On 27/05/16 01:07, Mark Galasso wrote:
> Where in that tutorial do they show you the overlapping image that will
> replace the masked portion of the picture?
> specifically, I did not see any overlapping image of the couple in the
> last part of that tutorial?
> All they did was show you the couple with the "red masked" outline and
> that's it? no demonstration of using an overlapping image without them
> in it?
Hi Mark,
I'll look into the tutorial and will see if this can be made more clear.
>
> Do you have anything else tutorial wise that will be more thorough about
> how to use the masking tool?
>
> Or maybe you could provide me with some examples?
Download the images and project files at the bottom of the Masking
tutorial and follow along with the tutorial. I'm sure you'll then
understand how it's meant to work.
>
> When you speak of an overlapping image to replace/fill in the Red mask
> in a photo... do you mean that you would need to carve out/mask out the
> same area in the overlapping image
> without the part you are masking out in the initial image being masked
> and then what? Can you explain this part of the process please?
You would just carve out the unwanted part. Then blender will then look
in the overlapping image(s) for content to fill the gap.
>
> Does this mean when you are doing a shoot and you see a reflection for
> example that you want to take out later, you will need to shoot that
> same shot, but without that reflection?
Exactly.
> And if you then have a shot without the reflection you don't want, why
> would you need to do a mask on the initial photo with the reflection you
> don't want if you already have another shot with the reflection not there?
The problem is that the only way to get rid of the reflection would be
to move the camera. And moving the camera will introduce parallax and
thus give stitching errors. So what you would do is first shoot the
complete panorama from the desired viewpoint. Then move the camera a bit
and take an additional image so that you have a view of the window
without the reflection. This extra image will be included in the project
using viewpoint correction, and masking will be used to ensure that it
will only be used to replace the reflections.
Exactly the same problem occurs with the camera's shadow: you cannot get
rid of this without moving the camera, but moving the camera will
introduce parallax.
For more about viewpoint correction see
http://www.ptgui.com/examples/vptutorial.html
As shown in the tutorial the other main use for masking is to correct
problems caused by moving objects or people. This doesn't require you to
shoot any additional images, you just place masks to tell the blender
which parts of which images should be used in the overlap area.
>
> Sorry to be confusing, but I am confused about how to do this..
>
> You mentioned earlier, that JPEG could not save transparent masks...
> When you shoot in RAW can you explain what advantages you are afforded
> with the masking tool?
I was referring to the panorama file format, not the source image
format. If you mask our parts of the source images and there's no
overlapping content from other images then the panorama would have a
transparent hole at that location. With Tiff or Photoshop output this
can be represented by transparency (you would see the grey/white
checkerboard when opening the image). But jpeg doesn't support
transparency so PTGui will need to use a color (black) to fill the hole.
Joost