On 29 Dez. 2011, 00:45, Monkey wrote:
> Hi all,
>
> I've developed an alternative to Enblend which I hope might find some
> interested users here. It's not a better blender, and it lacks many of
> Enblend's features, but it has two redeeming qualities
> ...
> It should work straight out of the box
2012/1/3 kfj <_...@yahoo.com>
> With this problem, plus the missing wraparound (which is bad but much
> less grave), plus the necessity of manually changing the --bgr
> depending on source material - I'd recommend to use multiblend with
> caution. I think it shouldn't be put into hugin bundles unt
A little more info on the bug Kay found:
It happens when an area of the mosaic is covered by two or more
images, and the particular layout of images around that area means
that multiblend decides that an image which doesn't actually cover
that area is the correct choice. I think (looking at the PT
On Jan 2, 7:51 pm, "Carlos Eduardo G. Carvalho (Cartola)"
wrote:
> Well, I don't know how to fix the CP so you get good CPs, but here I prefer
> to put them manually. If you choose to try this I suggest that you can do
> only 2 CPs between each pair. Try to put them far from each other. I
> usua
On 3 Jan., 10:16, Harry van der Wolf wrote:
> Does this issue persist when using reduced size images resulting in a
> smaller pano (as also enblend sometimes has some (other) glitches with big
> panos)?
If you think that 9 images of 12 MP each are big, what would you
consider small enough to jus
On 3 Jan., 12:08, Gerhard Killesreiter
wrote:
> I've tried adding CPs manually but the matching process wasn't really
> working.
Do you mean that you had difficulties placing the control points, or
that the optimizer failed to produce a useful result?
> My concern is that I either used the wron
I suggest you make your photos and the pto online so we can try to help.
Cheers,
Carlos E G Carvalho (Cartola)
http://cartola.org/360
2012/1/3 kfj <_...@yahoo.com>
> On 3 Jan., 12:08, Gerhard Killesreiter
> wrote:
>
> > I've tried adding CPs manually but the matching process wasn't really
>
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Am 03.01.2012 12:27, schrieb kfj:
> On 3 Jan., 12:08, Gerhard Killesreiter
> wrote:
>
>> I've tried adding CPs manually but the matching process wasn't really
>> working.
>
> Do you mean that you had difficulties placing the control points, or
> tha
Hi,
well, maybe someone else can help with the automated CPs finding, because I
really dont use it. I usually put them manually as I also has never liked
the automated results.
What I can see is that I have reduced the biggest CP distance to 14 doing a
"everything without translation" optimizatio
David,
I am curious of the algorithm you use to determine the seams. The code is well
organized but not well documented.
Does Multiblend look at the masks of the input images, or just the rectangular
bounding box the image? Once it has been determined that a pixel should come
from a particul
Kay,
why on earth would one want to use the Windows version of multiblend with
hugin on Linux?
David updated his website, the Linux source is available since this weekend:
http://horman.net/multiblend
Just do
g++ -msse2 -O2 multiblend.cpp -ltiff -ltiffxx -o multiblend
Then as root copy the new
Jim Waters wrote:
> I am curious of the algorithm you use to determine the seams. The code is well
> organized but not well documented.
I think it's pretty generous to even call it organized, but thanks ;)
> Does Multiblend look at the masks of the input images, or just the rectangular
> boundin
On 3 Jan., 16:53, Jan Martin wrote:
> Kay,
>
> why on earth would one want to use the Windows version of multiblend with
> hugin on Linux?
I thought it might be easier for the author initially to only get it
to run properly on one platform. I think this helps to narrow down
problems more quickly,
2012/1/3 Monkey
>
> Is there someone who can change the title of this discussion, as
> multiblend is no longer Windows only?
>
> David
>
>
This is a mailing list, not a forum. Just start a new mail (and somewhere
in your mail refer to this mail thread if necessary <
http://groups.google.com/group
On Jan 3, 5:31 pm, kfj <_...@yahoo.com> wrote:
> And with the native Linux version, I need
> the --bgr switch as well.
Do you mean that you *didn't* need the switch on Windows, with the
same images?
> I'd also like a fast alternative to enblend. So, David, don't be
> disheartened. I'm sure it's
Hi,
sorry for the delay :) I changed the subject.
I compiled it on FreeBSD. Got the source code today on the site. I had to
convert the file write.cpp to unix format in order to avoid g++ complaining
about a missing end of line at the end of the file and I had to use this
line to compile:
g++ -D
On 3 Jan., 18:18, Monkey wrote:
> > The two nadir images don't have the exact yrp, but even so, preference has
> > to be given to one
> > image over the other.
