Nadir Patching Techniques Discussion

[Marcus]

A question about Nadir patching techniques.

Using a Nikkor 10.5 fisheye lens, I have been using these 2 techniques
(brief description);

Technique One: Remove the camera from tripod and shoot a handheld
nadir. Patch in using Alpha masks and viewpoint correction.

Technique 2: Move the tripod aside and shoot another image (lets call
it the patch) which includes the down area obscured by Pan Head.
Defish, select an area and patch it into the bottom cubic face using
transform tools.

Question is, does anyone combine these 2 methods by shooting another
(patch) image (from the tripod, moved a meter plus to the side as in
Technique One), then using Alpha Masks and Viewpoint Correction (as in
Technique 2)? This would seem to be a much better method for shooting
HDR panoramas and removes the risk element of handheld shots.

If you do, can you tell me if it is best to defish the patch image
prior to applying the masks.

Any other advice most welcome as I'm still trying to groove a
foolproof method. I've never liked the 'tilted tripod' route and as I
mentioned, handheld is no good for HDR. So both these methods will do
the job but I think I prefer applying Alpha Masks to fiddling with the
Transform tool.

What do you all think?

Thanks,

Marcus

[Erik Krause]

Marcus wrote:

> Technique One: Remove the camera from tripod and shoot a handheld
> nadir. Patch in using Alpha masks and viewpoint correction.
>
> Technique 2: Move the tripod aside and shoot another image (lets call
> it the patch) which includes the down area obscured by Pan Head.
> Defish, select an area and patch it into the bottom cubic face using
> transform tools.
>
> Question is, does anyone combine these 2 methods by shooting another
> (patch) image (from the tripod, moved a meter plus to the side as in
> Technique One), then using Alpha Masks and Viewpoint Correction (as in
> Technique 2)? This would seem to be a much better method for shooting
> HDR panoramas and removes the risk element of handheld shots.
>
> If you do, can you tell me if it is best to defish the patch image
> prior to applying the masks.

I never defish. PTGui does this perfectly so why should it be necessary?
Patching technique depends on the floor structure. If it is even I
simply move the tripod aside a bit and shoot tilted down. In PTGui I use
viewpoint correction (after all images are aligned as good as possible
without viewpoint correction) and blend priority to use only the
necessary parts.

If the floor is uneven I try to shoot handheld from the same position as
the tripod shots where. In PTGui I use the same workflow as above.
Sometimes I need to remove my feet in photoshop afterwards (I always
output blended and layers PSB) - either by simply editing masks or (if
this isn't possible) some use of the patch tool or the duplicate layer,
mask upper and move lower layer technique on the flattened image.

With viewpoint correction and blend priority it isn't really necessary
to patch the tripod in photoshop any more...

--
Erik Krause
http://www.erik-krause.de

[John Houghton]

I use method 1, regardless of whether the nadir shot was handheld or
taken from a shifted tripod. There isn't really any difference. I do
sometimes output the nadir in a separate equirectangular image (after
optimization with VP correction together with all the other images).
I then extract identical rectilinear views (usually by PTEditor or
lately Pano2VR) for merging by hand in Photoshop. I never defish
images outside of PTGui.

This one includes 4 offset nadir shots, all with masks to eliminate
shadows from multiple light sources, stitched in with VP correction in
one PTGui run:

http://www.johnhpanos.com/ely40d.htm

John

[Keith Martin]

>(I always output blended and layers PSB)

I used to do this. Now I'm more likely to output as layered TIFF. The
upside of this is that each layer is generated as a separate file -
so I can open the flat blended image and just one image layer at a
time, if I need to blend things manually.

Of course, my setup doesn't come too close to TIFF's limits; I get
12kx6k equirects, which can approach 500MB as flat 16-bit images.

k

[Erik Krause]

Keith Martin wrote:

> >(I always output blended and layers PSB)
>
> I used to do this. Now I'm more likely to output as layered TIFF. The
> upside of this is that each layer is generated as a separate file -

you know that I have a Photoshop action to quickly load a TIFF with
alpha into a masked layer on http://erik-krause.de/ttt ?

> so I can open the flat blended image and just one image layer at a
> time, if I need to blend things manually.

That's a clever way, too. Especially since PTGui does a really great job
in blending nowadays and it's seldom necessary to manually blend in more
than one layer or two.

However, I like the possibility to switch on and off single layers
quickly in order to see which one contributes the relevant details. On
the other hand this can be done in PTGui pano editor as well...

[Keith Martin]

>you know that I have a Photoshop action [...]

Heh. I had no idea, but this could be rather useful. Thanks!

>Especially since PTGui does a really great job
>in blending nowadays and it's seldom necessary to manually blend in more
>than one layer or two.

Absolutely. What helps too is holding down the Shift key while
dragging the patch image into the full equirect image. Because
they're both the same pixel dimensions (even though the patch is
mostly 'empty') this makes it arrive at precisely the right location,
no nudging needed.

But now I need to try your action!

k

[1drey]

Marcus,

I am one of a few who uses T2 method.
I found it more convinient in situaitions when the nadir shot doesn't
have enough details for the placing of control points. Another
situation when patching technique is referable - when nadir image is
produced with rectilinear lens (necessary in some cases).

Andrey

[Marcus]

Thanks to everyone for the replies.I think I will do some further
experimenting with Blend Priority. This may be the answer to
successful Viewpoint Correction.

Marcus