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

Canon 200D/SL2 Nikon D5300 Comparative Review

  • Please log in to reply
34 replies to this topic

#1 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 02:56 AM

Finally, here is my comparative review of the Canon 200D/SL2 and Nikon D5300.  It will be split across a number of posts because of the size limit for images in a single post. 

 

If you want to skip to end, the conclusion is here:

https://www.cloudyni...view/?p=8903240

 

I bought the Canon 200D/SL2 and Nikon D5300 because I wanted an "entry-level" backup camera for astro-imaging.  The idea was to compare the two and then keep the "winner".  Why did I select these two cameras in the first place? The Nikon D5300 was chosen because it is inexpensive and seems to be one of the most recommended budget DSLR cameras for astro-imaging.  The Canon 200D was chosen because it is believed to be the first Canon entry-level camera that has on-sensor analog digital converters (ADC) instead of using a separate off sensor unit.  It was thought this should make it closer to being an ISO-less camera than any previous entry level Canon.  Unfortunately this turned out not to be the case but now I'm jumping ahead of myself.

 

My review is mainly focused on technical details pertinent to astrophotography that you won't find elsewhere.  So I won't review usability, touch screens, fold out LCD panels etc.  I didn't even notice that the Canon 200D had a touch screen until I brushed it by accident!

 

Let's start with read noise etc.

 

Read Noise, Gain etc.
 

Read noise at each ISO can be looked up at PhotonsToPhotos.net so I won't repeat it here:
http://www.photonsto...,Nikon D5300_14

 

This shows that the Nikon has slightly lower read noise.  My figures were similar to those shown except for Canon 200D read noise at ISO 100 and ISO 200. 

 

Canon 200D read noise:
                   Mine    PhotonsToPhotos
ISO 100       7.4e           4.9e
ISO 200       5.0e           3.6e

 

My figures showed a much steeper rise with a read noise of 7.4e at ISO 100 and 5.0e at ISO 200 instead of 4.9e and 3.6e respectively.  I've been sent bias frames from other 200D owners that agree with my figures.  Neither Bill Claff (the author of PhotonsToPhotos) nor myself understand this difference.  It remains a mystery.

 

Estimates of other parameters such as quantum efficiency (QE) and the ISO that gives unity gain can be found here:
http://www.photonsto...,Nikon D5300_14

 

My values for the ISO that gives unity gain were 215 for the Nikon and 240 for the Canon.  Again we are seeing a big difference on the Canon for some unknown reason.  It should be pointed out that one cause of slight differences is the calculation of gain. 

Neither camera has a completely linear curve of pixel variance (the square of standard deviation) against mean pixel value so one tends to obtain slightly different estimates of gain depending on how bright the pixel values are. 

 

Note also that the Nikon value is for the green pixels.  The Nikon red and blue channels are scaled differently, as is typical of Nikon cameras.  One noticeable effect of this is histogram combing in the red and blue channels.  However it's not thought to have any effect on astro-images.

 

QE will be touched upon later in the review when the spectral response is examined.


Edited by sharkmelley, 21 October 2018 - 04:40 AM.

  • tkottary and Shawnmac44 like this

#2 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 02:59 AM

Nikon D5300 Calibration Frames
 

Here are the master bias and master darks created at ISO 400.  A 2 minute exposure was used for the darks.  The same linear scaling has been applied to both to allow direct comparison.

 

NikonD5300_Bias.jpg NikonD5300_Dark.jpg

 

Amp glow is quite noticeable along the bottom edge.  This means it is potentially quite important to subtract a master bias during processing even if dark frames are ignored.  It will prevent a magenta band at the bottom of the image.


Edited by sharkmelley, 21 October 2018 - 03:00 AM.

