Forum:
Q & A
Thread:
Nodal Point of Sigma 8mm 1:3.5 EX DG Fisheye Lens
|
|
Forum:
Q & A
Thread:
How to find control Points On cloudy sky? Using PTGui
|
Re: How to find control Points On cloudy sky? Using PTGui
Posted: 4 hours ago
|
go to thread
|
Can you explain exactly how the images were shot for your panorama? i.e. is it a single row 360 degree panorama, or multi row , or a 360x180 (x around +zenith and nadir) etc. It's difficult to know how to advise in detail without knowing exactly what it is you are trying to stitch. If it's a single row, then one image would not be expected to be completely sky.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
How to find control Points On cloudy sky? Using PTGui
|
Re: How to find control Points On cloudy sky? Using PTGui
Posted: 13 hours ago
|
go to thread
|
Opticalmedia said: Now after this I usually get a great stitch and I'm sure that once I get my lens Calibrated properly for a/b/c - I will get an even more accurate stitch?
If you usually get a great stitch, then it's very likely you have a previous project file to use as a template. This will set up your lens parameters in lieu of calibrated values. (But why not calibrate your lens anyway)? Assuming you are using a pano head, you can initialize the y,p,r values to their known nominal values. You should optimize the main row of the panorama with y,p,r checked for those images only, but uncheck y on the first image, and uncheck all other images in the Use Control Points Of list, and check the Link boxes above the p and r columns. This will force the optimizer to keep the pitch and roll settings the same for all the images in the row. Having done that, you can uncheck the link boxes and proceed normally, with image 0 of that row as the anchor (y,p,r unchecked).
Usually, an unlinked image will have predictable y,p,r values based on neighbouring ones, so these can be assigned manually. In the case of a zenith sky image, you probably rotated the camera upwards from the last horizontal shot, in which case its yaw value will be the same as the previous shot. The pitch will be +90. That should be good enough, given that the sky will have changed anyway so a perfect match will be impossible.
Finally, the whole panorama can be levelled using t1/t2 points and the Level Panorama option on the Edit menu of the Panorama Editor window. Or if there are no verticals or horizon for t1/t2 points, you can level it visually by dragging the whole image or by making adjustments using the Numerical Transform option. These operations will not disturb the alignment of the images with each other and keep the unlinked image in place.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
Stitch Speed Test on different machines
|
Re: Stitch Speed Test on different machines
Posted: 19 Aug 2008 at 19:02 GMT
|
go to thread
|
------------------
System Information
------------------
Time of this report: 8/19/2008, 19:42:06
Machine name: MAIN-PC
Operating System: Windows XP Home Edition (5.1, Build 2600) Service Pack 3 (2600.xpsp.080413-2111)
Processor: Intel(R) Core(TM)2 CPU 6420 @ 2.13GHz (2 CPUs)
Memory: 2048MB RAM
Page File: 382MB used, 3560MB available
PTGui: V8.0 Beta 6
Enblend: V3.0
Run time: 68 sec.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
stitching--objects different sizes
|
Re: stitching--objects different sizes
Posted: 18 Aug 2008 at 7:46 GMT
|
go to thread
|
It would be nice if you could upload some images somewhere so we could see exactly what the problem is. If you stitch the images in the usual way and output the result as a layered PSD file, you should find that the images match up correctly, with no distortion issues. It's then easy to assess the parallax by switching the top layer on and off.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
2QTVR Quality setting problem.
|
Re: 2QTVR Quality setting problem.
Posted: 16 Aug 2008 at 5:23 GMT
|
go to thread
|
jimmyd said: It appears my only option is to save the cubics as tiff, then save them as jpeg with PS getting them down to their desired file sizes. This of course creates another "save" and thus more quality loss.
You are only losing quality once - when you save to jpeg. You can save to tiff as many times as you like and it will have no effect on quality.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
2QTVR Quality setting problem.
|
Re: 2QTVR Quality setting problem.
Posted: 15 Aug 2008 at 19:25 GMT
|
go to thread
|
When you generate cubic tiles, there is no quality seting option in the dialog panel. There is a quality setting on the dialog panel for generating a QTVR file (.mov, which has 6 cubic tiles in jpeg format contained within it). So varying the quality setting will vary the .mov file size. As there is no quality control provided when generating cubic tiles, it's not surprising that they always turn out the same size. I assume they are output at maximum quality and are intended for editing purposes, though tif, being lossless, is preferred.
Your arithmetic is suspect! 6272x3.142 is not 1971 as a moment's thought will confirm. It's 19706 approx. But this is not the sum you should be doing anyway. The tile size often recommended is panowidth divided by pi, which is 1996.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
|
Re: PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
Posted: 14 Aug 2008 at 14:04 GMT
|
go to thread
|
tturner said: If it ever happens again I will grab a screen capture for you.
No. If it happens again, undo the optimization (Ctrl+Z) and save the project file with new name. The key to investigating problems is being able to reproduce them. If you have the state of of the project as it existed just before the images flipped and can then simply load the project and run the optimizer to demonstrate the fault, it should be straightforward for Joost to investigate. I have no detailed knowledge of the inner workings of PTGui, but I would be very surprised if the setting of the rotate option on the Crop tab had any effect on an optimization run.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
|
Re: PTGui Pro - very odd and unpredictable
Posted: 14 Aug 2008 at 6:54 GMT
updated: 14 Aug 2008 at 6:57 GMT
|
go to thread
|
tturner said: I am referring to the rotate selector on the crop panel in the lower right corner.
That rotation control (which I had forgotten) does the same as the one on the crop tab. It has no effect on the stitching.
All these rotation effects are largely cosmetic. It's important to supply all the images to PTGui in the same orientation, and images are shown in this orientation on the Source Images tab. These thumbnails never change. It's quite possible to supply mixed orientations if you make use of individual lens parameters, but this complicates things unnecessarily.
I always generate my images from RAW and rotate them to be the right way up when opened by PTGui, i.e. in portrait orientation when taken on the pano head. The images can be supplied in landscape, but then roll settings of -90 degrees need to be set on the Image Parameters tab to bring them upright in the output area. The rotate buttons on the Project Assistant tab can make those changes with a single click and also rotate the thumbnails. If jpeg images are used, then the images will generally be in landscape orientation, though often with a marker in the exif data indicating the state of an auto rotation sensor in the camera. PTGui will act on this marker and automatically apply the appropriate roll values and rotate the thumbnails. The Source Images tab, however, will show the images in the native landscape orientation.
When setting control points on zenith and nadir images, there is no "up" or "down" aspect to the images. To match the parts of the images you are interested in with corresponding areas in the horizontal shots, it makes things a lot easier to rotate the image - purely for viewing convenience. This has no effect on the roll parameter of the image and hence it's orientation in the output area isn't affected. Nor is there any change to the orientation of the image on the Project Assistant tab, for the same reason.
All this automatic rotation stuff is intended to save you work and to get things right automatically for a novice user. Unfortunately, it is also apt to cause confusion if you don't understand what's going on.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
|
Re: PTGui Pro - very odd and unpredictable
Posted: 13 Aug 2008 at 18:40 GMT
|
go to thread
|
tturner said: I have had my source images flip after the first optimization. First pass through all is good and have gone back into control points to take out bad points like clouds and moving trees. It is during these additional optimizations that some of the images have rotated 90 or 180 degrees not from the start. After setting rotation to "0" from "Auto" I no longer have this problem. So is this a setup error or software problem?
I'm not altogether clear what the problem is. In changing the rotation to 0 from auto, I assume you are referring to the Control Points tab, or is there somewhere else that I have missed? The rotation settings there have absolutely no effect on the output panorama. They are purely for the convenience of the viewer to ease mental contortions when comparing images. As explained earlier, the orientation of the thumbbnails on the Project assistant tab reflect what is going on in the output area. If thumbnails there flip, it is because something has happened to the images in the output area. This does not necessarily indicate anything wrong with the optimization. If the optimization report is good then, really, that's the important thing. The entire panorama may be upside down or otherwise out of level, but that should not be regarded as a fault or a problem. It's just not in the ideal position on the stitching spherical surface. Essentially, PTGui has no means of knowing which way is "up". The panorama can easily be transformed into its correct orientation by telling PTGui which features should be upright.
If this doesn't help, maybe you could show some examples of images flipping.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
|
Re: PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
Posted: 12 Aug 2008 at 20:21 GMT
|
go to thread
|
I had no problems stitching it with PTGui. The optimization report was not very good, though the final panorama looks fine, with only a couple of tiny stitching errors. However, there are significant parallax errors, indicating that the head could be set up more accurately.
This probably accounts for the quality of the optimization.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
|
Re: PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
Posted: 12 Aug 2008 at 19:09 GMT
|
go to thread
|
jimmyd said: One must understand something about PTgui. It was not designed to be "user friendly".
On the contrary, that was one of its primary aims: to provide a friendly interface to the notoriously "user hostile" Panorama Tools. It was not designed to be a fully automatic stitcher, but the gradual introduction of (optional) automated features tends to nudge it in that direction.
I have no idea why the thumbnails rotate on the PA tab, but it is of no significance.
1. The thumbnails reflect the current orientation of each image in the output panorama area. Obviously, the roll value is rounded to the nearest 90 degree increment.
2. The rotation buttons at the right of the thumbnails rotate the entire panorama image, in the same manner as applying 90 degree roll adjustments via the numerical transform option.
As you rotate the output image up, down and around, the input images take up assorted ortientations, hence the thumbnail variations on the Project Assistant tab. Whether this is actually useful information or not is debateable. I think it best to ignore it. It doesn't normally indicate anything about how well or badly the images are aligned with each other.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
|
Re: PTGui Pro - very odd and unpredictable - due to my incompetence I guess...
Posted: 12 Aug 2008 at 17:26 GMT
|
go to thread
|
PTGui is fully independent of the Panorama Tools programs these days, though you can still optionally make use of them if you install them using Jim Watters installer program (on Windows). That will install a compatible pano12.dll automatically.
It is not difficult to get consistently good results with PTGui, but it does require a little experience and an awareness of what is going on. What can make all the difference is not deliberately withholding from PTGui all you know about the project. 360x180 panos are usually taken with standard configurations (e.g. 4 around + zenith and nadir, say). You know what the lens focal length and crop factor is. You know what the yaw, pitch and roll angles are. You've stitched several panoramas previously so there are probably some reasonably good sets of lens distortion parameters available in these old projects.
It's really not rocket science. Give all this information to PTGui instead of just chucking the images at it and asking it to work everything out from scratch every time. As Hans pointed out, there is an extremely convenient way of easily and rapidly setting up this information: you apply an old "good" project as a template (via the File menu). Seba evidently has some successful projects available for use in this way.
After setting up all the project parameters in this way, use the Control Points menu to (you've guessed it) assign control points. You can then optimize y,p,r,fov and shift parameters for all the images (except uncheck y,p,r on one image for it to act as an anchor). Use the option delete the worst control points and reoptimize.
This will give you a reasonable general alignment, with very little chance of any images going topsy-turvy.
To get the optimum stitch, you need to check the worst control points and correct them if badly positioned, and make good any deficiencies in the automatic control point assignment - i.e. improve the spread if there is clumping and of course add control points if none have been generated between some images where that is necessary or desirable. Of course, you can also fine tune the lens parameters too in the optimization, if that is considered advantageous.
With a little practice, all this can be done quite quickly. Sometimes very little manual input is needed, other times the images are such that more work is necessary.
Ideally, you will perform a careful calibration exercise to construct the best possible template using images shot in ideal conditions.
The above describes in general terms the workflow I adopt, and I seldom experience any of the nightmare problems I see reported from time-to-time here and in other places.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
Pano joining in the wrong place..? when shooting 360
|
Re: Pano joining in the wrong place..? when shooting 360
Posted: 10 Aug 2008 at 6:42 GMT
|
go to thread
|
Note that you may need to click the "Edit Entire Panorama" button to enable dragging of the whole panorama image. You can also use the "Set center point" button on the Panorama Editor window to bring a particular point to the centre. Or you can use the Numerical Transform option (123 button) to apply yaw adjustments to shift the whole image left or right, which won't risk upsetting the levelling of the panorama. Finally, if you have the already stitched 360 panorama image, you can adjust the centre in Photoshop by using Filter->Other->Offset with the Wrap Around option checked.
John
|
|
alert moderator
|
|
Forum:
Q & A
Thread:
Color matching on a handheld nadir.
|
|
Prev
