PTGui 7.7 beta 1
PTGui Support
A beta version of PTGui Pro 7.7 is available! This includes exposure
fusing, the same algorithm used by Enfuse and Tufuse.
The process is simple: create an HDR panorama as usual, but choose
'Method: exposure fusing' instead of 'True HDR'. A preview can be
generated and tweaked just like with tone mapping. The fusing parameters:
- Target Brightness: PTGui will choose the pixels with a brightness as
close as possible to this value
- Boost Shadows / Reduce Highlights: if nonzero, PTGui adds extra
'virtual' blend planes with reduced or increased brightness. Can be used
if the shadows are too dark even in the brightest exposure.
- Sigma: this is the sigma parameter in the fusing algorithm. It can be
seen as a selectivity setting: if sigma is nearly zero, the algorithm
will choose only the best exposed pixels from each layer. If sigma is
greater, the algorithm more gradually merges different blend planes. At
values close to zero, the effect can be a bit unpredictable.
Also, PTGui Pro is now available in a native 64 bit Windows version.
It's a bit faster (quick tests show about 20% speed increase) compared
to the 32 bit version.
Joost
Atomicmak
i am just about to produce one enfused output and will update once i
get it done.
this is my first post to group and i am really excited to see the
result of enfuse as i have manually enfused 7 brackets and i found
that enfuse is fast better than regular HDR methods.
so realisting tones.
keep up joost.
regards
mak
www.vox360.in
Miso Ristov
cheers,
m.
Sam Kittner
Secondly, forgive my ignorance, but can someone explain or point to a
link to a description of the differences of a "true hdr" vs. exposure
fusing?
I am curious as to not only the technical differences but general
expectations of how similar exposures would look run through each
process.
Many Thanks,
Sam
<http://www.kittner.com/panoramas>
Erik Krause
> Secondly, forgive my ignorance, but can someone explain or point to a
> link to a description of the differences of a "true hdr" vs. exposure
> fusing?
See http://wiki.panotools.org/Enfuse for a brief description or
follow the link to the mentioned paper for full technical details of
exposure fusion.
> I am curious as to not only the technical differences but general
> expectations of how similar exposures would look run through each
> process.
"True HDR" might or might not yield very similar results. The major
difference is, that exposure fusion uses the "best" pixels from each
image and blends them together much the same like enblend or PTGui
blender blend the overlap.
HDR first merges several images into a format where there is no
limitation to brightness values (they can reach from 0 to infinity).
You can use this HDR images directly f.e. for image based lighting or
to view on an HDR monitor and you can tonemap them back into viewable
(limited) brightness range by using several different tonemappers
which yield more or less natural looking results. See
http://wiki.panotools.org/HDR for more info.
The beauty of exposure fusion is that it always yields relatively
natural looking results, even without user interaction. Hence it
should be possible to avoid the "Fuse settings" step in PTGui at all
or adjust once and then use this fixed settings.
best regards
Erik Krause
http://www.erik-krause.de
Erik Krause
> A beta version of PTGui Pro 7.7 is available! This includes exposure
> fusing, the same algorithm used by Enfuse and Tufuse.
...but why do you first stitch then fuse? Did you solve the zenith
and nadir vortex issue in the fusion algorithm?
At least if the images are linked the fusing beforehand doesn't have
any disadvantages...
Atomicmak
i have question for you.
What would be the best and perfect method to use Enfuse with/without
ptGUI ?
let's for example
I shoot 6+1(zenith)
7 brackets of each shot
convert raw to tiff in photoshop
load all 49 shots in ptGUI and stitch them and give output with LDR
Blend Plane so i get all 7 spherical
HDR them in photomatix/photoshop (that adds colour cast and have to re-
correct them in photoshop OR lightroom)
Finally i produce different size tiffs from that final tiff and give
web output.
NOW if suppose i plan to use Enfuse what will be the good way to do
the work in less time ?
my systems are pretty good but the current working task taking so much
time the way from RAW to WEB output.
Please suggest
Regards
MAK
www.vox360.in
Erik Krause
> NOW if suppose i plan to use Enfuse what will be the good way to do
> the work in less time ?
> my systems are pretty good but the current working task taking so much
> time the way from RAW to WEB output
enfuse any bracketed series using same paramters for each and stitch
the result images. This way you avoid any zenith vortex problems and
handling of large images is reduced to a minimum.
Atomicmak
i was thinking the same way as you suggested as in most interior cases
i can use this process to make single 7 shot pano of enfused images
from all 7*7=49 images.
the only thing i am thinking is exterior in public areas as i am
taking 7 brackets people may move here and there and transition will
not come good in hdr OR enfused images and i may have to mask them
from good exposed spherical to enfused/hdred image so for that kind of
pano i think i have to stick to my old method of making 7 spherical
and maskout ghosted parts by enfuse/hdred image.
thanks a lot
will wait for morning to get to my office pc and check the new method.
MAK
www.vox360.in
Erik Krause
> the only thing i am thinking is exterior in public areas as i am
> taking 7 brackets people may move here and there and transition will
> not come good in hdr OR enfused images and i may have to mask them
> from good exposed spherical to enfused/hdred image so for that kind of
> pano i think i have to stick to my old method of making 7 spherical
> and maskout ghosted parts by enfuse/hdred image.
You can always apply the project file as a template to a single set
of images and choose those for output you actually need.
If f.e. the people who moved are in one exposure step load all images
of this step into PTGui, apply the template, in create panorama tab
disable the images you don't need and create (best as non-blended
layered PSD). This way you get exactly what you want with minimum
computing effort.
MacUser
You really did a great job with PTGuiFuser.
