Evaluation of Sensor Analysis & Smart Histogram

Anything that doesn't fit into any of the other forums
Post Reply
User avatar
turfpit
Posts: 1779
Joined: Mon Feb 13, 2017 8:13 pm
Location: UK
Contact:

Evaluation of Sensor Analysis & Smart Histogram

#1

Post by turfpit »

Folks

I am starting to think about writing the documentation update from SC 2.9 to 3.1. One of the major features in SC 3.1 is the introduction of the Smart Histogram which relies on having run the Sensor Analysis tool. I have been searching for material/examples that could be included in the documentation. There is plenty of evidence of people running the Sensor Analysis and posting results but very little evidence of outcomes from using Smart Histogram recommended settings.

So 2 questions:
  • Has anyone done any systematic comparisons between 'traditional' capture settings and those being offered by the Sensor Analysis/Smart Histogram?
  • Using the Sensor Analysis/Smart Histogram combination, what impact has there been on your post-processing techniques and final image outcomes?
Smart Histogram has great potential and in order for it to evolve I would have thought feedback would assist the development.

Dave
User avatar
oopfan
Posts: 1321
Joined: Sat Jul 08, 2017 2:37 pm
Location: New York
Contact:

Re: Evaluation of Sensor Analysis & Smart Histogram

#2

Post by oopfan »

Dave,

I hope this is helpful. Since the beginning of 2018 I used my kit five times to take astrophotos with the assistance of The Brain (i.e. Smart Histogram). That was from January 14th through February 12th. In March I was busy taking an online course on CCD Photometry. All astrophotos since then I've used what I learned in the course to set exposure, not using The Brain.

I want to say upfront something that might render the data I am about to present useless. The Sensor Analysis I used was dated January 9th using the version of SharpCap that was current as of that date. The image of M51 that I will present used the latest version of SharpCap as of February 12th but using that sensor analysis data from January 9th.

M51 was the last DSO I captured using The Brain (see attachments). I followed the standard procedure of defining a region around an area of the frame devoid of stars and nebulosity. I selected "Unity Gain" in the dropdown box. It recommended an exposure of 9.4 seconds. I applied the settings and then dialed in an offset of 20. I tried Live Stack for the first time. I asked SharpCap to save the individual frames to FITS files that I later processed using Deep Sky Stacker and StarTools.

I collected a total of 510 Lights, 50 Darks, 50 Flats, and 50 Bias frames. I asked DSS to keep the best 80% for a new total of 407 frames and a total integration time of 64 minutes. Post-processing in StarTools was very challenging due the raining noise in the faintest areas of the nebulosity that I wanted to bring out. If you've got your monitor's brightness and contrast turned up you might see the vertical stripes in the attached final image.

As you may know I guide passively. I rely on the accuracy of my polar alignment and my mount's Periodic Error Correction in right ascension. There is no motor control in declination. For exposures of 9.4 seconds I can get very round stars but over the course of an hour the image drifts in right ascension and declination by 20 to 30 pixels, sometimes more. This is no problem when stacking since DSS pulls all of the frames into alignment. The problem with the raining noise rears its ugly head when I try to extract detail from signal that is close to the noise level.

As I know now the raining noise problem can be solved through proper exposure. If you look at the FITS file attachment that I stretched in order to show what a frame might look like in SharpCap using the Histogram Stretch feature, notice that the faint wisps of nebulosity around the satellite galaxy is just not visible. Through the miracle of stacking I was able to bring that out but it was mired in raining noise. Now I know that the 9.4-second exposure was insufficient. I talk about this in greater detail in several posts I made recently in a topic called "Exposure" in the General Discussions section.

One common occurrence in all five images I've taken using The Brain is that I consider the frames to be underexposed especially for someone like myself who has problems with raining noise. If I actively guided or I used dithering then I would expect the problem to go away. My skies are quite dark here. I can tolerate much longer exposures before sky brightness becomes a problem. I am not entirely sure why The Brain thinks that my sky is brighter than it really is.

By the way the "maximum pixel value" of 26304 in the "Image statistics" section of FITS Liberator is for a star far out of the field of view. So even if The Brain doubled the exposure the star would not saturate.

