Solution to the Quiz

1) Both panoramas were done with the very same set, each of 8 "virtual" images (2000 x 2000 pixels) as input to PTGui 5 on Win XP. (BTW: It could also have been alternatively made with PTMac or HuginOSX).

Notes about the "virtual lens":

These 8 images (Zenith+ 6 horizontal + Nadir) were first "extracted" from an existing equirectangular image and radially "lens blurred" under a radial gradient mask in PS CS.

This simulates digitally (but the softness near the edge is willingly somewhat exaggerated) a 200 x 200 degrees (!) HFOV FF Fisheye image.

This "blur gradual defect" is the only one affecting the image. The Virtual Fisheye Lens/ Sensor combo is "Perfect" and therefore, of course, unrealistic: a= b= c= d= e= 0 and there is absolutely no light fall-off or vignetting. No color adjustment is necessary. The shots were also perfectly positioned on a Perfect virtual Tripod and Spherical Panohead. The angular distance between shots are exact and equal.

2) Both panoramas were made with the same basic project script in PTGui 5.

Only the HFOV (and correlated Cropping dimensions in PORTRAIT mode) were changed from one to the second:

88 degrees for the first panorama Vs 200 degrees (No crop) for the second panorama. Both for a FullFrame (rectangular) Fisheye.

Notes:

88 degrees is the HFOV of a Nikkor FE 10,5 mm clipped by an APS-C sensor.

200 degrees is said to be the circular VFOV of this same lens when it is not clipped by the sensor size and deprived of its sun shield. The HFOV is said to be about 140 degrees on a 24 x 36 mm sensor.

The popular Sigma 8 mm EX has a circular FOV of about 180 degrees. It is clipped to 113-118 degrees HFOV by APS-C sensors.

3) Both panoramas were blended with Smartblend and the Output setting (Create Panorama Tab) were exactly the same, except the file name, of course.

There is no other factor contributing to the difference than the Crop dimension (and the associated "v" setting in the "Lens Settings" Tab Windows).

Therefore the two rendered output Panoramas should have been exactly identical. The crop size should not change anything on this type of panorama/

They are obviously not identical and this is visible thanks to the very large HFOV of the source images that amplifies the phenomenon.

Notes:

Enblend produces seemingly the same output as does Smartblend i.e. different output panoramas.

Both output panoramas are (apparently) identical with PTGui 5 internal blender engine.

Conclusion:

One could naively attribute the much better aspect of the first panorama to an assumed better sharpness and contrast of the "first" lens (that has a smaller HFOV) and correlatively to the softness of the "second" (with a much larger HFOV). This erroneous statement is commonly repeated on some forums.

It's possible to personally experiment and verify all the above statements (amongst many other things): To do so, you may go to this page.

Back to first page of the Quiz