Welcome to Fractal Forums

Is it possible to apprach the subject without cutting it? Namely, I don't want to increase zoom and I just want to get closer to the subject. However, whatever value is used for far plane or min distance, the result does not change. At some distance (for given zoom) the closest subjects are cut. How can I change (decrease) this distance??  :sad1:

Example is here:

Mandelbulb3Dv17{
Y....MW...U1/...w....2....E3.RDdGMsyzuBy9okmfE3Ep.w8lAwefzn0Io6P4JqxzOhXqdhH9qzD
........................................kz1........Y./..................y.2..asE
..........E/..../M.0/.....Ei....B2...2E2.....USStDnq8zqD/..........m/dkpXm1....U
z.....kD12../..........wz..................................u1....y1...sD...../..
.z1...sD2Gq0fxwrhxH1ifX0RI.HzAs1PPwcX0pjY.D7sanlEunaA4Bnk13OzEerXETAvFqDtqC.ra5k
JxXWKhctxvnMzO9hoi7ENPqD......Y4..............kD.2....sD..kz0...................
.............oAnAr1...sD....zUkb/A.eT4k.6/O.1UCc/A.WV4k.c6O.1Ugc/A......P....k1.
.....EpsZz1.......kz.wzzz1.U..6.P....M4...EB....W....k1....F....8/...I1.....SE52
...U.qFG9yzb2zzzRYoWzz7lz16.mc..zXCc..k18XGQeGyD0sIRhrVAkz1..........2.28.kFrA0.
.Ub96aAIVz9.1se7Umvxz0........../6U0.wzzz1...........s/...................E.0c..
zzzz.................................2U.8.kzzzD.................................
/6U0.wzzz1...................................2CcN/UvNP6.eeWCNq0.yRii.EJJUk1f..XR
S1.mx3CcN/UvNP6.QsLsUa3.ibhV..bTV1OK.sSq40.ly3CcN/UvNP6.MwLsUa3.ibhV.kqTV1OK.sSq
40.kz3CcN/UvNP6...EsUa3.eeWCNq0.IJ36wk8.wyLsUa3.................................
E....2....E.....I....k.....FjFKNX34OZFaQjtKG4BJNs/........................U/4MU/
4M..0.....UaNaNaNaNyzMaNaNaNaNyDaNaNaNaNazXNaNaNaNatz...........................
.....................................QFqCTzIXbzD........Q.2.....................
........................}

Press F6 or uncheck "Fixed zoom and steps" in the navigator. Hope it helps ..

Edit: I guess you can not approach without increasing the zoom.

Exactly, lxh. Zoom is changing when unchecking that option.

Unfortunately, for stereoscopic pictures, the focal length (depends on FOV and zoom) determines perception of 3D (and roundness of objects). Ideal horizontal FOV is usually between 30-50 degrees. When having FOV=50 and about 10-times magnification, we should get approx. FOV=5, which is too small. Rounded objects might appear flat.

It's unfortunate that we cannot select the closest distance ("cut distance"). I just hope that some future version will enable selection of this parameter.

 :hmh:


 :hmh:


The closest distance begins with the starting plane, if you cut the object, you are to close   :verysurprised:

Zooming does not change the FOV but the size of the starting plane.
If you want common behaviour like most 3d programs you have to let the zoom increase with the distance to the object.
Set the "minimal distance" in the stereo tab to a lower value if you want more 3d effect in the near distance.

M3d is a bit different because you can also use this planes for 2d calculations, it got more options (negative FOV from a big starting plane is also possible), but this is not very intuitive.

This changes everything! If FOV remains the same, then everything is PERFECT.  :embarrass:

Why FOV is very important? Try to make stereoscopic picture of a ball with some background objects by using let's say FOV=50 and FOV=1.
FOV=1 should give you stereoscopic picture of the ball which should resamble a pancake (shallow depth) if you retain disparity of background objects smaller than 1/30 of picture width. This is a so-called cardboard effect in stereoscopy.


