Kolor Forums | Autopano, Panotour & Autopano Video Help

klausesser wrote:Ok - if you would use a custom-style preset: where would you get it from? How would you generate a pattern for your individual setup to feed it to the controller?

I'll take PW presets for Tokina 10mm and it's done. There are 4 of them:
4x-30° + 4x30°
6x -30° + 6x30°
6x 0° + Z + N (zenith and nadir ;-))
6x -10° + 1x60°

You are right, I don't need theese presets. BUT only in the case that TC calculates the preset RIGHT for my 10mm FE for 360x180 (at least 360x165 without the footprint).
I'll make a few test with different diagonal angle for 10mm.

gkaefer wrote:but saving the manual created pattern on the TC device (in form of xml ;-) ) and giving the template a useful name you ever can come back and choose your saved "pattern" to do the "same" pano on a new scene (without the need to remember the exact values you used the last time)

Hey Georg!

That´s a nice theory - but in real live your patern would only fit the ONE situation you generated it from! So recalling the pattern for tje use in a different surrounding at least would mean to EXACTLY reproduce the setup. Otherwise the pattern wouldn´t match!

So what would you win? Nothing.

You don´t need to "remenber the exact values used at the last time" when you let the TC do it´s work: it´s all there in your entries . . .

Either your setup is correct: then the TC works perfect, or the setup isn´t correct: then no custom preset will save you.

I shot in very different places: in an underground-tube, on a tower, inside a museum, inside a shower, in stores, hotels, landscapes, open air cinemas, i used
15mm fisheye, 20, 30,85, 300 and 400mm lenses, i shot in the rain, at night, with 3, 7, 9, 12 steps bracketing . .

I NEVER had a need for a preset-pattern. I ALWAYS thanked heaven for NOT needing presets. The TC is fast, small, reliable - reduced on what a panorama-photographer REALLY needs to shoot in an absolutely reliable and comfortable way. And it´s simple in a very clever way to avoid making mistakes which shows you having ruined the shot when you come back home and start stitching.

When i used the N800 and PW-presets is several times stood in the field and wasn´t able to use a different lens than which i intended to use before - because i had no preset for this lens . . Such things happen when you NEED to use presets!
Another time i decided on the location not to use a fullframe-camerea but my backup (crop) camera and another lens insted of the one on the fullframe - took me ecaxtly ONE touch on the display to select it instead of looking for the matching preset.

It´s widely misunderstood i mean: because you store your cameras and lenses (a lot of it) as well as their values ONCE in the controller at home you just need to recall it from the list - and have the values pre-set.

So in theory it might be sounding good to have custom-style presets . . but: for what? Not for real shootings - that´s for sure.

:cool:

When Josef started designing the controller he was going to adopt the preset-concept from PW. I stopped him. I told him about my experiences using it in real life producing panos . . . and i could convince him to go a way which definitely is much more clever under pano-production aspects.

So: blame it to me that the TC doesn´t provide "custom-style presets" . . and be glad not to need them :cool:

best, Klaus

FelixAutopano wrote:

klausesser wrote:Ok - if you would use a custom-style preset: where would you get it from? How would you generate a pattern for your individual setup to feed it to the controller?

I'll take PW presets for Tokina 10mm and it's done. There are 4 of them:
4x-30° + 4x30°
6x -30° + 6x30°
6x 0° + Z + N (zenith and nadir ;-))
6x -10° + 1x60°

You are right, I don't need theese presets. BUT only in the case that TC calculates the preset RIGHT for my 10mm FE for 360x180 (at least 360x165 without the footprint).
I'll make a few test with different diagonal angle for 10mm.

The TC for sure calculates them correct - if you feed it the correct informations . .

best, Klaus

FelixAutopano wrote:My lens is here - http://www.tokinalens.com/tokina/products/atx/atx107afdxfisheye/

I reprogrammed newly the TC controller and
at 10mm (168° diagonal) it begins on 0°
- -

"The Tokina AT-X 107 NH is a fish-eye lens that gives the digital photographer an approximate 180° field of view"
from Tokina.

http://www.tokinalens.com/tokina/products/atx/atx107afdxnhfisheye/

That means: it´s supposed to have 180° diagonally. So why did you set 168°?

best, Klaus

klausesser wrote:"The Tokina AT-X 107 NH is a fish-eye lens that gives the digital photographer an approximate 180° field of view"
from Tokina.
That means: it´s supposed to have 180° diagonally. So why did you set 168°?
best, Klaus

http://lenshero.com/lens/Tokina-AT-X-107-DX-10-17mm-f3.5-4.5-Canon-ef-lens
or better here
http://www.hdrlabs.com/tools/panocalc.html

But for required results from TC calculated preset I need less than 180°, yet less than 168° .

klausesser wrote:When Josef started designing the controller he was going to adopt the preset-concept from PW. I stopped him. I told him about my experiences using it in real life producing panos . . . and i could convince him to go a way which definitely is much more clever under pano-production aspects.

