This might be a little specialised for the CameraLabs forum but I thought I'd share.
When "astro" images are captured most of the detail (distant stars, nebulae etc,) is still very faint even after long exposures so those images are typically stretched to bring forward the faint stuff without totally blowing the highlights. This process often loses a lot of the colour information but it is possible, using a non-standard approach
, to retain all
the colour with, to many, surprising results:
It almost seems too colourful so I decided to see if my choice of RGB filters was having an effect. Here's a graph which shows the response of my "astro" RGB filters, as modified by the sensitivity of my camera sensor, compared with the typical response of a DSLR type Bayer matrix:
The Bayer matrix filters are designed to emulate, to some extent, the response of the human eye and there's a significant overlap in the (dashed) curves for all three colours. The astro filters I use have a degree of overlap between the blue and green (solid) curves and a well separated red response.
How does this affect star colours? Well, star colour is primarily affected by the star's surface temperature (hot things are "red hot", very hot things are "white hot", extremely hot things can even look blue-ish) and the distribution of colours radiated is well understood (link
). Here's the result of calculating the theoretical colours delivered by my astro filters (column A) and the Bayer matrix filters (column B) when fed such radiation:
I used the excellent SpectralCalc tool (link
), splitting the response curves into 10nm chunks and adding the photon counts for each of the red, green and blue filters. The astro filters I use deliver greater colour saturation/discrimination than the Bayer filters which, as mentioned, are designed to mimic the human eye's response to bright coloured light. By the way, the alphanumeric code in brackets next to each temperature is a rough approximation of the star's spectral class
associated with that temperature - the Sun, not shown, has a temperature of about 5800°K (G2V) and that temperature was used to set the white-balance. For my own benefit I experimented with a change in colour saturation of the colours in column A to try and replicate the colours in column B, The result is shown in the right hand pair of columns.
How does desaturating affect real images? Here's that shot of Albireo
again with the desaturation applied:
Whether or not to apply such a desaturation to emulate a more "natural" look is a matter of taste. It isn't appropriate for images which contain a lot of narrowband emission objects (just as glowing gas clouds of hydrogen) but then such images are often in false colour anyway.