Go Shopping Cart Site Map User Panel and Admin English Home

Home > Easypano Forum    Easypano Forum has been upgraded.

Easypano Forum


Welcome Guest Register Login Search The Forum Display List of Forum Members
 All Forums
  Panoweaver
 
Subject Topic: stitching error Post Reply Post New Topic
Message posted by pixelator on January-16-2006 at 2:09pm
View pixelator's Profile Profile   Search for other posts by pixelator Search   Visit pixelator's Homepage www   Quote pixelator Quote   Send Private Message Send Msg  
pixelator
Standard Member
Standard Member
United States
April-21-2005
502 Posts
Quote: Originally posted by 360texas on January-16-2006

First make sure you are stitching Drum or Full Frame (there is a personal joke here). 

 


LOL  
The personal joke that Dave is refering to is that I was using Full Frame instead of Drum.  My set up is a Canon 350D, Sigma 8mm wich in fact produces drum.  I am still trying to tweak my image with what seems to be stiching errors though.  I believe at this point I am going to try and reshoot, after shooting the 4 stops and top image, remove camera from tri-pod and shoot bottom.  This was suggested as an idea, so to leave no stone unturned I am going to give it a try.

Regards, Pixel


Message posted by spinem on January-16-2006 at 3:37pm
View spinem's Profile Profile   Search for other posts by spinem Search   Quote spinem Quote   Send Private Message Send Msg  
spinem
Gold Member
Gold Member

February-02-2005
180 Posts
Quote: Originally posted by 360texas on January-16-2006

With out seeing the affected images it is really hard to determine the issue.

First make sure you are stitching Drum or Full Frame (there is a personal joke here).

I think the reason PW 4 has X and y dimension adjustments is because sometimes inside the camera the SENSOR is not perfectly aligned with the perpendicular lens axis. Means the sensor is off center from the center of the lens. Even different Canon 20d camera sensors can be off set by a few microns.

PW4 allows you to adjust the center of the sensor/ image with the lens as well as.. making cropping adjustments to remove the black background and light fall off purple fringe the at the lens edge rim. Thats where the glass meets the metal lens housing.

If Drum.. check the image "sensor" alignment X and Y and R for radius by trying to make it larger or smaller radius.




Dude, I have to admit... you ROCK.
I just opened PW4, And looked at the image. It felt like the X center is a little bit off the center of the image. So I moved the FisheyeImageEnclosing to the left alittle bit. And now There is NO PROBLEM at all :D

great. Thanks a lot :)


Message posted by andy05 on January-19-2006 at 3:41am
View andy05's Profile Profile   Search for other posts by andy05 Search   Quote andy05 Quote   Send Private Message Send Msg  
andy05
Standard Member
Standard Member
 
December-18-2005
8 Posts

It must be something to do with how the Agnos bracket locates... i just tried the same pano with 4+T and the first one was unusable and the second one was exactly correct.  I changed nothing at all on the camera or the mount/tripod, but once I had down the "up" on the 1st set, I had the turn the camera back down for the next 4. I suspect that when I did that I inadvertantly hit a sweet spot on the bracket and it worked. Its so unpredictable though? General Lee, do you notice similar results in that some are miles better than others for no apparent reason?

Andy


Message posted by smooth on January-19-2006 at 9:21am
View smooth's Profile Profile   Search for other posts by smooth Search   Visit smooth's Homepage www   Quote smooth Quote   Send Private Message Send Msg  
smooth
Forum Moderator
Forum Moderator
Australia
November-23-2002
5401 Posts

Nodal Points "Correction" Entrance Pupils are NOT stationary within a Fisheye lens. It can vary up to 1/4" depending on the scene being shot. If objects in one shooting position are close to the lens when compared to the other shots in the panorama series you WILL always have some noticeable stitching errors if you are not helping the process along with control points. This is another reason why most images shot inside a room are taken from the middle and not to one side etc. (to keep equal distance! Thus keeping the nodal entrance the same as possible.) You will always notice these misalignments when objects are closer to the lens and not so when shooting outdoors where all things are basically in the distance.

Re-levelling each shot will add to this problem. Remembering we are talking about the Agnos panohead and it is recommended for good reason that you level position (1) only and then shoot your shots in rotation without adjusting levels.

