M31 again
11 views
Skip to first unread message
JR
Jan 24, 2021, 12:41:30 PM1/24/21
to croydo...@googlegroups.com
Back out again last night to set up in a brief interlude of clear skies about half 8, to try a few more experiments with the ZWO ASI533 cooled colour CMOS camera.
Getting everything set up and working proved more time consuming than expected and I didn't manage to get any organised imaging started before cloud reappeared.
I ended up with only three exposures suitable for stacking in DSS, one each at 1, 2 and 3 minutes exposure. Gain was 100, all taken at minus 10°C with the chip cooler. I used 40 darks and no flats or bias. Processing was limited to balancing the colour channels and stretching the image. Esprit 80ed f5/400 with flattener and uv/ir cut filter.
I've noticed a strange grid pattern effect that comes and goes when zooming in and out of the image. I'm not sure it's in the image but it may be. Maybe flats would reduce it. Does anyone know what causes the grid and how to get rid of it?
Despite the obvious faults in the image, I am impressed by what this camera can do.
James
Getting everything set up and working proved more time consuming than expected and I didn't manage to get any organised imaging started before cloud reappeared.
I ended up with only three exposures suitable for stacking in DSS, one each at 1, 2 and 3 minutes exposure. Gain was 100, all taken at minus 10°C with the chip cooler. I used 40 darks and no flats or bias. Processing was limited to balancing the colour channels and stretching the image. Esprit 80ed f5/400 with flattener and uv/ir cut filter.
I've noticed a strange grid pattern effect that comes and goes when zooming in and out of the image. I'm not sure it's in the image but it may be. Maybe flats would reduce it. Does anyone know what causes the grid and how to get rid of it?
Despite the obvious faults in the image, I am impressed by what this camera can do.
James
William Bottaci
Jan 24, 2021, 6:18:24 PM1/24/21
to croydo...@googlegroups.com
Hello James, nice result (though I'm not sure you needed the cooler to
get to minus 10°C last night :).
Just to clarify, did you take just three lights/subs in total, one
each for each exposure time? Also, the gain of 100, what is that as a
percentage of maximum (don't know your capture software)?
As you probably know DSS reduces noise as a (square root) function of
the number of subs (shot noise), and in addition by subtracting camera
noise using the calibration frames. Having just one each of the subs -
I don’t see how DSS can do anything with them. Also the darks have to
be of the same exposure as the stacked subs, but you have three
different times...
Am I missing something?
I've an idea on the grid pattern, but first need to know the format
and the pixel size/dimension of the capture files (are they FITS by
any chance)? It may be a debayering thing.
Also it'll help if you have the text file of the camera settings. For
SharpCap it's in the same folder as the subs, and the filename would
be the same as the first sub filename but with '.CameraSettings.txt'
appended.
For your camera do you know the arrangement of the photosite filter
colours, e.g. RGGB, RGBG etc?
I think ZWO cameras are likely to be RGGB but there's a complication
(there always is :). ZWO may read the arrangement from the top left
downwards, FITS from the bottom left upwards, and each processing
software has its own idea about what the pattern in the FITS file is.
However, if the colour looks as expected (and it does) it's not worth
worrying about, the colour.
For the grid pattern effect still need to know the format and pixel
size/dimension of the capture files, thanks.
You certainly have an enviable camera; the specifications and your
image bears that out. Yes, lots of noise but that's going to be down
to the (lack of) calibration frames. Your camera can go colder, that
should be worth it next time.
Just to mention, the cooling will affect thermal noise only, but
setting and knowing the temperature means your calibration frames will
be at matched temperatures. You can now have a practical set of master
calibration frames because they are all going to be the same
temperature. This will save you time and effort whilst imaging - you
only need to do these once (so you'll have no excuse for not using
them :).
Kevin and I were out too, some imaging (no results yet), but
fortunately no cloud for us to speak of.
We saw the stars Arcturus Spica Vega rising in that order. These are
summer stars. Nice thought, whilst scraping ice off the equipment.
Thanks for sharing.
William
get to minus 10°C last night :).
Just to clarify, did you take just three lights/subs in total, one
each for each exposure time? Also, the gain of 100, what is that as a
percentage of maximum (don't know your capture software)?
As you probably know DSS reduces noise as a (square root) function of
the number of subs (shot noise), and in addition by subtracting camera
noise using the calibration frames. Having just one each of the subs -
I don’t see how DSS can do anything with them. Also the darks have to
be of the same exposure as the stacked subs, but you have three
different times...
Am I missing something?
I've an idea on the grid pattern, but first need to know the format
and the pixel size/dimension of the capture files (are they FITS by
any chance)? It may be a debayering thing.
Also it'll help if you have the text file of the camera settings. For
SharpCap it's in the same folder as the subs, and the filename would
be the same as the first sub filename but with '.CameraSettings.txt'
appended.
For your camera do you know the arrangement of the photosite filter
colours, e.g. RGGB, RGBG etc?
I think ZWO cameras are likely to be RGGB but there's a complication
(there always is :). ZWO may read the arrangement from the top left
downwards, FITS from the bottom left upwards, and each processing
software has its own idea about what the pattern in the FITS file is.
However, if the colour looks as expected (and it does) it's not worth
worrying about, the colour.
For the grid pattern effect still need to know the format and pixel
size/dimension of the capture files, thanks.
You certainly have an enviable camera; the specifications and your
image bears that out. Yes, lots of noise but that's going to be down
to the (lack of) calibration frames. Your camera can go colder, that
should be worth it next time.
Just to mention, the cooling will affect thermal noise only, but
setting and knowing the temperature means your calibration frames will
be at matched temperatures. You can now have a practical set of master
calibration frames because they are all going to be the same
temperature. This will save you time and effort whilst imaging - you
only need to do these once (so you'll have no excuse for not using
them :).
Kevin and I were out too, some imaging (no results yet), but
fortunately no cloud for us to speak of.
We saw the stars Arcturus Spica Vega rising in that order. These are
summer stars. Nice thought, whilst scraping ice off the equipment.
Thanks for sharing.
William
JR
Jan 25, 2021, 5:35:25 AM1/25/21
to croydo...@googlegroups.com
Thanks for your comments and questions William. Dedicated astro cameras are an interesting expedition for me at the moment. DSLRs seemed a lot simpler. Any help is appreciated.
On the minus 10 I finally got myself rigged up with an active (has its own power supply) 12 metre usb extension lead from the garden into the house where I can stay warm apart from polar aligning. Alas they haven't yet invented anything automatic to tweak the bolts as needed with a setup nightly rig. The cable even seems to work fine taking a trial video of the moon. What a difference it makes. I don't mind scraping ice but not when its forming on me!
The asi533 has gain from 0 to somewhere around 400. ISO in general photography seems straightforward. Astro CCD cameras have a fixed gain setting. In CMOS you get to choose any number you fancy, but more systematically, between three main options: HDR, highest dynamic range in stops - gets you the subtle light variations, Unity gain (offering output = input, not sure why this is good but sounds 'pure'?), and LRN, lowest read noise.
You can't have HDR and LRN at the same time of course, though you'd like it. ZWO tackles this problem empirically by testing all three gain options with the same camera, same telescope and same object, presumably in the same place and at the same time (ie sky conditions). See the link Trevor very helpfully circulated.
As a rule of thumb, ZWO concludes that HDR is best for fewer longer exposures and LNR is best using many, often enormous numbers, of short stacked exposures, often implausibly short. ZWO camera drivers allow you to set your option choice up at a click, using gain numbers based on the characteristics of your particular camera chip. Dr Robin Glover works out the answer for you in his Sharpcap software.
There are practical considerations too. The LNR option has the advantage that you don't need an expensive mount or excellent guiding, not that you wouldn't like both. Quite possibly due to global warming, equally important is that long exposure opportunities have become an endangered species, especially using monochrome and filters.
The 533 chip is a bit more complicated yet, as it has circuitry which kicks in at gain 100 to increase dynamic range and reduce read noise to levels similar to gain of zero. See the graphs below.
There were only three frames. DSS just seems to stack them, though it did give a message about incompatible formats. I reckoned that adding up a 1 minute and a 3 minute frame approximates to two 2 minute frames in terms of photons, to add to the real 2 minute sub. I doubt it does with sq roots in the equation but I judge by results and the alternative of waiting here seemingly forever for a night long clear sky.
As far as darks go, I fudged it by taking 100 second exposures. Some software adjusts for darks exposures that don't match lights. Nebulosity by Dr Stark does. I somehow doubt that DSS does. I don't think it makes a lot of difference with low noise CMOS cameras especially when cooled. I also thought like you and went for the advantage of getting a library set of 40 at the same exposure and temperature. I have had DSS reject master darks, but that might be because they came from a different software.
I wondered about debayering but the chip is definitely RGGB and that's what I used in the DSS options. As you say, debayering softwares are not uniform in how they apply a given pattern. I'd be surprised if DSS has it wrong. ZWO must by now be the most numerous make out there and a new version of DSS came out recently which I was using. The best test would be taking some daytime shots and see how they come out debayered
I've managed to take a screen shot of the grid on my ipad. It comes and goes as the extent of zooming in and out changes. Maybe it's the ipad screen. I have had grids before but due to debayering and they are there all the time. Chip resolution is 3008 x 3008 square pixels of 3.76μm.
The 533 colour chip has some excellent features, a pixel scale good with small refractors, low noise and low noise reducing circuitry, high sensitivity back lit, and zero amp glow - no need for darks according to ZWO! Sony has abandoned CCD development and the 533 is no doubt the start of yet better to come. It's born of security camera usage (the square chip is designed to be used with fish eye lenses that produce circular images and don't need the wasted area from a rectangular format). A monochrome version would be quite something.
James
On the minus 10 I finally got myself rigged up with an active (has its own power supply) 12 metre usb extension lead from the garden into the house where I can stay warm apart from polar aligning. Alas they haven't yet invented anything automatic to tweak the bolts as needed with a setup nightly rig. The cable even seems to work fine taking a trial video of the moon. What a difference it makes. I don't mind scraping ice but not when its forming on me!
The asi533 has gain from 0 to somewhere around 400. ISO in general photography seems straightforward. Astro CCD cameras have a fixed gain setting. In CMOS you get to choose any number you fancy, but more systematically, between three main options: HDR, highest dynamic range in stops - gets you the subtle light variations, Unity gain (offering output = input, not sure why this is good but sounds 'pure'?), and LRN, lowest read noise.
You can't have HDR and LRN at the same time of course, though you'd like it. ZWO tackles this problem empirically by testing all three gain options with the same camera, same telescope and same object, presumably in the same place and at the same time (ie sky conditions). See the link Trevor very helpfully circulated.
As a rule of thumb, ZWO concludes that HDR is best for fewer longer exposures and LNR is best using many, often enormous numbers, of short stacked exposures, often implausibly short. ZWO camera drivers allow you to set your option choice up at a click, using gain numbers based on the characteristics of your particular camera chip. Dr Robin Glover works out the answer for you in his Sharpcap software.
There are practical considerations too. The LNR option has the advantage that you don't need an expensive mount or excellent guiding, not that you wouldn't like both. Quite possibly due to global warming, equally important is that long exposure opportunities have become an endangered species, especially using monochrome and filters.
The 533 chip is a bit more complicated yet, as it has circuitry which kicks in at gain 100 to increase dynamic range and reduce read noise to levels similar to gain of zero. See the graphs below.
There were only three frames. DSS just seems to stack them, though it did give a message about incompatible formats. I reckoned that adding up a 1 minute and a 3 minute frame approximates to two 2 minute frames in terms of photons, to add to the real 2 minute sub. I doubt it does with sq roots in the equation but I judge by results and the alternative of waiting here seemingly forever for a night long clear sky.
As far as darks go, I fudged it by taking 100 second exposures. Some software adjusts for darks exposures that don't match lights. Nebulosity by Dr Stark does. I somehow doubt that DSS does. I don't think it makes a lot of difference with low noise CMOS cameras especially when cooled. I also thought like you and went for the advantage of getting a library set of 40 at the same exposure and temperature. I have had DSS reject master darks, but that might be because they came from a different software.