Unfortunately, the current implementation of creation of stereoscopic pair is still not flawless, since it can create pairs with unaligned vertical positions.
The only way to get "perfect" stereoscopic picture is to use 3D navy and choose rectilinear lens and check "Fixed zoom and steps". After making the first picture, you should make a move to the left or right by "Sliding" buttons and then make another picture.
In StereoPhoto Maker you should adjust stereo window by horizontally adjusting the pictures. Of course, due to later horizontal cropping, the original horizontal sizes should be larger than desired (e.g. choose 2200x1080 for the final 1920x1080 stereoscopic picture).

The vertical position is not changed, create left-eye and right-eye with the newest testfile version and everything should be symmetric (else post sample parameters because i don't believe you  ^-^ ).
Only when creating a "very left-eye" image from default position there are asymmetrics.

Here is the first example with some (smaller) disparities. Have I made a mistake somewhere?

Mandelbulb3Dv17{
Y.....Y/...g2...w....2....ES6UkTtzLyzmbP.0SQoX1EDyG9MoUSPzP1tSh2v5Rrzyl/Lmlz8TyD
................................2mFRe.x6wz1........A./..................y.2..U2F
..........k0..../ME//....6ED/...c1...2E2.....2e4ziX/PXoD/..........m/dkpXm1/BnAH
y.EnAnYD12../2UaNaNaNaNmzAnAnAnAnAyD.......................u1....y1...sD...../..
.wHnAnYDRCJ9tvRNYxXakWJiU7tNyC/zwB6L8nnjuOdkLkCnvqPoQo.A3O9NzMyZBWlor1djzrct3mZm
wwndgc30JL9YyIBFrRjrZFqD......293.............kD.2....sD.2E.0...................
.............ckpXm1...sD....zUkbEA.eT0l.6/821UCcEA.WV0l.c6821UgcEA......w....k1.
.....EpsZz1.......kz.wzzz1.U..6.P....M4...EB....W....61....F....8/...I1....UJl52
...U.iVFwxDE./ozPM2Tzz7lz1..mc..zXCc..k18XGQeGyD0sIRhrVAkz1............28.kFrA0.
.Ub96aAIVz9.1se7Umvxz0........../6U0.wzzz1...........s/...................E.0c..
zzzz.................................2U.8.kzzzD.................................
/6U0.wzzz1...................................2CcN/UvNP6.eeWCNq0.yRii.EJJUk1f..XR
S1.mx3CcN/UvNP6.QsLsUa3.ibhV..bTV1OK.sSq40.ly3CcN/UvNP6.MwLsUa3.ibhV.kqTV1OK.sSq
40.kz3CcN/UvNP6...EsUa3.eeWCNq0.IJ36wk8.wyLsUa3.................................
E....6....E.....I....k....kLBxaMdJrQI/5S.IaQ....................................
...................yz..............................sz.........yD................
...sz0........yD...................sz...........................................
.....................2..........0....YYPoJqNZ756ExqRZ75.........................
8..............................E........kz1........wz.........zD..........2.....
................................................................................
................................}

Jesse, any news concerning the previous post? Should I send some more data?

Greets!

I was trying out the 360 panorama setting with stereoscopic animation yesterday and found some big disparity errors.
Went back to reproduce from parameters and found that a smaller Stereo minimal distance would produce a greater disparity. (probably makes sense...)

Hope this might help in locating the problem.

