My Autopano Pro forum

Thanks for the Mac build. Looks good so far.

Very nice Alexandre.

Congratulation Alexandre,

The Linux version rocks. So far the most stable "of them all",  at least on SuSE10.1! Finally free of Wine and/or Windows. :-)

Thank you, Tomas

awesome!  I'll check it out!  Tomorrow at work I'll give it a few rounds on some fisheyes!  (1Ds2 and the new Sigma 8mm)  I'm excited about that, fisheyes area  quick and easy way for us to capture an environment!  And now that HDR stuff has been fixed, autopano is going to work out great for us!  Great update Alexandre!

A quick question on the image positioning, previously when I would move an image around (an image that could not grab any control points, and currently no way to manually add points), the moved images (that had no CPs) would all snap back to their origin point instead of staying where I moved them when I would re-calculate the stitch (with other images that had new/updated Control Points)  Does the moving around of images still work in this way?  Will 'permanent' moving of the images that stick in their place be a seperate feature from adding manual control points?  Or will adding manual control points be the only way to make a rouge image stick in place if it can't ever find CPs on it's own?

Last edited by Daniel_Buck (2007-07-02 01:51:56)

You have access to a context menu on picture with no link in the control point editor to snap them to a neightboor. This way, you lock the picture by adding false cp and it will move during optimization.

hankkarl wrote:

Do the advanced lens model parameters also apply to rectilinear lenses?

We also have an advanced lens model for standard lenses, but for the moment, I've desactivated it to concentrate on fisheye. But yes, it will available for any lens and on some tricky case, it gives good results on standard lens.

AlexandreJ wrote:

You have access to a context menu on picture with no link in the control point editor to snap them to a neightboor. This way, you lock the picture by adding false cp and it will move during optimization.

What about sub-panoramas?  I forced all pictures to the same panorama, and got an unlinked pano with two pictures and a larger pano.  Can these also be locked?

hankkarl   wrote:

APP 1.4 A2 seems to detect panos with a lot more accuracy than before.  Its not just a lower RMS--the panos look better and don't need nearly as much work after detection.

Same feeling there!

APP can stitch some 360x180 panos I was very proud I had managed to stitch thanks to accurately (and manually) positioned CPs,  wisely conducted step by step optimizations and very complex multi-layered PSD.

shame on you

Alexandre, did you test support of Nikorr 10mm fish-eye?

When do you plan to fix bug with Zenitar 16?

OK. I will ask man with Nikor to send you his project, after he test A2 and confirm his bugs

Will wait for next alpha, thanks ... I afraided to wait for 1.5 alpha

AlexandreJ wrote:

I have the strong feeling that it mainly comes from the fact that I solved a bug relative to multiple CP on same location.

I was quite surprised you didn't give more attention to this "horrid bug" (possibly removing unwanted CPs afterward!) BTW, that few peoples complained about it was a surprise, too: probably few APP users are taking care of CPs at this level of details and I suppose this is the right way to follow when using APP. As an example, I can see a moderate number of CPs on moving clouds in the panorama I'm presently stitching but this results in no visible harm in the final image. That SIFt places CPs in locations where I can't decide whether they are right or wrong and don't place any at locations where I could place good ones is disturbing.  Changing my "PT habits" needs time, happily this looks like being still possible

Will there be a windows x64 build of 1.4?

To anna. Do not hesitate to share the picture with me. I'll check the software against them and tell you how to do stitch them with autopano.
---
To foundation : yes, v1.4 will be build in x64 for windows and linux.