I wondered about debayering but the chip is definitely RGGB and that's what I used in the DSS options. As you say, debayering softwares are not uniform in how they apply a given pattern. I'd be surprised if DSS has it wrong. ZWO must by now be the most numerous make out there and a new version of DSS came out recently which I was using. The best test would be taking some daytime shots and see how they come out debayered
I've managed to take a screen shot of the grid on my ipad. It comes and goes as the extent of zooming in and out changes. Maybe it's the ipad screen. I have had grids before but due to debayering and they are there all the time. Chip resolution is 3008 x 3008 square pixels of 3.76μm.
The 533 colour chip has some excellent features, a pixel scale good with small refractors, low noise and low noise reducing circuitry, high sensitivity back lit, and zero amp glow - no need for darks according to ZWO! Sony has abandoned CCD development and the 533 is no doubt the start of yet better to come. It's born of security camera usage (the square chip is designed to be used with fish eye lenses that produce circular images and don't need the wasted area from a rectangular format). A monochrome version would be quite something.
James
JR
Jan 25, 2021, 5:38:27 AM1/25/21
to croydo...@googlegroups.com
The ZWO graphs don't seem to be emailable. The link is here
Trev S
Jan 25, 2021, 8:48:44 AM1/25/21
to croydonastro
Hi James, that's certainly an improvement.
With regards the grid problem, what capture software are you using? If Sharpcap, do you have the Colourspace set to "RGB24"? I find that the RAW8/16 settings with my ASI290MC give a grid pattern something like the one you are experiencing but RGB24 gives a good colour image. Otherwise, as William says it does look like a debayer problem.
Trev S
Mar 26, 2021, 4:58:04 PM3/26/21
to croydonastro
Hi James, I don't know if you got anywhere with it, but looking into this debayer issue with RAW files, AstroPixelProcessor has a set of options to apply the debayer pattern (e.g. RGGB) to RAW images in order to process them in colour. I assume PixInsight has something similar. It seems that you can get better quality by capturing RAW images, then process them afterwards. The main issue seems to be that they cannot be viewed in most image software without showing the chequerboard pattern. Some more explanation can be found here
JR
Mar 27, 2021, 6:45:31 AM3/27/21
to croydo...@googlegroups.com
Hi Trev
Thanks very much for sending that useful link on debayering and (also on what the different formats are and their advantages, having wondered about that for ages). Just goes to show one should read the instructions.
I haven't got any further on the grid problem, but I have noticed it occasionally on other photos I've downloaded not taken by me. I wonder if it's something to do with the way the ipad displays images. It dates from 2015 and must be old tech. The grids happen only at particular enlargements. Even slightly smaller and larger and there is no grid. I also don't get the effect at all on my laptop.
I have had grids before notably when I started taking solar images with the ZWO 120, but discovered then from the web that it was to do with debayering, as mentioned by the Sharpcap link. I use DSS for stacking and debayering, though there I failed at first to realise you had to tell the software that images need debayering. After that no problems.
Pixinsight does have an integration script, ie stacking, where again you have to specify the Bayer matrix layout. You can do all sorts of fancy things like give colour channels different weights. It's very good at identifying lights, darks, flats etc automatically. You just point it to a Windows directory with all the data and it does the rest. So far I have stuck to the default settings.
I then got very strange results and it turned out there was some bug in the latest update of the Pixinsight integration process. That was fixed very quickly but found Pixinsight took considerably longer to do what DSS did in 5 minutes, same data, same laptop. It seems to do a great deal of statistical estimating to line things up. The results did look slightly better, but nothing major compared to the differences in end product achievable by different processing settings.
James
Hi James, I don't know if you got anywhere with it, but looking into this debayer issue with RAW files, AstroPixelProcessor has a set of options to apply the debayer pattern (e.g. RGGB) to RAW images in order to process them in colour. I assume PixInsight has something similar. It seems that you can get better quality by capturing RAW images, then process them afterwards. The main issue seems to be that they cannot be viewed in most image software without showing the chequerboard pattern. Some more explanation can be found here
--
You received this message because you are subscribed to the Google Groups "croydonastro" group.
To unsubscribe from this group and stop receiving emails from it, send an email to croydonastro...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/croydonastro/6928dcf1-8f60-48e1-ae6f-ce372892da73n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages