Loading ...
Sorry, an error occurred while loading the content.

Re: [DeepSkyStacker] Re: Improving the way that DSS handles dark frames

Expand Messages
  • Kevin
    The problem is that you can t really perfectly match dark and light frames. Ahh but you can Jerry, have you forgotten the long exposure noise reduction
    Message 1 of 34 , May 1, 2010
    View Source
    • 0 Attachment
      "The problem is that you can't really perfectly match dark and light frames."



      Ahh but you can Jerry, have you forgotten the "long exposure noise reduction" option? : )

      I know, a bane for all but the most fortunate of us sitting out at the Atacama desert with virtually cloud free skies year round and endless stretches of capture time :D

      Back to reality ....



      At 12:57 PM 4/30/2010, you wrote:
      >Hi Jerry,
      >
      >Indeed the signal is theoretically linear with exposure time, as
      >long as the maximum values are not reached.
      >It starts with hot pixels, then with individual less hot pixels all
      >around the image, then the ampglow parts...
      >
      >I wouldn't use a 30 minute dark of a summer night to scale it down
      >if I could help it ;)
      >
      >Then of course it depends of the sensor. Some are more 'perfect'
      >than others and less prone to deviate from the linear.
      >I think Christian Buil did some tests quite some time ago. It might
      >still be available on his website.
      >
      >Anyway, that's why even if it's an oversimplification, it's easier
      >to match the dark and light frames exposures.
      >
      >Clear skies,
      >Luc
      >
      >--- In DeepSkyStacker@yahoogroups.com, Jerry Lodriguss <lists5@...> wrote:
      > >
      > >
      > > Hi Josh,
      > >
      > > I guess we could be arguing semantics here... as far as I know, dark
      > > signal is linear with exposure time.
      > >
      > > I think it's also linear with temperature (that's what the
      > > dark-signal doubling temp is about).
      > >
      > > So, it's basically linear, but if you change the temp during a
      > > sub-exposure run, like what happens in a DSLR, then the thermal
      > > signal is different from frame to frame (until it stabilizes in a
      > > couple of hours in the camera, and provided the ambient temp doesn't
      > > change) but that is because the temp is changing and not because
      > > thermal signal is not linear.
      > >
      > > Jerry
      > >
      > >
      >
      >
      >
      >
      >------------------------------------
      >
      >Yahoo! Groups Links
      >
      >
      >
      A Beginner's Guide to DSLR Astrophotography:
      http://www.astropix.com/BGDA/BGDA.HTM

      An Advanced Guide to DSLR Astrophotography:
      http://www.astropix.com/GADC/GADC.HTM

      .
      .
      .
      .
      .
      .
      .
      .
      .
      .
      .





      [Non-text portions of this message have been removed]
    • DaveH
      Blair Honestly...I keep looking at my DSLR...sitting forlornly on my table here...but it just sits. So I will not put you to more trouble..(!) 99.9% of my
      Message 34 of 34 , Jun 27, 2010
      View Source
      • 0 Attachment
        Blair
        Honestly...I keep looking at my DSLR...sitting forlornly on my table here...but it just sits.
        So I will not put you to more trouble..(!)
        99.9% of my understanding of SNR issues comes from sites like this;
        http://www.starrywonders.com/

        And frankly...(I am always frank..!)...I find that my opinions on SNR are not always in accordance with others..
        I was mostly going on about LENR to tweak our intrepid colleague JL..
        Because of a prior falling out we had on the subject.