  • bobzeq25 likes this

#3 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:02 AM

Canon 200D Calibration Frames
 

Here are the master bias and master darks created at ISO 400.  A 2 minute exposure was used for the darks.  The same scaling used for the Nikon darks has been used again here to allow direct comparison between Nikon and Canon:

 

Canon200D_Bias.jpg Canon200D_Dark.jpg

 

There is no amp glow on the Canon master bias.  It shows some vertical striations but these are at the sub-electron level and will not pose a problem.  The bigger issue is the fixed pattern noise in the master dark. Since this will also appear in the light frames, it makes it more necessary to calibrate light frames with a master dark.


Edited by sharkmelley, 21 October 2018 - 04:09 AM.


#4 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:06 AM

Dark Current Charts

A comparison of dark frames is very interesting. Here are some image crops giving a side by side comparison.  This is the extracted undebayered green channel because otherwise the Nikon scaling of the red and blue channels would distort the statistics.

 

CanonVsNikonDarks.jpg

 

They are 5 minute darks taken at ISO 800 at an ambient temperature of 20C using cameras that had acclimatised to ambient temperature for at least a couple of hours while switched off.  So, for instance, there was no sensor heating resulting from using live-view.

 

I've converted the standard deviation of the noise to electrons using the gains I calculated.

 

It can be seen that the Canon 200D dark frame is much noisier than the Nikon D5300. But to be fair it should be pointed out that the Nikon does apply spatial filtering to longer exposures which removes the obvious bright pixels. However, it definitely demonstrates why many people don't see the need to use dark calibration frames on the D5300 - because of those low noise levels.  But I consider that calibration with darks is pretty much essential on the Canon 200D to remove the thermal fixed pattern noise (FPN).

 

The bottom two panes of the above diagram show the result of subtracting one dark from another and scaling the result by 1/sqrt(2). Even with dark subtraction the 200D is still not down to the low noise levels of the D5300. That being said, the Nikon spatial filtering does not remove all fixed pattern noise, so there is still a small benefit to be obtained from dark calibration. 

 

This benefit may or may not be noticeable in real long exposure images, depending on many factors such as ambient temperature, speed of optics, light pollution levels, dithered acquisition etc.


Edited by sharkmelley, 21 October 2018 - 04:12 AM.


#5 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:10 AM

Dark Current Charts continued

 

One other thing is that in continual use, taking long exposures typical of deep sky imaging, the internal camera temperature (and therefore the thermal noise) climbs quite sharply. Again the 200D is worse than the Nikon D5300. The rear panel of the Canon 200D (behind the articulating display) becomes quite warm to the touch.

 

I've added both the Canon 200D and Nikon D5300 to my graphs of other cameras I've tested:

 

DarkCurrentGraphs5.png

 

The noise figures were generated from subtracted pairs of dark frames and then converted to dark current.  The per pixel dark current (and hence per pixel noise) seems to be average or better than average compared to other cameras.  However, remember that their respective pixel sizes (3.7microns for the 200D and 3.9microns for the D5300) are much smaller than the other cameras. 

 

Therefore my preferred chart adjusts for the pixel size and shows dark current per unit area of sensor:

 

DarkCurrentGraphs5_v2.png

This gives a better comparison of total thermal noise across the sensor taken as a whole.  On this measure the Canon 200D comes out worse than any camera I've ever tested and the Nikon D5300 is a very average performer.  For both of them it reflects a big increase in internal camera temperature during the course of a long imaging session.  This temperature increase is not caused by reading the data from the sensor, as is often misreported.  For instance if the camera is just sitting there taking a super long exposure then the temperature is still climbing just as fast.



#6 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:15 AM

Canon Banding Issue
 

I took a test image of NGC7000 which will be presented later. However I noticed some coloured horizontal banding in the final stacked image.  Looking at individual subs I found nasty coloured horizontal bands that moved around from exposure to exposure. 

 

This single exposure is a typical example (2min at ISO 400):

 

Canon200D_light_ISO400_2min.jpg

 

Over on the DPReview forums, Roger Clark suggested to me it might be the battery voltage - apparently he never lets batteries drop below 50% on his Canon cameras because of banding issues.  Just to be sure, I fully charged a battery and then took a 2minute dark, this time at ISO 800 in case the higher ISO helped. 

 

Here is the result:

 

Canon200D_ISO800_2min_dark.jpg

 

The banding is still quite obvious. It's possible that this problem can be mitigated by using an even higher ISO e.g. 1600 or 3200 and ensuring the back of camera histogram peak (due to sky fog) is 1/3 from the left hand side in an attempt to make the signal swamp the banding.

 

However the point here is that taken together with the increasing read noise at low ISOs this is not a camera that considered ISO-less in any sense of the word.


Edited by sharkmelley, 21 October 2018 - 04:14 AM.


#7 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:19 AM

Spectral Response
 

It's interesting to see the differences in spectral response from the two cameras because I have always suspected a big difference between Sony sensors and Canon sensors.  Here are two spectra taken minutes apart using the same equipment.

 

They were opened in Photoshop/ACR with fairly neutral processing:

 

SunlightSpectra.jpg

 

I obtained almost identical colours using PixInsight and applying the White Balance and sRGB Colour Matrices from from DXO site, followed by a constant gamma of 2.2

 

The vertical lines are of course the Fraunhofer lines in the sun's spectrum  - I used those to align the spectra.  The Canon has a wider response down past blue and up beyond red.  However the main difference from an astro-imaging viewpoint is near the OIII emission. 

 

The Canon gives a cyan colour for this emission whereas the Nikon is much bluer.  This will affect the colour of any emission nebulosity containing significant OIII emissions.

 

Using the gain factors I calculated for each camera and allowing for the pixel size difference between them (3.7microns for the Canon and 3.9 microns for the Nikon) it was possible to convert the raw digital values to electrons per unit area of sensor. Note that I have also taken account of the fact that the Nikon scales the red & blue channels differently to the green channel.

 

Here is the resulting plot:

 

Canon200D_vs_NikonD5300.png

 

For convenience I have marked the position of the common hydrogen and oxygen emission lines. The sharp dips on the curves correspond to the Fraunhofer lines.  You can also see how the relative sensitivities of the camera at each wavelength translate into the colours seen on the first chart.  The difference in green sensitivity near the OIII lines is quite pronounced.

 

The Nikon has higher peaks in each colour channel, which suggests either the sensor itself is more sensitive or the colour dyes on the CFA matrix have a higher transmission. Note that the vertical scale on this chart is completely arbitrary.


Edited by sharkmelley, 21 October 2018 - 04:17 AM.


#8 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:38 AM

Spectral Response continued

 

To demonstrate how different emission nebulae look using the two cameras I took a real astro-image.  I deliberately imaged a nebula that everyone is familiar with, the North America Nebula NGC7000.

 

The same Skywatcher ED80 scope was used for both and in both cases I collected 30x2min exposures at ISO 400. Unfortunately sky conditions changed slightly between the two images as you can see by the haze around the bright stars in the Nikon image.

 

I used the same processing sequence for both images in PixInsight with lights, darks, flats and bias.  The white balance and sRGB colour matrix were taken from the DXO website and a constant gamma of 2.2 was applied.

 

Here is the Canon 200D image:

 

Canon200D_NGC7000_v2.jpg


Edited by sharkmelley, 21 October 2018 - 03:38 AM.


#9 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:39 AM

Spectral Response continued again

 

Here is the Nikon D5300 image:

 

NikonD5300_NGC7000_v2.jpg

 

In retrospect, a better example might have been the Dumbbell Nebula. But in any case it is still very apparent how differently the two cameras present the colours. 

 

However, note that the colours of the main stars are almost identical in each image.  This is to be expected because stars are broad spectrum emissions and all DSLR cameras should render them fairly accurately because the (compromise) colour matrix is calibrated to broad spectrum sources such as a Gretag Macbeth colour chart.


Edited by sharkmelley, 21 October 2018 - 03:41 AM.

  • tkottary likes this

#10 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:43 AM

QE Estimate
 

Quantum efficiency is usually calculated from the green channel.  Although I'm not sure how to calculate the absolute QE, by looking at the number of electrons recorded by the green channel in sunlight I estimated the D5300 to have a green channel QE 1.05x higher than the 200D.  This is the area under the green curve

 

Bill Claff on the PhotonsToPhotos site shows a QE of 55% for the Nikon and 49% for the Canon, using DXO data:

http://www.photonsto...,Nikon D5300_14

 

Bill also gave an unpublished figure of 54% for the D5300 and 51% for the 200D.  This gives a factor of 54/51 = 1.06x.  That's pretty good agreement with my ratio given that inaccuracies can come in from many different places.

 

In any case, the difference in QE is pretty small and won't make a noticeable difference in any image.  Also a single QE figure certainly doesn't represent the whole story because of significant differences in the shapes of the curves.


Edited by sharkmelley, 21 October 2018 - 04:24 AM.

  • tkottary likes this

#11 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:48 AM

Nikon Spatial Filtering
 

The Nikon D5300 applies a spatial filtering algorithm to all exposures of 1/4sec and longer.  It's known as hot pixel suppression (HPS) and its purpose is to remove hot pixels and colour noise from an image.

 

This digital filtering has  both advantages and disadvantages.  The advantage is that it leads to less noisy images and reduces the need for dark frames. The disadvantage has only recently come to light: it has a rather peculiar effect on the colour of tightly focused stars, in some cases turning a large number of them predominately green. 

 

Here's an example of the effect when imaging a test pattern of tiny white dots:

 

CanonNikonSpatialFilteringV2.jpg

 

The various star colours in the first two panes are quite normal - it is a kind of colour Moire effect caused by an individual pixel value being a lot brighter than the neigbours.  This pixel then dominates the star colour.  It's not a problem because the star colour will average out over many dithered exposures.

 

You can see that as the Nikon D5300 exposure changes from 1/5sec to 1/4sec the spatial filtering kicks in and the stars turn predominately green.  Unfortunately it's likely they will still be green when averaged out over many dithered exposures.

 

However, this has been happening for years and no-one seems to have noticed it until recently.  I judge from this that for one reason or another it is not a big practical issue for most imagers.

 

At the present time the effects and consequences of this are not completely understood.  It's still a developing story which can be followed here:
 

https://www.cloudyni...d-star-colours/


Edited by sharkmelley, 21 October 2018 - 04:26 AM.


#12 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 03:54 AM

Conclusion
 

If you are choosing between these two cameras from an astro-imaging perspective, here are some criteria you might use:

  • Read noise - the Nikon has lower read noise, especially at low ISO.
  • Thermal noise - the Nikon has lower thermal noise.  The Canon noise is high.
  • Banding - the Canon can exhibit obvious banding issues.
  • Low ISO performance - the Nikon can be used effectively at low ISO. It's behaviour is far closer to being ISO-less than the Canon.
  • Quantum efficiency - the Nikon has slightly higher QE.
  • Colour Accuracy - for emission nebulae you might prefer the colours from one camera over the other.  I believe the Canon colours are the more accurate ones. This could be a very important issue for some people.
  • Spatial Filtering - the Nikon spatial filtering can be seen as an advantage (lower noise) or a disadvantage (star colour issues)

I guess a conclusion is incomplete without a final recommendation wink.gif  

 

Before the issue of spatial filtering came up, the very clear choice was the Nikon D5300.  It's amazing to see a 5 year old camera still ahead of a recent offering from Canon, especially one assumed to have Canon's more recent technology - on sensor ADC.

 

However the Nikon spatial filtering issue means this is no longer an unreserved recommendation.

 

That being said, both cameras are capable of producing very good astro-images and most users of either camera will be very pleased with the results they obtain.


Edited by sharkmelley, 21 October 2018 - 04:26 AM.

  • ccs_hello, Miguelo, Jon Rista and 4 others like this

#13 Merk

Merk

    Viking 1

  • -----
  • Posts: 902
  • Joined: 19 May 2014
  • Loc: Cyprus, Nicosia

Posted 21 October 2018 - 07:27 AM

Very nice comparison! waytogo.gif

Seems that canon has better colors and  is sharper on the ngc7000 pic? or is it because of seeing?

 

I would like to see a fully processed target from both in the futuresmile.gif (I ask a lot!!)



#14 JukkaP

JukkaP

    Viking 1

  • -----
  • Posts: 636
  • Joined: 14 Aug 2016

Posted 21 October 2018 - 07:44 AM

Clearly there is seeing factor as you see from the big stars. Or nikon has serious blooming problem ;).
  • Merk likes this

#15 bobzeq25

bobzeq25

    Hubble

  • *****
  • Posts: 17,419
  • Joined: 27 Oct 2014

Posted 21 October 2018 - 10:27 AM

Can't one tweak color to be anything one wants it to be?  PixInsight has a multitude of tools for adjusting color, hue, and saturation.  "Multitude" is not really an exaggeration.  Among many things, you can do curve and histogram adjustments on a color by color basis, so you can vary your color adjustments by brightness level.  There's much more.

 

One possibility why green stars with a 5300 is not a problem that has been prominent here, even among some very experienced imagers.  Green is an artifact often seen in astrophotography in general.  Some with DSLRs (50% of the pixels are green), but even with mono cameras and RGB filters.  StarTools has a tool to change green to yellow or brown.  There's some kind of green suppression as a third party tool for Photoshop (green pixel gun?).  PixInsight has selective chromatic noise reduction.

 

Bottom line.  People often find themselves reducing excess green, and changing it to something else.  More tweaking. 

 

I've often found discussions about "accurate" color somewhat strange.  Our eyes are both imperfect, and highly variable from one person to another.  "Color blindness" is not a simple crossing a sharp line, there are all sorts and levels of it.  The recent "is this dress purple or green" deal (something like that) plays on our variable perceptions.

 

Accurate color in astrophotography would be gray.  From Lessons From the Masters (not an exact quote):

 

"I used to worry about color a lot, sure that the color police would arrive and confiscate my equipment.  I worried about calibrating color to G2v stars.  These days I just please myself."


Edited by bobzeq25, 21 October 2018 - 10:47 AM.


#16 Merk

Merk

    Viking 1

  • -----
  • Posts: 902
  • Joined: 19 May 2014
  • Loc: Cyprus, Nicosia

Posted 21 October 2018 - 10:52 AM

Can't one tweak color to be pretty much anything one wants it to be?  PixInsight has a multitude of tools for adjusting color, hue, and saturation.  "Multitude" is not really an exaggeration.  Among many things, you can do curve and histogram adjustments on a color by color basis, so you can vary your color adjustments by brightness level.  There's much more.

I don't believe you can, except if your tweaking also has layer masks so you apply some colors in specific areas, but that is almost impossible work.

 

Try matching the colors from the nikon d5300 to the 200d in the ngc7000 picture. You will see that is not an easy task or it is not possible. I tried and failed.



#17 bobzeq25

bobzeq25

    Hubble

  • *****
  • Posts: 17,419
  • Joined: 27 Oct 2014

Posted 21 October 2018 - 11:01 AM

I don't believe you can, except if your tweaking also has layer masks so you apply some colors in specific areas, but that is almost impossible work.

 

 

Photoshop uses layer masks.  PixInsight has none (basically just a single layer of masking), can vary color adjustments by brightness levels instead.  Using pixel math is a possibility.  That multitude of available tools should make a difference in difficulty.


Edited by bobzeq25, 21 October 2018 - 11:02 AM.


#18 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 12:23 PM

Very nice comparison! waytogo.gif

Seems that canon has better colors and  is sharper on the ngc7000 pic? or is it because of seeing?

 

I would like to see a fully processed target from both in the futuresmile.gif (I ask a lot!!)

