The best possible average control point distance

111 views
Skip to first unread message

Michał Niedźwiecki

unread,
Aug 16, 2021, 10:09:45 AM8/16/21
to PTGui Support
Hi,

What are the best possible average control point distance? How You achieve?

I have Sony A7III with Sigma A 24-70mm lens and Nodal Ninja 6 P1 C2 Mecha head on  Benro Rhino 14C tripod. When I using 24mm I achieve average control point distance (ACPD) about 2.0 - 3.5 (this is good) - near objects[1]. Sometimes I get less than 2.0 (this is very good) - far objects [2]. So, sometimes I must use Mask to fix small stitch error on straight lines. It extends my working time. I set nodal point with 0.5mm precision on head scale using focus on 2m distance.

It is possible adjust better? The head and body+lens is heavy and make stresses on my tripod. So, nodal point orbits around physical center. I tries move 0.5mm and 1mm in both scales and both directions but I get ACPD worse. So my nodal point is set in the best as possible.

It is possible to to improve ACPD? Maybe another lens?

Earlier, I used Mengs PH-720B head and Sony FE 28-70 mm lens (kit). But it extends my working time during taking photo series and move tripod risk. However if I didn't kick the tripod I achieve ACPD about 1.8 or better and no stitch errors [3]. But the kit lens is worse. Sharpness at the edges looks like astigmatism. I must use bigger zoom (i.e. 35mm) and bigger overlap to achieve good quality. 

Can You propose another not expensive good lens for Sony E or Canon (I have adapter) for 20K panorama? Any 24mm or 28mm without zoom and fully manual? I found some rankings on internet but it presents fish-eye lenses. 

I have Samyang 14mm but the resolution is too small for me however is excellent when they are many moving people in near [4]. I have Canon 50mm 1.8 (mounted by adapter). The resolution is amazing but it significantly extends my workflow time [5].

[5] https://download.nkg-mn.com/gigapan/50mm/ - Canon 50mm

-- 
Best regards
Michał Niedźwiecki

Erik Krause

unread,
Aug 16, 2021, 11:57:18 AM8/16/21
to pt...@googlegroups.com
Am 16.08.21 um 16:09 schrieb Michał Niedźwiecki:

> It is possible to to improve ACPD? Maybe another lens?

The vast majority (by far) of stitching errors is caused by parallax.
First adjust your panoramic tripod head thoroughly. Since you have a
zoom lens, pay attention to the zoom setting. The no-parallax-point
might move for different focal lengths.

--
Erik Krause

Michał Niedźwiecki

unread,
Aug 16, 2021, 12:11:53 PM8/16/21
to pt...@googlegroups.com
I found a nodal point for 24mm and focused on 2m and f/16. I take photo series for near objects with the same lens settings and I get ~2.5 average control point distance. Is it possible to get better?
I tried to move 0.5mm and 1mm in both scales (horizontal and vertical) and both directions but I got worse distance. So I have the best nodal point setting.

Is it possible to appear parallax error on the edge of my lens and sensor? The Sigma A 24-70mm has 82mm diameter. Kit lens has only 55mm diameter.

--
You received this message because you are subscribed to the Google Groups "PTGui Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ptgui/34f2e94b-b228-593b-0487-8a49c02f16d4%40gmx.de.


--
Pozdrawiam
Michał Niedźwiecki

John Houghton

unread,
Aug 17, 2021, 4:33:28 AM8/17/21
to PTGui Support
On Monday, August 16, 2021 at 3:09:45 PM UTC+1 nkg...@gmail.com wrote:
It is possible adjust better? The head and body+lens is heavy and make stresses on my tripod. So, nodal point orbits around physical center. I tries move 0.5mm and 1mm in both scales and both directions but I get ACPD worse. So my nodal point is set in the best as possible.

 There is a third test that is worth performing:  Take two shots with the camera pointing vertically down, panning the head around through 180 degrees between them.  Align the images in PTGui with control points only on the floor in the background.  The centre rotation point of the head should be coincident in the two images.  Any offset in the direction parallel to the lower rail should be corrected by an appropriate adjustment of the lower rail setting.  Correction for an offset in the direction perpendicular to the lower rail may be tricky.  It requires either a rotation of the camera on its mount to the upper rail or a shift at right angles to the upper rail.  This example shows the sort of view you get if you generate a layered PSD file in PTGui and load in Photoshop with the top layer at 50% opacity.

nad-cali.jpg

John

Hugo Rodriguez

unread,
Aug 17, 2021, 5:59:08 PM8/17/21
to PTGui Support
Going below 2-3 pixels average is difficult; you need a quite sturdy head and a precise adjustment. Even with those, the lens shift that all lens have (a small deviation of the lens center off the sensor's center, due to the lens mount mainly) can make going below almost impossible unless you find the exact point where the lens is perfectly mounted aligned with the sensor's center.
If you ask me, I usually achieve 1-1.5 avg in my panos thanks to a careful set-up and processing in PTGui.
But these small deviations can be corrected with the lens shift adjustment of PTGui, and John can explain to you very well.
Regards,

Michał Niedźwiecki

unread,
Dec 30, 2022, 4:21:53 AM12/30/22
to PTGui Support
Problem solved. After 1.5 year ;)

I bought new lens: Sony 50mm f/1.8. Cheap, small distortion and other cons on f/11-16, light weight.

Just stitched 4 spherical panos with 64K resolution (86 photos + nadirs). 
My typical results of optimisation: avg 1.5-1.7, always less than 2 :)

So, the lens is very important. 
ptgui.png

Erik Krause

unread,
Dec 30, 2022, 5:42:50 AM12/30/22
to pt...@googlegroups.com
Am 30.12.22 um 10:21 schrieb Michał Niedźwiecki:

> I bought new lens: Sony 50mm f/1.8. Cheap, small distortion and other cons
> on f/11-16, light weight.

Yes, most 50mm f/1.8 are like that, and for a reason. The construction
of the lens is quite simple and almost symmetrical, and there is hardly
any entrance pupil shift. And it is a prime lens, after all. Zoom lenses
have many more glass elements and therefore many more problems and
compromises in lens construction.

--
Erik Krause

Michał Niedźwiecki

unread,
Dec 30, 2022, 6:38:12 AM12/30/22
to pt...@googlegroups.com
Can you recommend a good and not expensive 28mm or 35mm lens for Sony E mount?
50mm requires too many shots. 

--
You received this message because you are subscribed to the Google Groups "PTGui Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ptgui+un...@googlegroups.com.


--
Pozdrawiam
Michał Niedźwiecki
Reply all
Reply to author
Forward
0 new messages