Boya


Parameters from Mutant Mushroom by HalTenny

Mandelbulb3Dv17{
Y.....s/...C2...M7...2A....pw25tGkl4.BN8SivcJ75Er2sIEG9/K.AsaDMK6Zk6.Zuij8YW//0E
................................RMROLX1l4.2........Y./.................Uq/2...5F
..........k2..../M.1/....6Uc/...X3...6E4.....A0jqE90OanD/.EewsFEbJ3I0/......lJJp
z.EnAngD/2../2koj3qMhCcwzGSr1hcjl2.k6t3/K/1luz9...........U0.....y1...sD...../..
.w1...sDkNp2SE7ZfwfK.b8EGqF3zm4U8X/3eHnDV6aoX/1GUwfKk3VcjOPBzaNtc/SPmKmjwUkZpgNy
nwnSmoJScnD8zGk0UBbq6EmD...../2S..............kD.2....AE/6E.....................
.............A6U.i1...sD....zU8is3.GtWL.cb9S/Ucis3.8vWL.6j9S/U4js3EewsFEe....UZ.
.....EpsZz1.....6yzz.wzzz1kN..E......M2....C....G1...63....C....8/...sox...UJVr.
...U.W1KEyzlciszzzzzzzzzz16.oc..zzSi...MarH7iXyD0oa4dabdnz1..........2k.8.kXWF1.
.sM93P58iz9.MmnWK2zwz0........../6U0.YnFP/..fjrwdvgqzs/Fbf24LNyD..........E.mcEz
Qw/6...Fbf24LNyD.A72QifFjz9.VZPxp0FszInA8Yj9Wg/..IYOt77cK.A.s8ea00e3./kPLyRqkSlj
pAX0tvW6P...DAF3F0e3.1Uy2KVVUO/E.MGu.wrzMw1..wzzz1.E./2.zzlzzzD../2E.szDzzzz..2E
./EzTxzzz1.E./2.QsrzzzD../2E..bT9L9m..2E./.lyhQh61.E./2.MwrmpWA../2E.kqT9L9m..2E
./.kzhQh61.E./2....cU08.zz/cU08.yz1cU08.xz3cU08.................................
E....E..........I....A....kL4x4PYZaPbFJNo7LM934PZZ4N..................U.0.......
...................2./........zD...................0./.........E................
..cF.1.......E0E........kz1.....................................................
...........................3.........wZFjl4NdtqNDB4R.2qGVlKNdF4.................
...............................E........kz1.............................6.2.....
...../..........................................................................
................................/....E....k...../pKMuZaPb/WEjV5.m/..............
............5g..................OaNaNaNSoz9........szEpsZi7lU5zD................
................................................................................
..........................................E.....I....A....kLGx4RVFLN............
.....................MU/4.........................D5.HnAnAnA1m/E.......uW.2.....
................................................................................
........................................................}

Here´s another one from Hal that does not want to align.
These parameters create a lot of disparity in several places for me.

The picture is a crop from the left bottom.

Boya


Mandelbulb3Dv17{
Y.....s/...C2...ET...2A...U8Cl10ZuUD.z8m7Qb1jQ1kFN/SL19QzzHpGdfP2EIczmSaUQ2r/s1k
................................rgq8aqAjG/2........Y./..................y.2...5F
..........E9.....M.1/....6EL0...V3...2E0.....kXx70xGPwiD/.U0RC4E...m.dkpXm1.BnAn
z.EnAnYD12../2kVsKBL9t02.hMR6qXn40zD.......................u1....y1...sD...../..
.y1...sDyGOhGIxlGunNJl1Qy7ZUy0eIDOdWNwiD7bPdVHl6hvH87U2mcFQpyYpG0ZUJ5ncjKrEPzzEk
JvP5OpT4QCGvyYvxG6ppBtcD......oD/.............kD.2....sD.AE.....................
.............ckpXm1...sD....zU8is3.GtWL.cb9S/Ucis3.8vWL.6j9S/U4js3U0RC4EZ.....x/
.....EpsZz1.......kz.wzzz1Ee.M7......M2...EB....m....c3....9....l.....In..k4Z353
PU/U.W1KEyzlciszsU3YzTAe90..oc..zzSi...MarH7iXyD0oa4dabdnz1..........2U.8.UQgk0.
.wUmc2beYz1.dA8E5Exwz0........../6U0.YnFP/Upfjrwdvgqzs/Fbf24LNyD..........E.lckz
uUXA..EWD7yHGpzD............meZpwzPpzKEA8oDU.06..YKaZ6Aeiz1.58nQJ6Tnz.ERpBK.iAwj
/6V0z5FAZ.............klrlkCx9yj.6wTJHz2Vz1...8cU0.E./2.zz/cU08../2E.szDU08c..2E
./EzT/8cU0.E./2.Qs5cU08../2E..bTU08c..2E./.ly/8cU0.E./2.Mw5cU08../2E.kqTU08c..2E
./.kz/8cU0.E./2....cU08.zz/cU08.yz1cU08.xz3cU08.................................
E....I.....3....2....A....EEh3aSdtqNU6oPs/UQ..........................k/9.......
..........UNaNaNaN0wz...........aNaNaNaNiz1.....................................
................................................................................
...........................3....3....w3IjlKS4x4PYZaPb/..........................
.............................E/E................................................
................................................................................
................................0....E/...k0....BZ4SEZaPnhKOo.EOiRKA............
.................k.1Ak.1A......./......6mz1........wzoAnAnAnAnzD........Uz1.....
...sz...........................................................................
..........................................E.....I....c....EHZtqNZ7rA.I5PV/.IjR5.
..........................U/4MU/4M..0.....UaNaNaNaN/.dNaNaNaNC/E........E.2.....
...../..........................................................................
.....................................................2.....3....6....2YEjVrIhxqP
oVaFjl4N.........................Q...6..0.U.0........YNaNaNaNC.E........Uz1.....
...wz...................M.YNaNaNaNq../......../EnAnAnAnAHz1.....................
................................................................}

Maybe the only way is to make separate views in 3D navy by clicking left and right "sliding" buttons (with rectilinear lenses and checked "fixed zoom and steps") and then adjust horizontal alignment in some stereo software like StereoPhoto Maker? I have attached picture below by using such procedure.

I only hope that we are mistaken and that Jesse will find our errors.

There is no bug, it is normal that if the object is on the right side the right eye becomes a bigger projection, if you look in the middle part for a vertical shift, i bet there is no.
And 360 pano is not really working i guess, but who gots a full sphere monitor?  The FOV should fit the actual situation you are viewing...

Some years ago I thought the same. However, it does not hold for perfect lenses (no barrel or pincushion distortion). If lenses are perfect (read: rectilinear - straight lines on the object appear as straight lines on the picture), and you make horizontal and parallel shift of the camera (according to camera original position), the size of the object cannot change on the picture. Namely, camera with perfect lens (without barrel or pincushion distortion) resembles "camera obscura".

Some time ago I thought that if the absolute distance to the object becomes smaller, object must appear larger on the picture. This is not correct and I almost mislead MSc student to make huge mistake. Namely, the size of the object on the PICTURE depends only on the distance to the object in DIRECTION of the lens axis! See the first attached picture. On camera 1 and on camera 2, the size of the object on the picture is exactly the same even though the absolute distance to the object on the first camera is much shorter. However, the distance in direction of the lens axis (d) is the same!

In order to explain that, you can observe the next picture (view from above). If you parallelly change camera position, the picture width remains the same, even though in changed camera position, the ABSOLUTE distance to the object is higher. Similarly it holds for the side view, where you change e.g. camera parallel toward the viewer (you). Vertical size must remain unchanged.

This is the reason why, in stereoscopic photography and film, we must have relatively high-quality lenses with negligible barrel and pincushion distortions. In this case, the same objects must have the same sizes on both cameras without exceptions (if the cameras are parallelly displaced).

Contra-intuitive at first, isn't it? You can make the following experiment: take a camera with wider lenses (e.g. 28mm or smaller 35mm equivalent) with relatively small barrel distortion and take a picture of one vertical stick placed in the middle of the picture. Take a picture. Then, only horizontally rotate the camera (without approaching to the stick) so as that stick appears somewhere close to left or right border of the picture. Take another picture. Observe both pictures and measure the length of the stick on both pictures. They are not the same even though the absolute distance to the stick did not change! However, the distance to the stick in direction of lens axis is significantly different.

Crunchy, I have to agree with you.
In my real life 3d camera aligning I´ve never experienced that an object grows with greater interoccular distance.

Jesse. I found a motif that shows disparity near the centre.
Could you please check it out?

