shots of each pic because I knew I could not get the framing and zoom perfect.  I hoped at least one of the multiples would suffice.  Either that or I would let the analysts pick the best of the bunch.  The zoom on the Dimage is a touchy thing and it was near impossible to get it correct.  Sorry.

Edit to add.  The later series I did not focus so much of exact placement.  Look at the earlier ones.  The first four are for pic 16 and a bunch after for pict17.

« Last Edit: January 10, 2009, 02:25:20 AM by 10538 »

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: New Raj shadow analysis
« Reply #52 on: January 10, 2009, 02:54:33 AM »
And then, numbers mentioned to me that there are a few "pink" images interspersed with his photos. I think that these two images show distinct colouration   

Here's three pics taken in succession that clearly illustrate the odd behavoir.
One bluish, then reddish, then gray.



« Last Edit: January 10, 2009, 02:56:13 AM by 10538 »

Offline nekitamo

  • Administrator
  • Major Dronie
  • **********
  • Posts: 224
  • Karma: +28/-0
Re: New Raj shadow analysis
« Reply #53 on: January 11, 2009, 11:35:15 AM »
Here are some quick measurements I did to correct my 3D model of the pole, just in case somebody else needs them (approximate values, not very precise):


Offline HPO

  • Administrator
  • Hero Dronie #1
  • **********
  • Posts: 316
  • Karma: +27/-0
Re: New Raj shadow analysis
« Reply #54 on: January 11, 2009, 02:37:34 PM »
Here are some quick measurements I did to correct my 3D model of the pole, just in case somebody else needs them (approximate values, not very precise):

Thanks nekitamo for sharing !

Today I received the DimageX Nemo send me, but can it still be of any use after Neki found the real pole?, and I think it's proven now that the camera wasn't a DimageX.

Did we exhaust the ideas of vignetting and pinkish imaging in the DimageX.

Now that we have these cameras, can more tests be done by HPO and numbers?

Also, have all the ideas of PICT16 being faked been exhausted, I realize that no one has reproduced the shadows in a 3D program, but is there still some testing that can prove this fake with your model and the real camera?

Edit to add: that the vignetting issue was, no vignetting in Raj's pictures, but PICT13 does show vignetting, and I am wondering if zooming the camera in would alter that result.

I can of course do the tests to confirm the vignetting and pinkish imaging, but the FOV problem overrules these IMO.
Like nekitamo showed in this picture.



And can the FOV really be tested with a scale model and a scale pole?

Offline onthefence

  • Administrator
  • Hero Dronie #3
  • **********
  • Posts: 1045
  • Karma: +50/-0
Re: New Raj shadow analysis
« Reply #55 on: January 11, 2009, 03:25:03 PM »
I can of course do the tests to confirm the vignetting and pinkish imaging, but the FOV problem overrules these IMO.

And can the FOV really be tested with a scale model and a scale pole?

Thank you all, I don't think I need any more analysis, your point is made with me.

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: New Raj shadow analysis
« Reply #56 on: January 11, 2009, 05:41:56 PM »
I understand too.  Do you both come up with the same FoV values?

Edit to add:  I just checked it with my own crude method and came to the same conclusion.
« Last Edit: January 11, 2009, 06:04:44 PM by 10538 »

Offline nekitamo

  • Administrator
  • Major Dronie
  • **********
  • Posts: 224
  • Karma: +28/-0
Re: New Raj shadow analysis
« Reply #57 on: January 11, 2009, 06:53:13 PM »
Just noted your last posts, but already made this animation so here it is...

http://www.youtube.com/watch?v=NYrLtS02rUU

That ugly white thing on the bottom is angle meter "glued" below the camera, I marked 55° (green) and 49° (blue) FoV angles to be sure the camera is set properly.

Offline 10538

  • Administrator
  • Hero Dronie #2
  • **********
  • Posts: 826
  • Karma: +33/-0
Re: New Raj shadow analysis
« Reply #58 on: January 17, 2009, 10:51:50 PM »
OTF has already brought up the possibility that Raj may be shorter than me.  I have noticed that when I compare shots of the pole I took in the morning with those taken in the afternoon (same focal length), the later (more light) shots appear to have more FoV.  Does aperture affect this?   Taking shots in the middle of May would yield much different aperture settings.

Offline nekitamo

  • Administrator
  • Major Dronie
  • **********
  • Posts: 224
  • Karma: +28/-0
Re: New Raj shadow analysis
« Reply #59 on: January 18, 2009, 12:20:16 AM »
Perhaps you mean DoF = depth of field or range of sharpness?
Aperture affects depth of field in a way that smaller aperture means greater DoF due to longer exposition required for the same image brightness:


With bigger aperture the exposition can be shorter and you get smaller DoF:


Btw. small F numbers (i.e. F2.4) are used for big aperture sizes and big F numbers (i.e. F16) for small apertures.

As for the FoV (or more precisely - horizontal angle of view in my calculations), it is only affected by focal length (assuming constant lens/sensor sizes) no matter how much (or little) light enters the lens.This is because the aperture mechanism is physically placed in such position that its "hole" size doesn't affect viewing angle, but total amount of light reaching the sensor - like some kind of "optical" valve.