Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Tony Hallas on Using a DSLR

This topic has been archived. This means that you cannot reply to this topic.
63 replies to this topic

#1 TimN

TimN

    Gemini

  • *****
  • Moderators
  • topic starter
  • Posts: 3,415
  • Joined: 20 Apr 2008

Posted 15 November 2014 - 11:09 AM

Shawnhar posted an interesting Youtube video - in Beginning and Intermediate Imaging - from Tony Hallas describing his method of processing using A DSLR. Its very different from what you usually see posted. In case you are interested its here:  http://www.cloudynig...n-using-a-dslr/



#2 jerry10137

jerry10137

    Viking 1

  • *****
  • Vendors
  • Posts: 783
  • Joined: 26 Jan 2013

Posted 15 November 2014 - 12:21 PM

He has a lot of really good information in there.  He's right about how to clean up DSLR images.  Dithering aggressively is the only way to fly....



#3 APshooter

APshooter

    Skylab

  • *****
  • Posts: 4,294
  • Joined: 14 Oct 2013

Posted 15 November 2014 - 06:41 PM

Nice presentation.  Thanks for the heads up.



#4 Roodyj

Roodyj

    Sputnik

  • -----
  • Posts: 26
  • Joined: 23 Oct 2014

Posted 15 November 2014 - 08:01 PM

I like the presentation and the results one can achieve. Tony uses Registar, instead of DSS or other such software. Unfortunately, Registar is $150. Is there a good reason to spend that much on a "stacking" program? I am at a loss to understand the greater value.



#5 jerry10137

jerry10137

    Viking 1

  • *****
  • Vendors
  • Posts: 783
  • Joined: 26 Jan 2013

Posted 15 November 2014 - 10:53 PM

Registrar will also register multiple images. It's really good at that. However, $150 is a bit steep. Why not go a little further in price and have a full suite of tools within PixInsight. Tony likely used Registar in the talk simply because that just what he uses. The technique is available elsewhere and you do not need Registar.



#6 Roodyj

Roodyj

    Sputnik

  • -----
  • Posts: 26
  • Joined: 23 Oct 2014

Posted 16 November 2014 - 11:12 AM

Thanks for the suggestion, Jerry. My level of astrophotography at this point is way below the quality that PixInsight would treat appropriately. Maybe in time I'll get to the expert level. lol



#7 Footbag

Footbag

    Cosmos

  • *****
  • Posts: 9,115
  • Joined: 13 Apr 2009

Posted 16 November 2014 - 11:53 AM

I was also surprised at how different his methods are to what you typically see. 

 

I'd completely disagree with his points on calibration frames and even some other aspects, if his results didn't speak for themselves. 



#8 Agnotio

Agnotio

    Ranger 4

  • -----
  • Posts: 387
  • Joined: 29 Aug 2008

Posted 16 November 2014 - 01:53 PM

Totally agree with his thoughts on color mottle.  I too mistakenly thought it was hot pixels, but it cannot be since as he points out the color mottle is 5-10 pixels in size, or even bigger.

 

Here an example from a stack of 3 images with no dithering (Canon 450D).  On the left is without noise reduction, on the right after Noel Carboni's actions (noise reduction and less crunchy).

 

 

Attached Thumbnails

  • mottlenoise.jpg


#9 TimN

TimN

    Gemini

  • *****
  • Moderators
  • topic starter
  • Posts: 3,415
  • Joined: 20 Apr 2008

Posted 16 November 2014 - 03:19 PM

I certainly agree with his points on dithering. I started to do dithering awhile ago and even dropped darks as I noticed they added no value when I was dithering. However, I never got rid of the mottling effect with as little as 9 frames the way he can. Also, I have never heard it suggested to drop the other calibration frames - flats and bias. Also, actually modifying the subs before stacking was supposed to be wrong as well. I will see if there is any way to increase my dithering and play with his other suggestions. As they say - "whatever works". 



#10 nickatnight

nickatnight

    Skylab

  • *****
  • Posts: 4,374
  • Joined: 13 Apr 2008

Posted 16 November 2014 - 03:49 PM

I'm sure Tony has some pretty clean optics and sensor to get away without shooting flats. My flats do a lot more than just help vignetting, such as dust modes, optical flaws and other imperfections from my self mod hack job. As for darks and bias, I'm gonna stick with those as im still manually dithering every 4-6 frames, which may not be enough. I did find it interesting that he applies noise reduction to the raw files before stacking, although that may be how he gets away without calibration frames.

#11 mmalik

mmalik

    DSLR camera modifications

  • *****
  • Vendors
  • Posts: 11,235
  • Joined: 13 Jan 2012

Posted 16 November 2014 - 04:13 PM

Good presentation but I think dark calibration is still the best way to reduce noise/color mottle; I especially like the multiple LENR... feature and has served me well. Regards



#12 Agnotio

Agnotio

    Ranger 4

  • -----
  • Posts: 387
  • Joined: 29 Aug 2008

Posted 16 November 2014 - 04:50 PM

Good presentation but I think dark calibration is still the best way to reduce noise/color mottle; I especially like the multiple LENR... feature and has served me well. Regards

 

Your comment made me realize that my previous one was wrong.  I think the color blotches are actually dark (thermal) signal.  Theoretically, this signal should be subtracted through dark calibration.  I think one of the issues in the photo I linked above is that the darks were not properly calibrated to the temperature of the lights (there was a break between my imaging session and darks).  That being said it's hard to perfectly match the temperature.

 

Here's the stretched dark signal in a 3-minute exposure image, it looks like the color mottle (10 pixel block for comparison):

 

Attached Thumbnails

  • darknoise.jpg

Edited by Agnotio, 16 November 2014 - 06:59 PM.


#13 John_K

John_K

    Viking 1

  • -----
  • Posts: 766
  • Joined: 28 Apr 2005

Posted 16 November 2014 - 08:45 PM

I like's Tony's presentation, although it would have been better if it was relevant more to taking images through the telescope rather than with a lens, but still very relevant.

 

On the issue of Dithering, if we were to adopt Tony's philosophy on setting dithering as high as possible, does Backyard EOS go as far as this?

 

  

Attached Thumbnails

  • BYE.jpg

Edited by John_K, 16 November 2014 - 08:45 PM.


#14 TimN

TimN

    Gemini

  • *****
  • Moderators
  • topic starter
  • Posts: 3,415
  • Joined: 20 Apr 2008

Posted 16 November 2014 - 09:10 PM

If you want to know how much you are dithering this may help. I believe PHD2 has 5 levels of aggressiveness. They are  +/- 0.5, 1.0, 2.0, 3.0, 5.0 pixels. However, these are guide camera pixels and PHD2 will multiply them by the the dither scale factor - as indicated on the Global tab in the brain. So, although I'm not familiar with BYEOS, I assume it will have 5 levels of dither ( as does SGP) which represent the five levels in PHD2. So, if you chose the maximum dither or highest amount in any program using PHD2 you should be getting 5 x (the dither scale factor) equals the total number of pixels dithered.



#15 guyroch

guyroch

    Vendor (BackyardEOS)

  • *****
  • Vendors
  • Posts: 3,682
  • Joined: 22 Jan 2008

Posted 16 November 2014 - 09:43 PM

I like's Tony's presentation, although it would have been better if it was relevant more to taking images through the telescope rather than with a lens, but still very relevant.

 

On the issue of Dithering, if we were to adopt Tony's philosophy on setting dithering as high as possible, does Backyard EOS go as far as this?

 

If you are going to use maximum dither aggressiveness in BackyardEOS, as you have in your image set to 5, you better increase the calm down period.  You have set at 2, that is way way to low for maximum aggressiveness.  The default is 10 seconds when aggressiveness is set to 2.  If you set it to 5 you need at least 30 seconds IMO... otherwise you risk your next image to start before PHD as resumed stable guiding again.

 

Lets not go nuts here and push dithering to the max... go slowly and inspect your result.  Most people will guide at shorter focal length so a single pixel of dithering on the guide camera will be more pronounced on your imaging camera.

 

Hope this helps. 

 

Guylain


Edited by guyroch, 16 November 2014 - 11:30 PM.


#16 PongosDad

PongosDad

    Vostok 1

  • *****
  • Posts: 139
  • Joined: 25 Sep 2010

Posted 16 November 2014 - 10:28 PM

Interesting stuff for sure, so I decided to give the process a shot. The major difference being that I'm using DSS for stacking.  I learned pretty quick that these pre-processed subs don't get along well with sigma clipping, so I tried to combine with "average" with better results.

I was pleasantly surprised with some clean results on the Eastern Veil.

But then I got into some data for M31  (25 subs @ 600s) and have had some major issues with this ACR (Adobe Camera Raw) method of pre-processing.  In short, the resulting stack has had major gradient issues.  After effectively dealing with them (I'm using PhotoShopCC wtih my main approach being Carboni's Actions for gradients) there are still some artifacts that I can't quite explain.

