Darks or Bias or both, dslr vs other.

[Malpi12]

Darks or Bias or both, dslr vs other.

I am a bit puzzled.

A couple of posts recently, (by @Budgie1 and @alacant) suggest not to use darks, instead use bias only with dslr.
I have found this when starting out in dslr (Canon 60D unmod with 50mm and 135mm lenses) I played with stacks in DSS of 20 x 1, 2, or 4sec exposures on fixed tripod: then doing light only, light+dark, light+bias, then L+D+B.
L only was of course poor, L+D and L+B mostly equal, and L+D+B being often the same as L+B sometimes marginally better.

The 'marginal' didnt seem worth the lost imaging time if the sky was still clear.

So I am puzzled on two fronts,
most online tutorials and 'tubes insist on darks.
secondly, dont most dedicated astro cams use the same chips as produced for dslr (more recently for mirrorless) so why not bias only for dedicated (but uncooled) astrocams aswell ?

I originally thought my findings were because I was inexpert, that my ambient temperature allowed great mismatch in exif temperature (not necessarily sensor temperature)  and also using very short exposures. Then I read that proper imagers with longer tracked exposures may also not be using darks !

Anyone got some good links to exposures of the truth
Is it just down to the inability of a dslr to get darks sufficiently close in temperature to the lights ?

 

[Astro Noodles]

I'm certain that those guy use dithering instead of darks.

[rickwayne]

> Is it just down to the inability of a dslr to get darks sufficiently close in temperature to the lights ?

No,  although that can affect the results. DSLRs use different sensors, in the main, than astro cameras, and modern ones have very little dark current. My Pentax K-5iis is one of these, I don't bother with darks. Astro  camera sensors are optimized for different parameters, and so are comparatively weak where DSLRs excel. Google "IMX183 amp glow" to see examples of my astro camera's dark current, for example!

Software can compensate for differing temperatures to a degree, if the actual sensor temperature is recorded in the EXIF data from a DSLR ("dark scaling").

Bias is a bit of current added to the photosite wells to ensure that all of them report an at-least-slightly positive result. If the actual light fluxes onto two adjacent photosites differ by a magnitude that the sensor would represent as seven electrons, one might generate 5 electrons and the other 0 (it's really at a level that would result in -2 electrons, but that's a physical impossibility!). But when you read the sensor, the difference is only 5. By adding 10 electron's worth of bias current, you raise the laggard above the unknowable floor of zero, so they're  15 and 8, restoring the relative values.
 

A bias frame records this as well as any "floor" current that the sensor produces even when no light is on the sensor. Subtracting the bias frame value lets you correctly extrapolate where zero is for that photosite. It's especially helpful with flat calibration frames.

Speaking of which, you should definitely be shooting  and processing with those. Vignetting, dust, and other systemic problems boil right out of your photos that way.

Edited July 26, 2021 by rickwayne

[AstroMuni]

20 hours ago, Malpi12 said:

Darks or Bias or both,

The experts use Darks, Biases and Flats. For modern CMOS cameras I have been told that its enough to have just Darks & Flats.

Oops...forgot to mention Dark-flats

Edited July 27, 2021 by AstroMuni

[alacant]

20 hours ago, Malpi12 said:

L+D and L+B mostly equal, and L+D+B being often the same as L+B sometimes marginally better.

Perfect. So, like us you worked out for yourself what was best. You didn't get bogged down in theory.

There will be the armchair theorists who will tell you it is totally insane not to include dark frames in your calibration; they can't cope with pragmatism.

But wait. Can we generalise? No, of course we can't:

• Your 60D: bias only works fine.
• Our 700d: 1200d, 2000d and 4000d: bias only works best
• Our 450d: the only way we can control the banding is to use dark frames
• Someone else's dslr: until we've done your L+D etc, we don't know.
• My mate's asi294 at -5º: darks are essential

Hands on is perhaps the only way to go.

Just our €0,02

Cheers

[Budgie1]

21 minutes ago, alacant said:

Hands on is perhaps the only way to go.

I couldn't agree more because I don't think it's a case of use or don't use Darks. It's about what works best for your set up and camera.

When I read about not using darks with a DSLR, I didn't just stop using them, I tested the theory on my own camera with various images stacked in different ways and, for me, I didn't notice any difference between using & not using Darks with my EOS 1300D. So I stopped using them and spent the time I would have been taking the Darks to take more Subs.

