software debayering vs real mono

[vertigo262]

Thats interesting. On everything I've researched mono is far better. Just takes a lot more time to do. 

The question is even though osc is lower, will it be high enough in the future where people won't want to do mono.

so that 10 mp camera, would need to be a 30 mp camera just to match the resolution, but the issue of getting enough data and sensitivity is still an issue.

Will be interesting to see what happens when the canon 5dr comes out with 50 mp. But it still isn't cooled, and gathering the photons will be more difficult in rgb,

I think my future is narrowband and mono, but a difficult decision. But I'm not sure i've seen osc imagery comparable to narrowband. But then, there could be some osc's doing almost as good. But probably really expensive

[vertigo262]

I just thought about this.

even  though you would have a loss in resolution, and the lower QE. What if you took an slr, turned it to mono in it's firmware. provided it had that feature. And then used narrowband filters? would this work? 

mostly for cheap learning purposes?

Or would the Bayer matrix screw things up

[NickK]

I just thought about this.

even  though you would have a loss in resolution, and the lower QE. What if you took an slr, turned it to mono in it's firmware. provided it had that feature. And then used narrowband filters? would this work? 

mostly for cheap learning purposes?

Or would the Bayer matrix screw things up

You'd need to remove the bayer matrix (a thin set of colour filters) from the front of the sensor. The problem is that the firmware is used to compensating for the bayermatrix so you may find it messes things up unless you rewrite/alter the firmware.

In the end - it's actually cheaper (time/effort) to buy a mono chip..

[IanL]

Resolution is a bit of a red herring to be honest. Firstly in human vision the luminance component is much more important for detail. This means that then you shoot LRGB is is common to bin the pixels for the colour filters, i.e. each block of 4 is combined in to one bigger pixel during readout (not in firmware or software). The final image will have less colour resolution but you won't really notice it, and you get to the same SNR quicker by binning due to reduced readout noise. It is also why interpolation during debayering of OSC data works better than you might think as the loss of true colour resolution matters less.

If you are oversampling the you can also use drizzling to recover the lost resolution provided you use lots of subs that are dithered. I have tried this with DSLR data in PixInsight with a lot of success, basically you can double the resolution with no appreciable reduction in quality of the image.

You certainly can remove the Bayer filters from a OSC chip, and there is a mega thread in the DIY forum on it. Not for the faint hearted though!