Hi Ahmedmostafa,
Sorry, that is not how it works. In short:
The geometry is dependent on the correct camera position [POV = point of view] the Target [POI = point of interest, and the field of view [FOV = typically expressed in mm for the lens, in conjunction with the “sensor size /film gate’, or just in degree, horizontal, vertical or diagonal]. Beside the banking [dutch-angle] is needed if the image taking camera was not horizontal. Not to forget that the image format is key, as well it should not be a cropped image, so the Camera Calibrator has trouble with the wrong middle point. If one of these measurements or parameters is not set up exactly (I really mean exactly), then one should NOT start modeling.
The emergency solve would be: Solve the camera in a new file, set up a zero (XYZ point) with a measurement used for the other scene. The calibrated scene needs then to go on the one before, and be adjusted as far as possible. It is a mess to be honest to work that way. Again, this is not the way it is supposed to work, nor even remotely suggested. If you ignore the steps suggested (above), you’re on your own.
Here is an examination of your file vs a solve:
https://www.amazon.com/clouddrive/share/kxUZ3KNVneJsuE9iIJfwq2SwyTnQP27Ql1BIf7dxcgC
—-
The camera in the provided scene, while used as camera projection with the image link you provided above, is far away from even fitting format wise to the scene. I have no idea what you have used, as no material tag is in the scene at all.
The scene that you have used to set it up was provided in the thread here only to train your eye! To see how much POI, POV, FOV and all other parameter of the geometry will influence the result of an photo/image. It was not intended to be used as starting point. It was supposed to be a exploration tool, not a tool to ignore the Camera Calibration tool. Yes, things can be set up that way, and I have done it in the ‘90s that way, but I have a lot of experience with Perspective, hence why I stressed that point of my education. The Camera Calibrator simplifies this, and followed with an analytical eye, what is parallel to the axis (X, Y,Z0 and what not, will lead pretty much always to a solution, if the image has enough information.
Ahmedmostafa, I’m not sure how to solve this for you: I can explain what the tools are for, I can even go the extra mile and explain the techniques, I have tried to teach some simple photography material, on top of that, I shared a link with you, that contains many Camera-Projection tutorials. Most of that is way above and beyond what this Forum is about, content and time wise. I want that you understand these things, but I have the feeling, I don’t get through with my information, or it is not used at all.
https://help.maxon.net/us/#TCAMERAMAPPING
Read this carefully. Each step is crucial. Nothing can be changed in sequence for this, the camera is the crucial part here. To add this later, again: not possible.
Make certain that any axis in the image is really parallel to each other x to x, y to y, and z to z. If that is not done nor understood, then any modeling will not work.
If the camera is not calibrated, do not start with the modeling. There is no magical trick to fix it then, especially if the underlaying geometry is not understood.
One tip, start with a little card box, something simple, shoot it on your own. (If this example sounds not familiar, then it indicates something, so check the tutorials!) Measure where the camera was and measure the box. Measure the distance between the box and the camera. Write everything down, and get all the data of that camera and lens. This is called Set-Survey. This will help you to control your results. Do this again and again, and when it is really in your muscle memory, then try something more complex. Until that muscle memory is not there, do yourself a favor and avoid any image download from the web, where you don’t even know if it is not cropped, distorted or “whatever”. Just don’t.
To not have the camera calibrated and start modeling, with a scene that was for demo purposes only with no change,... I don’t know what to say. It seems that most of what I have shared has been misunderstood. Honestly, how often have I written that the left wall is not parallel to the right one. In your “please fix” file, the walls are parallel, no adjustments at all. There is a communication problem, and with the amount of repetition so far, I don’t know where I can improve things at this time. So, I stop here with this cat-game example. I feel really uncomfortable to use 3rd parties art work here already, this is not how it should be to begin with. If you can’t get something done with a camera, set up a scene in C4d, render it, and explore the redering if you can get the camera “back”, try all kind of situations. You need to do that.
I understand, you feel more comfortable with modeling. It is a common practice to add everything in a scene that is comfortable, but it is often only slowing one down, or disables even to get things working in the first place. I say that with certainty, having inspected a five digit amount of scene files from others so far. There is a pattern. I write here to enable for people to get things done instead of drowning.
Start with simple things. Anything else, I have run out of ideas: Except, ask MAXON for an 1on1 training so it can be explored what step is not clear to you. I guess all the tutorials here and on my YouTube site have not helped, so the 1on1 will be my last suggestion here on this Cat Game image, OK?
You have shown a good grasp of the modeling part, but something stops you with this technique, that little things needs to be found, an interactive session seems here the best idea, so far I can tell, following your progress since a while. It is doable. Start simple, that is my advice and why I think this cat game image is not a good idea
My best wishes
edited