Perhaps I need to get more experienced in using PTGuiFuser but
whatever settings I tried the output is always giving poor details in
the dark areas.
The dark areas are also darker and wider then when using Enfuse.
I compared the PTGui blended planes output, processed with XFuse 0.5
(default Enfuse settings), with the direct output of PTGUiFuser and
the differences in the dark areas are huge. I tried the default
PTGuiFuse settings and I tried with some "boosted darks" settings.
Q) Is it possible to get the same results in PTGuiFuser compared to
Enfuse ?
Best,
Wim.
Hans
On Mar 11, 8:37 pm, MacUser <mvo...@gmail.com> wrote:
> Hello Joost,
>
> You really did a great job with PTGuiFuser.
>
> Perhaps I need to get more experienced in using PTGuiFuser but
> whatever settings I tried the output is always giving poor details in
> the dark areas.
> The dark areas are also darker and wider then when using Enfuse.
> I compared the PTGui blended planes output, processed with XFuse 0.5
> (default Enfuse settings), with the direct output of PTGUiFuser and
> the differences in the dark areas are huge. I tried the default
> PTGuiFuse settings and I tried with some "boosted darks" settings.
the preview looks great.
I tried both a full 11500 pixels which took 1,5 hours and I got this
result http://www.panoramas.dk/test/panofused.jpg
Tried again with a small 2000 pixels pano with same result
Preview is perfect http://www.panoramas.dk/test/ptgui-fusepreview.jpg
and a HDR output is acceptable after photoshop correction.
Hans
MacUser
When saving a project the customized settings of PTGuiFuser are not
stored and are disappeared after re-opening the project.
After a lot of fiddling with the sliders of PTGuiFuser (especially the
Sigma setting) I can get closer to the results of Enfuse but I can't
get the same results.
The preview on my machine (MacBookPro+OSX Leopard) is much lighter
then the fused output so for me it is hard, if not impossible, to
predict the output based on the preview.
I am confident that all of this will eventually be solved in next
betas so in the end we will have a valuable tool for fusing bracketed
shot panos and that is great.
Best,
Wim.
MacUser
I did some tests with extreme PTGuiFuser settings and when comparing
the output they all looked the same for me......
MacUser
Found the reason why the output looks all the same despite different
and extreme PTGuiFuser settings.
As it turns out the Batch Stitcher isn't capable of handling
customized PTGuiFuser settings and uses the default settings instead.
When using the Create Panarama button all PTGuiFuser output is fine
and then the output looks the same as the preview window.
Best,
Wim
PTGui Support
> ...but why do you first stitch then fuse? Did you solve the zenith
> and nadir vortex issue in the fusion algorithm?
>
> At least if the images are linked the fusing beforehand doesn't have
> any disadvantages...
In theory the results of fusing part of an image are influenced by high
contrast areas in the neighbourhood (this is due to the low pass
filtering in the algorithm). If such a high contrast area is visible in
only one of the two bracketed sets there would be brightness differences
in the fused images.
In practise it's probably not very bad, since the blender will equalize
this.
Anyhow, being able to do HDR/fuse first and then blend is definately
something I plan for a future version but it's a significant rewrite of
the code.
Joost
Atomicmak
http://www.vox360.in/mak/neemrana_fort.mov
http://www.vox360.in/mak/neemrana_fort2.mov
regards
MAK
www.vox360.in
Chris Thomas
Appears very smooth and "real" on my calibrated display.
Do you think 7 brackets are necessary or maybe a little "overkill"?
Are you bracketing 1 or 2 stops?
Cheers
chris
Chris Thomas
Photographer
cell... 403-615-1212
In North America
call... 1-800-870-5110
http://www.christhomas.com
-----Original Message-----
From: pt...@googlegroups.com [mailto:pt...@googlegroups.com] On Behalf Of
Atomicmak
Sent: Thursday, March 13, 2008 3:45 AM
To: PTGui
Subject: [PTGui] Re: PTGui 7.7 beta 1
here are my two first results of ptgui fused stitch from 7 brackets.
http://www.vox360.in/mak/neemrana_fort.mov
http://www.vox360.in/mak/neemrana_fort2.mov
regards
MAK
www.vox360.in
[Chris Thomas] snip
Atomicmak
well i do take 7 because sometimes here in india the client may not
satisfy with the 3 dimensional light on photo which is realisting so
we do work on postproduction to enhance darker areas with filling of
light from 7th bracket and lighter area patching with -7th bracket
which helps a lot.
shooting in outdoor with 7 brackets is not much problem just two
clicks more but the experience says that 5 brackets are good to go
with except one need special masking treatments for their job.
hope this helps.
mak
www.vox360.in
On Mar 13, 10:20 pm, Chris Thomas <ch...@albertalocations.com> wrote:
> Very nice work MAK !
> Appears very smooth and "real" on my calibrated display.
>
> Do you think 7 brackets are necessary or maybe a little "overkill"?
> Are you bracketing 1 or 2 stops?
>
> Cheers
> chris
>
> Chris Thomas
> Photographer
> cell... 403-615-1212
> In North America
Erik Krause
> In theory the results of fusing part of an image are influenced by high
> contrast areas in the neighbourhood (this is due to the low pass
> filtering in the algorithm). If such a high contrast area is visible in
> only one of the two bracketed sets there would be brightness differences
> in the fused images.
>
> In practise it's probably not very bad, since the blender will equalize
> this.
In practice the zenith vortex (or cone) is far more of a problem than
the above mentioned effect.
The fusion and the blending algorithm both use a Burt and Adelson
multi resolution blending. If the number of levels for the blending
step will be same or higher than for the fusing step (which certainly
will be the case), won't the effect of different brightness will be
completely removed?