Re: Enblend 3.0 memory problem
- --- In PanoToolsNG@yahoogroups.com, "AYRTON - avi" <avi@...> wrote:
> On 12/1/07, Hans Nyberg <hans@...> wrote:
> > With PTGui you can get a 50% boost in speed by using Enblend 3.0 and with mostly
> > same identical blending result.If you have a new Intel machine it may be much larger than 50%
> So Hans you're saying that it's probably a good idea to use it with my
> ptgui pro instead of the 1.3 or the 2.5 version that I'm using now
Enblend 3.0 for Mac is universal.
Here are some new speedtests I just made.
MacBook Pro 2.16 dualcore 2gb ram
Standard 5400 RPM Harddisc
8 images Canon 5D + 15mm fisheye
Full output 11500x5750
Speedtest at 6000x3000 16 bit output as blended tif.
PTGui Warp + PTGui Blend
10 minutes (average of 3 times)
PTGui Warp + Enblend with 1500 mb Ram applied
4.40 minutes (average of 3 times)
With a 7200 RPM disc both get a 20% boost
The only major problem with Enblend is thelack of blending over the edge at zenith (and
nadir of couse but thats not important.)
This is only a problem with full fisheye 180 degree images without a zenith image.
Or if you do multirow and take the top row images with blending over the zenith.
This has been a known problem all years I have known Enblend. It could without doubt
have been solved but this is an Open source project and nothing happens really.
There is almost no activity around enblend development.
The special 1.3 version which is very fast compared to the 2.5 is slower than 3.0.
I have done tests on the full 11500 pixels output from the same project but only on my G5
PTguiWarp + Blend 31 minutes
PTGui Warp + Enblend 1.3 26 minutes
PTGuiWarp + Enblend 3.0 19.20 minutes
One of the main reason why PTgui is slower is the colorcorrection which is made together
with its own blending. Especially with 16 bit images it increases the warp time with more
When you export to Enblend it does just the warping.
- On Saturday, December 01, 2007 at 21:06, Hans Nyberg wrote:
> This has been a known problem all years I have known Enblend. It couldAs far as I heard Andrew Mihal got a job soon after he released
> without doubt have been solved but this is an Open source project and
> nothing happens really.
version 3.0 and since then only the hugin community felt responsible
for the project. However, remember the story of smartblend? Mike
originally wanted to make it commercially available and no one was
interested: http://www.panotools.org/mailarchive/msg/29296 If he gets
a good job, smartblend will be lost to community, since the source
code is not available, so be grateful enblend is at least open
Could very well be enblend would never have been written if it was
intended as a commercial project. So please instead of to nag about
those lazy open source developers lets go ahead and encourage them to
improve their programs not only by using them but by giving
productive feedback and by supporting their efforts wherever
possible. Especially you, Hans, shouldn't have been missing on the
panotools google summer of code supporters list:
may be next year...
- Enblend can be found at <http://sourceforge.net/projects/enblend/>
Compiled versions for Windows, plus source.
Click "Download" and choose "Browse All Files." Then the page loads,
scroll down a bit.
The newest files in this package are dated January 27, 2007.
MacPorts has version 2.5.
Fink has a version called "3.0-11"
On Dec 1, 2007, at 3:26 AM, dmgalpha wrote:
>> Well Thanks anyway, at least I know the author now.
>> I have to say I find it very weird that the development of a
> component like Enblend is
>> taken care of by just a couple of people who just do it for fun.
>> And it is even difficult for people without insight in it to find
> the download.
> I find it weird that you find it weird. A lot of open source and free
> software is build in the spare time of their developers.
> I spent several hours today trying to compile the latest enblend under
> OS X. I now have an info file for fink and a tar source file with the
> latest CVS if anybody wants to compile it (and has a properly
> configured unstable branch of fink 10.4).
> daniel m. german