Ditch the darks?

[david_taurus83]

Can I ditch the darks?

 

It' been almost 4 weeks since I've last had the scope and camera out. Clear sky is a precious commodity here at 52N 2W! I captured over 90 minutes of dark frames last time. Admittedly, I done it wrong. As soon as I set up and polar aligned with Polemaster, checked drift align which was ok, the clouds rolled in. So I captured my darks first, 45 x 2 min exposures. I know your not supposed to but damned if I was wasting clearer skies later on. I eventually captured 2 hours worth of subs on M101, guiding was working well and I thought I would get a really good result. I was eventually disappointed to say the least as the combined image is awash with severe gradient. Likely because I was shooting from my light polluted garden with 3 streetlights close by, though I have had better luck with M45 and M81.

 

Anyway, I've been thinking about driving to a darker location not far from home. I don't mind working from the car for a few hours capturing lights but I've been thinking if I can do away with the dark frames. I've  been reading that if I dither between shots, (I use APT) I can get away without darks. Is it really that easy? Do I need to replace the darks with something else? Forgive my naivety as I'm still a rookie on the subject. I do have a copy of Making Every Photon Count but I've not long started it and haven't got that far yet! There's a (hope I'm not jinxing it) clear sky "promised" this weekend and I'm fancying my chances on M51 from around 11pm onwards. 

 

Kit being used will be a modified 600D on a 200p, AZEQ6 Pro, PHD2 guided and all processing done in Pixinsight. I've been using the Light Vortex Astronomy tutorial for manual pre processing. Now there' a learning curve! Thanks.

 

David

[Tommohawk]

I dont see why you would want to ditch the darks? They should help reduce the dark signal from the light frames. Obviously you have to be really sure that its totally dark!

Bear in mind that with your 200p if you cover the scope with the end cap some light may leak in from behind the mirror. Been there - done that! It's safer to remove the camera and put the camera cap on. I think you can do darks first - they just need to be at approx the same temperature. The sensor temperature may rise with long exposures - it might be wise to leave a gap between. That said, the lights may also alter as the sensor heats. Not normally too much of an issue this time of year.

Does your processed pic show more gradient than the unprocessed lights? Do the darks show any gradient - they can with light leakage.

Are you doing flats? That will help with vignetting which might compound gradients. You can usually process out gradients - have a google - Doug German has a tutorial on one method.

[michael8554]

Dither and use Bias as Darks when stacking has been suggested for Canon ds!r's.

 

 

[bobro]

Dithering plus flats and Kappa Sigma clipping works pretty well for me with my EOS1000D. The lack of darks can mean some amp glow in one corner for longer exposures (over about 4 or 5 min) but that isn't too difficult to fix in processing.

[david_taurus83]

@Tommohawk Reason I want to ditch darks is the saving in time! I'd rather be capturing light frames with that time. 

 

I've also seen that some people create a darks library. Is this a selection of stacked master darks, taken at various external temps and various exposures, to be used later on for processing multiple images that have light frames taken with similar temps and exposures? 

 

Is it beneficial or detrimental to both dither and add darks? 

 

So many combinations!

[Tommohawk]

5 minutes ago, david_taurus83 said:

@Tommohawk Reason I want to ditch darks is the saving in time! I'd rather be capturing light frames with that time. 

 

I've also seen that some people create a darks library. Is this a selection of stacked master darks, taken at various external temps and various exposures, to be used later on for processing multiple images that have light frames taken with similar temps and exposures? 

 

Is it beneficial or detrimental to both dither and add darks? 

 

So many combinations!

I've never been a ditherer! For darks you only need to do them once per ISO setting, though as you say ideally you have different darks for different temps - but you cant control the temp so this is tricky.

I'd suggest winter darks, spring darks, and summer darks -  thats all I ever did. You can read the temps from the EXIF data, but remember thats the processor temp not the sensor temp - gives you an idea though.

Also you dont need tons of darks - 20 should do.

Did you do darks with lens cap or scope cap?

[Tommohawk]

And yes, once the darks are stacked, just keep the stacked dark in a library and use that in future - makes processing faster.

[david_taurus83]

13 minutes ago, Tommohawk said:

I've never been a ditherer! For darks you only need to do them once per ISO setting, though as you say ideally you have different darks for different temps - but you cant control the temp so this is tricky.

I'd suggest winter darks, spring darks, and summer darks -  thats all I ever did. You can read the temps from the EXIF data, but remember thats the processor temp not the sensor temp - gives you an idea though.

Also you dont need tons of darks - 20 should do.

Did you do darks with lens cap or scope cap?

I take darks with the scope cap on. Its an old blue 200p so has the steel plate on the back behind the mirror. I believe the newer ones dont have this cover any more to make collimation easier? I've drilled small holes where the screws and grub screws are for easy access!

 

Below is a strecthed 2 min dark and below that is my light polluted stretched stacked image:

 

[Demonperformer]

I believe @ollypenrice uses the no darks, but dither+master bias approach.

So, if the results he gets are good enough for you ... 

[david_taurus83]

^^^The difference is Olly knows what he's doing, whereas I on the other hand...?

[happy-kat]

Have you tried processing your image without those darks and compared the noise?

With my dslr it's flats and dark flats that make a difference I don't do darks.

 

[ollypenrice]

44 minutes ago, david_taurus83 said:

^^^The difference is Olly knows what he's doing, whereas I on the other hand...?

Who told you that? They've obviously never met him!

[Tommohawk]

That stacked image does look pretty polluted - does that red "banding" show in the individual subs? Just wondered if it might be some kind of processing artifact.

I always got pretty good results without dithering, and just used darks flats and bias for calibration with my Canon 550d. I think I've always been lucky in that until now I haven't suffered with light pollution, and I think this probably makes all the difference so my experiences may not be a fair comparison.

[david_taurus83]

I haven't tried stretching the individual lights to see how bad it is. I did think about maybe removing the earlier lights as the target was passing directly overhead later into the session. I think I may have been pointing too close to a streetlight beforehand. 

[carastro]

David, you are imaging from Birmingham which is heavily Light polluted.  I image from SE London - same problem.

I think you will find there are some targets that really need a dark location and from my experience M101 is one such image.  Are you using a LP filter? 

As regards darks, cameras have moved on a bit since I used a DSLR, but I never used to take darks at the sacrifice of clear skies and image data.   The most I would do is a couple before I started, 2 midway (or when clouds are passing over) and two when packing away.  Unless you are certain of the temperature I don't see how a darks library will work.

I have heard that you can ditch the darks if dithering.  I would try Olly's method of adding a master bias instead of taking darks and dither and see if that helps, plus a LP filter from Birmingham.

Carole 

[david_taurus83]

Yes, sorry, should have added. I have a Skytech CLS CCD clip filter in the camera. I will try dithering anyway and see how it goes. 

 

Regarding a dark library, I have seen a thread on cloudy nights where someone used an IR thermometer to measure the sensor temp in their Canon camera and compared the results to the exif temp reported by the processor. The results were within a couple of degrees F. Seems like the processor temp might be close enough to go by.

[Tommohawk]

1 hour ago, david_taurus83 said:

Regarding a dark library, I have seen a thread on cloudy nights where someone used an IR thermometer to measure the sensor temp in their Canon camera and compared the results to the exif temp reported by the processor. The results were within a couple of degrees F. Seems like the processor temp might be close enough to go by.

Thing is, even if you can control the temp of the darks somehow, maybe by just deleting those not within a couple of degrees, you're not going to be able to control the lights temperature. I think you're stressing about this too much!

[jjosefsen]

Doesn't DSS support dark scaling?

PixInsight does and from what I gather, it's all math. It scales the dark noise in your darks to match the noise in your lights, and then subtracts it. As long as you don't reach a scaling factor above 2x you should be fine.

Admittedly I don't know how DSS does it..

[david_taurus83]

Well the forecast for tonight has, predictably, turned sour ?

 

So I most definitely will not be driving up to the local hills. At best, there may be some gaps in the clouds to coax me to set up in the back garden, try out guiding the new mount. I think the happy medium approach I need to take should be to dither, but also take some darks. Certainly not 90 mins worth. Maybe 3 x 3 mins while I set up and a couple every half hour after that. Thanks for all the input folks!

[happy-kat]

If you do that medium approach then you will ideally test processing with or without darks so you can identify that darks are adding noise or not.

[StuartJPP]

Once I started dithering I ditched darks. I did a comparison with the same data with and without darks and to be honest the ones with darks ended up slightly (subjectively) worse since they weren't exactly temperature matched. Plus it just saves a lot of time.

But those were just my own findings.

[Gerry Casa Christiana]

Did exactly the same. Now use dithering and only flats and bias and I don't really see any difference. I think darks can help when the outside temperature is higher but even then I don't use it anymore. Give it a try. You'll get more subs too of course