I've never suggested abandoning the built-in on-the-fly calculation of shooting patterns, I'm sure it's very useful; I've merely advocated adding the ability also to be able to use Papywizard-style custom presets to satisfy the irrational whims of loonies like me.

But I think you've already explained that the T&C controller doesn't have enough RAM to be able to do that now.

So: blame it to me that the TC doesn´t provide "custom-style presets" . .

Oh, I do, I do.

FelixAutopano wrote:

klausesser wrote:"The Tokina AT-X 107 NH is a fish-eye lens that gives the digital photographer an approximate 180° field of view"
from Tokina.
That means: it´s supposed to have 180° diagonally. So why did you set 168°?
best, Klaus

http://lenshero.com/lens/Tokina-AT-X-107-DX-10-17mm-f3.5-4.5-Canon-ef-lens
or better here
http://www.hdrlabs.com/tools/panocalc.html

Why does Tokina´s own page mention 180°? I mean, at least THEY should know it best, don´t they?

FelixAutopano wrote:But for required results from TC calculated preset I need less than 180°, yet less than 168° .

Yoe need what the lens provides natively!

Basing on this informations everything gets calculated. No matter which way - TC or otherwise. Read the specs!

best, Klaus

klausesser wrote:Yoe need what the lens provides natively!

I've entered angle what the lens provides natively. The angle is THAT what it covers for real not what the manufacturer has on its web.
On 10mm it is hard to estimate the real angle, but I'm sure that is NOT 180°. On 17mm I clearly see that diagonal angle is little over 90°.
Don't you tell us, that the TC knows that Tokina has on its website diagonal angle 180° and thus calculates non-real-world preset .

Why from 12mm is it correct?

FelixAutopano wrote:

klausesser wrote:Yoe need what the lens provides natively!

I've entered angle what the lens provides natively. The angle is THAT what it covers for real not what the manufacturer has on its web.
On 10mm it is hard to estimate the real angle, but I'm sure that is NOT 180°. On 17mm I clearly see that diagonal angle is little over 90°.
Don't you tell us, that the TC knows that Tokina has on its website diagonal angle 180° and thus calculates non-real-world preset .

The Tokina-Website says 180°@10mm, doesn´t it? I guess you´re puzzling things here

best, Klaus

FelixAutopano wrote:I've entered angle what the lens provides natively. .

How do you know what the lens "provides natively"?

I mean: you no way should rely on the values the manufacturer provides . . . :cool:
How would he know the focal-length of a lens that he makes . . no, you´re right: don´t believe him . .

Better enter the "native angle" you read on some funny websites . .

ok - i´m out. Good luck.

Klaus

I'd rather enter what the lens really shows in real conditons on my body = what I see through the lens on my body.
I know that on 10mm it isn't 180° diagonal, I figure that is around the 168° maybe less and on 17mm I'm sure it isn't 100° but the little 2° over the 90°.

Why the TC DOESN'T calculate with the submitted angle?!

:):) Why aren't you driving through the oneway street when on cityplan (or in apple maps or other gps device) it isn't oneway? Do you see what I implie :):)

FelixAutopano wrote:Why the TC DOESN'T calculate with the submitted angle?!

because it´s too narrow. Josef told you already - as well as i did.
With 168° diagonal fov you can´t get sufficient overlap to cover a sphere.

You will be able to cover the sphere when you shoot at least two rows of 6 plus Zenith -
or use the 180° diagonally fov which Tokina claims to provide.

Klaus

I've played with diagonal angles for 10mm and here are the results:
From diagonal angle 169° to 161° TC calculates preset 6 shots at 0° + zenith (I cannot see the base plate).
For diagonal angle 160° TC calculates best preset :D:lol: 6 shots at -9° + 6 shots at +1 + zentih.

On the image down you can see the difference between 6x0°+Z and 6x-9°+Z (at 81° - by moving the initial zero to -9° on TC) = this is solution for me.

But I must again write that TC calculates the preset WRONG for this wide fisheye.

FelixAutopano wrote:But I must again write that TC calculates the preset WRONG for this wide fisheye.

Could you be precise, please? What do you mean in detail? Can you prove your statement mathematically?

The TC calculates on the base of informations YOU give it as well as on the base of the common use of terms.

So: either Tokina is lying about the lenses values or you don´t know what you´re talking of. To be honest: i don´t think
Tokina is lying . . .

Again: i´m using the TC for years now with two very different heads: Merlin and Panoneed. I NEVER had an error with a mis-calculation of a shooting.
Neither using a Nikon 10,5mm fisheye nor using a Canon 15mm fisheye nor using a Nikon 16mm fisheye - nor using 20, 35, 50, 85, 105, 200, 300, 400 and 600mm
rectilinear lenses.
Afaik there are more than 200 TCs in use around the world. Did someone of the users complain about non-precise calculations?
Not to my knowledge.

best, Klaus

P.S.: it´s commonly known that the Merlin´s very large base makes it a bad choice for being used with fisheyes. Better use a manual head providing
a small footprint.

klausesser wrote:

FelixAutopano wrote:But I must again write that TC calculates the preset WRONG for this wide fisheye.

Could you be precise, please? What do you mean in detail? Can you prove your statement mathematically?

The TC calculates on the base of informations YOU give it as well as on the base of the common use of terms.

So: either Tokina is lying about the lenses values or you don´t know what you´re talking of. To be honest: i don´t think Tokina is lying . . .

I don't think Tokina's data can be entirely correct, because surely the diagonal FOV could not be 180 degrees on both a Canon 1.5x cropped sensor and a Nikon 1.6x cropped sensor.

For example, I know that the VFOV (camera in portrait orientation) of the Sigma 8mm f3.5 fisheye is 180 degrees on a Nikon 1.6x cropped (DX) sensor but smaller than 180 degrees on a Canon 1.5x cropped (APS-C) senor.

Again: i´m using the TC for years now with two very different heads: Merlin and Panoneed. I NEVER had an error with a mis-calculation of a shooting.
Neither using a Nikon 10,5mm fisheye nor using a Canon 15mm fisheye nor using a Nikon 16mm fisheye - nor using 20, 35, 50, 85, 105, 200, 300, 400 and 600mm
rectilinear lenses.
Afaik there are more than 200 TCs in use around the world. Did someone of the users complain about non-precise calculations?
Not to my knowledge.

Well, although I don't have any hands-on experience of the T&C controller as you are so keen to remind me, I wouldn't be happy with 6-around at zero degrees, and a +90 degrees zenith if that's what the T&C would calculate as the only 'optimal' pattern for a Nikkor 10.5mm FE on a Nikon DX sensor body.

P.S.: it´s commonly known that the Merlins very large base makes it a bad choice for being used with fisheyes. Better use a manual head providing a small footprint.

True - but irrelevant to this dicussion, the T&C claims to be able to shoot spherical panos using fisheye lenses automatically calculating an 'optimal' shooting pattern.

As I understand it, the issue in question is simply that the user concerned doesn't believe that the pattern calculated by the T&C is 'optimal ' for his needs. And he wishes that the T&C also offered the 'flexibility' of supporting Papywizard-compatible custom presets for shooting patterns.

...........

I doubt there would be any issue with the shooting patterns calculated by the T&C for shooting spherical panos using rectilinear lenses of any focal length.

klausesser wrote:Could you be precise, please? What do you mean in detail? Can you prove your statement mathematically?

I've measured the horizontal view angle (one point at left - the same point at right and PW showed 128.8° and from known sensor sizes I've counted the diagonal 159.5°. All rough.
For this value TC calculates preset: 6 shots at -9° + 6 shots at +1 + zentih. But I have much better presets: 4 or 6 at -30°° + 4 or 6 at +30°. Eight or twelve against thirteen with stupid -9° and +1° rows.

klausesser wrote:So: either Tokina is lying about the lenses values or you don´t know what you´re talking of. To be honest: i don´t think
Tokina is lying . . .

Yes. Tokina is lying to me. I know sure that I don't see through the lens on my body 180° diagonal. Sure.

I've solved it provisory myself for 10mm. I start TC with -9° as zero.
I don't say a word against TC for rectilinear lenses. But it has some reserves by using with the fisheye lenses.

