M74 asi1600 (now with full process and final colour)

[stanlite (Grady)]

Hello All,

Thought i would share my experiment with a low surface brightness object with my new ASI1600 MM camera. The following image was captured over three nights due to limited imaging window.

This is a Lumanance image shot from the Northern gold coast under moderate to heavy light pollution (sky mag 19.1 or worse) Using a TS APO 102mm F7, imaging scale 1.09 arc sec per pixel

The first image is the total integration capture of 290 X 60 second subs with no bias or flat frames applied, dark's have been applied.

The second image is 210 X 60 second subs representing subs within a standard deviation of the mean. (representing subs between 1.8 arc seconds FWHM and 2.7 arc second FWHM with the mean being approximately 2.3 FWHM).

The 290min integration includes approximately 60mins of subs under heavy half moon.

No processing has been done apart from a simple DBE and histogram transformation using Pix in sight.

Images have been cropped one to one in pix in sight to allow highest resolution within limits of these forums.

Let me know what you think, i am capturing RBG tonight provided the mount holds together.

[glend (Glen)]

Grady what gain and offset settings were used, and temperture you were running? Honestly it looks fairly grainy for a ASI1600, suggesting skinny data, but 290x 60sec should be enough. Perhaps its downsize compression for the forum, do you have it in a larger size on Astrobin or Dropbox?

[RickS (Rick)]

Impressive for a dim object in less than ideal conditions, Grady. The second image has clearly better detail as you'd expect. You could always do a FHWM biased weighting and use more of the subs to get a balance between detail and SNR.

Originally Posted by glend

Honestly it looks fairly grainy for a ASI1600, suggesting skinny data, but 290x 60sec should be enough.

The ASI1600 has low read noise and decent QE. It doesn't have the power to break the laws of Physics, so shot noise from the target and sky glow is the limiting factor here.

Cheers,
Rick.

[Atmos (Colin)]

Very nice job Grady considering the conditions. You're braver person than I to attempt such a dim target for a light polluted area. I wouldn't touch it unless I'm at a dark site

Rick is correct, no camera can make it less noisy. In this case we're talking signal (Galaxy) to noise (sky glow) unrelated to camera or equipment.

[stanlite (Grady)]

Originally Posted by glend

Grady what gain and offset settings were used, and temperture you were running? Honestly it looks fairly grainy for a ASI1600, suggesting skinny data, but 290x 60sec should be enough. Perhaps its downsize compression for the forum, do you have it in a larger size on Astrobin or Dropbox?

Sorry glen I don't have Astrobin i have to get round to signing up soon. I was using unity settings, I am trying to keep everything as simple as possible setup wise to see what is achievable, once i have a benchmark i will start playing around with camera settings. As for the graininess of the image, Visually there seems to be limited downsample error from the XISF file to the Jepg mostly i assume because of the heavy crop. I did this stretch very quickly as a demonstration of two things firstly lucky imaging and secondly that this camera can perform from light polluted skies on dim targets using broadband filters. I am not surprised there is a bit of grain in the image really i am impressed it captured this much detail. I ran image states on a random raw image. The difference in average mean ADU from the background to the target is only 240 in a 16 bit image format with the background ADU being 1800 and the target being 2040. Narrow band background ADU i get with this camera is around 8-900 so that show the level of light pollution getting through.

Originally Posted by RickS

Impressive for a dim object in less than ideal conditions, Grady. The second image has clearly better detail as you'd expect. You could always do a FHWM biased weighting and use more of the subs to get a balance between detail and SNR.



The ASI1600 has low read noise and decent QE. It doesn't have the power to break the laws of Physics, so shot noise from the target and sky glow is the limiting factor here.

Cheers,
Rick.

Thanks Rick i an quite pleased with this myself actually given the conditions. I agree with the second image however, i have no idea how to perform biased weighting haha. I suppose some research is in order.

Originally Posted by Atmos

Very nice job Grady considering the conditions. You're braver person than I to attempt such a dim target for a light polluted area. I wouldn't touch it unless I'm at a dark site

Rick is correct, no camera can make it less noisy. In this case we're talking signal (Galaxy) to noise (sky glow) unrelated to camera or equipment.

Atmos my local BDSM club is shut for the week so i had to do something to keep me sane.

[SimmoW (SIMON)]

I like your humor Grady!

Thanks for posting, will review on PC when home. Yes you picked a hard target for Lum. Good results considering the target and location.

Rick, what do you mean by this? 'You could always do a FHWM biased weighting and use more of the subs to get a balance between detail and SNR.'

[Placidus (Mike and Trish)]

Originally Posted by RickS

Impressive for a dim object in less than ideal conditions, Grady.

Couldn't agree more. Very difficult goal to do a faint galaxy under moonlight and light pollution. Excellent work and well done. If you can get more data, it will be brilliant.

