Few Questions

[ZiHao]

Hello!

Have a few questions here:

Is DSLR darks waste of time? I am currently using D5300, and since taking darks really consumes a lot of time, I would be happier to take many flats and bias frames only  

Is PHD2 drift alignment tool accurate using a ST4 connection? I have been getting +-5 arc seconds when drifting. Normally I will do DARV and then move on to PHD2 drift tool. My declination RMS value was always pretty low ( about 0.50" ), I supposed that this means my polar alignment is pretty good already?

Thanks and Clear Skies!

 

[geordie85]

To answer your first question, I never used to take darks when imaging with a dslr. I just use a big dither after each sub. 

I cannot answer your second question as I've never done this before 

[cjdawson]

Darks are never a waste of time.  The thing about them though is that you need to make sure that the camera is at the same temperature as when the light frame was taken.   This make capturing good darks with a DSLR harder.   I managed to create a library of dark images, which did improve the output when I processed an image compared to processing the same without darks.

 

Not sure about your second question as I don't use the drift alignment tool in PHD2.  Shouldn't make much difference if you are using an ST-4 or other connection, actually what I understand is that PHD2 was designed to work without the need for an ST-4 connection and will actually work better by controlling the mount directly.  That said, if you don't have the option of controlling the mount directly, the ST-4 connection is perfectly fine.

[ZiHao]

12 minutes ago, cjdawson said:

The thing about them though is that you need to make sure that the camera is at the same temperature as when the light frame was taken.

I will never bother with darks frames then since I am not using a cooled camera and because of the high temperature here in Malaysia. But will still try to take darks next time out to see if they improve things or not. Thanks for your help.

[cjdawson]

4 minutes ago, ZiHao said:

I will never bother with darks frames then since I am not using a cooled camera and because of the high temperature here in Malaysia. But will still try to take darks next time out to see if they improve things or not. Thanks for your help.

It doesn't matter what temperature the camera is, just that the darks are taken at the same temperature that the image was taken.

 

I'll give you an example.

when I use a Canon 70D, my camera warms an stays at a temperature (I'm not sure what temp, it doesn't matter)  once I finish taking the light frames, I take darks as well, this way they are at the same temp as the light frames.   Does mean that I can't image all night long though.

 

With my ZWO ASI1600-MM-Pro   I can cool the camera to -20°C.  this means that i can take a set of dark frames during the day, meaning that I can take images all night long.  As the camera is always cooled to -20°C is keeps things the same.

[ZiHao]

4 minutes ago, cjdawson said:

t doesn't matter what temperature the camera is, just that the darks are taken at the same temperature that the image was taken.

Oh... so they only thing that matters is the temperature of the surroundings, not the temperature of the camera sensor...and as long as the temperature outside the camera is constant, the darks will be effective, so DSLR users can only image for about 2 hours as the temperature does not change much right?

[vlaiv]

Darks are important if you also use flats.

In order to properly calibrate with flats - you need dark current removed. Sensor needs to be at exact temperature for lights and darks, for darks to work properly. You can get away with same ambient temperature provided that sensor ends up being in thermal equilibrium with surroundings.

Taking darks with camera that does not have set point temperature control is going to be tricky, because ambient temperature changed thru the course of the night.

There is algorithm that you can use to compensate for this. It's not ideal, but actually works in most cases. It's dark frame optimization algorithm. It tries to guess scaling factor for dark frame and each light frame that it is calibrated with by some clever statistics - proper scaling factor is one that yields the most smooth calibrated frame. This is particularly helpful if you have amp glow - because sub with amp glow removed will be "smoother" (or maybe better term is flatter?).

In order for this algorithm to work - you need stable / usable bias frames. Some CMOS sensors have issues with bias frames, and they should be skipped for calibration, but if your DSLR has good bias, then it's worth checking out this algorithm.

In DSS there is option for that under "darks" tab:

Even sensors with problematic Bias can use this algorithm, but not as implemented in DSS - it needs a bit of change - by using two sets of dark frames of different exposure length (and extracting bias from that, so it's a bit more involved and I'm not sure if it is implemented in any software).

[cjdawson]

13 minutes ago, ZiHao said:

Oh... so they only thing that matters is the temperature of the surroundings, not the temperature of the camera sensor...and as long as the temperature outside the camera is constant, the darks will be effective, so DSLR users can only image for about 2 hours as the temperature does not change much right?

it's the opposite.  It's the temperature of the sensor that is important, only the sensor the rest of the camera can be anything.

[Z3roCool]

If taking darks the next day with  a set point camera do you need to keep the camera at the same angle in the scope?

[vlaiv]

17 minutes ago, Z3roCool said:

If taking darks the next day with  a set point camera do you need to keep the camera at the same angle in the scope?

No, you can even take camera off the scope - it does not really matter as darks are completely void of light. Only thing that matters is that you have your camera at exact same settings like you did for lights (same exposure length, same gain/offset/iso, ...).

With well behaved set point CCD you don't even need to take darks of the same exposure length, but in that case you need to do something called dark frame scaling (again there is option in DSS in form of dark multiplication factor). This also requires bias removal, so master bias needs to be included.

[Stub Mandrel]

Before I cooled my DSLR  I just set up three sets of darks - for cold (0 or less), cool (1-10 degrees) or warm (10-20 degrees) evenings. Rare evenings hotter than 20 degrees it was questionable if it was worth imaging.

Before this I sorted lots of darks according to Exif temperature (generally goes in 2-degree steps on the 450D) and matched the range the imaging session covered to the darks used. I decided this extra faff wasn't worth the bother.

[rickwayne]

Best way is to be scientific about it -- take a set of darks and process with and without them, see what yields better results for you. Some DSLRs actually benefit very little from them. Most of the Pentaxians I follow have given up on them. 

You should be able to extract the sensor temperature from your raw images with EXIFtool or something similar. If you're going to do it right, you'll build up a library of darks at a variety of temperatures, then choose the correct one based on the temperature of a given evening's "lights", as Stub suggests. It's often inconvenient to replicate field temperatures, and sensors change slowly over time, so that a dark library you shot two years ago might not be very useful today. That's why most folks who shoot them do so in the field.

But it does significantly cut into your imaging time. I used to pull the camera off the scope after I was done imaging, cap it, and then set it to shoot as many darks as I could get while I was tearing down the mount and scope and packing everything up. I only almost left the camera and intervalometer on the ground once. ?

[Alien 13]

It is simples, just turn on the in camera long exposure noise reduction then you will get a perfectly matched dark for every light, only problem is that it halves your imaging time.

Alan