Two more experimental images with ASI1600mm-cool CMOS

[Shiraz]

The M16 in Ha was 24 subs of 200seconds each. No calibration of any sort was used - it was dithered. The seeing was pretty bad (the guiding was running at about 1.2arcsec RMS_ - normally not suitable for imaging). The gain was set at 100 and an initial sub was 300seconds - however, the atmosphere was so turbulent that it messed up the stars and I changed back to 200 sec to give something like roundish stars. Although it is quite soft, was pleasantly surprised at the smoothness of the result. This is the full image - 3meg so a bit slow. Wanted to show how the 1.25 filters affect star shape in the far corners - the full 16mpix will not be available with the small filters. I am coming to the conclusion that, with a bit of care, calibration is not essential for good SNR with this camera - a little bit more light exposure may be a better use of time than calibration of large numbers of subs. Plan to do some tests and quantify how much benefit there is from calibration.

full size: http://www.astrobin.com/full/252173/0/?real=&mod=


The moon fits easily in the field of the 1600 camera with a 1000mm fl. An AVI sequence was taken at 8 bits, 5ms?, gain10? and about 1-2 frames/sec full frame. The best 25 of 100 frames were stacked in AS!2 (lucky imaging software) and finished off in PixInsight. Again, the seeing was not good, but the very short exposures had the effect of freezing the occasions of better seeing and the result is not too bad. This is not a DSO, but it seemed appropriate to post here, where current users of the camera post. Fan was running - no obvious sign of vibration effects.
small version http://www.astrobin.com/full/252168/0/
full size crop:http://www.astrobin.com/full/252168/0/?real=&mod= NB you might have to pan around to find it - there is a lot of black sky.

By the way, taking sky flats with this camera is a doddle - with a download time of about 1 second on my system, the whole process (under bright sky) can be over before the sky changes enough to require a new exposure setting.

thanks for looking - these are experimental images, so would appreciate feedback. Regards Ray

[gregbradley]

Fabulous Eagle. Impressive.

I can't see the moon image. Is the link OK?

Greg.

[Shiraz]

I think so - it is a big field and the moon is in the centre - you may need to look around a bit. Will put a link to a smaller version as well

edit : done - sorry bout that

[Atmos]

The Pillars have resolved quite well! Needs longer subs but as you've mentioned, at the mercy of conditions :)

At 8-bit, I thought this camera was capable of ~23 FPS at full res (thinking your moon shot).

[codemonkey]

Nice work Ray!

Re the 1.25" filters - what's your spacing like? I didn't realise they could affect the shape like that, I thought it was just a matter of vignetting. Is that affected by f-ratio, as vignetting is?

[Shiraz]

Like everything to do with USB3, the 23 fps is what you will get on a good day, with a tail wind and a downhill slope. What I got was a fallback to USB2 (don't know why), windows10 intent on doing it's own thing when it wanted and a few other devices sharing limited USB resources on a moderately old machine. Higher frame rates will only be available on this setup with subframe selection - at this stage have not investigated.

[Shiraz]

I had to put a shim in to keep the filter holders from rubbing against the input window of the camera - so as close as is possible. I think the effect is worse on fast scopes, but I only have f4 scopes. However, my interpretation is that the pacman shape comes about because the vignetting modifies the light cone to the extent that it no longer focuses to a fine point - it probably will not be as bad with a slower scope.

[codemonkey]

Thanks Ray. Sorry, to clarify, are you using the black adapter that's on the camera by default, which I believe makes the back focus 17.5mm, or did you remove that so you get the male M48 thread & 6.5mm back focus?

[Shiraz]

removed the adapter ring and screwed the camera straight into the wheel - with a shim.

[TR]

Very respectable Ray. You captured the structures above the beak. Very nice for a test shot.

Terry

[Paul Haese]

Nice on both counts Ray. Excellent sharp detail in both images.

[RickS]

Impressive sharpness, Ray!

[clive milne]

Ray,
I understand that you are competent with optical theory so forgive me if this sounds like a lesson in sucking eggs, (it is more for the benefit of others)





Hmmm... difficult to say, but I suspect that there might be more than vignetting and bad seeing going on there.

If it were the filters (and let's assume that the filter wheel accurately centres the filters over the chip) then any distortion would be rotationally symmetrical about the optical axis. (it isn't)

If it were solely down to seeing, wind buffeting and/or tracking error, then the distortion would be more or less uniform across the field (with the obvious aplanatic patch disclaimer) {it isn't}

As for vignetting... assuming that you are NOT sampling at a level where the airy disk is clearly resolved, and that the system is essentially anastigmatic, then the star shapes shouldn't be expanded or extended (by removing any portion of the light bundle) stars should still focus to a point (or an approximation of one at least)
By definition, vignetting is subtraction - not addition.
Ignoring the wave nature of light for a bit, consider the following allegory:
Imagine 50 target shooters set up in a line... the shooters all have the same skill or accuracy. If you drove a bull dozer on to the field, blocking any number of rifle men from seeing the target, the only difference will be that the target has less bullet holes at the end of a round... but the percentage of bullseyes and spread pattern will be statistically identical.

I guess that means that is is a function of collimation, focal plane tilt, field curvature and corrector spacing.... plus a combination of the above. So let's assume as a priori that not one of these parameters is ever ideal in the real world, the task then reduces to minimising the contribution of each one.