Brian

PS:

My kit consists of a William Optics 71mm f/5.9 refractor, Altair 290M camera, a Unitron Model 152 equatorial mount, no active guiding, just Periodic Error Correction (PEC).
Attachments
M51_darkss_00000.CameraSettings.txt
(465 Bytes) Downloaded 155 times
M51 subframe stretched.jpg
M51 subframe stretched.jpg (223.42 KiB) Viewed 3468 times
M51 processed cropped.jpg
M51 processed cropped.jpg (32.12 KiB) Viewed 3468 times
User avatar
turfpit
Posts: 1779
Joined: Mon Feb 13, 2017 8:13 pm
Location: UK
Contact:

Re: Evaluation of Sensor Analysis & Smart Histogram

#3

Post by turfpit »

Thanks Brian. Useful info including the Image Statistics for a FITS frame plus a nice M51 with no core burnout.


Dave
User avatar
admin
Site Admin
Posts: 13177
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Evaluation of Sensor Analysis & Smart Histogram

#4

Post by admin »

Just a quick note to remind people what the smart histogram is designed to do...

* It works out the exposure time and gain combination (between user specified exposure limits) beyond which it's not worth increasing the exposure further based on calculation of the total noise level in the final stacked image assuming a fixed total stacking time.

What that means is that if the smart histogram says 'use 120s exposures at 100 gain' then you can be sure that you won't see fainter details or less random noise in the final image if you decide to go for 12x300s exposures over an hour instead of the 30x120s exposures suggested.

Does that mean that you always have to use the exact value suggested? No, not at all...

* You might have good reasons for selecting a longer value (such as Brian's situation above) or if Smart Histogram suggests 1s exposures but you don't want to save and then stack 3600 frames!

* You might select a shorter exposure for reasons such as tracking issues or wanting to reduce satellite/plane trail issues by taking shorter frames.

In either case you should use the minimum and maximum exposure dropdowns in the brain to specify the limits that you want to set on the exposure time - for instance if your mount can only handle 30s exposures due to poor tracking then set a maximum of 30s. If you want at least 120s exposures to limit frame count then set a minimum value of 2 minutes. If you do this then the calculations can take account of your limits and give you the optimum combination of gain and exposure within the range you have chosen (in particular this might suggest a different gain than if you did not apply an exposure range).

Hope this helps,

Robin
GregL
Posts: 68
Joined: Sun Apr 30, 2017 6:19 pm

Re: Evaluation of Sensor Analysis & Smart Histogram

#5

Post by GregL »

I have been playing with it briefly using my new QHY183c. Did the analysis using RAW16, tried to measure sky background with 'brain' but never got a result. Am curious as to whether there is anything I can do to speed things up... ?

Also, curious as to why, when one opens the histogram with the RGB24 profile as to the point of suggesting I should do a sensor analysis for all the profiles I intend to use? If I have anything but RAW the sensor analysis refuses to run.

greg latiak
avalon observatory
In perpetually cloudy Ontario
User avatar
turfpit
Posts: 1779
Joined: Mon Feb 13, 2017 8:13 pm
Location: UK
Contact:

Re: Evaluation of Sensor Analysis & Smart Histogram

#6

Post by turfpit »

Greg

See http://docs.sharpcap.co.uk/3.2/#Running ... 20Analysis. RAW mode is required for colour cameras.

Dave
GregL
Posts: 68
Joined: Sun Apr 30, 2017 6:19 pm

Re: Evaluation of Sensor Analysis & Smart Histogram

#7

Post by GregL »

I understand that. The question was why does the software suggest I run a sensor analysis when I am using an RGB profile with a colour camera? I have a raw analysis for this camera already. And if the analysis process requires 'raw' then what is the point of suggesting I do a sensor analysis for all the modes the camera supports? That message should not be there...

gregl
User avatar
turfpit
Posts: 1779
Joined: Mon Feb 13, 2017 8:13 pm
Location: UK
Contact:

Re: Evaluation of Sensor Analysis & Smart Histogram

#8

Post by turfpit »

Greg

I think only Robin can answer that one. Probably just the text that needs refining.

Dave
Post Reply