search

cta-observatory / cta-lstchain

LST prototype testbench chain
https://cta-observatory.github.io/cta-lstchain/
BSD 3-Clause "New" or "Revised" License
24 stars 77 forks source link

Calculate image parameters in telescope frame and use effective focal length #988

Open maxnoe opened 2 years ago

maxnoe commented 2 years ago

This will properly calculate hillas and other parameters in degrees and if used together with the effective focal length applies the coma correction.

This should improve the off-axis performance a lot.

maxnoe commented 2 years ago

Related: https://github.com/cta-observatory/ctapipe/pull/1910

moralejo commented 2 years ago

This will properly calculate hillas and other parameters in degrees and if used together with the effective focal length applies the coma correction.

ok!

This should improve the off-axis performance a lot.

I do not think it will be the case for mono analysis. In stereo the main problem is you get the geometrical reconstruction wrong, i.e. get wrong impact parameters & Hmax. That then enters all the other reconstructions.

In mono we will get a shift in the reconstructed direction but as long as we are doing the same in data and MC (and we are, i.e. using the nominal focal length for the transformations) that will be the only bad effect. I don't think RF performance (for energy reco & g/h separation) will be affected. It is just like using different units for image parameters.

The shift in reco direction can be taken into account at higher level in the analysis and is anyway pretty small compared to angular resolution for the standard wobble offset (0.4 deg, hence 0.4*1.0466 = 0.4186, or 0.0186 deg shift)

maxnoe commented 2 years ago

@moralejo Yes, agreed for mono, but the dl1 files are also used by the combined magic + lst analysis, right?

moralejo commented 2 years ago

Yes, indeed! Also, in mono the gamma spot offset will be more significant if you go e.g. 1 deg off-axis or more. I was thinking on the effect in our standard observations, and that one will be small.