Out of memory during texturing

Hi all,

I’m experiencing some problems texturing a model as I’m getting a “run out of memory” message.

The model has 20M faces computed using a laser scan and I’m using 220 images (6Kx4K) for the texturing step.

My computer has 32GB RAM. However, something sounds weird to me as I’ve already create other models with same number of faces, a higher number of images of the same size, and the same unwraping configuration (adaptative texel size).

Where do you think the problem can be? Any idea on how to solve it?

Thanks in advance!

Hi Imanol, 

Please try close the application, and clean cache and try again with the texturing.
Please check your mesh topology as in some cases misalignment caused similar error. 

Is there also something added to this error message or is it just out of memory?

Can you try these texturing settings:

Hi Ondrej,

I’ve checked your configuration after cleaning the cache and I got the same result. No other message is obtained (even in the console), only the application has run out of memory.

Checking the alignment, there are some cameras that are aligned weirdly (like out of the model). I’ll disable them and texture again.

I’ll come back with news.

Imanol

The cache requirements for texturing are 500 bytes per triangle, so the 20 million triangle model should only need 10 GB of cache. The problem cold be the windows virtual memory is too small and that is causing the application running out of memory. Try to raise the page file to higher value.

Hi,

I’ve disabled the cameras that were aligned wrongly (far away from the model) and now it works with your parameters… So I will try to realign these images. I used control points as they were part of a small component but, as I can see right now, this is not working…

Imanol

Hi, that is a good new. Try to use there more control points. Are these images taken with another camera? Or are they blurry?

Hi,

I will try it this afternoon. 

They are taken with the same camera but very near to the surface. They are very sharp. However, there isn’t much overlapping between them (the photograph ignore our capturing protocol)…

Thanks!