PDA

View Full Version : An elementary question on overlap



enrico_953
11-09-2009, 03:51 PM
Hello all,

I'm new to this forum and considering the acquisition of an NN5/RD8 setup.
Hopefully, someone can enlighten me.

I have taken 'typical' panos by simply rotating my rig on the tripod,
and arbitrarily adding about 10-20% overlap between my images.
Then of course, post process/stitching.

When I look at the rotator specs I see numbers like 18 degrees/ 20 stops,
30 degrees/12 stops, etc.
They all work out to an exact 360 degrees and I fail to see where the 'overlap' comes into play.

My only guess is that when using the settings in the camera/lens database,
the nodal point setup accounts for an overlap to be added among the image sets taken.
Also, when setting up a 'non-standard' body/lens with Fanotec's instructions,
the same 'overlap' will be taken into account.

Am I on the right track, or am I just a dimwit.
(I'd ask my wife the same question, but I already know her answer)

Any feedback is appreciated...thanks for your patience.

cheers,
Enrico

"paddle faster...I hear banjo music"

hindenhaag
11-09-2009, 11:40 PM
Hi enrico,

In the database all the data is inclusive the overlap.

Use this link, http://www.frankvanderpol.nl/fov_pan_calc.htm. I prefer 30% overlap to be more secure. Depends on the shooting site. You have to use the amount of pictures you have to take to find about the degrees you have to set on the rotator.


Heinz

markkuk
11-10-2009, 04:07 AM
When I look at the rotator specs I see numbers like 18 degrees/ 20 stops,
30 degrees/12 stops, etc.
They all work out to an exact 360 degrees and I fail to see where the 'overlap' comes into play.
The overlap comes from using a lens that has a wider field of view than the angle between click stops.

2BeCalifornian
12-12-2009, 02:00 AM
Hello Enrico,

To me, overlap is needed for your stitching software to correlate common patterns between two pictures.

You already found out that your rotator has "stops" and "degrees" and sure enough, stops x degrees = 360?. So, in your example, 18 degrees / 20 stops means that tour rotator will allow you to stop 20 times, every 18 degrees in a 360? cycle.

Heinz is right, 30% is a good start. The keyword though is "start". Your rotator has a finite amount of stops so in the end, it's the number of stops you key into your rotator that will set your amount of overlap.

Let me show you how:
Here's a link to an excellent book that has a pano calculator http://www.hdrlabs.com/tools/panocalc.html And yes, Heinz gave you the link to the creator of that pano calculator!
Section 1: set your sensor data and the lens focal (the wider the lens, the less stops on your rotator)
Section 2: function of your focal length and sensor dimensions, you get the Horizontal and Vertical Field of Vision (FOV). Note: the Vertival FOV of a portrait picture equals the Horizontal FOV on a landscape picture.
Section 3: Now that you know the FOV, enter the desired overlap and the calculator gives a theoretical number of images.

Now, this number of images is rarely an integer. You have to round it up....to match the number of stops on your rotator.

The whole purpose of Section 4 is to recalculate the overlap given the fact that your rounded up the amount of images...

I went through this exercise not too long ago for my D700 + RD16 rotator. Here's the result:I've added the theoretical number of images, and the rounding up to accommodate that the RD16 can do...

I hope I did not go too deep... in summary:
- overlap is necessary for panoramic softwares to stitch images properly
- the actual overlap is driven by a combination of your lens orientation (portrait/landscape), focal and rotator setup
- too little overlap will cause an improper stitching, too much overlap isn't good either (especially for your point of no parallax isn't set properly/close objects). Keeping overlaps values in mind is a must (in fact, I laminated the table in the PDF attached - and carry this table with my pano head!)

Cheers!

Philippe

YvesG
12-14-2009, 03:11 PM
Interesting link you suggested.
On my side, I have worked on excel format to calculate it regarding the position on the sphere, the angular step on the panoramic head and of course the minimum overlap you want.
The only issue is that it is in french for the time being : Panorama calculator (http://photos.yves.over-blog.com/article-panorama-par-assemblage-panorama-calculator-41157419.html)

2BeCalifornian
12-14-2009, 04:00 PM
Hi Yves,

I found your spreadsheet equally interesting. Do you need help translating it? Let me know!

Bonjour Yves,
J'ai trouve votre feuille de calcul forte interessante. Avez vous besoin d'aide pour la traduction? Dites mpi quoi.

Thanks/Merci!

Philippe

YvesG
12-15-2009, 12:29 AM
Thank you for your support.
In fact, I just need time to translate it it and I had some people asking for that new file.
That is why I decided to put it online without the english and turkish translation.
I guess that by the end of this week I will add the english translation.
Thank you again.

YvesG
12-17-2009, 07:48 AM
OK, translation in english is done :
Panorama calculator (http://photos.yves.over-blog.com/article-panorama-calculator-file-and-tutorial-in-english-41324537.html)

nick fan
12-17-2009, 08:46 AM
OK, translation in english is done :
Panorama calculator (http://photos.yves.over-blog.com/article-panorama-calculator-file-and-tutorial-in-english-41324537.html)


Hi YvesG,

Thx for the calculator and video.
I don't understand the 2 results in your "results details" table. what is +/-15 deg and +/- 60 deg? what is 60 deg and 130 deg?