Here's a 200% crop of the background which shows the issue.

ACR stack with artifacts.jpg

Possibly a product of my dithering?  I only had PHD2 set to a dither scale of 1.

 



#17 shawnhar

shawnhar

    Cosmos

  • *****
  • Posts: 9,452
  • Joined: 25 Jun 2010

Posted 16 November 2014 - 10:59 PM

If I had to guess, your dithering is not working. That looks like the "walking noise" I had before dithering, the diagonal red streaks. Once I started dithering and using Sigma in DSS, it completely went away. Your red dots don't seem to match that though, it looks like 35 or so individual dots. Very weird



#18 PongosDad

PongosDad

    Vostok 1

  • *****
  • Posts: 139
  • Joined: 25 Sep 2010

Posted 16 November 2014 - 11:12 PM

Makes sense Shawn.  There's always the chance I had managed not to toggle dithering via BYEOS for that run.  And if I had turned on dithering for the Veil run, that would explain the difference in results.

Thanks

Jim



#19 Ivo Jager

Ivo Jager

    Vendor ( Star Tools )

  • *****
  • Vendors
  • Posts: 380
  • Joined: 19 Mar 2011

Posted 17 November 2014 - 02:24 AM

This is a bit of a cross post from Reddit, but I feel it's important enough to repeat here;

 

This is a downright dangerous video/presentation, which will put a lot of people on a dead-end track. He is doing a lot of budding astrophotographers a great disservice here.

This is going to be a bit of an essay, but bear with me. And please do understand I do not mean to slag anyone off, but it really needs to be said that there are a great number of issues with the material in this presentation.

 

  1. He starts with asserting that a DSLR is not a CCD camera. I could be accused of being pedantic here though; I think he's merely trying to say make a distinction between off-the-shelf consumer cameras and cameras optimised for scientific data acquisition, but they largely use the same hardware; the QHY8 and Nikon D50 even used the same CCD chip for example. However, noise considerations are pretty much the same and what he calls "Color Mottle" is not an inherently physical phenomenon or an attribute of a CCD chip, DSLR or otherwise; Color Mottle is introduced by Tony and his use of Adobe Camera RAW. The debayering algorithms and noise reduction filters in Adobe RAW look for correlation and introduce this correlation which we perceive as mottling. This is because debayering algroithms are interpolation algorithm that make up plausible data for missing data; 75% of the red, 50% of the green and again 75% of the blue pixels in the full frame image are completely made up! Give me a RAW frame and I can show you a nice random-like poisson noise (not correlated noise) guaranteed. Mottling is not inherent to DSLRs. Period.
  2. Dithering is only effective to reduce fixed pattern noise; his color mottle is simply the result of transformed (by the debayering) random shot noise; the 'Color Mottle' is never in the same place, so dithering isn't required to fix that. Tony's solution to use dithering is just a very roundabout way of fixing up his initial error of debayering his frames with Adobe Camera RAW. It's really a less effective way of applying Bayer Drizzle (scroll down the full page) of the RAW frames. The latter procedure doesn't introduce the mottling in the first place.
  3. It gets worse with Tony's 'pre-processing' in Adobe Camera RAW. For starters it's already debayered and color balanced, which increases noise and sows the seed of his 'Color Mottle'. Then he applies chromatic aberration correction (making the data no longer usable for deconvolution in a mathematically sound way), applies noise reduction (increasing 'Color Mottle' even more and destroying faint detail, while making the data no longer usable for deconvolution or stacking in a mathematically sound way) and applies non-descript 'sharpening' (making the data no longer usable for deconvolution or stacking in a mathematically sound way). Worse, he's doing this on a per-frame basis, completely ignoring the fact that a stacked image with higher SNR would at least help the smarter noise reduction and sharpening algorithms (let alone decon) to distinguish between real detail and noise.
  4. All the while it seems he's under the impression that his data is 16-bit. It's not. It's 14-bit for the D6 that he used (most of which goes to waste, but I digress). Because he hasn't stacked yet (which at 16 frames would net him an effective 18-bit data set).
  5. But hey, let's make it worse. Let's stretch everything, making any non-linear process (such as STACKING) completely useless. You cannot simply average (stack) data that has been stretched by a multi-point histogram stretch and get a statistically meaningful result when averaging. Let's say we have two linear readings 0.1 and 0.5 for a pixel. The average while linear is obviously 0.3. Now let's say that we have stretched it with a simple 2.2 gamma correction curve after taking that average. That would give us 0.3 ^ 1/2.2 = 0.578532609. This is the right way of doing it. Take averages (e.g. stack) first, then stretch. Now let's see what happens when you first stretch and then take the average; ( 0.1 ^ 1/2.2 + 0.5 ^ 1/2.2 ) / 2 =0.540429613. They are very different numbers and they'll be more different the more you stretch.
  6. At this stage he's basically now shooting in JPEG, perfectly mimicking the way on-board processing makes data unusable for AP, his only saving grace being that his data is still 14-bit instead of 8-bit, that this data has not been compressed with a lossy codec and that he's using a top-of-the range D6 with great optics attached. I wrote this conclusion on the ST forums, but trying to find out more about this 'technique' sent me to the PI forums where Carlos Milovic drew the same conclusion almost a year ago.
  7. Stacking after all this has no mathematical bearing on reality anymore. At 19:30 he seems proud he's obliterated the last bit of what is probably recoverable detail. This made me really, really sad. He claims he's averaged the signal of the 16 frames. This is demonstrably not true.

TL;DR: He is advocating some very, very nonsensical and destructive things. This video makes me very, very sad.... :(


Edited by Ivo Jager, 17 November 2014 - 02:32 AM.


#20 garret

garret

    Surveyor 1

  • *****
  • Posts: 1,656
  • Joined: 26 Jul 2009

Posted 17 November 2014 - 04:44 AM

 

TL;DR: He is advocating some very, very nonsensical and destructive things. This video makes me very, very sad

 

On his own website you can order several DVD's about astronomical image processing made by himself...  :hmm:   

The DSLR DVD is yours from  $60,00 to $80,00 :shocked:

 

Garrett van der Veen



#21 John_K

John_K

    Viking 1

  • -----
  • Posts: 766
  • Joined: 28 Apr 2005

Posted 17 November 2014 - 05:48 AM

Well, I am confused now.

 

So is the consensus that Drizzle works but what Tony recommends post that can be more effectively controlled simply by the standard DSS stacking process without the pre-precessing need?


Edited by John_K, 17 November 2014 - 05:48 AM.


#22 shawnhar

shawnhar

    Cosmos

  • *****
  • Posts: 9,452
  • Joined: 25 Jun 2010

Posted 17 November 2014 - 11:29 AM

IVO!!!! I did not know about Bayer Drizzle, you are awesome.



#23 nodalpoint

nodalpoint

    Apollo

  • *****
  • Posts: 1,332
  • Joined: 03 Jun 2013

Posted 17 November 2014 - 12:44 PM

It would be wise to follow Ivo's advice on this! Thanks for posting Ivo.

 

Anyone with access to Astronomy magazine can check back issues from fall of 2013 where Tony Hallas describes this process in several of his columns.



#24 TimN

TimN

    Gemini

  • *****
  • Moderators
  • topic starter
  • Posts: 3,415
  • Joined: 20 Apr 2008

Posted 17 November 2014 - 01:22 PM

Wow, thanks Ivo for taking the time to post this. It will certainly save me trying this stuff. Your advice makes complete sense. Other than dithering - which I do anyway - I will not try what Tony suggests.



#25 Agnotio

Agnotio

    Ranger 4

  • -----
  • Posts: 387
  • Joined: 29 Aug 2008

Posted 17 November 2014 - 07:42 PM

Dithering is only effective to reduce fixed pattern noise; his color mottle is simply the result of transformed (by the debayering) random shot noise; the 'Color Mottle' is never in the same place, so dithering isn't required to fix that. Tony's solution to use dithering is just a very roundabout way of fixing up his initial error of debayering his frames with Adobe Camera RAW. It's really a less effective way of applying Bayer Drizzle (scroll down the full page) of the RAW frames. The latter procedure doesn't introduce the mottling in the first place.

 

Hi Ivo, great post, thanks for that.

I don't know if it's entirely accurate though to say that color mottle isn't in the same place.  As I understand it, color mottle is largely the result of dark thermal signal, which is a property of the sensor.  Sure, additional color mottle can be created by debayering prior to stacking, but most people avoid this.

Essentially, Tony was right to advocate for dithering, but it was overblown in his examples because of the debayering problem that he introduced in his workflow (using Adobe Camera RAW).  Dithering is still very useful in getting rid of color mottle regardless.


Edited by Agnotio, 17 November 2014 - 07:50 PM.



CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics