Nikon lens 14mm or 14-24m

[Rob C]

I got no responses to my test image?

I just purchased a

Nikon D700
RRS Ultimate-Pro Omni-Pivot Package

and still need to see what lens works. I posted a test with the 14-24

sample 360

Sorry to be a pain, but what am I supposed to be seeing? All I have is a rather poor quality wide-angle image.

Rob C

[BernardLanguillier]

I got no responses to my test image?

A few small issues:

- There appears to be a stitching error in the upper part of the colored wall.
- Flare is a problem around the windows.
- zenit needs to be shot as well

Cheers,
Bernard

[pixjohn]

I am not sure what your seeing? the mov worked for Bernard. The image is just a test.

Sorry to be a pain, but what am I supposed to be seeing? All I have is a rather poor quality wide-angle image.

Rob C

I saw the stitching error after I uploaded the mov. As stated above the mov is just a test and they do not need the zenit. I will do a new test with the D700 and shoot it to play with the image.

A few small issues:

- There appears to be a stitching error in the upper part of the colored wall.
- Flare is a problem around the windows.
- zenit needs to be shot as well

Cheers,
Bernard

[OldRoy]

I am not sure what your seeing? the mov worked for Bernard. The image is just a test.

I saw the stitching error after I uploaded the mov. As stated above the mov is just a test and they do not need the zenit. I will do a new test with the D700 and shoot it to play with the image.

Hi again
I just came back to this thread and I'm having trouble recalling what it was about... But, and it's a big but, I seem to remember pointing out that the D700/14-24 combination (which I have) is, emphatically, total overkill for VR panos, indeed completely unsuitable.

The posted sample test is supposed to demonstrate what exactly? I don't mean to be impolite, but it just shows a badly stitched (at least two errors), badly exposed incomplete "VR" (this sort of shot - which I've done ****loads of times needs HDR: I've never tried using lighting for VR panos and don't intend to try). It's actually a cylindrical 360.

As far as I'm concerned doing a VR pano well requires a fully stitched 360 x 180 including zenith and nadir. It may be expedient to exclude the latter and incorporate a mirror-ball effect with advertising content, but if you want to do VRs properly, you need a setup that works with all the elements in place. In the case of the rig in question, this means at least two rows of six shots. Time consuming in PP and generating a huge file, even at 8 bit. For final web display as a Flash or QT file at (usually) <1Mb...

I haven't used the lens, but I'd guess that a 16mm Nikkor FE would be the ideal for use on an FF body. Personally I use a 10.5 FE which has been shaved so it can be used on the D700 as well as my D200, which is my preferred body for this purpose. I shoot it at base ISO and bracket +/- 3 stops, then use Enfuse for HDR of 3 selected exposures. I've used the lens on the D700 as a circular FE but on balance prefer to shoot with the D200 (6 round + zenith + 3 x nadirs; 70 shots in all - that's enough work for me, thanks) which results in a final equirectangular 8 bit tif @ just over 200 Mb after a few tweaks in PS.

Is anyone using the 14-24 for fully stitched VR panos? I'd be interested to know what the advantages are.

Short version for all this waffle is: VR panos require a FISHEYE!

Roy

[pixjohn]

I purchased the nikon 10.5 and starting over testing the lens.

[OldRoy]

I purchased the nikon 10.5 and starting over testing the lens.

For non critical use I convert with ACR, correcting vignetting (even at F8) and ca (approx +30 / -30 but I can't recall which way round) but the best results are from NX2 (switch off defishing!) although it takes a little longer.

Have you had the lenshood shaved, or are you using the D700 in DX mode?

Good luck.
Roy

[pixjohn]

The brand new lens met a hacksaw last night  So now its a shaved 10.5

[OldRoy]

The brand new lens met a hacksaw last night  So now its a shaved 10.5

More nerve than me. I got it done by someone with a lathe (and the skills to turn it).

[Stephane Desnault]

When you're trying to print or make really high-res panos, the 14-24 really gets you the best results, definitely recognizable over some other combinations. I do agree though that for basic QTVRs with the cube faces at 1024x1024, it won't make much of a difference.

On the other hand, when you're also using your panos to print and sell posters like I do...

[pixjohn]

It took 15 min to make the cut. It was easy

The nikon is not my main system, If I want high end files I use my Leaf Aptus 75 on a cambo wide DS or H2

When you're trying to print or make really high-res panos, the 14-24 really gets you the best results, definitely recognizable over some other combinations. I do agree though that for basic QTVRs with the cube faces at 1024x1024, it won't make much of a difference.

On the other hand, when you're also using your panos to print and sell posters like I do...

[Stephane Desnault]

It took 15 min to make the cut. It was easy

The nikon is not my main system, If I want high end files I use my Leaf Aptus 75 on a cambo wide DS or H2

The 10.5 is a great lens too  - I was replying to the other poster who seemed to think that the 14-24 was an unworkable solution for some reason.

[OldRoy]

The 10.5 is a great lens too  - I was replying to the other poster who seemed to think that the 14-24 was an unworkable solution for some reason.

And I still think it is.

You say elsewhere that you print, and even sell, prints. From VR 360 x 180s? Well oddly enough so do I, sometimes. Unaccountably some people seem to like the wierd perspective of an equirectangular projection (I actually crop them a bit for prints.) But we were talking about VR panos, weren't we? For which the whole raison d'etre is the web.

I stand by my original assertion that it's an unsuitable lens. It requires two rows of six shots at minimum, and if you do viewpoint correction nadir patching (I do) another three shots. Total, 15. And you get diagonally linked nuisance control points created if you use auto cp generation. Its sheer bulk and weight makes big demands on the pano head too. Plus with a zoom there's always a chance of accidentally altering the fl whilst working. File sizes are huge too.

I have never heard of anyone else routinely using a 14-24 for VR pano creation Stephane. The full frame 16mm fisheye would be the lens of choice, I'd have thought, for optimal resolution, manageable, 360 full-frame panos - not a shaved 10.5 (which, again, I have) - it leaves unused a large proportion of the sensor (or requires DX mode...) Of course there are people out there who like to create huge resolution panos using longer focal lengths and good luck to them. You're obviously in this category. But 360 x 180s for web use?

I think it would be very bad advice to recommend this huge, expensive, inappropriate lens to someone inexperienced in creating VR panos primarily for web use.

Roy

[Stephane Desnault]

Hi OldRoy,

I think we agree, I'm just a bit less sanguine in the way I express my view .

I certainly wouldn't recommend the 14-24 as a "first lens" for aspiring panoramists. I successively used a D50 with the Sigma 10-20mm, then a D80 with either the Sigma or the 10.5, and then switched to a D3 with the 14-24 (I bought the D3 before the D700 was announced).

And as I progressed, each combination was very definitely "good enough" for what I was doing. I think I got lucky with my instance of the Sigma 10-20: it was just razor sharp on my D80, much more than the reviews led me to hope for. In the end though, it's the 14-24 + D3 that brings out the most details in the image.

I haven't tried the 16mm yet. I certainly will, knowing that it's the lens of choice for high end 360 automated gear (the $30k rotating thingies with a sensor that is only a narrow strip) - so it obviously qualifies for the precision. I'm just waiting for a good second hand one to cross my path.

You can see a small sample set of my panos at http://www.360cities.net/search/desnault - the 10 first ones are shot with than 14-24. I think the difference of quality is visible, even on the web.

One last thing: I process all lens distorsion and vignetting with DXO before feeding my pics to the stitcher (except when using a fisheye of course), and that way I avoid some vexing issues with both the Sigma and the 14-24.