1. CAMERA SETUP
- I’ve completed the camera matching and have the perspective etc solved (see attached)
- I’ve unlocked ‘Keep in foreground’ so I can see 3D elements over the back plate
- But if I click-off the Calibrator tag or add an object, my back plate disappears
Questions:
- How do I keep the background plate visible while adding/editing stuff?
- How do I switch between camera views after calibrating?
e.g. If I choose ‘Default Camera’ I’m shown the same view as the render camera but without the background image - plus I can’t pan/zoom etc. The same when toggling the Camera frame icon from the Objects manager - weird right?
However, if I switch to ‘All Views’, then switch back to ‘View 1’, I see the standard perspective view I’m looking for plus I can pan/zoom/edit etc, and now the camera toggle button in the objects manager works as expected.
But why can’t I switch between the render camera and the perspective camera without first going ‘All Views’ then ‘View 1’? Is this something to do with the Camera Calibrator?
2. EDITING A CUBE
- A cube has been added and has been made editable (see attached)
- I’m in Poylgon mode but hovering over the faces doesn’t make them selectable
Questions:
- How to I select different faces of the cube while seeing the backplate so I can edit the cube to form the room?
Please try the following, cmd-drag the camera (Object Manager, to get a copy of the camera and delete from the new camera the calibration tag. This should take all the problems from the scene if this new camera is active.
The default camera should show the background image as well.
From the image, I can’t tell more. I can provide an upload link if the scene needs to be kept private. With that, I can say more. Images (in most cases) do not provide enough data).
The cube is not limited by the Background Object. I have no idea what the problem is, without checking the Project (save with assets).
Please try the following, cmd-drag the camera (Object Manager, to get a copy of the camera and delete from the new camera the calibration tag. This should take all the problems from the scene if this new camera is active.
Yes, so as soon as I delete the Camera Calibrator tag from the duplicated camera, I can rotate/pan/zoom the view.
Dr. Sassi - 19 April 2022 01:48 AM
The default camera should show the background image as well.
I don’t see the background image unless the render camera is toggled on, and the camera matcher tag is clicked. As soon as the camera matcher tag is deselected, the image disappears. Is this normal? It’s a lot of pointing and clicking while trying to model.
Dr. Sassi - 19 April 2022 01:48 AM
I have no idea what the problem is, without checking the Project (save with assets).
I can’t upload the project with images because there seems to be a 4MB limit sorry.
However, I can no longer replicate the camera toggle not working so have assumed this was a glitch of some kind. Have also tried creating a new project with camera matcher but clicking the camera toggle goes to the default camera which is the expected behaviour so I don’t understand why I wasn’t able to do this previously.
QUESTIONS
Is there a way to stay in the camera view (with the matcher image) and model?
I currently switch cameras so I can select faces to model if that makes sense - can you select through faces to select faces behind front faces? Or do you always have to rotate the view to find the face to select?
Perhaps we have a different idea about the Camera Calibrator. Perhaps here we find an answer to the problem. This might get way beyond the simple use of the Camera Calibrator, so bear with me; I will try to help here. Perhaps you see it the same way, going by the solution you showed in the image. But since I write in a forum, I will answer more broadly.
Any image that wasn’t cropped at any time and which shows enough “edges” that allow for X, Y, and Z alignment (while preferably a lens profile is included) will allow recreating the “Field of View” and the position of the camera in space based on the data given into the tag.
The key idea is NOT to move the calibrated camera at all, as that will destroy the relationship to the image. In short, taking an image reduces the 3D space into a 2D image. This can be re-engineered back to provide the point of view (the position and orientation of the camera, hence the name Camera Calibrator. This point of view is crucial to creating either camera mapping or placing objects in the space.
Optically, the rotation or the change in Field of view should be possible, but not with the background object/material, as this virtual construction fills the background.
In short, the image and the camera built a unit that should not be changed. My suggestion to use a copy was more to get out of the problem that you have described, not to move the camera around. The point of interest is always in the center of the image in an uncropped image, except you use a tilt/shift lens, of course. One might even agree that it is the middle (shifted and tilted), but that is not the theme here.
The view that holds the main camera should be the render view. The “Default” Camera has no idea about the Camera Calibrator.
If you use it for camera projection, the calibrated camera will most likely be used for the projection material only and should serve as a render Camera.
The key idea is NOT to move the calibrated camera at all, as that will destroy the relationship to the image.
Yep, I definitely don’t want to move the camera after the image is matched, I understand this (I’m coming from Modo). But once the camera is matched and locked, I was trying to understand the best way of modelling to match the image (hence my confusion when toggling the matched camera didn’t work).
However, modelling within the locked camera view, often requires selecting rear faces that are obscured by faces in front.
Is there a way of selecting through faces in C4D or is the workflow to switch cameras so I can see the faces I need to select?
Ans should the camera matcher but know where the floor is and place the grid on the floor? Or is there a process for telling the matcher where the floor is?
I understand coming from a different app where you were fluent and the claustrophobic feeling while entering a new environment.
To select parts of the object, use the four-panel view or set one editor view to a “work camera”. Work camera means here a normal camera that you use to look around. Each window can represent a camera you have in the Object Manager, so one view could show the view from the back, so you can select and work in the main view again.
With small polygon numbers, you could also use the Structure Manager, set it to Polygons and click on the one (the entry line) you need.
Another way is to select all polygons in the “way” and use Select> Hide Selected, then select what you need.
I’m on L.A. time, and it is getting late, take care.
But when I add a wall (240cm high), it’s a very different size to the camera match: See here
I also received an error: ‘The file you are attempting to upload has invalid content for its MIME type’ when trying to upload a zip containing the project and image so here’s a dropbox link. Hope that’s OK.
At first, I thought this was because of the wall’s distance from the camera but no matter where I move the wall, it doesn’t match the size of the image.
Thanks for your help and time, very much appreciated.
Dr. Sassi - 20 April 2022 01:31 AM
You have set the “Pin” and with that the “0, 0, 0” point of the scene. The Cube would be in the garden, behind the large window.
OK, totally confused sorry! The tutorials I watched (and even the C4D help page) made no mention that the pin sets the 0,0,0 point in the scene. Instead, the pin is used to set the known length - at least that’s what the info I found said.
QUESTIONS
1.
So if the pin is actually used to set the 0,0,0 point, how do you set it when the 0,0,0 point is hidden behind the bed - like in the photo I have?
2.
How do you set the known length when the pin is set on the 0,0,0 point, and not on an object where you know the length?
3.
And if you use two pins (one to set the 0,0,0 point and another to set the known length) how do you tell C4D which one sets the 0,0,0 point?
Dr. Sassi - 20 April 2022 01:31 AM
I have set up the scene as camera mapping. Press play to see how that works. Frame Zero would be the camera match. If you uncheck the window wall, you can see where your cube is (red) and where it would be inside.
Thank you. It looks cool but I don’t understand how you’ve set this up sorry.
QUESTIONS
1.
I only see the calibrator tag I setup. So where are the lines and pin used to set-up the 0,0,0 point and know length?
2.
How is the picture appearing without a calibrator tag?
I feel like I understand the basic principles (have used similar features in Blender, fSpy and Modo) but am totally confused at how C4D implements them sorry.
1)
If you can’t see the 0,0,0 point in the image, you can’t calibrate that point. The point needs to sit somewhere in a spatially established and visible location. If the 0,0,0 point is 200 cm away from an, e.g., established X edge/point, then you need to make an entry accordingly to that,
2)
The Pin has only a single position; any edge in the scene on an X, y, or z-axis needs that measurement.
3)
There are no useful things to do with the two points.
1)
The camera mapping is a translation of the results of the Camera Calibration. Once translated into Camera mapping, there is no need for this Calibration Tag anymore, hence no need for the Pin.
2)
That is the material set to camera projection. As mentioned, check out the videos (link below about Camera projection).
I’m obviously missing something basic as I’m still confused sorry.
Dr. Sassi - 20 April 2022 08:40 PM
If you can’t see the 0,0,0 point in the image, you can’t calibrate that point. The point needs to sit somewhere in a spatially established and visible location. If the 0,0,0 point is 200 cm away from an, e.g., established X edge/point, then you need to make an entry accordingly to that,
But isn’t that what I did originally? I set the pin on a point in the scene I could see, and where I knew the known Y length - right? Let’s turn this around, what specific steps did you take to set my pin to the correct location?
cmscss - 20 April 2022 08:40 PM
How do you set the known length when the pin is set on the 0,0,0 point, and not on an object where you know the length?
Dr. Sassi - 20 April 2022 08:40 PM
The Pin has only a single position; any edge in the scene on an X, y, or z-axis needs that measurement.
I don’t understand how that answers the question sorry. If you can’t see the 0,0,0 point in the photo, and you have to drop the pin on a Y line that has a known length, how do you set this up properly? Hopefully your answer to the question above will solve this!
OK but that’s the page I originally followed under ‘Example of how a camera can be calibrated based on a photo’. Sorry, I feel like an idiot but I can’t see any info in those steps with regard to setting the 0,0,0 point - specifically…
“Click on Add Pin. In the Viewport, drag the Pin onto an end of one of the Y axis lines. Click on the Y axis line, which will select both the line and the Pin. Enable the Known Length Y option and enter a value of 178 cm (the squash serve line lies 1.78 m above the floor. Done.”
Dr. Sassi - 20 April 2022 08:40 PM
The camera mapping is a translation of the results of the Camera Calibration. Once translated into Camera mapping, there is no need for this Calibration Tag anymore, hence no need for the Pin.
So where is that info in the user manual? If I search for ‘camera mapping’ under R25 nothing comes up - I see the camera calibration page I’ve been using but nothing about camera mapping?
Dr. Sassi - 20 April 2022 08:40 PM
That is the material set to camera projection. As mentioned, check out the videos (link below about Camera projection).
OK, will watch tutorials on camera projection instead but ‘Camera Projection’ means something different to me.
Just to clarify, I’m not looking to project textures and move around a scene. I’m just trying to model a room/environment that matches the angle/lighting of the bed so I can keep the bed in Photoshop, and replace the room.
Do I use Camera Projection for this because I don’t want the room from the photo (or the bed) in the final render.
The 0,0,0 is the default setting for the Pin; if the Line or Grid is in the distance to your desired zero point, then the values in the Pin> Attribute Manager needs to be changed accordingly.
I did not change the Pin; what makes you think I did?
When you uploaded your version of the scene, you said… “You have set the “Pin” and with that the “0, 0, 0” point of the scene. The Cube would be in the garden, behind the large window.”
I interpreted this to mean that my cube wasn’t in the correct place (in the garden instead of in the room) because I’d set the 0,0,0 point of the scene incorrectly. Therefore, I thought you changed this which is why your geometry was in the correct place.
But if you didn’t change my camera matcher, then I don’t understand why my geometry didn’t match the scene but your did.
Yes, those are the instructions I used to set the pin at the end of the Y line.
Dr. Sassi - 20 April 2022 09:21 PM
The 0,0,0 is the default setting for the Pin; if the Line or Grid is in the distance to your desired zero point, then the values in the Pin> Attribute Manager needs to be changed accordingly.
What do you mean by “in the distance”?
As far as I can tell, I calibrated the camera using the steps outlined in the help page and if I understand your previous comments correctly, you’re saying that you didn’t change that setup when uploading your version of the scene - is that correct?
The Pin locates the solution in a specific place in the world system of Cinema 4D.
At the moment you place that on zero while that point is as well the window frame, then things behind it, going by the camera position, will look smaller.
You did everything right, except the cube was behind the window, but since the image is not physically there, it looked like the cube was inside the room. Yes, that is confusing. Hence my idea is to turn the image into a camera projection setup. So, I hoped that the behind the window (cube) would be clearer.
Distance. The Pin has 0, 0, 0 as default. If you need something to sit on zero, but you have no data on that part of the image, then you look for something that has data (e.g., a line). The distance to the zero point and where the Pin will sit needs to be reflected in the X, y, and Z parameters of the Pin. That’s why they are not hard-wired to zero and allow to be adjusted.
Again, the scene I sent back was based on the things you shared.
Thank you again for your time and the info - very much appreciated.
OK, so if you’re saying that I set the Camera Calibrator correctly, I’m way less confused.
However, if I understand you, in order to model while looking at the image (as seen here: https://youtu.be/uLEhR2hw_No?t=144) I have to perform extra steps - is that right?
I had trouble working out what the extra steps are based on the video (https://youtu.be/uLEhR2hw_No?t=144) because there’s no talking - plus it looks to be an older version of C4D.
So in R25, I create a Background Object but do I need to setup Camera Projection if all I want to do is model as per the video?
Or are you saying that the model created by the artist in that video won’t be to scale because the artist didn’t setup a Camera Projection?
You’re very welcome, and please continue to ask until this feels like muscle memory to you. What I try to do is to find ways to make it click for you. Since I know very little about you, I might fail with some explanation attempts. What works for someone will not work for someone else. But: Sometimes, the summary of failed attempts helps get an obstacle out of the way. As I mentioned, that is sometimes not an easy task, as the image is in the viewer, and one can place objects outside of the image space while it stays in front of the image. Weird, I know.
In the first Youtube link, I also had no audio, but clicking through it, it seems that the author recreated the room with a single cube. Then had the image projected onto the walls and the window.
When taking an image with a camera, we reduce the 3D space to a 2D representation. E.g., rectangles become a rhombus. If we use a slide projector and project the image to a very similar geometry from the position where the camera has initially been, the perspective distortion that turns a rectangle into a rhombus is projected into a rectangle again.
The Camera Calibration has not seen a lot of changes, so there is no problem with the version.
The function of camera projection is not dependent on a specific size, but once you have defined the size of one line in the Camera Calibrator, size is determined.
The critical relationship is the camera position that took the image and the geometry of the captured object. This creates a particular perspective. This perspective defines and requires a specific geometry. The size is convenient; otherwise, you will have trouble placing anything into the scene other than by eyeballing.