Originally Posted by RickS

The ASI1600 has low read noise and decent QE. It doesn't have the power to break the laws of Physics, so shot noise from the target and sky glow is the limiting factor here.

At last, someone said it! The Emperor is scantily clad !!!

Older chips like the 16803 might have a QE of around 50%, and more is always better, but Sony can't bring out a chip with a QE of more than 100%, so you can never improve shot noise on a dim object by more than a factor of two over a 16803.

Even a zero readout noise chip with 100% QE won't help by more than a factor of two or so on a very faint object. If there is air glow or moonlight or light pollution, the improvement may be negligible.

As Rick said, you can't break the laws of physics.

What a very low readout noise chip does let you do is to take relatively shorter exposures of blindingly bright things, like the core of M42 or the Hourglass in the Lagoon. That is why these chips seem at first glance to be producing miracles. The first things that people try them on are blindingly bright, air glow and shot noise are not the limiting factors, and if they can only do relatively short exposures, they get great benefit from the reduced readout noise.

Furthermore, if one's gear is capable of doing one-hour subs with no field rotation and no guiding glitches and no wind buffet, there is no compelling reason to start doing 60 second subs, no matter how low the readout noise.

But back to your excellent image, Grady, well done under difficult circumstances!

[RickS (Rick)]

Originally Posted by SimmoW

Rick, what do you mean by this? 'You could always do a FHWM biased weighting and use more of the subs to get a balance between detail and SNR.'

Hi Simmo... the SubframeSelector script lets you calculate and attach a weight (in a FITS keyword) to each sub based on the parameters for each sub (FWHM, Eccentricity, SNR, etc.) David Ault wrote a nice spreadsheet that simplifies the process, but it's still at a typical PI level of difficulty

The spreadsheet is here: https://drive.google.com/file/d/0B2J...BiZDlHR00/view

You weight the subs after they have been calibrated but before they are registered. After registration you can tell ImageIntegration to use the calculated weight instead of the default noise weighting. That means you can give priority to better FWHM and/or Eccentricity or any other parameter(s) you like.

Cheers,
Rick.

[Shiraz (Ray)]

stars are beaut Grady, but it looks like you have some fixed pattern noise that is walking across the background. maybe have a close look at your dark and bias process to see if there is something imprinting itself on the lights. It could even be pixel to pixel sensitivity variations that will respond to accurate flats, but have not observed anything like that on mine.

M&T, agree that these chips do not bypass basic physics, but they do allow some things that are simply impossible with other chips. Grady's perfect stars at 1.09 arcsec from an EQ6 are an example of what is possible when you can reach sky-noise-limited performance in 60 seconds rather than 25 minutes (for a 16803 for example- there is about a 25:1 difference in required sub lengths for the two chips, all else being equal). Of course Grady will need 25 x as many subs to get the same SNR, but he will have the resolution advantage of throwing out any short term poor seeing data and will never have any problems with field rotation if he moves his system. The 1600 ushers in a new way of doing business that changes many of the old assumptions about how good the mount must be and how long the subs must be to detect dim signals. Of course it not yet time to throw out existing high end systems, but the 1600 camera allows "high end" equivalent results to be produced with much lower spec systems (mounts in particular).

[stanlite (Grady)]

Well i finally got the colour data and i spent some time processing the image properly and i am pleased with the results (although i think the colour balance still needs a bit of work.)

I applied a new bias frame and dark's (still no flats) and this seems to have removed most of the patten noise we were seeing the colour image has had the full gambit of pix in sight processing so i will include the new mono lum as well.

I also ran the weighting you suggested Rick (changed your equation a bit though haha) and now the total integration is 261 out of 291 LUM frames without a drop in the fine details. Very impressive. so now the new image data is as so.

Lum = 4hrs 21mins at 261 X 60sec
RBG = 1hour 50mins at 120sec X 18 per channel
total = 6hours 10mins
all at bin 1 at unity on the ASI1600MM cooled Drizzled times 2 in processing

Total stacking time and pre-processing time 12hours 52mins on an i5 4670@3.4GHZ with 16Gb DDR3 RAM. Raw file size for total integration 24GB. Final image XISF file 650MB.

I couldn't get a full image on Astrobin but this is a high resolution here

http://www.astrobin.com/265761/

I have also included a new copy of the LUM to show the improved background shot noise. I like the fact i can see at least 6 background galaxies in the close up LUM image

Regards, Grady

[Atmos (Colin)]

Looking good Grady! Who says you need a large aperture telescope to do galaxy imaging

[stanlite (Grady)]

Originally Posted by Atmos

Looking good Grady! Who says you need a large aperture telescope to do galaxy imaging

sane people

[Geoff45 (Geoff)]

Very nice Grady. The colour has come up well. Pretty good for a small APO.
Geoff

[RickS (Rick)]

A very nice result, Grady!

[Slawomir (Suavi)]

A sensational image IMO