I'm still trying to work out why the Canon looks sharper, shows fainter stars and fainter nebulosity. Maybe it was focusing, maybe it was tracking, maybe it was high level cloud or maybe it was Nikon's spatial filtering issue.

 

Can't one tweak color to be anything one wants it to be?  PixInsight has a multitude of tools for adjusting color, hue, and saturation.  "Multitude" is not really an exaggeration.  Among many things, you can do curve and histogram adjustments on a color by color basis, so you can vary your color adjustments by brightness level.  There's much more.

 

One possibility why green stars with a 5300 is not a problem that has been prominent here, even among some very experienced imagers.  Green is an artifact often seen in astrophotography in general.  Some with DSLRs (50% of the pixels are green), but even with mono cameras and RGB filters.  StarTools has a tool to change green to yellow or brown.  There's some kind of green suppression as a third party tool for Photoshop (green pixel gun?).  PixInsight has selective chromatic noise reduction.

 

Bottom line.  People often find themselves reducing excess green, and changing it to something else.  More tweaking. 

 

If you're asking whether or not the Nikon D5200 colours be transformed into Canon 200D colours (or vice versa) the answer is no.  There's no way of doing it

 

The question about green artifacts is an interesting one.  Yes it's a very common step in astro-processing to kill green casts.  But why is it necessary in the first place?  Where does this green come from it come from if it is not dominant part of the spectrum arriving at the camera?

 

Mark


Edited by sharkmelley, 21 October 2018 - 12:24 PM.


#19 bobzeq25

bobzeq25

    Hubble

  • *****
  • Posts: 17,419
  • Joined: 27 Oct 2014

Posted 21 October 2018 - 01:03 PM


The question about green artifacts is an interesting one.  Yes it's a very common step in astro-processing to kill green casts.  But why is it necessary in the first place?  Where does this green come from it come from if it is not dominant part of the spectrum arriving at the camera?

 

Mark

That's an excellent theoretical question.  I'm guessing that how RGB is produced in the camera (be it a DSLR or other) has the biggest role.   And note that most all imagers adjust color based on what their specific eyes see (which, as I said, varies), not by some theoretical analysis of the spectrum received by the camera.

 

But the main operational thing for many is that it's usual, almost universal, to reduce green, which is why the tools exist.  They work well to address the issue of excess green, which is what most people really care about.  You wondered why people seem to not care about star color.  The answer is they do care, but find they have adequate tools to deal with it.

 

Even with a 5300.  <grin>  People use them, happily, and produce good looking images that do not overemphasize the green in the stars.  Below is M31, D5500, which presumably has the same issue?  Are the stars too green for you?

 

I can make them less so, if need be.  <smile>  Or reduce/transform green anywhere.  What I cannot do is make your eyes see what mine see.

 

And we have not even introduced the effects of monitors.  <smile>

 

Minor point.  I believe a major reason that some people (like me) tend not to do darks with 5300/5500/5600s is that it's hard (well nigh impossible) to match sensor temperature precisely, and the resulting imprecise darks can do more harm than good.

 

m31final_smaller.jpg


Edited by bobzeq25, 21 October 2018 - 01:23 PM.

  • Jim Waters likes this

#20 Merk

Merk

    Viking 1

  • -----
  • Posts: 902
  • Joined: 19 May 2014
  • Loc: Cyprus, Nicosia

Posted 21 October 2018 - 01:26 PM

I'm still trying to work out why the Canon looks sharper, shows fainter stars and fainter nebulosity. Maybe it was focusing, maybe it was tracking, maybe it was high level cloud or maybe it was Nikon's spatial filtering issue.

 

 

Well stars look the same in both pictures, maybe slightly smaller on nikon d5300, but nebulosity is better and sharper in Canon. Was it on a different night? If so seeing can be a reason.

 

The only thing that would really solve this problem is to  image a target with both cameras several hours exposure. That way the seeing will be averaged out.

