Photo Match improvements

utiler

Hi guys, I could be wrong but to my knowledge there hasn't been any improvements to Photo Match since its release in SU8 yet there are a few things that I think it would improved accuracy.

Not sure how easy it would be to implement them but I'm putting them out there for discussion

1. Option to set the horizon line to 'horizontal.
Whenever I've received a photo that I want to use in Photo Match I generally take it into an editing program and ensure the horizon is level. [see it is from the naked eye]. I don't see why we the user cannot set the horizon line in photo match to be level.

2. Option to set the Blue axis to vertical [two point perspective]
As with the horizon line, I think we should be able to set the blue axis to vertical seeing SU allows us to toggle two point perspective to the regular. We all know that Photo Match works best if the vertical axis is set close to the centre of the image but why do we accept this? Why can't we manipulate the perspective environment?

3. Scaling the blue axis
This is a real 'hit and miss'. Recently I measured then modeled a small project and when I tried to scaled the photo match to it nothing would work. I ended up fudging the model just to get it to work.
I'm wondering if we were able to use the measure tool and actually set a measurement to set the grid size? I think this would see a lot more success with matching a photo into a scalable environment in SU....

Thoughts?

JQL

I would add 1 point perspective.

Both 1 point and 2 point perspective could be set by scaling the axis perpendicular to camera or the 2 axis that hit the horizon, leaving the rest of the axis with a static scale.

This would be like manipulating FOV.

What you ask for blue axis on point 3 should be available for all 3 axis.

utiler

True, JQL. We can only hope that Photo Match gets a face lift in the next release.

JQL

I haven't used it very much latelly, I've been finding other ways of modelling my buildings from photos.

What I now do is adjust the crop photos in Krita and deform them to confine to rectangles. They don't even have to have the right proportion. They just have to be vertical and horizontal.

Then I import them to sketchup draw a rectangular shape in top of them wich I know the correct dimensions, scale the texture to that shape and sample the texture to a wider face.

Then I have a correctly scaled texture wich might be a façade or simply a door or window.

I then compose my building with that method for all sides. It's much more accurate, easy and controlable, as you don't need to work inside a static "2D" scene and in the end results you can make it look better:

Rua da Padaria - Proposta_05.jpg