        So there...I have fessed up,and feel better.
        As to the broader subject of noise reduction in DSLR's..
        well..IMHSHO..it is never very successful..no matter how the darks are subtracted..
        COOLER weather makes the main difference..
        As I may have mentioned months ago..at minus 20C (in Ottawa) I no longer saw ANY real benefit to dark subtraction..
        That was when I switched to a cooled/mono camera.
        And yet with a CCD camera I still do,..(a LOT)
        So this is enough for me...I do not feel the need to "bog down" on the math so much...Its the principals that interest me.
        By this I mean filters (less sky noise),signal (longer exposures) and cooling.
        I cannot change my (bad) sky in any event.
        The default position of the DSLR crowd (for faint signal) is always that if you stack enough frames...anything is possible..
        But this is only a THEORETICAL truth.
        The square root of N has a way of getting harder and harder to increase above 100 frames..
        Sure..in a zero noise camera thousands of brief images can be stacked...(lucky imaging)
        http://www.ast.cam.ac.uk/~optics/Lucky_Web_Site/

        But for DSLR's noise is too high.And cannot be subtracted enough.
        Let me close with a set of some 50 images of M57,taken with a DSLR,and with a mono CCD...

        http://www.flickr.com/photos/daveh56/sets/72157620881341430/

        In the DSLR images,good as I feel some are,NO outer/faint detail can be seen,on my or any other results I have seen from similar sky.
        However more recently I have STARTED to be able to get the fainter signal; (mono camera)
        http://www.flickr.com/photos/daveh56/4736718965/
        How..?
        Narrowband filters,and long/binned exposures.(720 seconds here)

        And yet I would ask how many DSLR images with a Ha filter can show this..?

        Is this chip QE..?
        I think not.
        Is it cooling..?
        No..we have minus 20C often up here..

        So it can ONLY be due to the MORE subtractable read and dark noise in mono chips.

        So..in closing..(!!);
        I think my spending more time understanding dark subtraction in DSLR's is not well spent...

        BECAUSE EXPERIENCE SHOWS IT DOES NOT WORK VERY WELL IN ANY EVENT...
        (sorry for that..)

        If you want to reply off list feel free...although I do not feel this is some kind of "inappropriate" topic here..

        And in defense of DSLR use...I am a faint signal freak,and have bad sky...
        If not, I would still be enjoying the 10 million pixel FOV and 600$ cost of my Canon.The ease of image processing and ability to use it for family photo"s. (pinkish ones)

        DSLR imaging remains a good thing....But I think we have WAY less control of SNR than we think (hope ?) we have.
        Dave