Mandelbulb3Dv17{
Y.....s/...C2...ET...2A...k5kvDQJ61E.P8iidPYGs1k/Pn9t3.b/.263MQ43dJozi0xtA4Oj.2k
................................3Rd7VxFEa/2........Y./..................y.2...5F
..........U8.....M.1/....6kM0...V4...2E0.....g9LQ.XqkchD/.U0RC4ElJJVz.......pW6W
y.EnAnYD12../2kVsKBL9t02.hMR6qXn40zD.......................u1....y1...sD...../..
.y1...sD/QG4h7ZQ4vfRTP6XaSTnyE5.xt0yE7hDqFn5GGxMKvfYV.17vz7NyuWzIloqgggjSVaGAc7d
yufVgq2HGgLpy4W7mYLf/bgD......2m/.............kD.2....sD.AE.....................
.............ckpXm1...sD....zU8is3.GtWL.cb9S/Ucis3.8vWL.6j9S/U4js3U0RC4Ef.....x/
.....EpsZz1.......kz.wzzz1.L..6.0....M2...UE....m....o8....F....8/...w22....SF52
...U.W1KEyzlciszzzzzzTAe906.oc..zzSi...MarH7iXyD0QwRAgHzmz1..........2k.8.kXWF1.
.sM93P58iz9.MmnWK2zwz0........../6U0.wzzz1...........s/...................E.0c..
zzzz.................................2U.8.kzzzD.................................
/6U0.wzzz1....................................8cU0.E./2.zz/cU08../2E.szDU08c..2E
./EzT/8cU0.E./2.Qs5cU08../2E..bTU08c..2E./.ly/8cU0.E./2.Mw5cU08../2E.kqTU08c..2E
./.kz/8cU0.E./2....cU08.zz/cU08.yz1cU08.xz3cU08.................................
E....I.....3....2....A....EEh3aSdtqNU6oPs/UQ..........................k/9.......
..........UNaNaNaN0wz...........aNaNaNaNiz1.....................................
................................................................................
...........................3....3....w3IjlKS4x4PYZaPb/..........................
.............................E/E................................................
................................................................................
................................0....E/...k0....BZ4SEZaPnhKOo.EOiRKA............
.................k.1Ak.1A......./......6mz1........wzoAnAnAnAnzD........Uz1.....
...sz...........................................................................
..........................................E.....I....c....EHZtqNZ7rA.I5PV/.IjR5.
..........................U/4MU/4M........UaNaNaNaN/.dNaNaNaNC/E........E.2.....
...../..........................................................................
.....................................................2.....3....6....2YEjVrIhxqP
oVaFjl4N.........................Q...6..0.U.0........YNaNaNaNC.E........Uz1.....
...wz...................M.YNaNaNaNq../......../EnAnAnAnAHz1.....................
................................................................}

Maybe the procedure used for creating stereo picture in Mandelbulb 3D is converging the cameras and therefore making keystone distortions?
The only way without generating distortions is to move the camera parallelly and afterwards crop left or right parts of the picture to set stereo window.

I understand what bothers you. Make another experiment (it will take less time). Take a 2D photography of some subject and look on that subject by arbitrarily changing the distance and orientation of the photography. Do you notice that you don’t have any problems watching the subject on photography even though the distance from the subject to both eyes differ and the subject’s physical size on the photography is the same for both eyes?

Stereoscopic picture is no exception. Any vertical/size misalignment is problematic.


I hope that you'll find enough time to make a new version.

Maybe just another picture to show relation between making pictures and viewing pictures. Note that pictures of the object (rectangle) are the same size when taking and showing.

I would think Jesse knows if his program is converging or not...
And a convergence would not yield the resulting vertical shift the near the centre of the picture in my last post.

I made a movie yesterday with the same base parameters, and all the "statues" display all have the same disparity problem.

This is true. However, convergence of cameras lead to keystone distortion. Only parallel shift is OK for stereoscopy.
It's true that keystone distortion might be corrected later on by using some special filters. However:

- you won't get perfect correction in all the cases, and
- correction will cause degradation of picture quality and sharpness.

Could you please try the new 1.798 as i told, those corrupt example pars are all made with 1.796!

Yes, it seems that version 1.798 works properly (I've tested it on 1.796). At least I cannot find any vertical disparities on my example given on the first page of this thread.

Thank you very much, Jesse!!!

Woohaa..

My first tests with 98 are all fine, also for the animation maker.

Thank you!
...

Moderator: Maybe I´m missing something, but how should I have known there is a new release?

Crunchy: I´m coming to Slovenia with a theatre peice in 3D this August.
Pm or whatever if you´re interested.

In the upper Menu (Home, Help, Search,...) there is "Downloads" button as well. I also agree that the new versions should be "advertised" more aggressively.  :dink:


Great!! Will PM you...

Sorry, i really thought that i have told it in this thread, but it must be in another one or whatever...