When is it best to mask?

[Quantos]

Okay, thanks

[cam3d]

The benefit to masking the for the alignment stage (sparse generation) is to eliminate anything which is in subjective motion to the actual target - The only time you might want to turn off masking in the initial stages of reconstruction is if your subject is feature poor and is stable in relation to it's feature rich environment. This way, if the cameras won't align based on the subject alone, they might still align with the stable environmental context.

[Quantos]

Is there any benefit to doing masking during the sparse or dense cloud generation? Or is it better to wait for texture?