        --- In DeepSkyStacker@yahoogroups.com, "Blair MacDonald" <b.macdonald@...> wrote:
        >
        > Glad to hear you had no major damage!
        >
        > I can send you more details on the noise calculations off list (so as not to
        > bore everyone) if you want. One interesting thing that comes out of them is
        > that for light polluted skies there will be little difference between the
        > techniques. I noticed I said the images were taken from my yard in error,
        > they were taken from the St. Croix Observatory with very dark skies.
        >
        > Blair
        >
        >
        >
        > From: DeepSkyStacker@yahoogroups.com [mailto:DeepSkyStacker@yahoogroups.com]
        > On Behalf Of DaveH
        > Sent: June-27-10 2:30 AM
        > To: DeepSkyStacker@yahoogroups.com
        > Subject: [DeepSkyStacker] Re: Improving the way that DSS handles dark frames
        >
        >
        >
        >
        > Blair
        > Sorry...meant say KB...(!!)
        >
        > LENR file was 108 KB
        > nonLENR was 55KB...
        >
        > Had a few (small) cracks in my ceiling (plaster) and found all my kitchen
        > drawers open..
        > All cell phones died...because everyone called home at the same time...
        > etc,etc.
        >
        > Worst thing was I had to realign my mount...!
        > Not sure WHAT happened there...but all is OK now.
        > I have to say you 100% lost me on the noise math...but I will look at it
        > again.
        > I find my (limited and mostly zen like) understanding of SNR is much better
        > when I deal with a (mono) camera where every pixel gets the same
        > signal...and the same noise.
        > That alone,plus the cooling, was worth the switch.
        > Mostly I go by my eyeballs about what is working better/worse...
        > And the LENR sure looks worse.
        > So I will concede the idea as a bad one.
        >
        > It got foggy out,so I get to go to sleep at 2 AM.
        >
        > Dave
        >
        > --- In DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com> , "Blair MacDonald"
        > <b.macdonald@> wrote:
        > >
        > > Dave:
        > >
        > > Megabytes? There is a difference in size but both files are under 1 MB.
        > The
        > > in camera file is 948 kB while the master flat is 748 kB. The difference
        > is
        > > due to the noise in the image. JPEG compression will produce different
        > sizes
        > > for files, the more detail the larger the file. In this case the
        > additional
        > > noise looks like detail to the algorithm. Your comments about the colour
        > are
        > > spot on. Generally I fully colour correct the DSS output files, but I
        > chose
        > > not to do it on this one so I would not corrupt the data.
        > >
        > > Hope you didn't have any damage for the quake the other day! Good luck
        > with
        > > the processing.
        > >
        > > Blair
        > >
        > > From: DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > [mailto:DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com> ]
        > > On Behalf Of DaveH
        > > Sent: June-26-10 10:47 PM
        > > To: DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > Subject: [DeepSkyStacker] Re: Improving the way that DSS handles dark
        > frames
        > >
        > >
        > >
        > >
        > > Blair
        > > I have to admit that's quite a difference noise wise...
        > > Interestingly one file is 50MB and one is 100MB...but doubt that means
        > > much..
        > > Also color is better with LENR...(odd)
        > >
        > > Thanks for posting that...I am up to my A$$ in processing stuff tonite..
        > > Brain fog...you know how it goes..
        > > I will mull all this over !!
        > > :)
        > >
        > > Neil
        > > Congrats on the cooling box !!
        > >
        > > Dave
        > >
        > > In DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > , "Blair MacDonald" <b.macdonald@> wrote:
        > > >
        > > > I see the images didn't make it through so I've posted them on my PBase
        > > > site. The InCamera.jpg is at
        > http://www.pbase.com/astronut/image/125982611
        > > > while the stack processed with a master dark made from four different
        > dark
        > > > frames can be found at http://www.pbase.com/astronut/image/125982613.
        > > >
        > > > Blair
        > > >
        > > > From: DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > [mailto:DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com> ]
        > > > On Behalf Of Blair MacDonald
        > > > Sent: June-26-10 9:35 PM
        > > > To: DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > Subject: RE: [DeepSkyStacker] Re: Improving the way that DSS handles
