Pano2VR keeps shutting down!!

[thauzar]

Hello everyone,

This is my first time using this software and let me say that I am completely dissapointed. It looks nice, I was sold with the samples from the web site, but in execution it's a pain in the ass.

First, IT KEEPS CRASHING DOWN!!! All the time, everytime, for no reason. You want to add zone hotspots, you need to click the little box there, it asks to create the .png because they are not there yet, bang it crashes.

You managed to add hotspots by points because you just got tired of trying the zone ones, then you want to save your files, it crashes.

Another time I only wanted to add the skin to the flash save thing, I clicked the preference icon, it crashed...


I just put 6 hours of useless work in the trial of this software which promised amazing and easy to make virtual tours... seems impossible to even generate 1 damn quicktime movie or flash with a friggin hotspot in it.

I have a brand new studio xps with a 1Gb 5730 graphic card, i5 quad core, 6Gb of ram, I doubt that it's my computer that's the problem.

If you care about having new customers, someone better get in touch with me asap else if I can't get a simple demo file out because it keeps crashing, I'll just go to the competition.


Please help me Obi Wan Kenobi, you are my only hope...

[360Texas]

Ok you are using a PC. Do you know if the operating system is 32bit or the 64bit version ? During the crashes - did you notice any error messages ? Might help if we knew more about your situation.. maybe screen capture might be good to see.

This is a monitored forum. After dinner (its 6:45pm Central Texas Time) tonight I visited the forum and approved your comment. Sorry for the delay. Yes there are 3 or 4 people in different time zones around the world.

After re-reading your second comment.... I should mention that we do watch for improper language.

[smooth]

You are using "Area" hotspots and if your image is largish you must use cube faces not equirectangular.
But! if the cube faces are also large the program will (always has) crash and burn. Some tests would prove what is the maximum size of cube face you can to work with Area hotspots.
Anything over 1800 pixels crashes the program everytime for me.

Yes, I would like to see this fixed also.

I simply avoid Area hotspots and use Point hotspots. But this is a work around to a problem that should have been rectified a long time ago.

Regards, Smooth

[thauzar]

Thanks 360texas and smooth,

I tried both 32 bits and 64 bits versions. I also tried administrator mode, and compatibility with XP mode. They both crashed in the same way. I would love nothing more than having a crash error number to give you but it simply says that Pano2VR had to quit and that windows will check on this and tell me if there is any fixes for it (but I dont get any follow up on this from windows). Only 2 or 3 times did it crash with a message saying out of memory.

Reading smooth, I guess it must be something related to the size of the faces. I noticed that Pano2VR tends to go and use ALL the ram. It's very rare that a program really needs 99% of the ram, it usually tries to manage it to leave some for other programs and also for himself in case he needs to execute another simple task, you know like opening a menu. Maybe Pano2VR does not manage the ram so well, charges everything on ram, then runs out of memory and crashes?

But even with point hotspots, when I try to save the files, it crashes without warning at 66%, be it quicktime or flash. Some partially made .mov or .swf file shows up on the desktop but no html page or anything else got created.


Just this morning, I tried saving a .mov, without any options, no html, just default settings. I got a file!! yes!. Then I clicked on the preference icon to test with html checked, just clicking on the preference button send the program crashing, I did not even see the menu open.

So yeah, this looks promising but it would be nice if I could at least open the preference pannel, play with the option, and generate some tests files to check out how it works and how I can make it work for my project.

[thauzar]

Hey again.

I tested the .mov I got and it is broken even if Pano2VR did not crash while creating it. On the other hand, I did a flash file which worked very nicely. Again, just default settings for the flash test, and the program did not crash while saving, but again, it crashed as soon as I clicked anything. So it's not only the preference button, it's any next action that you want to do.

Also, the ram usage does not go down after saving a file, may be related to the crashes?

Am I the only one with those problems?

Gotta go to work but I hope to read more answers tonight and I'll be able to do more tests then.

[thomas]

What is the size of your image? Is it equirectangular or cube faces?

[smooth]

I have done some testing with interesting results.