FelixAutopano wrote:I don't say a word against TC for rectilinear lenses. But it has some reserves by using with the fisheye lenses.

No. The calculations base on the maths. If you lens doen´t have 180° but something else the calculations for that focallength
are correct.

The difference is - as it is in Andrew´s model - in the personal preferences to shoot as well as in the geometry of the head.
As i mentioned several times: the Merlin has a too large base for being used with fisheyes.

See: when i use my Canon 15mm fisheye on my 5D2 the base of thge head is extremely small because of the spin-axis
is very close to one side of the horizontal arm and the vertical arm is rather long - so the camera sits rather high. This provides a very
steep downlook angle - which results in a small footprint = small Nadir-hole.

here´s a pano i recently shot using 15mm FE on Panoneed-head and the TC for controlling:
http://360impressions.de/Wohnung/

The stitch was done nearly automatically. Are you still telling me the TC can´t calculate fisheyes correctly? :cool:

best, Klaus

realy dumb question:

why dont you use on your TC the angles preset instead of the spherical preset?.
using the sperical preset it gives you some faulty Nadir - whoever to blame for that...
using the angles preset and using 180 for vertical and 360 for horicontal angles you get what you want: a sphere where 100% of your 360"x180" is covered by the Images!

Georg

FelixAutopano wrote:I've measured the horizontal view angle (one point at left - the same point at right and PW showed 128.8° and from known sensor sizes I've counted the diagonal 159.5°. All rough.
For this value TC calculates preset: 6 shots at -9° + 6 shots at +1 + zentih. But I have much better presets: 4 or 6 at -30°° + 4 or 6 at +30°. Eight or twelve against thirteen with stupid -9° and +1° rows.

Wow - now that is really weird.

I wonder with how much entusiusm everybody is discussing this subject. As I mentioned, the maximum down angle for the Panogear is 70°, so the maximum vertical angle is 160°. If the diagonal value is 168° for the 10mm lens, then the vertical angle must be 139° (rule of Phytagoras and your sensor size), then 0.5° would be missing above the base of the Panogear theoretically. The only reason for a smaller angle can be, that the diagonal angle at 10mm is maybe in reality smaller than 168°, or the head and the camera are not adjusted horizontally. The diagonal angle of 168 is very much at the limit for a 25% overlapping. If you go down, then the overlapping gets smaller than 25%. As I offered to you, I can change the software to every down angle you would like. It would be your risk when the stiches then have problems. The discussion about presettings wouldn't be, if you would have 4 or 5 different lenses together with different bodies with different crop factor. Because then you must take a booklet with you to make shure you remember all these settings in the field. The nadir hole is 40° by panogear and the 10mm has a vertical angle of 139° and an horizontal of 93°, so one shoot would completly close the nadir hole, why are we discussing this missing little angle so much?

mediavets wrote:Well, although I don't have any hands-on experience of the T&C controller as you are so keen to remind me, I wouldn't be happy with 6-around at zero degrees, and a +90 degrees zenith if that's what the T&C would calculate as the only 'optimal' pattern for a Nikkor 10.5mm FE on a Nikon DX sensor body.

Andrew, i - of course - honor your opinion. But what you claim definitely is the way of your personal use. It´s not what it really needs to be. If you like you also might take 25 images to shoot a sphere wih 10,5mm on DX. No problem.
As i told before: in earlier days i used a 10,5mm Nikon on a 20D (1,6) and a D300 (1,5). In both cases the TC calculated perfect shots for 20% (5+1) or for, more average, 25% (6+1) - of course i told it when i used 1,6 and when i used 1,5. This setting is relevant.

The TC doesen´t calculate basing on the crop-factor but basing on real sensor-size. So you need to tell it the exact sensorsize of your 1,5 or 1,6 crop camera - as well as the exact focal-length and the lens-model. When a lens has 180° according to the manufacturer - then set 180° and nothing else.

mediavets wrote:

P.S.: it´s commonly known that the Merlins very large base makes it a bad choice for being used with fisheyes. Better use a manual head providing a small footprint.

True - but irrelevant to this dicussion, the T&C claims to be able to shoot spherical panos using fisheye lenses automatically calculating an 'optimal' shooting pattern.

Right. Using the Merlin you have a very large area of silver-grey then. Did you do it using the TC? I did. That´s why i know the issue must be elsewhere - but not in the calculations!
The calculations simply means using the maths.

mediavets wrote:As I understand it, the issue in question is simply that the user concerned doesn't believe that the pattern calculated by the T&C is 'optimal ' for his needs. And he wishes that the T&C also offered the 'flexibility' of supporting Papywizard-compatible custom presets for shooting patterns.

I doubt there would be any issue with the shooting patterns calculated by the T&C for shooting spherical panos using rectilinear lenses of any focal length.

If you insist on "user-preset": use PW on a PC, Mac or Nokia. Don´t buy a TC if you feel better dealing with redundancy. Only buy a TC when you need it to get your work done straight, precise and comfortable.

Believe me: i know very well what i´m talking of from my (well, almost ) every-day-work.

I ALWAYS use 6+1 (25% overlap) resp. 5+1 (20% overlap) shots with fisheyes - DX:10,5mm and FX:15mm.

The TC´s precise calculations allows 20%=5+1 in 90% of the time with the 15mmFE on a 5D2. I use 25%=6+1 only when i shoot in very small places like a bathroom or in an elevator or so.
The reason is that on VERY close distances you "by the book" would need a new alignment of the NPP due to the close focus-distance. But setting the overlap to 25%
compensates that perfectly.

A result of using 6 shots @-12° + Zenith you can see here: http://360impressions.de/Wohnung/

best, Klaus

mediavets wrote:

FelixAutopano wrote:I've measured the horizontal view angle (one point at left - the same point at right and PW showed 128.8° and from known sensor sizes I've counted the diagonal 159.5°. All rough.
For this value TC calculates preset: 6 shots at -9° + 6 shots at +1 + zentih. But I have much better presets: 4 or 6 at -30°° + 4 or 6 at +30°. Eight or twelve against thirteen with stupid -9° and +1° rows.

Wow - now that is really weird.

Weird? Yes - at least . . :cool:

best, Klaus

PS: on my manual head i use the same "pattern" as on the Panoneed/TC combination: 6 shots @-12° and 1 Zenith.
Result: http://360impressions.de/ZuerichBar.html

FelixAutopano wrote:

I've measured the horizontal view angle (one point at left - the same point at right and PW showed 128.8° and from known sensor sizes I've counted the diagonal 159.5°. All rough.
For this value TC calculates preset: 6 shots at -9° + 6 shots at +1 + zentih. But I have much better presets: 4 or 6 at -30°° + 4 or 6 at +30°. Eight or twelve against thirteen with stupid -9° and +1° rows.

Answer:
I calculated the diagonal angle too and found 154.8° if your horizontal angle is 128.8 and your sensor 22.3 x14.9
Have you ever made a full sphare with a lot of details in the zenit? I gues not, as we found out, that if there are a lot of details and there is no zenit shoot, you get a lot of stitching errors. For that we make for every full sphare a zenit shoot. If you have that in mind, the calculation of the T&C makes a lot of sense. Just make as much as possible full sphares with a lot of details on the ceiling, I bet you have no stitching problems. But if you run your settings with 6 pictures 30° down und 6 pictures with 30° up, I expect a lot, as in the zenit you get so much overlapping, that the stitcher will have problems.

klausesser wrote:

mediavets wrote:

FelixAutopano wrote:I've measured the horizontal view angle (one point at left - the same point at right and PW showed 128.8° and from known sensor sizes I've counted the diagonal 159.5°. All rough.
For this value TC calculates preset: 6 shots at -9° + 6 shots at +1 + zentih. But I have much better presets: 4 or 6 at -30°° + 4 or 6 at +30°. Eight or twelve against thirteen with stupid -9° and +1° rows.

Wow - now that is really weird.

Weird? Yes - at least . . :cool:

best, Klaus

PS: on my manual head i use the same "pattern" as on the Panoneed/TC combination: 6 shots @-12° and 1 Zenith.
Result: http://360impressions.de/ZuerichBar.html

Klaus your pattern is fine.; 6-around at -12 and 1 zenith.

Except that I have often found that shooting outdoors with a manual head (no XML) that a +90 zenith shot would often bw omitted from the stitch becasue it lacked any matching features with images in the main row of 6 images.

The 'solution' presented by Hans Nyberg was to shoot the 'zenith shot at about +60-65 which covers the zenith well but when the yaw is also chosen carefully will almost always mean you can get an good link between the 'zenith' shot and the main row of 65.

....................

But felix said that the T&C calculated that weird 6 shots at -9° + 6 shots at +1 + zenith pattern.

I can't see that that would ever be 'right'/optimal.

So how did it arise?