To Drizzle? or not to Drizzle?

[RikM]

I agree, but I think in any discussion we need to be clear about whether we're talking about the theoretical results or the empirical ones. In theory moving by a whole number of pixels won't help for the reasons already stated. Empirically it might, because it may well be impossible to guide to the level of accuracy required though different mounts may give completely different levels of accuracy. Without qualification, a statement such as "I dither guide to an exact number of pixels" might give completely the wrong impression about what is actually happening. I'm reminded that "in theory practice gives the same results as theory, but in practice it doesn't"

James

Understood and I agree entirely. I did use the word 'about' but you are right that it might have given the wrong impression Sometimes I wish I knew more about what goes on inside the software but in practise...I just press the buttons

[JamesF]

Also, although I am a bit ashamed to ask because it is mention so much, so i should know, but what is the bayer matrix? is this to do with the lay out of the sensor.

Yes, it's related to the sensor layout. In general (and there are no set rules for this) a colour camera sensor will have filters applied to each of the pixels in the sensor such that any given block of four pixels will have two that are sensitive to green light, one to red and one to blue. This is the Bayer matrix. To create a full colour image the missing colour components for each pixel are synthesised from those nearby, so if you're considering a pixel that has a green filter for instance then the red and blue components might be calculated as the average of the surrounding red and blue pixels. By this method you create a full colour image that is the same resolution as the camera sensor but without actually sampling all colours of light for every single pixel.

This is what makes the resolution of the camera more difficult to be certain of. If you think about the red or blue components of the light arriving from the target then only one pixel in four will receive that light whilst for green light only one pixel in two will receive it. Also because the colour components for the "wrong" colour pixels are generated from the surrounding data that may compromise the way that drizzling works. Drizzling relies on being able to pick "real" data out of misaligned images, but in a colour sub some of the signal has already been interpolated to fill out the missing data. On the other hand it may be that whilst there may be degenerate cases where drizzling wouldn't work, in the real world where colour transitions may be fairly smooth it works out ok. This is a point at which my "layman's understanding" of signal processing falls in a very large heap

James

[JamesF]

Sometimes I wish I knew more about what goes on inside the software but in practise...I just press the buttons

This isn't at all a bad thing. It often results in getting more work done than those who feel compelled to take everything apart to find out how it works in the first place

James

[lensman57]

This isn't at all a bad thing. It often results in getting more work done than those who feel compelled to take everything apart to find out how it works in the first place

James

Hi to James, Olly and the rest of you good people,

I have been reading this thread with great interest as I really don't much about all this and it has been quite educational, so my thanks you all and the OP. In general am I correct in assuming that drizzling would benefit or at worst it will not do much but prolong the processing time ? That is my understanding of what I have read on the Stark Lab's site. The other point that puzzels me is why is there an option to drizzle in planetary satcking software such as Registax or AS!2? Surely the consensus of opinion is that planetary images are usually oversampled by the cameras.

Regards,

A.G

[ollypenrice]

I haven't really worked out why over- or under-sampling should matter so much because in my tiny mind both ought to work. I'm a little disturbed that such a luminary as Craig Stark thinks it does matter, though.

One thing is clear with planetary; you get natural dither via the seeing. I use resample-drizzle on my solar images and the bonus is prodigious.

Olly

[JamesF]

The other point that puzzels me is why is there an option to drizzle in planetary satcking software such as Registax or AS!2? Surely the consensus of opinion is that planetary images are usually oversampled by the cameras.

This is an exceptionally good question and one that has been bothering me for some time as I'm struggling to come up with what seems like a reasonable answer. I've not even yet discounted the possibility that in such cases drizzle is actually working as a "smart" rescale that perhaps makes detail more apparent to the human eye when it's actually there all along, but just too small to pick out easily. Or perhaps as I hinted at in an earlier post, drizzle doesn't just "stop working" as you oversample, but just has a reduced effect as the sampling rate increases. If we were oversampling four or five times during planetary imaging perhaps it wouldn't make that much difference at all, but that's not exactly easy to achieve.

I find Craig Stark's assertion that undersampling is required for drizzle to actually pick more data out of the subs a reasonable one and I'm prepared to stretch the definition of undersampling to include anything that isn't at least two times oversampling because of Nyquist, but beyond that I'm rather grasping at straws.

James

[M00NMonkey]

Hi Guys,

OK, so I managed to get a bit of data Monday, only 13 x 80sec but I did a standard settings stack in DSS with bias, flats and darks. The image wasn't amazing, but I was happy with the result.

I then used exactly the same subs using the saved txt file and set 2x drizzle to on. (all other settings as standard)

When I got the result it was a, bright off white colour with streaks of colours all across it, clearly meaning something has gone wrong in the process. I played around in the DSS setting post stack and all that did was change the off white colour to which ever colour I played with.

I am sure I have missed something here and I am sure you realise that I am not familiar with the software......though I am trying to understand it all.

Any tips or advice as to what I have done wrong here?

I just wanted to compare the difference between the 2 stacked versions and see how well drizzle works. Clearly at the moment it is destroying the limited data I have.

Thanks

Jez

[M00NMonkey]

OK, well, I re-tried stacking, this time I used kappa sigma for the darks, flats and bias and I got a result. Yet to see if it better with or without 2x drizzle, once I get round to having a look I will post comparisons.

Does anyone know why it didnt work the first time i stacked using 2x drizzle, when I didnt use kappa sigma and just had the settings as normal?

Also has anyone any book suggestions on AP processing etc? explaining in depth what these settings mean, and also delving into PS too?

I had a look for books but the best i could find is dated 2007. and is fairly basic stuff.

I do have 'making every photon count' and this is a great book, but where do i go from there?

Thanks

Jez