First, when you "agree" to Pano2VR to create the .png files (it should make them at that moment) but doesn't!
If you are lucky enough to get out of that screen without the program crashing you MUST click save and likewise - if you are lucky the 6 cube .png files are then created.
Because if you don't save it then and "prove" you have created the 6 .png files you can waste a hell of a lot of time building your work up only to have the program fail to create the .png file or simply crash without warning and leave you with a FAT "0" after all your time and effort.

pano2vr_crash.jpg (18.94 KiB) Viewed 5494 times

I get consistent good results with 1900 pixel cubes. I get consistent crashes if I up the cube face to 1950 pixels (but can get a good result "if" the wind is blowing to the south at 22MPH) I.E: Sometimes!
It 100% has to do with memory management. Pano2VR doesn't release it like it should.

I tried again at 2000 pixel cubes and it program crashes everytime. No warning, just Pano2VR has stopped working and needs to close.

May I suggest, we don't always need the whole 6 cube faces. So unless we actually add an Area hotspot on a cube area - don't create the extra cubes! This might help with memory....

Regards, Smooth

[visual360media]

It very rarely crashes on me, and i am usuing an old hp pc, 2gig pentium 4 and 500mb ram and an old nvidia geforce 4 graphics card.
(mind you, this old thing is years old, and is solid, runs all my adobe stuff albeit slow! )
My images are equi, cubes, .mov or qt that i drag in and alter. Only just started with hotspots and did have it freeze up a couple of times
so it looks like a hotspot bug, maybe to be fixed in next version soon.

[thauzar]

Had to take a look here on my lunch break.

@Thomas, I tried with .mov but it said that hotspots could not be saved on .mov so I converted it to cubic faces. The final output .mov I get from AutoDesk Stitcher Unlimited is about 27Mb, I never took the time to weight each of the individual cubic faces that either Stitcher or Pano2VR creates though.

@smooch The product I want to offer is 360 spheric views of rooms, linked together by hotspots, with maybe a plan or something like I saw in the "park" flash exemple on this site, so even though I admit getting rid of the floor and ceiling could save memory, it's not what I want to offer as a final product. I will look at the image sizes tonight when I get back but they are probably (most definitely) very large since I want to offer the best quality product. On the other hand, I guess that since most people dont even own 1080p computer screens, lowering the resolution to 1950 pixels can't do too much damage right (and if it helps Pano2VR from crashing it's something good right there)? I'll have to do some comparative tests.

[smooth]

@smooch

That would be Smooth not smooch!

The product I want to offer is 360 spheric views of rooms, linked together by hotspots, with maybe a plan or something like I saw in the "park" flash exemple on this site, so even though I admit getting rid of the floor and ceiling could save memory, it's not what I want to offer as a final product.

It is not my idea to get rid of image tiles. I suggest to Thomas to not generate hotspot .png tiles if there is no hotspot on them. I would never suggest you limit image view.

I will look at the image sizes tonight when I get back but they are probably (most definitely) very large since I want to offer the best quality product. On the other hand, I guess that since most people dont even own 1080p computer screens, lowering the resolution to 1950 pixels can't do too much damage right (and if it helps Pano2VR from crashing it's something good right there)? I'll have to do some comparative tests.

Please let us know what size files you are actually playing with.

Regards, Smooth

[visual360media]

Good answer Smooch

[360Texas]

I agree with SMOOTH,

We use Canon T1i to take 4 raw images. Typically a raw image is 18 - 21mb each image. We convert raw's to 16bit RGB TIF's. The Tiffs are 50 - 60mb each. Using PTgui we stitch the 4 16bit images. The final stitch is is around 150mb 16bit RGB 6000 x 3000 pixel equirectangle image. We take that image straight into Pano2vr.

We cube face convert them to 1910 pixel (6000 / Pi (3.1415) = 1909.9 to place the tripod cap in cube #5.
We convert cube face back to equirectangle (change 5999 width) width to 6000 pixels wide.

We then convert the new equirectangle image to flash using viewer dimension 800 x 500 pixels. And tweak the settings to create a final 1.5mb flash .swf

So we converted a 150mb 16bit Tiff down to a 1.5mb .swf flash file. We figure we really got an acceptable final panorama.

Here is an example of the finished panorama in the World Wide Panorama Event Sept 2010 "Crossroads".
Under the small viewer.. look for graphic icon called "OPEN FULL SCREEN". There you will viewing the small 1.5mb pano on a 24" monitor 1920 x 1200 res. Pat's image is of a Bird Island scantuary near Holbox, Mexico. Canon t1i raw, Sigma 8 f3.5 on a hand held monopod.

http://www.worldwidepanorama.org/worldw ... -6358.html

[thauzar]

Sorry SmooTh for messing your nickname hehe
My original images were raw .nef files, 9Mb files, unedited 2848x4288, that I convert to tiff (default settings from lightroom3) and end up with something 69Mb. Then I take the 36 images and stich them in AutoDesk Stitcher and I get cube faces with a size of 6656x6656 and 253Mb. So definitely, II was overdoing it if your tests show that over 1950 pixels Pano2VR crashes. I'm still surprised that a software developped to do that kind of work was not designed to handle large files with lots of pixels. I mean, I did not try to crash it on purpose by giving it too much information, I just processed my files with default settings in all those softwares.

I haven't bought my pano lens yet, I'm still unsure about if it's best to get something like the nikon fisheye 10,5mm DX or the standard 18mm rectilinear. I'm want to avoid the dome distortion effect that most fisheye give on the edges. My pictures are taken with a D90 and the basic 18-105mm lens that comes with it, set on 18mm. (with conversion I end up taking my pictures at 27mm aprox.)

I look at 360's panorama and wow I'm surprised of the image quality for something that is 800x500! Except for some artefacts in the sky on the top it looks good. It seems complicated to do though, all this switching from equirectangular to cubic and back and forth. Can't import your cubic faces, use the "field of view" option to set the tripod at the bottom and the starting point of the panorama at the same time, and then use the patch tool (that I still have to try but looked nice in theory ) to extract the part of the bottom face with the tripod to edit in photoshop then put back in place?

Thanks for all your infos I will resize my pictures to 1950x1950 and resume my tests.

[thauzar]

w00t
everything went like a charm without crashing, I even added zone hotspots to my cubic faces.

I now have to watch those tutorials about how to properly put the info so I have working hotspots

Thanks everyone especially Smooth for your precise info about the maximum working pixel size.

[smooth]

w00t
everything went like a charm without crashing, I even added zone hotspots to my cubic faces.

I now have to watch those tutorials about how to properly put the info so I have working hotspots

Thanks everyone especially Smooth for your precise info about the maximum working pixel size.

I'm glad you have a working result and are about to learn the great features of Pano2VR.

I get cube faces with a size of 6656x6656 and 253Mb. So definitely, II was overdoing it if your tests show that over 1950 pixels Pano2VR crashes. I'm still surprised that a software developped to do that kind of work was not designed to handle large files with lots of pixels.

That would equate to something like 20000x10000+ equirectangular. That is fine if you have the webspace and are using the multi-resolution feature and so long as you don't use "Area" hotspots. Link your hotspots with "Point" hotspots. But remember to give consideration to download time.

I haven't bought my pano lens yet, I'm still unsure about if it's best to get something like the nikon fisheye 10,5mm DX or the standard 18mm rectilinear. I'm want to avoid the dome distortion effect that most fisheye give on the edges. My pictures are taken with a D90 and the basic 18-105mm lens that comes with it, set on 18mm. (with conversion I end up taking my pictures at 27mm aprox.)

The first thing you should know is that "all" fisheye distortion is removed when the images are stitched. You shouldn't be concerned about this whatsoever. The Nikkor 10.5mm Fisheye is a fine lens and on your D90 you can expect a maximum file size of about 11600x5800+ equirectangular or cube face of 3692 pixels. Most non multi-res Internet displayed panoramas do not exceed 8000x4000 or cube face of 2546 pixels. The "norm" having cube face size 1600+ to cater for higher resolution screens. Using a Nikkor 10.5mm Fisheye you would only need to shoot 8 images in total. 6 rotational shots at -10 degrees, 1x Zenith shot at +65 degrees and a hand held or offset/tilted Nadir shot to fill the Nadir hole (if required).

..... all this switching from equirectangular to cubic and back and forth. Can't import your cubic faces, use the "field of view" option to set the tripod at the bottom and the starting point of the panorama at the same time, and then use the patch tool (that I still have to try but looked nice in theory ) to extract the part of the bottom face with the tripod to edit in photoshop then put back in place?

You don't need to be back and forth. If you have cube faces to start with just leave it that way. You extract whatever area you wish to deal with, from anywhere within the sphere using the Patch tool. Do your editing and them bring the patch back in and publish and save as required.

Regards, Smooth