Your nodal camera positioning should not move if you have the MrotatorT set up correctly. Though I do advise you use a Twin Axis Hot Shot Level to make 100% sure that your lens in not dropping. Again these levels are set only in position (1)  with is the 0 or 360 degree mark.

Regards, Smooth


Message posted by andy05 on January-19-2006 at 5:47pm
View andy05's Profile Profile   Search for other posts by andy05 Search   Quote andy05 Quote   Send Private Message Send Msg  
andy05
Standard Member
Standard Member
 
December-18-2005
8 Posts

Hi Smooth

A really dense question here, but I'm good at those! I think you've hit on something I've asumed didn't matter - I just shoot four shots on the Agnos head starting from a random click point and just count round from there. Must the first shot always be at 0/360 for it to work out?  At the moment, I just start from the nearest click that suits which will hardly ever be 0/360 - although I do make sure its level for the first one?

Andy


Message posted by smooth on January-19-2006 at 11:23pm
View smooth's Profile Profile   Search for other posts by smooth Search   Visit smooth's Homepage www   Quote smooth Quote   Send Private Message Send Msg  
smooth
Forum Moderator
Forum Moderator
Australia
November-23-2002
5401 Posts

Can't see it makes a differences which position you start from! So long as you class it as the starting point and it's 100% level both horizontal and vertical and you are shooting in a clockwise rotation. If shooting an up and down shot these must be taken at the very last click stop.

Regards, Smooth


Message posted by Gen. Lee on January-22-2006 at 1:41pm
View Gen. Lee's Profile Profile   Search for other posts by Gen. Lee Search   Quote Gen. Lee Quote   Send Private Message Send Msg  
Gen. Lee
Platinum Member
Platinum Member

May-15-2005
372 Posts

Hi andy05,

You asked above if some of my panos are way off while othere are quite good and the answer is yes. But I only have this problem on the bottom and top shot. The bottom shot is always the worst. This is only a problem when shooting things like tile floors or docks. Things with lots of horizontal/verticle lines.

Smooth is definitly correct about the nodal point "moving". In fact we are really rotatating around the nodal point and not "exactly" on it for the reason he stated.

Normally the bottom shot is the closest to the camera. The top shot is further away.

The Angnos rotator is a wonderful tool but it does have a few faults. The small arm that rotates up and down for the nadar and zenith shots does not snap into the 90 / -90 position very well. In fact mine does not snap at all. I have "feel" it and the screw has to be just right or it is too loose or too tight.

It uses a friction screw and a spring loaded steel ball to pop into the hole on the arm. This is a little "sloppy" when trying to get it to aling correctly. You have to be very gentle and feel it slide into place.

There is quite a bit of play in it and you have to mess with it to make sure it is in that position. When the camera is rotated down I can't tell sometimes if it is at exactly -90 deg. I have to eyeball it and i usually make it look exactly verticle. But since we are not supposed to re-level it is important that the down shot be aligned with the other shots. So the down shot must me at "exactly" the -90 deg position on the arm and NOT exactly verticle.

After reading this thread a few more times and thinking about it, I believe this to be the problem. If you don't get the down shot to correctly snap into the correct position then it makes the errors worse. Sometimes I get it right and the errors are very small. Other times they are huge.

I have been sketching a few modification plans for the little arm. We need a more accurate method of rotatoing the little arm. The current design is a little "sloppy". What I mean by sloppy is difficult to get exact. I mean there is play in it. It is not accurate every time. I don't mean "sloppy" construction.

There needs to be a different kind of locking design. One that will pop into the hole and secure the lens from moving. I have problems keeping the camera level at horizontal at times becasue the friction screw doesn't hold very good and if you screw really tight when you get to top/bottom shot the screw is so tight it just shakes the whole rig trying to loosen it enough to rotate.

I have looked at the little spring ball and tried to see if I can push it in with a tool but I cant. It is soooo tight I don't think it is moving at all. On the main verticle arm that holds the little arm there is a deep groove worn into the arm. That tells me there may be something wrong with the tension spring on it.???

So my conclusion is the top and bottom shots MUST be at the exact 90 / -90 position on the rotator and NOT actually verticle.

General Lee


If you wish to post a reply to this thread you must first Login
If you are not already registered you must first register

Forum Jump Page of 2 Post Reply Post New Topic
Printer Friendly Version Printable version

Powered by: - Web Wiz Guide Discussion Forums