But surely colors look better on Canon (for my taste)

 

I suspect both cameras are un modded 



#21 Merk

Merk

    Viking 1

  • -----
  • Posts: 902
  • Joined: 19 May 2014
  • Loc: Cyprus, Nicosia

Posted 21 October 2018 - 01:30 PM

Photoshop uses layer masks.  PixInsight has none (basically just a single layer of masking), can vary color adjustments by brightness levels instead.  Using pixel math is a possibility.  That multitude of available tools should make a difference in difficulty.

Tried also Lightroom which you can color adjustments in only a specific color. Better than PS for that job but still not close enough to replicating the colors on the 200d.



#22 bobzeq25

bobzeq25

    Hubble

  • *****
  • Posts: 17,419
  • Joined: 27 Oct 2014

Posted 21 October 2018 - 01:40 PM

 

The only thing that would really solve this problem is to  image a target with both cameras several hours exposure. That way the seeing will be averaged out.

That won't happen, is not a "real solution".  All of us know there are good nights and bad nights, with way different results.

 

Two setups operated simultaneously could work.


  • sharkmelley and Merk like this

#23 sharkmelley

sharkmelley

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,535
  • Joined: 19 Feb 2013

Posted 21 October 2018 - 01:50 PM

That's an excellent theoretical question.  I'm guessing that how RGB is produced in the camera (be it a DSLR or other) has the biggest role.   And note that most all imagers adjust color based on what their specific eyes see (which, as I said, varies), not by some theoretical analysis of the spectrum received by the camera.

 

But the main operational thing for many is that it's usual, almost universal, to reduce green, which is why the tools exist.  They work well to address the issue of excess green, which is what most people really care about.  You wondered why people seem to not care about star color.  The answer is they do care, but find they have adequate tools to deal with it.

 

Even with a 5300.  <grin>  People use them, happily, and produce good looking images that do not overemphasize the green in the stars.  Below is M31, D5500, which presumably has the same issue?  Are the stars too green for you?

 

I can make them less so, if need be.  <smile>  Or reduce/transform green anywhere.  What I cannot do is make your eyes see what mine see.

 

And we have not even introduced the effects of monitors.  <smile>

 

Minor point.  I believe a major reason that some people (like me) tend not to do darks with 5300/5500/5600s is that it's hard (well nigh impossible) to match sensor temperature precisely, and the resulting imprecise darks can do more harm than good.

 

attachicon.gif m31final_smaller.jpg

I don't have all the answers.  I can't say if the D5500 has the same issue or not because Nikon seems to tailor the algorithm to each individual camera. For instance the D810A doesn't have the issue as far as I can tell.

I think star colours are really interesting. They are a result of many different physical processes at work - star temperature, reddening by dust, absorption of wavelengths by various compounds.  For me "kill green" processing is not acceptable because it can never give me the "correct" colour. 

 

If astro-imagers find themselves performing "kill green" on most of their images then they really ought to be asking why. There are probably multiple reasons and the Nikon spatial filtering is probably not the main cause.

 

Mark


  • Merk and bobzeq25 like this

#24 Merk

Merk

    Viking 1

  • -----
  • Posts: 902
  • Joined: 19 May 2014
  • Loc: Cyprus, Nicosia

Posted 21 October 2018 - 11:05 PM

I had the greenish pictures before I modded my camera, after removing the filter that problem is gone but I can't seem to get the reds correct. I have a naked sensor mode so I have to always use Idas lps

#25 whwang

whwang

    Mercury-Atlas

  • *****
  • Posts: 2,922
  • Joined: 20 Mar 2013

Posted 21 October 2018 - 11:26 PM

I never “kill green” in my digital images. I don’t know why this is needed by some.

I know the PixInsight NR that can “kill green” existed sine the very early version of PI, when a lot of people (like me) still use film. For reason I don’t understand, film images can often show green grain/noise. So such an algorithm is well justified for film images. But it shouldn’t be the case for DSLR images.

There are no green stars because the stars that emits the strongest in green (like our sun) is defined by our eyes to be white.

Edited by whwang, 21 October 2018 - 11:27 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