calibration frames with asi1600mm cooled

[Wirral man]

just wondering how the calibration files work on a cmos sensor and how do i go about it.

if the camera is cooled are darks the same or can i alter the times to make a set of different times,

bias do i still use the fastest exposure if so how do you do that with the 1600 as no shutter speed?

flats again how do you go about doing these with gain and offset etc..

how many frames do you take or need and also with lrgb do you have to use all calibration files with each channel for stacking also does that apply to narrowband images as well can this all be done in apt 

thanks in advance

John  

[geordie85]

For flats I use the flats aid in APT then select run, I then run the same routine but with my scope covered and capped for dark flats.

For bias you just set the exposure time for as short as possible.

For darks you just run the same routine as you did on the night of data capture. The added bonus of having set point cooling is that you can build dark libraries for all the exposure times you'll use. I usually do this on cloudy nights.

When using my mono atik 383l I'd only take 2 sets of calibration frames. 1 through my luminance filter and the other exactly the same, except they'd be 2x2 binned.

I never had any issues using luminance flats cor my coloured filters.

[vlaiv]

With ASI1600 you need following calibration files:

- darks - taken at exact same settings as light subs - Gain, Offset, Duration, Temp

- Flats - depending on your flat panel / source - go for lowest exposure time that will get you to about 80% of histogram. You can change Gain, Offset settings, but I prefer to use the same as for lights. My typical exposures are in milliseconds (~5ms for Lum) - I have really strong flat panel.

- Dark flats - exact same settings as your flats.

Don't use bias frames with ASI1600 - bias is not stable, you will only have trouble calibrating with it and you can do full / proper calibration without it.

Calibration would be: master dark - stack darks, flats - stack flats, flat darks - stack flat darks, master flat - subtract stack of flat darks from stack of flats. Calibrated frame - subtract master dark from sub and then divide by master flat.

If you include this files only DSS will do this for you automatically. I suspect other software will be able to properly handle this calibration scheme.

Not using bias files means you can only use matching darks. So build your dark library out of exposures you are likely to use, and calibrate with exact matching master dark (like 1 minute dark at Gain X, Offset Y, or 4 minute master dark at Gain Z, Offset W). You can't do dark scaling because of this, and you can't use dark optimization. If you choose to use new exposure length for a particular night - you can shoot darks of matching length later (just keep everything else the same for those darks).

Cooled cam means you can take your darks separately - just make sure you set temperature that you are going to use for lights (maybe even have two sets of darks, one for summer, one for winter, or better re shoot your darks every 5-6 months). You can do it during the day, or on cloudy night. If you take your camera off the telescope, make sure that you prevent "unseen" light that can penetrate plastic cap (like IR and UV) - place aluminum foil over top, or place camera cap down on some transparent surface like wood / stone table.

If you have filter wheel and permanent setup - you can shoot flats and have flat library (one master flat per filter). If you have "mobile" setup or you change filter orientation / fiddle with tube you need to reshoot your flats. I take my flats for every filter change because I use filter drawer and I'm "mobile". It is fairly quick because fast download times (10-15 mins per flat + flat darks - if I image with more than one filter per night - I use only one set of flat darks when applicable - exposure for flats must be good across filters).

As for number of each of calibration files - I'm a bit fanatical. The more - the better applies here and I do take it seriously - I shoot 256 of each.

[wimvb]

With my ASI174MM Cool, I don't bother with bias frames. I shoot darks that match the lights in temperature, gain and time. I shoot flats at 0 gain, but the same temperature as the lights. Then I shoot dark flats at the same settings as the flats. I calibrate the flat frames with the dark flats, integrate to create a master flat. I integrate the dark frames to create a master dark. Finally, I calibrate the light frames with the master dark and flat.

I take 30 flats, and 30 dark flats (these are fast, because the exposure time is at most 1.5 seconds). The flats have an average signal of 25000 ADU (even if the camera is 12 bit, the files are 16 bit). I take about 20 darks, usually the next cloudy night after I do the light frames. With a cooled camera, you can create a dark library, but I haven't standardised exposure times and gain yet. So I shoot darks after each imaging session.

Btw, the fastest exposure time of most if not all ASI cameras is 32 microseconds. That information is in the specs on the ZWO website.

[Wirral man]

think i need to take my time reading these answers lol

ok will get these done yet again thanks 

[Wirral man]

1 hour ago, vlaiv said:

With ASI1600 you need following calibration files:

- darks - taken at exact same settings as light subs - Gain, Offset, Duration, Temp

- Flats - depending on your flat panel / source - go for lowest exposure time that will get you to about 80% of histogram. You can change Gain, Offset settings, but I prefer to use the same as for lights. My typical exposures are in milliseconds (~5ms for Lum) - I have really strong flat panel.

- Dark flats - exact same settings as your flats.

Don't use bias frames with ASI1600 - bias is not stable, you will only have trouble calibrating with it and you can do full / proper calibration without it.

Calibration would be: master dark - stack darks, flats - stack flats, flat darks - stack flat darks, master flat - subtract stack of flat darks from stack of flats. Calibrated frame - subtract master dark from sub and then divide by master flat.

If you include this files only DSS will do this for you automatically. I suspect other software will be able to properly handle this calibration scheme.

Not using bias files means you can only use matching darks. So build your dark library out of exposures you are likely to use, and calibrate with exact matching master dark (like 1 minute dark at Gain X, Offset Y, or 4 minute master dark at Gain Z, Offset W). You can't do dark scaling because of this, and you can't use dark optimization. If you choose to use new exposure length for a particular night - you can shoot darks of matching length later (just keep everything else the same for those darks).

Cooled cam means you can take your darks separately - just make sure you set temperature that you are going to use for lights (maybe even have two sets of darks, one for summer, one for winter, or better re shoot your darks every 5-6 months). You can do it during the day, or on cloudy night. If you take your camera off the telescope, make sure that you prevent "unseen" light that can penetrate plastic cap (like IR and UV) - place aluminum foil over top, or place camera cap down on some transparent surface like wood / stone table.

If you have filter wheel and permanent setup - you can shoot flats and have flat library (one master flat per filter). If you have "mobile" setup or you change filter orientation / fiddle with tube you need to reshoot your flats. I take my flats for every filter change because I use filter drawer and I'm "mobile". It is fairly quick because fast download times (10-15 mins per flat + flat darks - if I image with more than one filter per night - I use only one set of flat darks when applicable - exposure for flats must be good across filters).

As for number of each of calibration files - I'm a bit fanatical. The more - the better applies here and I do take it seriously - I shoot 256 of each.

you take 256 of each!!! so how many lights would you take? what ratio is the normal ie say 100 light 25 calibration 25% or more ?

[wimvb]

Dark frames contain noise, just as any other image. The noise stems from random variations in the dark current. By increasing the number of subs that make up the master dark, you decrease the noise. This scales as 1/sqrt(N), where N is the number of subs that make up the master dark. So, as usual: the more the merrier. But imo, 256 is serious overkill. If your light frames have an exposure time of 1 minute, you would be shooting 4 hours of darks, and if your light frames are 5 minute exposures, you're looking at 20+ hours of darks. I don't think you'll see any difference in the light frames when calibrated with a 20 - 30 subframe master dark, compared to calibration with a 256 subframe master dark. Neither do I bother with using a certain percentage of the subframes to make up a master calibration image. I just shoot about 20 - 30 darks and 30 flats.

[vlaiv]

18 minutes ago, Wirral man said:

you take 256 of each!!! so how many lights would you take? what ratio is the normal ie say 100 light 25 calibration 25% or more ?

It's not about ratio, it is about additional (mostly read noise) that comes from calibration files.

I aim for 4h per channel currently, and for LRGB that means around 240 ish subs (just to answer your question) - this is mainly down to heavy LP, for NB I aim at 60 subs per imaging night (4 minute sub).

Take one light sub - you have around 1.7e of read noise in it. Calibrate with one dark file (and for argument sake let's neglect dark current noise - and just look at read noise) - you don't add any signal, but you add noise, so after calibrating you will end up with light frame containing 1.7e * 2^0.5  = ~2.4e of read noise.

If you add 256 darks to create master dark end result will look like this: (1.7e^2 + (1.7/16)^2)^0.5 = 1.7033 - so we almost introduced no additional noise (there is some but it is really small, difference to original being 0.033...).

Things get worse if you have perfectly aligned frames (perfect tracking/guiding) and you don't dither - this noise that you introduced is not random between subs - you add same calibration file to each sub, so same noise, and noise calculation in stacking is different resulting in higher noise in stack - that is the reason you should dither as much as it is feasible - every couple of frames for 1 minute frames or even every frame for longer (4 minutes or longer).

Point is that for me taking calibration frames wastes almost no time - I build my dark lib during the day in my basement, and flats because of strong flat panel take less than 15 mins for 256 frames.

[Andyb90]

2 hours ago, vlaiv said:

- darks - taken at exact same settings as light subs - Gain, Offset, Duration, Temp

Do you think the darks and dark flats would need the same USB limit setting as the lights? I use v2 of the camera.

Andy.

[vlaiv]

Just now, Andyb90 said:

Do you think the darks and dark flats would need the same USB limit setting as the lights? I use v2 of the camera.

Andy.

Not sure which one is v2 - probably the last before the pro version? If so, and I have the same one - I used to use USB setting of 40 (default and the lowest one) earlier, and did all subs with that. After trying to optimize download time in SGP (second or so seems far too much for camera capable of 30fps in full res) I fiddled with USB setting and now I'm at 64 (just my binary numbers OCD) - it looks like it provides less dark current amp glow (I have not measure it but just by looking at master dark it looks like there is less amp glow) - so I left it like that and redone my darks - now I use that setting for everything, I just don' touch it. Mind you, I'm using USB3.0 hub and cables.

[wimvb]

5 minutes ago, Andyb90 said:

Do you think the darks and dark flats would need the same USB limit setting as the lights? I use v2 of the camera.

Andy.

According to zwo, part of the amp glow is caused by the slow readout. That's why they added the buffer in the "pro" versions. If you don't have the pro version, it would make sense to use the same readout speed, ie the same usb limit, for darks and lights. Better safe than sorry.

[Wirral man]

so to summarise you could take say 50 subs on each channel say 2 minutes and use the same amount of calibration frames also and this would help the final result ...i dont have a flat panel i used to just use the white tshirt over the lens method when i had my dslr i will setup tomorrow and get some library's built...looks like i have plenty to do now i have also upgraded my processing software and am learning that also gone from cs2 to cs6 extended and managed to acquire pixinsight 1.8 going to need a science degree to use it.

[spillage]

You might also want to try using calibration frames from just the L filter and then make the decision if you feel the need to produce frames for the other filters!

 

Mark.

[geordie85]

13 hours ago, Wirral man said:

 

[kens]

All I can add to all the good info above is to not make your flats and flat darks too short. I keep mine above 1s though a bit less is safe. The idea is to avoid the same issues that affect bias frames with this camera. 0.1s or more should be ok.

Since you are aiming for an ADU level (I use 20000) you might need more T-shirts to increase the exposures if the light source is bright. What is your light source for flats?

[brrttpaul]

interested in this as i have the same camera but i have been using a superbias so will try without any bias, also noticed that  its mentioned to shoot flats at 0 gain , I am not quite sure of darkflats though, i just do bias, darks and flats never tried dark flats dont really know how i make them but will look into it

[kens]

Shoot all calibration frames at the same gain, offset and temperature as your lights

Dark flats (or more correctly flat darks) are darks that calibrate your flats. So same gain, offset, temperature and exposure as your flats

[Wirral man]

7 hours ago, kens said:

All I can add to all the good info above is to not make your flats and flat darks too short. I keep mine above 1s though a bit less is safe. The idea is to avoid the same issues that affect bias frames with this camera. 0.1s or more should be ok.

Since you are aiming for an ADU level (I use 20000) you might need more T-shirts to increase the exposures if the light source is bright. What is your light source for flats?

i used to just point at the clear blue sky with a white tshirt around the lens

[wimvb]

For flats, I use an IKEA FLOALT led panel set at its dimmest setting. I point the scope at zenith, cover it with a few layers of white cloth, and put the floalt on top. My exposure times at lowest gain are from about 0.15 (L) to 3 (Ha) seconds. Flats correct optical issues, such as vignetting and dust bunnies (but not those ON the sensor). As such, there's no need to shoot them at the same camera settings as darks.

Dark and bias frames correct electrical camera issues. For ASI cameras this is mainly defect pixels and amp glow. Amp glow is temperature and time dependent, and doesn't scale linearly with either. That's why you can't use dark optimisation or scaling. Any darks for ASI cameras need to match the light frames (including flats) they are to correct. Even if flat frames are bright and have short exposure times, they can still contain a small amount of amp glow. I'd recommend usind dark flats rather than bias frames to calibrate them.

[vlaiv]

12 minutes ago, wimvb said:

Flats correct optical issues, such as vignetting and dust bunnies (but not those ON the sensor)

Can you elaborate on this please?  Do you mean dust on sensor or optical issues on sensor?

[andrew s]

Just some comments on this thread.

14 hours ago, vlaiv said:

Don't use bias frames with ASI1600 - bias is not stable

If this is true then all frames will be subject to the unstable bias and will be adding noise.

Christian Buil did a study of this and related cameras here http://www.astrosurf.com/buil/atik_vs_zwo/ he did not note any issues with bias (offset) stability but he noted that the neither the PRO version or the Atik Horizon had reduced amp glow compared to the MM.

While I have not tested this camera I have found that downloading images in quick succession can have an adverse effect on the output levels e.g. Bias/Offset or short exposure flats. I now always have a 10 s delay (determined by trial and error) between images to ensure the camera has stabalised.

Regards Andrew

[wimvb]

14 minutes ago, vlaiv said:

Can you elaborate on this please?  Do you mean dust on sensor or optical issues on sensor?

Dust directly stuck on the sensor. Where they completely block any light, no calibration frames can correct them.

[vlaiv]

3 minutes ago, andrew s said:

Just some comments on this thread.

If this is true then all frames will be subject to the unstable bias and will be adding noise.

Christian Buil did a study of this and related cameras here http://www.astrosurf.com/buil/atik_vs_zwo/ he did not note any issues with bias (offset) stability but he noted that the neither the PRO version or the Atik Horizon had reduced amp glow compared to the MM.

While I have not tested this camera I have found that downloading images in quick succession can have an adverse effect on the output levels e.g. Bias/Offset or short exposure flats. I now always have a 10 s delay (determined by trial and error) between images to ensure the camera has stabalised.

Regards Andrew

Probably wrong term on my part, but here is what I've found:

Short exposure darks like 10-30s, or 1 minute have greater average value compared to respective long exposure darks.

Bias only files tend to have the highest mean value of all.

This should not be the case and only explanation that I have for this is that manufacturer implemented some sort of variable bias level depending on exposure duration to "combat" dark current signal when sensor is not cooled (and regain some of dynamic range).

Between power cycles for same settings darks tend to have same signal value (both bias and dark current). Same is true for bias files - so if you take two sets of bias files and stack them - difference will have mean value of 0.

That is what I've meant by instability of bias level - it prevents one from subtracting bias only from dark subs.

[wimvb]

13 minutes ago, andrew s said:

While I have not tested this camera I have found that downloading images in quick succession can have an adverse effect on the output levels e.g. Bias/Offset or short exposure flats. I now always have a 10 s delay (determined by trial and error) between images to ensure the camera has stabalised.

Interesting. I usually take flats and dark flats without a delay between exposures. I've noticed an increase in temperature during a sequence, but never bothered to look for any other differences. Thanks for pointing this out.

[andrew s]

2 minutes ago, vlaiv said:

Probably wrong term on my part, but here is what I've found:

Short exposure darks like 10-30s, or 1 minute have greater average value compared to respective long exposure darks.

Bias only files tend to have the highest mean value of all.

If this is the case then I think there might be something wrong with your camera.  I can't think Christian would have missed that as it would have shown up in his dark current measurements.

Regards Andrew