        > dark
        > > > frames
        > > >
        > > >
        > > > Dave:
        > > >
        > > > You had better check your math. You have to account for noise in all the
        > > > frames. If you use 10 light (science) frames with a DN value of 4000 for
        > a
        > > > particular pixel in the sky background (a totally arbitrary value) then
        > > the
        > > > photon noise is 63.25. Assuming this is a sky limited exposure then we
        > can
        > > > ignore camera noise (fewer steps in the math). Now assume that the
        > > > equivalent dark has a value of 400 (both the light and dark have already
        > > > been bias corrected) so the dark noise is around 20 (again assuming that
        > > the
        > > > camera noise is negligible). The total noise in the result is then 66.33
        > > > while the light minus dark is 3600 for an SNR of 54.27. The sum of ten
        > > such
        > > > images produces a final noise of 209.76 and a signal of 36000 for an SNR
        > > of
        > > > 171.62. This example is the equivalent of in camera processing.
        > > >
        > > > Now if we use the same data, but instead make a master bias corrected
        > dark
        > > > of the ten frames we get a total noise of 201 because the master dark
        > has
        > > a
        > > > noise of 20/sqrt(10) or 6.32. This then yields a final SNR of 179.1.
        > This
        > > > represents a slight improvement in SNR, equivalent to what you might see
        > > in
        > > > light polluted suburban skies. Now let's see what happens from truly
        > dark
        > > > skies.
        > > >
        > > > With very dark skies, let's assume that the background reduces to 800.
        > > Dark
        > > > signal and noise remains the same and the photon noise in the image is
        > now
        > > > 28.28 (for convenience we are still ignoring camera read noise). This
        > > yields
        > > > a final SNR of 36.51 for the in camera case and 43.64 for the master
        > dark
        > > > case. In fact the darker the background the better the effect of the
        > > master
        > > > dark.
        > > >
        > > > I got a chance to shoot two test sequences from my backyard the other
        > > night.
        > > > Both are attached if the list will let them through. The InCamera.jpg
        > file
        > > > is a stack of four images taken with LENR on and the AverageDark.jpg
        > file
        > > is
        > > > a stack using the average of four darks for correction. The difference
        > is
        > > > quite clear with the LENR set clearly inferior as the math predicts. The
        > > > images are 100 percent crops of M20 with identical linear stretches
        > > applied
        > > > and converted to jpg's. If the images don't make it through I'll post
        > them
        > > > in the files section.
        > > >
        > > > Blair
        > > >
        > > > From: DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > [mailto:DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com> ]
        > > > On Behalf Of DaveH
        > > > Sent: June-24-10 5:58 PM
        > > > To: DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > >
        > > > Subject: [DeepSkyStacker] Re: Improving the way that DSS handles dark
        > > frames
        > > >
        > > >
        > > >
        > > > John
        > > > Move to Ottawa...!
        > > > You NW coast types are struggling,I know.
        > > >
        > > > I have been reviewing what I said today...the thing about patience keeps
        > > > coming up...
        > > > I get impatient for a "result"...EVERY time I start imaging...I do a
        > full
        > > > processing on a (preliminary) stack of the first two images...!
        > > >
        > > > I am about to shoot some 30 minute exposures...I can tell you that is a
        > > LONG
        > > > 30 minutes....I keep feeling "bumps" on my roof (obs),etc.
        > > > But it has to be done...or at least attempted.
        > > >
        > > > If I lived in Seattle I would probably just use NB filters,mono
        > > > camera,ignore the moon,and post a lot of Ha images !
        > > >
        > > > I also look forward to someone posting an actual image/comparison of the
        > > > LENR and standard master darks.
        > > >
        > > > Dave
        > > >
        > > > --- In DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com> , "W. John Tipton" <tiptonwj@>
        > > > wrote:
        > > > >
        > > > > Hi Dave,
        > > > >
        > > > > I'm sure everyone would like to know where you live. Dark skies every
        > > > night of the year where you can do astrophotography must be great. I'm
        > > lucky
        > > > to get 10-15 nights a year with clear skies and little or no moon glow.
        > > > That's the problem with living on the coast and on the Earth with that
        > > pesky
        > > > Moon about. Seriously though, it would be interesting to actually see
        > > > comparison images (not just numbers) as you suggested.
        > > > >
        > > > > John Tipton
        > > > > Ocean Shores, WA
        > > > >
        > > > >
        > > > > ----- Original Message -----
        > > > > From: DaveH
        > > > > To: DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > > Sent: Wednesday, June 23, 2010 7:41 PM
        > > > > Subject: [DeepSkyStacker] Re: Improving the way that DSS handles dark
        > > > frames
        > > > >
        > > > >
        > > > >
        > > > >
        > > > > Why doesn't someone "conduct that (LENR) experiment"...?
        > > > >
        > > > > Why not keep an open mind ?
        > > > > The best way to improve DSLR images (from non dark skies) is to
        > increase
        > > > (individual) exposure signal,(with exposure time) and reduce sky noise
        > > (with
        > > > filters)...and wait for cooler weather.
        > > > >
        > > > > Stacking and dark subtraction also help,clearly.But they are
        > > secondary,in
        > > > my opinion.So in that sense whatever way you try to reduce dark
        > noise...it
        > > > is not adding much.
        > > > > But the LENR idea remains valid,in my mind.I tried to discuss it in
        > the
        > > > DSLR group....(ouch !!)...but no one has ever bothered to actually post
        > a
        > > > comparison set of images.
        > > > > Like exposing 10 minutes instead of 5...it seems a simple enough
        > > > experiment.
        > > > >
        > > > > Sorry,but I was just "passing through"....
        > > > >
        > > > > As far as the extra imaging time required...well,it gets dark every
        > > night
        > > > of the year...so "wasting time" is not a real strong argument in this
        > > regard
        > > > IMHO.
        > > > > Shooting 150 exposures of an object (!) suggests that many would use
        > > LENR
        > > > if it could be shown to be superior...
        > > > >
        > > > > Patience IS a virtue.
        > > > > Theory without experimental testing is speculation.(or dogma)
        > > > > Non critical discussion is the key.
        > > > >
        > > > > It (LENR) or ICDS as I called it back in the day, always struck me as
        > a
        > > > good one.
        > > > > On switching gear I have lost the motivation to test it.
        > > > >
        > > > > Dave
        > > > >
        > > > > --- In DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com> , Jerry Lodriguss <lists5@>
        > > wrote:
        > > > > >
        > > > > >
        > > > > > Hi Ray,
        > > > > >
        > > > > > I'm a little bit unclear as to your methodology... did you in fact
        > > > > > shoot a set of lights with in-camera long-exposure noise reduction
        > > > > > turned on, and then another set of lights with it off, and then a
        > set
        > > > > > of separate darks?
        > > > > >
        > > > > > The fact that your numbers yielded EXACTLY the same results to 8
        > > > > > decimal places makes the results highly suspicious. There is no way
        > > > > > the temperatures could have been exactly the same between these
        > three
        > > > > > sets of frames on a DSLR, so I would say this result was impossible.
        > > > > >
        > > > > > But, regardless of that, here is the big point that you are
        > > > > > missing... you are wasting incredibly valuable clear dark sky time
        > > > > > shooting in-camera darks when you could have been using this time to
        >
        > > > > > gather more signal to improve the signal-to-noise ratio. In fact,
        > > > > > there is no other way to improve the signal side of this equation
        > > > > > except through more exposure time in your light frames.
        > > > > >
        > > > > > Also, it is incredibly easy to shoot your darks on a cloudy
        > > > > > night. That way, you can shoot a whole bunch of them, and make a
        > > > > > really good master dark frame. What is so hard about this? It's
        > > > > > cloudy, and you have nothing else to do!
        > > > > >
        > > > > > The real experiment you should have done would have been to compare
        > > > > > the final signal-to-noise ratio in, say, 10 light frames with
        > > > > > in-camera long-exposure noise reduction (in-camera dark frame
        > > > > > subtraction) to 20 light frames with no LENR but calibrated with a
        > > > > > separate master dark.
        > > > > >
        > > > > > There is no way you would get equivalent results with this
        > experiment
        > > > > > because you would have twice the signal, and less noise injected
        > with
        > > > > > an excellent master dark.
        > > > > >
        > > > > > Therefore, the bottom line is... you will get better results and a
        > > > > > better signal-to-noise ratio in a given amount of clear dark sky
        > time
        > > > > > by continuously shooting light frames, and shooting separate darks.
        > > > > >
        > > > > > Isn't this what is really important?
        > > > > >
        > > > > > I guess one of the complaints you have is that your master dark is
        > > > > > still not well matched even with DSS's dark optimization... I can't
        > > > > > really speak to that, but I can tell you that this is not a problem
        > > > > > with automatic dark frame matching in Images Plus.
        > > > > >
        > > > > > Jerry
        > > > > >
        > > > > >
        > > > > >
        > > > > >
        > > > > > At 09:19 AM 6/23/2010, you wrote:
        > > > > >
        > > > > > >Well I started this thread, and now I'm coming back to it after a
        > > > break...
        > > > > > >
        > > > > > >Jerry, what you wrote below is also what I really expected to be
        > the
        > > > > > >case; but after taking some images at the weekend, I was mulling
        > > > > > >over it again last night and I wasn't really satisfied that the S/N
        >
        > > > > > >case was proven.
        > > > > > >
        > > > > > >So I did some tests in Excel today. And what I found was very
        > > > > > >interesting for Deep Sky Stacker users in this situation!
        > > > > > >
        > > > > > >The question I was examining is this: is it really worse from a S/N
        >
        > > > > > >perspective for a DSS user to employ "long-exposure processing"
        > > > > > >(in-camera dark frame taking and subtraction), instead of taking
        > > > > > >dark frames separately?
        > > > > > >
        > > > > > >In other words, is it really worse to take (say) 10 science frames
        > > > > > >with 10 in-camera dark frame subtractions; instead of 10 unaltered
        > > > > > >science frames, along with 10 dark frames of equal duration, which
        > > > > > >are separately stacked?
        > > > > > >
        > > > > > >We all understand and believe (correctly) that it's better to
        > > > > > >subtract a deep, averaged dark frame from a single science frame,
        > > > > > >than to subtract a short single dark frame from that single science
        >
        > > > > > >frame. But what about ensembles of science frames? Does this change
        > > > things?
        > > > > > >
        > > > > > >So I formulated the question thus: which has lower noise:
        > > > > > >the average of many [single frame - single dark]
        > > > > > >OR [average of many single frames] - [average of many single darks]
        > > > > > >...or are they actually the same?
        > > > > > >
        > > > > > >I've pasted my Excel work here. The numbers can be changed to
        > > > > > >anything; but the result is the same.
        > > > > > >
        > > > > > >----------------------------------------------------------
        > > > > > >NB From Excel Help: If you type in any cell the formula
        > > > > > >NORMINV(rand(),mu,sigma), you will generate a simulated value of a
        > > > > > >normal random variable having a mean mu and standard deviation
        > sigma.
        > > > > > >
        > > > > > >science pixel expected mean 100
        > > > > > >science pixel expected stddev 10
        > > > > > >dark (readnoise) expected mean 5
        > > > > > >dark (readnoise) expected stddev 3
        > > > > > >
        > > > > > >Science Pixel value Dark value Sci-Dark
        > > > > > >110.1708117 6.345658485 103.8251533
        > > > > > >85.76714854 5.430469359 80.33667918
        > > > > > >73.63786159 0.553469799 73.08439179
        > > > > > >83.88366663 8.263831047 75.61983558
        > > > > > >107.8439403 10.6056668 97.23827352
        > > > > > >95.84700254 7.365840233 88.48116231
        > > > > > >126.7948701 3.599151613 123.1957184
        > > > > > >107.0393959 8.510845687 98.52855017
        > > > > > >120.5378321 2.318231539 118.2196006
        > > > > > >79.69591954 2.981439116 76.71448042
        > > > > > >95.91562105 3.337949254 92.5776718
        > > > > > >104.3548672 6.88346532 97.47140186
        > > > > > >123.5308385 6.382802492 117.148036
        > > > > > >86.9848109 -2.471265402 89.4560763
        > > > > > >99.62030586 5.560751275 94.05955458
        > > > > > >89.52160833 3.369305426 86.15230291
        > > > > > >97.62052207 8.360958106 89.25956396
        > > > > > >
        > > > > > >AVERAGES
        > > > > > >Avg(Sci) 99.33923664
        > > > > > >Avg(Dark) 5.141092362
        > > > > > >Avg(Sci-Dark) 94.19814428
        > > > > > >Avg(Sci) - Avg(Dark) 94.19814428
        > > > > > >
        > > > > > >----------------------------------------------------------
        > > > > > >
        > > > > > >Thus proving that Avg(Sci-Dark) is EXACTLY the same as Avg(Sci) -
        > > > > > >Avg(Dark). The same output signal value implicitly means the same
        > > > > > >noise.
        > > > > > >
        > > > > > >The question matters because since they are the same, and if you
        > are
        > > > > > >using DeepSkyStacker (which demands dark exposures of equal shutter
        >
        > > > > > >speeds as the science frames), then there is NO S/N disadvantage to
        >
        > > > > > >using in-camera "long exposure processing" with cameras/MFDBs
        > > > > > >...unless you spend more time on darks (take more dark frames) than
        >
        > > > > > >the time on the science images.
        > > > > > >
        > > > > > >If you were using something other than DSS, you would probably take
        >
        > > > > > >longer dark frames rather than more of the same shorter ones, but
        > > > > > >you still need to spend more time on darks to see any difference.
        > > > > > >
        > > > > > >And since the in-camera "long exposure processing" option also
        > > > > > >achieves a much better temperature match of science and dark frame,
        >
        > > > > > >with non-temperature-regulated cameras, it will actually deliver
        > > > > > >superior results.
        > > > > > >
        > > > > > >I had to investigate this because although I was trying hard this
        > > > > > >weekend to get darks at the same temperature as my Comet McNaught
        > > > > > >science frames (by doing a batch of darks in between each batch of
        > > > > > >science frames), they still were not that well matched (even after
        > > > > > >"dark optimization"), and I was convinced that the in-camera dark
        > > > > > >subtracted ones I took in the same session were better - they were
        > > > > > >certainly squeaky clean of dark current, but I also sensed that
        > they
        > > > > > >were no worse in terms of readnoise as well. And now I know that
        > > > > > >they indeed were precisely the same in terms of readnoise.
        > > > > > >
        > > > > > >So from now on, unless and until I crack the issue of perfect
        > > > > > >thermal matching with hundreds of library darks or dark
        > > > > > >optimization, I will be going back to in-camera "long exposure
        > > > > > >processing", with no more feelings of guilt about it!
        > > > > > >
        > > > > > >Ray
        > > > > > >
        > > > > > >
        > > > > > >--- In DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com> , Jerry Lodriguss <lists5@>
        > > wrote:
        > > > > > > >
        > > > > > > > In-Camera long-exposure nose reduction is a very poor solution.
        > > > > > > >
        > > > > > > > It's not only losing clear darky-sky time for gathering photons
        > to
        > > > > > > > improve the s/n ratio...
        > > > > > > >
        > > > > > > > It's also that since you're only using a single dark frame, so
        > you
        > > > > > > > are actually adding a lot of noise to the light frame when you
        > > > > > > > subtract the thermal signal. You add 4x more noise with a single
        > > > dark
        > > > > > > > than, say, a stack of 16 darks.
        > > > > > > >
        > > > > > > > Jerry
        > > > > > > >
        > > > > > > >
        > > > > > > >
        > > > > > > > At 03:41 AM 5/1/2010, you wrote:
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >"The problem is that you can't really perfectly match dark and
        > > > > > > light frames."
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >Ahh but you can Jerry, have you forgotten the "long exposure
        > > noise
        > > > > > > > >reduction" option? : )
        > > > > > > > >
        > > > > > > > >I know, a bane for all but the most fortunate of us sitting out
        > > at
        > > > > > > > >the Atacama desert with virtually cloud free skies year round
        > and
        > > > > > > > >endless stretches of capture time :D
        > > > > > > > >
        > > > > > > > >Back to reality ....
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >At 12:57 PM 4/30/2010, you wrote:
        > > > > > > > > >Hi Jerry,
        > > > > > > > > >
        > > > > > > > > >Indeed the signal is theoretically linear with exposure time,
        > > as
        > > > > > > > > >long as the maximum values are not reached.
        > > > > > > > > >It starts with hot pixels, then with individual less hot
        > pixels
        > > > all
        > > > > > > > > >around the image, then the ampglow parts...
        > > > > > > > > >
        > > > > > > > > >I wouldn't use a 30 minute dark of a summer night to scale it
        > > > down
        > > > > > > > > >if I could help it ;)
        > > > > > > > > >
        > > > > > > > > >Then of course it depends of the sensor. Some are more
        > > 'perfect'
        > > > > > > > > >than others and less prone to deviate from the linear.
        > > > > > > > > >I think Christian Buil did some tests quite some time ago. It
        > > > might
        > > > > > > > > >still be available on his website.
        > > > > > > > > >
        > > > > > > > > >Anyway, that's why even if it's an oversimplification, it's
        > > > easier
        > > > > > > > > >to match the dark and light frames exposures.
        > > > > > > > > >
        > > > > > > > > >Clear skies,
        > > > > > > > > >Luc
        > > > > > > > > >
        > > > > > > > > >--- In DeepSkyStacker@yahoogroups.com
        > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com>
        > > > <mailto:DeepSkyStacker%40yahoogroups.com> , Jerry Lodriguss <lists5@>
        > > wrote:
        > > > > > > > > > >
        > > > > > > > > > >
        > > > > > > > > > > Hi Josh,
        > > > > > > > > > >
        > > > > > > > > > > I guess we could be arguing semantics here... as far as I
        > > > know, dark
        > > > > > > > > > > signal is linear with exposure time.
        > > > > > > > > > >
        > > > > > > > > > > I think it's also linear with temperature (that's what the
        > > > > > > > > > > dark-signal doubling temp is about).
        > > > > > > > > > >
        > > > > > > > > > > So, it's basically linear, but if you change the temp
        > during
        > > a
        > > > > > > > > > > sub-exposure run, like what happens in a DSLR, then the
        > > > thermal
        > > > > > > > > > > signal is different from frame to frame (until it
        > stabilizes
        > > > in a
        > > > > > > > > > > couple of hours in the camera, and provided the ambient
        > temp
        > > > doesn't
        > > > > > > > > > > change) but that is because the temp is changing and not
        > > > because
        > > > > > > > > > > thermal signal is not linear.
        > > > > > > > > > >
        > > > > > > > > > > Jerry
        > > > > > > > > > >
        > > > > > > > > > >
        > > > > > > > > >
        > > > > > > > > >
        > > > > > > > > >
        > > > > > > > > >
        > > > > > > > > >------------------------------------
        > > > > > > > > >
        > > > > > > > > >Yahoo! Groups Links
        > > > > > > > > >
        > > > > > > > > >
        > > > > > > > > >
        > > > > > > > >A Beginner's Guide to DSLR Astrophotography:
        > > > > > > > >http://www.astropix.com/BGDA/BGDA.HTM
        > > > > > > > >
        > > > > > > > >An Advanced Guide to DSLR Astrophotography:
        > > > > > > > >http://www.astropix.com/GADC/GADC.HTM
        > > > > > > > >
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >.
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >[Non-text portions of this message have been removed]
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >------------------------------------
        > > > > > > > >
        > > > > > > > >Yahoo! Groups Links
        > > > > > > > >
        > > > > > > > >
        > > > > > > > >
        > > > > > > > A Beginner's Guide to DSLR Astrophotography:
        > > > > > > > http://www.astropix.com/BGDA/BGDA.HTM
        > > > > > > >
        > > > > > > > An Advanced Guide to DSLR Astrophotography:
        > > > > > > > http://www.astropix.com/GADC/GADC.HTM
        > > > > > > >
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > > .
        > > > > > > >
        > > > > > >
        > > > > > >
        > > > > > >
        > > > > > >
        > > > > > >------------------------------------
        > > > > > >
        > > > > > >Yahoo! Groups Links
        > > > > > >
        > > > > > >
        > > > > > >
        > > > > > A Beginner's Guide to DSLR Astrophotography:
        > > > > > http://www.astropix.com/BGDA/BGDA.HTM
        > > > > >
        > > > > > An Advanced Guide to DSLR Astrophotography:
        > > > > > http://www.astropix.com/GADC/GADC.HTM
        > > > > >
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > > .
        > > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > >
        > > > > [Non-text portions of this message have been removed]
        > > > >
        > > >
        > > > [Non-text portions of this message have been removed]
        > > >
        > > >
        > > >
        > > > [Non-text portions of this message have been removed]
        > > >
        > >
        > >
        > >
        > > [Non-text portions of this message have been removed]
        > >
        >
        >
        >
        > [Non-text portions of this message have been removed]
        >
      Your message has been successfully submitted and would be delivered to recipients shortly.