ZW0 1600 LRGB

[Aidan]

Can I just clarify a couple of points with LRGB, Gain & ADU on the 1600 pro.

Just setup and set it off running for the first time doing LRGB.

Gain at 76   Offset at 50

Lum @ 35 secs is giving me a mean reading of 1442 in SGP

Does that sound about right?   

RGB are set for 60 sec exp with the same gain & offset.  Should I be going for the same mean reading?   Just started to pull the reds in... ADU is at 1088.

[Dr_Ju_ju]

I usually set gain @ unity i.e. 139, no off set, with rough timings of 3 mins for luminance\green & 5 mins for red\blue, all handled by SGPro...

[Aidan]

Its the ADU which i am really trying to figure out... and checking i'm around the right numbers.   I can alter the exposure times accordingly.  I did try 139, but 76 seemed to bring in a better image for me

[alexbb]

I use the _half_unity_ too, but I set the offset to 15. I expose 120s for each LRGB. If there's anything very burned, I also expose maybe 20-30 frames at 20s for each LRGB which I then apply later to the burned areas.

[symmetal]

For my 1600 I aim to expose so the background median value is around 10 X Read Noise ^ 2.  I use unity gain for RGB and 1/3 unity gain for L so that the exposure times are the same and the L has a higher dynamic range. At unity gain (gain 139, offet 21) the 10RN^2 ADU is 826. At 1/3 unity gain (gain 44, offset 12) the 10RN^2 ADU is 553. (using my dubious calculations)

For me 120 seconds gives me background ADUs around these figures. I have pretty dark skies, though normally cloudy. 

With you getting 1442 ADU at 1/2 unity for L in 35 seconds you must have more light pollution. It's generally accepted that exposing longer than the 10RN^2 gains little benefit. For you that gives exposures around 20s which does seem rather short, though several people on here take exposures around these figures and take loads of frames.

Alan

[Aidan]

Ok,  starting to understand that a bit....   I need to sit down and really get to understand the equations!     So, at the moment it looks like i am over exposing - a quick test showed that at 20 secs brings the ADU down quite a bit.   I'm going to leave the current image run going on the figures i have at the moment and next run ill aim for shorter subs etc.

Just out of interest, where you say about changing the gain for L and RGB, is that something i can automate in SGP?   At the moment i am just setting the values using the "spanner" in the main sequence. 

[symmetal]

1 hour ago, Aidan said:

Just out of interest, where you say about changing the gain for L and RGB, is that something i can automate in SGP?   At the moment i am just setting the values using the "spanner" in the main sequence. 

You can kind of automate it in SGP. On the Sequencer Panel if you right click on the cogwheel in the Event column you can set the camera gain for that event. You can't set the offset though. I assume it uses the offset set in the 'Equipment Spanner'. In the 'Control Panel' Camera tab you can set global gain settings for each binning option. I think if you set a value in the Sequencer Panel it overrides this global setting. If set in the Control panel or Sequencer Panel I assume it overrides the 'Spanner' gain setting though I'm not totally sure. I don't use them.

I do set it only in the Equipment spanner though like you, and have the Control panel Camera tab settings left at 'Not Set' as well as the sequencer event gain options. As you can't change the offset I don't find it all that useful, although the latest thoughts are to leave the offset at 50 for all gain settings and lose a bit of dynamic range at low gain settings. Not me yet though.  I put the gain and offset values I select in the 'Suffix' box ( eg. g139-21)and include that in the file name so it's easy to see what the settings are for each exposure. You have to remember to set them in the 'Spanner' first though for that event.

Alan

[Aidan]

Thanks Alan,  I’ll see how things go with offset settings.  Perhaps I’ll post up on the SGP forum to see if they can implement offset switching.

Great tip on the suffix for recording the settings in the file name.!

[symmetal]

6 hours ago, Aidan said:

Perhaps I’ll post up on the SGP forum to see if they can implement offset switching.

The offset settings are not available via the Ascom driver so until the driver gets updated there isn't much SGP can do.

I believe that the Ascom framework which the drivers are based around doesn't implement Camera Offset settings so it's not just a case of modifying the ZWO driver.  There was talk of possibly using an unused Ascom camera parameter and treating that as the offset but that gets a bit messy. I recall reading about that on a forum so am not presenting it as gospel.  

Alan

[symmetal]

There are quite a few topics on the CN forum discussing the ASI1600 read noise and background ADU. This one should get you started on the theory Aidi, and the John Rista calculations are what I use for my settings.

Actually, reading John Rista's post 25 on this thread on CN it's looks better to use higher gain to reduce the quantization noise, particlarly for luminance where the posterisation of background noise is particularly noticeable compared to a 16 bit CCD. So for me, 3 exposures at 40s (unity gain 139) gives lower quantization noise than 1 exposure at 120s (1/3 unity gain 44). Dynamic range is reduced at unity gain though I personally find the background 'posterised' noise more objectionable on the 1600 compared to a 16 bit ccd. Hence why so many subs are taken to reduce this 12 bit quantization noise. From John's post to reduce the quantization noise to the equivalent of a 16 bit camera needs 250 to 300 subs. 

It's all a trade off as to what option you choose but I still find the images from my Atik CCD much cleaner background noise wise than the ASI1600. 16 subs of 10 mins on the Atik is much less stressful than 250 subs of 30 seconds on the 1600 for equivalent background noise. I was taking 64 or so subs with the 1600 of a minute or so assuming that was enough but it seems it's not. The background noise was still blocky. 

I've been wondering whether it could be beneficial to add 4 bit noise to the 12 bit subs converted to 16 bit and before stacking,so that less subs can be taken for a subjectively 'cleaner' image.  Similar to adding 'dither' noise to the input analog signal to Analog-to-Digital converters to reduce the quantization error. I'll have to do some tests. 

Alan

[Aidan]

Thanks again Alan,  Ill start digging into this tonight.  Looking clear for some imaging, so a good time to start getting my head around it!     Having said that, i run the first set of LRGB i took of M31 the other night through PI and i was pretty pleased with the results to be honest.   I was suprised at the level of detal with just the following amount of data:

L @ Gain 76 Offset 50 - 30 subs at 30sec each
R @ Gain 76 Offset 50 - 20 subs at 90 sec
G & B @ Gain 76 Offset 50 - 20 Subs each at 60 sec

Probably could be much better when i understand more on the technicalites. 

[MartinFransson]

This is an interesting discussion! My problem is I´m using Astro Photography Tool and if I try to change the settings (gain etc.) it needs to disconnect the camera first. When doing so it warns about thermal shock since the cooling is turned off. So if I understand it right, I can only use one gain setting per imaging session.

Guess I´ll stick to unity...

[blinky]

dont worry about thermal shock, somebody asked ZWO and he said it was not an issue

[symmetal]

Actually Aidi reading through this lengthy thread on the SGP forum it looks like gain 76, offset 14, (half unity) is considered optimum for LRGB imaging and gain 200, offset 50 (twice unity) is considered optimum for narrow band imaging. This balances dynamic range, camera noise, read noise and quantization noise (12 bit). At low gain quantization noise becomes very noticeable unless you take 100s of subs, which for long exposure narrowband would take forever. At gain 200 the quantization noise is much reduced per sub at the expense of dynamic range so less subs need to be taken. There are some images in the thread illustrating these factors.

For me the 12 bit  A-D quantization noise is the most annoying, giving blocky posterised backgrounds and at gain 76 it is significantly improved over gain 0. You can take many short L subs relatively quickly to improve this further (long enough to get the background ADU up to around the recommended values) .

I'll have a go at using these settings to see how it works out. Your using offset 50 at gain 76 just causes a little loss of dynamic range compared to offset 12 so if you want to keep it at 50 all the time no problem though from my post below the background ADU values are different.

Martin, using these settings, unless you're switching between LRGB and narrowband in the same session you won't need to change gain during your run. 

Alan

[symmetal]

I made this Excel sheet calculating the background ADU levels for various gain, offset settings using different noise swamping calculations. 20 x RN^2 is commonly used but others are presented for comparison. I've also redone it using offset 50 for all gain settings so as you can see the background ADU settings need to be higher.

Constant offset 50

At offset gain 76 offset 50 the figure you had in your first post Aidi  on this thread of 1442 wasn't too far off the mark. 

Alan

 

[Aidan]

Hi Alan,

Just after a brief look at the charts, i'm starting to make sense of things....   Might take a bit longer than most to digest, but ill get there!