>
> Ideally yes, but the choice may have to be arbitrary (as it would be
> in Enblend, since it would be determined by input order).
pr
Hi Carlos,
2012/1/3 Carlos Eduardo G. Carvalho (Cartola)
> Hi,
>
> sorry for the delay :) I changed the subject.
>
> I compiled it on FreeBSD. Got the source code today on the site. I had to
> convert the file write.cpp to unix format in order to avoid g++ complaining
> about a missing end of li
On 3 Jan., 19:59, "Carlos Eduardo G. Carvalho (Cartola)"
wrote:
> I already didn't found the smartblend site, for
> example... :(
no idea where the site is, but you can download the Windows executable
from
http://wiki.panotools.org/wiki/images/3/37/Smartblend_1_2_5.zip
no source available, as
Thanks Kay, I thought I tried to access smartblend site, but I made a
mistake, I was clicking in an image example mentioned in
http://wiki.panotools.org/SmartBlend and got a "Not Found". That's what
happens when you're in a hurry... mistakes :)
Carlos E G Carvalho (Cartola)
http://cartola.org/360
-O3 causes a segfault on Linux in the same place - I wonder if it has
something do with optimising out some instructions which don't appear
to be used. In the mask section I make use of int* and float* pointing
to the same address, and it seemed to be around this point that
multiblend was failing.
Am 03.01.2012 19:59, schrieb Carlos Eduardo G. Carvalho (Cartola):
I already didn't found the smartblend site, for example... :(
There is none. The panotools wiki page is all that exists. smartblend is
closed source and out of development, which is a pity, since it is the
second best blender
Thanks Harry, it worked!
But the resulting image was not good...
Here is the multiblend generated image (reduced and converted to JPG)
http://cartola.org/arquivos/multiblend.jpg
and here the enblend version made with the same source remapped images
http://cartola.org/arquivos/enblend.jpg
The te
Hi Carlos,
Looks like you've found another example of the overlap problem - I
don't quite call it a bug, since the implementation of the seaming
algorithm is correct, but the algorithm itself obviously has failure
cases that I didn't anticipate. It occurs when images overlap in
certain ways that e
2012/1/3 Monkey
> -O3 causes a segfault on Linux in the same place - I wonder if it has
> something do with optimising out some instructions which don't appear
> to be used. In the mask section I make use of int* and float* pointing
> to the same address, and it seemed to be around this point tha
Thanks Erik! I just don't know if I understood correctly: the better blend
of all don't do 360º blending? Well... never mind, I don't have photoshop...
Thanks!
Carlos E G Carvalho (Cartola)
http://cartola.org/360
2012/1/3 Erik Krause
> Am 03.01.2012 19:59, schrieb Carlos Eduardo G. Carvalho
Hi guys! I'm trying to use hugin 2011.4.0 to generate a panorama.
In the assistent tab, I load the images, I align them, and then I
press the "create panorama" button: hugin asks me the project file
name, then the output prefix, and then nothing happens!
With previous hugin versions the creat
Hi guys!
I have the photos of a church facade, I've uploaded them zipped at
http://owncloud.cathopedia.org/apps/files_sharing/get.php?token=0e3e4716359456f31d428fce66f39d0416079353
, but after aligning them I'm not able to choose a projection which
could satisfy me.
Anyone could help me, download
Hi David,
I used 10 images to do this pano, as I do most of the times. I make an
overlap of about 50% of the images or even more, because usually I could do
it with half of the images at the horizontal line. At the horizon I usually
make 8 images and could do 4. I usually prefer to have more infor
Well, if you have aligned them you could have sent the PTO file together :)
Anyway, I don't have time to mount them now (have to put children in bed)
but I suggest you do a rectilinear or an equirectangular projection and
crop the final image. You can adjust its position in the fast preview
windo
On Tuesday, 3 January 2012 at 18:53:46 -0200, Carlos Eduardo G. Carvalho
(Cartola) wrote:
> 2012/1/3 Monkey
>
>> -O3 causes a segfault on Linux in the same place - I wonder if it has
>> something do with optimising out some instructions which don't appear
>> to be used. In the mask section I mak
On Tue, 3 Jan 2012 11:45:21 -0800 (PST), Monkey wrote:
> -O3 causes a segfault on Linux in the same place - I wonder if it has
> something do with optimising out some instructions which don't appear
> to be used. In the mask section I make use of int* and float* pointing
> to the same address, and
Am 03.01.2012 22:05, schrieb Carlos Eduardo G. Carvalho (Cartola):
I just don't know if I understood correctly: the better blend of all
don't do 360º blending?
At least photoshop CS3 autoblend didn't. No idea whether CS5 does. I
don't have it either...
--
Erik Krause
http://www.erik-krause.d
33 matches
Mail list logo
