Stacking TIFFs in DSS - Better for StarTools?

[Scottk]

On the Deep Sky Stacker site I have seen a discussion about converting DSLR RAW to TIFFs before stacking. It's apparently
related to how pixels are interpolated while debayering. (1) Has anyone here done it this way and (2) does it yield any benefits in ST?

[Mike in Rancho]

Hi Scott,

Do you have a link to the discussion? I tried to google it, but the first thing that popped up was this "article": http://www.ianmorison.com/deep-sky-stac ... -ngc-2158/

Holy smokes, that strikes me as a whole lot of wrong going on there. Every paragraph is full of red flags.

Are you actually contemplating these to be debayered (by a different program) tiffs before stacking them in DSS? It seems that would throw off the entire calibration process - which really should be applied in the bayered state.

Of course a tiff can hold a bayered dataset as well (DSS' own masters are the obvious example), though I don't see any advantage there. I believe RAW files, FITS files, and so forth, are really just packaged tiffs at heart anyway.

[admin]

Hi,

On the Deep Sky Stacker site I have seen a discussion about converting DSLR RAW to TIFFs before stacking. It's apparently
related to how pixels are interpolated while debayering. (1) Has anyone here done it this way and (2) does it yield any benefits in ST?

Interpolation is best kept as "dumb" as possible to avoid noise bleeding into neighbouring pixels, and to avoid the creation of multi-pixel artifacts/grain.
"Reconstruction" of "detail" that isn't actually there by "smart" algorithms, greatly hurts StarTools ability to model and perform noise reduction.

DSS should now be capable of creating stacks that are suitable for StarTools from most DLSRs.

The only useful intervention I am currently aware of, is the amelioration of compression artifacts in some Nikon 5000-series cameras.

Do you have a link to the discussion? I tried to google it, but the first thing that popped up was this "article": http://www.ianmorison.com/deep-sky-stac ... -ngc-2158/

Holy smokes, that strikes me as a whole lot of wrong going on there. Every paragraph is full of red flags.

Eek! Getting some strong Roger Clark vibes from this.