Nick

YvesG
12-17-2009, 10:52 AM
That means you need to shoot a first line at +15? and a second line at +60?.
And then you have to do the same at -15? and -60?.
Those angles are the one you can read on the arm that is supporting the camera (I am calling it the upper arm).
I don't know if I am clear enough.

nick fan
12-18-2009, 12:22 AM
That means you need to shoot a first line at +15? and a second line at +60?.
And then you have to do the same at -15? and -60?.
Those angles are the one you can read on the arm that is supporting the camera (I am calling it the upper arm).
I don't know if I am clear enough.


thx. that makes sense. what does 130 deg refer to?


Nick

YvesG
12-18-2009, 04:19 AM
The second angle is linked to the horizontal rotation.
That mean you will have to take a picture every 130?.
Last information under this angle, is the number of pictures to be taken.

nick fan
12-18-2009, 06:19 AM
The second angle is linked to the horizontal rotation.
That mean you will have to take a picture every 130?.
Last information under this angle, is the number of pictures to be taken.


I am lost. Take a picture every 130 deg for 17mm eq?
Besides why are the results different in French and English spreadsheets?


nick

YvesG
12-18-2009, 08:46 AM
There has been a bug when I translated the file.
To have same results, you have to select again the camera position.
Put the camera in vertical position and you will see that the final result is updated then.
Sorry for that, I am immediately updating the file to be downloaded.
Thank you.

nick fan
12-19-2009, 02:42 AM
There has been a bug when I translated the file.
To have same results, you have to select again the camera position.
Put the camera in vertical position and you will see that the final result is updated then.
Sorry for that, I am immediately updating the file to be downloaded.
Thank you.



Hi YvesG,

Puzzle solved. Is there a way to make a drop down list of sensor size? 35mm equivalent only works for 3:2 format sensor. It does not work for 4:3 or 16:9 sesnor.

Great Job!

Nick

YvesG
12-19-2009, 10:49 AM
In fact as far as I understood the crop factor is the ration between 36mm and the horizontal dimension of the sensor.
I guess I can make the sensor dimension to be visible and possible to be documented, so that the focal length written will not have to be recalculated.
What do you think about this approach ?

YvesG
12-19-2009, 10:55 AM
I just have verified the formulas and it possible to make the modification in the file to fit any camera.
I will just add the horizontal and vertical sensor size input and remove the comment on focal length.
That means, after documenting the sensor size, you will just have to tell what focal length you use, without any crop factor.

I am going to do that immediately and post the new file.

John Houghton
12-19-2009, 11:02 AM
In fact as far as I understood the crop factor is the ration between 36mm and the horizontal dimension of the sensor.
The crop factor is usually taken to be the ratio of the diagonal sizes.

John

YvesG
12-19-2009, 11:45 AM
No problem.
In fact with the formulas I use, I just need the sensor size now.
I have modified the file, so that you can document it freely and it is now on line.

Panorama calculator for all (http://photos.yves.over-blog.com/article-panorama-calculator-for-all-41476199.html)

I hope it will work correctly.
:biggrin:

YvesG
12-19-2009, 05:01 PM
The crop factor is usually taken to be the ratio of the diagonal sizes.

John


I have read my book again about photographic mathematics and I confirm that :
the crop factor is the ratio between 36mm (argentic) and the larger size (width) of the sensor, not the diagonal.

YvesG
12-19-2009, 05:29 PM
And here is what I found that puts me back to the beginning of the way to calculate crop factor :

Some camera companies use the diagonal field of view measurement method. That is, if they describe a camera as having a particular 35mm equivalent focal length, it means the camera produces the same field of view along the diagonal as a 35mm camera with a lens having the stated focal length.

Other companies use the horizontal field of view rather than the diagonal field of view. This produces a slightly different result from using the diagonal because the aspect ratio of the digital image (usually 4:3) is different from the aspect ratio of 35mm film images (exactly 3:2). When a digital camera produces the same horizontal field of view as a 35mm camera with a lens having a particular focal length it actually has a larger diagonal field of view.




At the end, it is not an issue as formulas are based on sensor size and not the crop factor.
But I am still questionning myself on what is the official definition (not wikipedia).

nick fan
12-19-2009, 07:39 PM
No problem.
In fact with the formulas I use, I just need the sensor size now.
I have modified the file, so that you can document it freely and it is now on line.

Panorama calculator for all (http://photos.yves.over-blog.com/article-panorama-calculator-for-all-41476199.html)

I hope it will work correctly.
:biggrin:


great thanks.

Nick

YvesG
12-20-2009, 04:11 AM
Concerning crop factor :

For a sensor that has the same format ratio than 24x36, you can use either the diagonal or the width of the sensor to calculate this ratio. It is giving the same result.
For other formats, you have to consider the width and not the diagonal.

So, as a general rule, to calculate the crop ratio, you should apply : Crop ratio = 36 / Sensor width(mm)

YvesG
12-20-2009, 10:08 AM
I have just reworked the results presentation area in order to simplify the reading.
I hope it will bring something and that you will like it.

http://idata.over-blog.com/2/52/89/18/Articles1/Pano-calc-low.jpg (http://photos.yves.over-blog.com/article-panorama-calculator-results-new-presentation--41528748.html)

nick fan
12-20-2009, 10:27 AM
I have just reworked the results presentation area in order to simplify the reading.
I hope it will bring something and that you will like it.

http://idata.over-blog.com/2/52/89/18/Articles1/Pano-calc-low.jpg (http://photos.yves.over-blog.com/article-panorama-calculator-results-new-presentation--41528748.html)



That is a nice tweak!
THx

Nick