Now I mostly use a OSC ASI294MC Pro and with that I have to use Darks to remove the amp-glow. But even with this camera I tested what I'd read and did notice a difference. With ASI294MC, if I used both Darks and Bias in the stack then I got a residue of the amp-glow remaining in the completed image. By removing the Bias frames from the calibration, this amp-glow was gone and the background was improved.

So, I use Bias not Darks with my DSLR and I use Darks not Bias with my ASI294MC. I don't have the foggiest about the science behind it, I just know it works for me.  

[Malpi12]

5 hours ago, alacant said:

Perfect. So, like us you worked out for yourself what was best. You didn't get bogged down in theory.

Just our €0,02

 

And worth every cent  👍

Being new to the camera - at first I just did lights to find out what all the modes were, find my way round the stars, see what exposure duration/startrail was acceptable, how to drive DSS and a host of other things to learn (like, my old fslr 50mm lens isnt much good ) 
( and like, dont walk away from the camera with the intervalometer in pocket but still attached, if you do, be ready to catch the camera+tripod as they fall over ! Phew, I did, but it was a near thing !!)

Then dutifully I moved on to darks, as the tutes said to do, yep good.
Then, whats this bias stuff all about ,,
that is when the head-scratching started, Thought at first that I wasnt understanding bias and/or DSS.

Thanks for all your thoughts everyone. 
 

Edited July 27, 2021 by Malpi12

[Malpi12]

5 hours ago, Budgie1 said:

I use Darks not Bias with my ASI294MC. 

That is interesting thanks, a note to be memorized for when my thoughts will in the future turn to proper astro cams, "watch out for reports of amp glow."
I think I would need a cooled one that can have a library of darks, It would grieve me to spend our few nights of stars doing ambient darks
 

Edited July 27, 2021 by Malpi12

[AstroMuni]

On 26/07/2021 at 20:32, Malpi12 said:

Anyone got some good links to exposures of the truth

Here is a good explanation....https://practicalastrophotography.com/a-brief-guide-to-calibration-frames/

[wxsatuser]

On 27/07/2021 at 16:55, alacant said:

Perfect. So, like us you worked out for yourself what was best. You didn't get bogged down in theory.

There will be the armchair theorists who will tell you it is totally insane not to include dark frames in your calibration; they can't cope with pragmatism.

But wait. Can we generalise? No, of course we can't:

• Your 60D: bias only works fine.
• Our 700d: 1200d, 2000d and 4000d: bias only works best
• Our 450d: the only way we can control the banding is to use dark frames
• Someone else's dslr: until we've done your L+D etc, we don't know.
• My mate's asi294 at -5º: darks are essential

Hands on is perhaps the only way to go.

Just our €0,02

Cheers

Modern cameras have on sensor dark current suppression.

The 450D does'nt have it.

[rickwayne]

That's a  great little tutorial on calibration frames. I would add one explanation to it. Dark flats are mentioned, but there's a bit of hand-waving WRT whether you need them. DFs are an alternative to bias frames.

Since flats are done with a fairly bright light source, their exposure times are quite short so there's little to no thermal current in play. That's why you don't need to use dark frames to calibrate them. However, there are pixel-to-pixel variations, as well as the applied bias current, so you DO need some sort of baseline zero-light image.

Bias frames can be used for this quite nicely. Advantage: You only have to shoot one set per gain/ISO that you use, and you can easily do it during the day, or cloudy time. However, some cameras (IIRC the 294 sensors in particular,  maybe the 1600?) deliver inconsistent results at such short exposure times.

The alternative is to use dark flats, which use the same gain/ISO and exposure as your flats, but with no light. Disadvantage: If your flats exposure changes, you have to reshoot your DFs. (Indeed, for us mono imagers, you have to shoot and store a set for every filter, since the transmissivity and so exposure differs between  them. For me, that would be  Ha, OIII, SII, L, R, G, and B!) I don't know if any of the common programs can scale one set of dark flats to different flat-frame exposures; if so, that would  lessen the  problem  considerably.

So if your camera takes consistent short-exposure images, bias + flats is a simpler way to go. If not, dark flats +  flats  is every bit as good. DLSRs  are designed to deliver great images at very high shutter  speeds, so I'd personally go bias; Your Mileage May Vary.

Edited July 29, 2021 by rickwayne