Hey All,
I am a 3d printing enthusiast, but have never been great at modelling anything organic. However, using VAM I can make almost any version of a human I care to. That said, I know I can export the edited model as an obj at low poly resolution, then upscale with smoothing in something like blender, which I have done, but the export doesnt have hair, clothes etc.
I was thinking about using reality capture (a program you can feed a bunch of images of something 3d and it will reconstruct the high res mesh) - has anyone tried this?
If someone can validate the theory, I want to try to make a plugin. Essentially load the mode, pose, features you want, the plugin will render it at a high resolution (like supershot) and rotate around taking a number of frames the user can set as a variable (along with angles variables and what not). Once done you would have a sufficiently large dataset to import into Reality Capture and turn into an HD mesh. The benefits over standard OBJ export are:
- Higher quality (Extremely high quality is possible with enough resolution and frames)
- Inclusion of hair and clothes
- Malleable enough to take advantage of your PCs processing (more frames = more data = higher quality - the sky is the limit)
- Can include other objects model is interacting with (holding items, sitting on items etc)
So, my 2 questions:
- Has anyone done this or something similar? Did it work? Additional considerations I should make?
- If I pursue this, is anyone interested in a plugin? Is anyone interested in getting access to the OBJs/STLs that I make?
Thanks in advance
I am a 3d printing enthusiast, but have never been great at modelling anything organic. However, using VAM I can make almost any version of a human I care to. That said, I know I can export the edited model as an obj at low poly resolution, then upscale with smoothing in something like blender, which I have done, but the export doesnt have hair, clothes etc.
I was thinking about using reality capture (a program you can feed a bunch of images of something 3d and it will reconstruct the high res mesh) - has anyone tried this?
If someone can validate the theory, I want to try to make a plugin. Essentially load the mode, pose, features you want, the plugin will render it at a high resolution (like supershot) and rotate around taking a number of frames the user can set as a variable (along with angles variables and what not). Once done you would have a sufficiently large dataset to import into Reality Capture and turn into an HD mesh. The benefits over standard OBJ export are:
- Higher quality (Extremely high quality is possible with enough resolution and frames)
- Inclusion of hair and clothes
- Malleable enough to take advantage of your PCs processing (more frames = more data = higher quality - the sky is the limit)
- Can include other objects model is interacting with (holding items, sitting on items etc)
So, my 2 questions:
- Has anyone done this or something similar? Did it work? Additional considerations I should make?
- If I pursue this, is anyone interested in a plugin? Is anyone interested in getting access to the OBJs/STLs that I make?
Thanks in advance