The best way to resolve it is iteratively, and by addressing one (and only one) variable at a time. <--- A hard earned piece of wisdom learned from many years working as a process analyser tech.


Aside from that... inspiring work Ray.
I suspect this camera might be a game changer, not withstanding its limitations.

best
~c

[Peter Ward]

The extreme corners of the field show plain old coma...or maybe field curvature (it's hard to tell them apart on simple inspection)...I certainly would not be suspecting the filters

Speckle-like noise from the sensor is also evident, not sure what steps you can take to mitigate that.

That said, the M16 resolution is remarkably good. :thumbsup:

[gregbradley]

I agree its not just the filters although the top right corner stars look clipped perhaps by the filters. The right hand side top and bottom show star distortions caused by significant tilt on the right side of your camera looking at it from behind. This sensor is about 50% larger than the 694.

I don't see any speckle like noise. Is it in the moon image which I can't open? The Eagle image looks remarkably clean on my monitor.

Greg.

[rustigsmed]

looking good Ray!
loving the detail especially!

[Shiraz]

thanks very much for all the feedback - very helpful. It has provided some impetus to chase down the issues in the field corners- although I could just crop to 14mpix and leave it at that. Agree that exactly what is going on isn't clear for now, although there definitely is some focal plane tilt. I suspect that the filters are are removing some of the light cone, so what is left is no longer from the full circular aperture and diffraction then creates the odd shapes. There could also be some coma from the CC, but I have used it with an APSc without seeing any. Mis-collimation is also a possibility that will need to be looked at, but star testing shows that to be pretty good for now.

That is the problem with f4 and 3.8 micron pixels - there is nowhere to hide any minor distortions.

Greg, did you try the lunar image again - you may need to pan about - there is a lot of black sky with nothing in it.

Peter, I had not noticed any abnormal noise in the camera even at high gain. It may be that calibration will help with the speckle that you see (yet to assess), but good old longer exposure will always help with noise - and I haven't had a clear window to do a long exposure yet (good night tonight, but family issues intrude)

thanks for the advice Clive - I will test each possible contributor individually... and, I agree that this camera may be a complete game changer, not particularly from a sensor perspective, but from the way it relaxes requirements for guiding and opens up the possibility of better resolution. Pretty good results have already been obtained by another user with a Dobsonian on equatorial platform using 1 second subs.

regards Ray

[el_draco]

Mary, mother of someone else.... What a FABULOUS moon shot!

[gregbradley]

Ah yes I can see the Moon shot now. Wow, razor sharp.

Greg.

[Peter Ward]

Hello Ray...the speckle is there alright...if it was just in the shadows I would not be fussed, but it is in the bright regions as well in your M16 shot.

Oddly... the Lunar image is very clean.

I've exaggerated it (with a high pass filter ) in the attached image.

[Shiraz]

thanks - yep that is my favourite image to date with the 1600 - when I had finished processing it, I spent about 20 minutes just looking around - had forgotten how interesting the moon is..

Ah - OK, that is the fault of the processor of the image. In order to enhance the meager high frequency components in the scene (seeing was poor), I pushed the fine wavelet sharpening hard. Thought I had kept noise under control, but not quite enough by the look of it.
regards Ray

Edit: the enhanced noise can be a easily removed from your jpeg - actually that's not a bad enhancement technique - excess high pass followed by a non-linear filter to get rid of the excess noise that the high pass pulls out.

[Shiraz]

Further to Peter's concern re the nature of the noise in the 1600 images, decided to have a closer look. The attached crops show part of a sub with linear "stretch" so that the bottom 0-4000ADU fill the screen DNR and also a 0-2000ADu version.

Then the light suddenly went on - this noise looks different to that which we normally see because there is practically no camera noise at all. This is what shot noise looks like in a narrowband image - more signal gives more noise and there is no camera noise (dark, read, FPN) to degrade the darker areas. I didn't think that I would ever see the day when we would have access to a high QE, essentially noise free camera - but here it is. And it is affordable. woohoo

[Peter Ward]

Now that's better.... looks much more like what I'm used to seeing in a well calibrated sub :thumbsup:

[strongmanmike]

So Razzor, does this mean we are moving into the era where darks and dithering will no longer be necessary...?

I want a large (30mm+) high QE, tiny pixel, noiseless one shot colour camera that produces good colour and doesn't require 100's of frames to go deep and stars don't saturate easily. I want the option of using no filters and to say good bye to multiple night imaging on a single object....please? :)...a return to the days of film when I could image as deep as anyone else in the world and on several fields a night.

Mike

[Shiraz]

but there was no calibration Peter - this is a raw sub with no corrections applied at all. Scarily, the camera really does not have any noise to speak of.



Not quite sure yet, but maybe dithering will still be necessary to get the utmost sensitivity and flats will help if you have vignetting or dust, but darks/bias etc will possibly be a waste of time.

the problem with the new cameras is that we seem to have almost reached the theoretical limit - there is nowhere else to go but bigger, but that may not be quite so bad if you only have to use 20 second subs and don't need to make 50kg of scope guide perfectly.

One possibility that might be worth pursing is to set up a second automated system to take colour while you get luminance on your main scope - an EQ6 with a medium level small Newtonian and a 1600mc should be able to give round colour stars with 30 second subs.

regards Ray

[Peter Ward]

Yes, I got that. :thumbsup:

I was referring to what my 16803 subs look like after calibration.

[Shiraz]

:thumbsup: