Description
This pull request adds the possibility to oversampling to the ParallelProjectionForm.
Motivation
The parallel projection for the probes was based on a simple algorithm that just integrated the quantity to be projected along a single line from the centre of each pixel of the array in the projection plane. This is not ideal if the pixels are relatively large compared to the size of the particles (in the case of smoothed particles). The new version allows for oversampling: instead of integrating along a single line, each pixel is now subdivided into N x N subpixels, and the result is the average of the integrations along all these lines. The oversampling rate N is a parameter that can be chosen by the user.
Tests
Tests were done with a simple model (a single SPH particle, no medium) and with a pixel size that is relatively large compared to the particle smoothing length. In the ideal world, an image with many photon packets and a parallel ImportedSourceLuminosityProbe should give the same result. With N=1 the agreement is not perfect, as N is increased the agreement increases.
Description This pull request adds the possibility to oversampling to the ParallelProjectionForm.
Motivation The parallel projection for the probes was based on a simple algorithm that just integrated the quantity to be projected along a single line from the centre of each pixel of the array in the projection plane. This is not ideal if the pixels are relatively large compared to the size of the particles (in the case of smoothed particles). The new version allows for oversampling: instead of integrating along a single line, each pixel is now subdivided into N x N subpixels, and the result is the average of the integrations along all these lines. The oversampling rate N is a parameter that can be chosen by the user.
Tests Tests were done with a simple model (a single SPH particle, no medium) and with a pixel size that is relatively large compared to the particle smoothing length. In the ideal world, an image with many photon packets and a parallel ImportedSourceLuminosityProbe should give the same result. With N=1 the agreement is not perfect, as N is increased the agreement increases.