Nikon D750 for Astrophotography

Started 7 months ago | Questions
atigun Contributing Member • Posts: 774
Re: Nikon D750 spatial filtering causes magenta stars

sharkmelley wrote:

rnclark wrote:

1llusive wrote:

elgol20 wrote:

1llusive wrote:

Did you make a decision?

The D750 has moved into my sights as well due to its excellent combination of low cost and high performance.

But I do worry about one thing: large pixels and the anti-aliasing filter. For sure, image detail and sharpness won't hold a candle to the D810 or later cameras. You can also buy two or three of them for the price of those.

There doesn't seem to be any talk of this camera eating stars or turning them odd colors, which is good.

you might see here:

a simple comparison between an unmodded D750(left) with the D810a, shot at the same time with same settings, both 135mm, one with Sigma Art the other Zeiss Apo, no post, RAW in LR, screenshots.

not sure if this helps. Pics may be not both in exact focus. It was a test that I did last October, decided to use my older D800E after that.

Thanks, but unfortunately the comparison has limited usefulness when using different lenses.

Both lenses were used at the same f-ratio. In my view, the 5.9 micron pixels (D750) versus 4.9 micron pixels (D800) are not that different.

The largest difference between the two images is the much brighter sky background in the D750 image.

But more important is the filtering seen in the images from both cameras is quite significant and includes mazing and other unnatural patterns. Is that from the raw converter or in the raw file?

Roger

I agree with you that there is a filtering issue but only on the Nikon D750. If you look at the D750 star colours in the images above, there is strong tendency for the stars to turn magenta. The reason for this is that there is something really weird about the Nikon D750 spatial filtering that I haven't seen before. It's something I'm still analysing in an attempt to work out what's going on. I see the same effect happening in D750 dark frames.

The headline summary is that the Nikon D750 spatial filtering algorithm preferentially attacks the green pixels in a star, thus altering the colour balance towards magenta. But why it does this, I'm not yet sure. It could potentially be a firmware bug.

The same magenta effect on the Nikon D750 can be seen in Alan Dyer's comparison - look at the "De-bayering Star Artifacts" section at the bottom of this article:

https://www.diyphotography.net/in-depth-testing-of-the-sony-a7iii-for-astrophotography/

If you are worried about the colour of small tightly focused stars then the Nikon D750 certainly looks like a camera to avoid.

Mark

It looks like purple fringing?  It can appear purple or green, depending on which side of the ideal focus point, as it is some kind of longitudinal fringing. Two different lenses where used that could interact differently with the sensors in this respect.

-- hide signature --

Atigun valley, a place north in Alaska

1llusive
1llusive Senior Member • Posts: 1,556
Re: Nikon D750 spatial filtering causes magenta stars

sharkmelley wrote:

sharkmelley wrote:

I agree with you that there is a filtering issue but only on the Nikon D750. If you look at the D750 star colours in the images above, there is strong tendency for the stars to turn magenta. The reason for this is that there is something really weird about the Nikon D750 spatial filtering that I haven't seen before. It's something I'm still analysing in an attempt to work out what's going on. I see the same effect happening in D750 dark frames.

The headline summary is that the Nikon D750 spatial filtering algorithm preferentially attacks the green pixels in a star, thus altering the colour balance towards magenta. But why it does this, I'm not yet sure. It could potentially be a firmware bug.

The same magenta effect on the Nikon D750 can be seen in Alan Dyer's comparison - look at the "De-bayering Star Artifacts" section at the bottom of this article:

https://www.diyphotography.net/in-depth-testing-of-the-sony-a7iii-for-astrophotography/

If you are worried about the colour of small tightly focused stars then the Nikon D750 certainly looks like a camera to avoid.

Mark

As evidence, here's one of my plots of the raw values in a Nikon D750 dark frame:

Every pixel in the raw data has been plotted against the maximum value found in the 24 neighbouring pixels in the 5x5 surrounding block.

You can immediately see that no green pixel is brighter than its neighbours. On the other hand there are a significant number of red and blue pixels that are significantly brighter than their neighbours. This is purely a result of the spatial filtering algorithm.

In the context of a small tightly focused star this means that on average the red and blue pixels will tend to dominate, leading to a magenta/pink star colour.

Mark

Very good, Mark. This must be laborious.

In the grand scheme of things, red/blue/magenta/pink stars are much better than green. At least they are close to actual star colors, unlike green, which never occurs in nature. Images from a D750 exhibiting this effect may blend in with others showing chromatic aberrations or those using certain filters. However since we have many choices in photographic equipment, we can use this information to make sound choices.

sharkmelley
sharkmelley Senior Member • Posts: 2,116
Re: Nikon D750 spatial filtering causes magenta stars
1

atigun wrote:

It looks like purple fringing? It can appear purple or green, depending on which side of the ideal focus point, as it is some kind of longitudinal fringing. Two different lenses where used that could interact differently with the sensors in this respect.

Sure. That's probably why these effects have been going on for so long without the real cause being properly understood.  The user blames their optics or their focusing technique.

Of course I can't say for certain that all the magenta/pink seen in that particular image is caused by the spatial filtering. All we can say for certain is that the poorly designed D750 spatial filtering will cause a magenta/pink hue to small tightly focused stars. If we can find further examples or if more testing can be performed then it will build up a better overall understanding.

It's frustrating that when Nikon first fixed the infamous star eater problem (e.g. on the D7000) they used an excellent algorithm. Since then they have released many cameras with inferior algorithms.

I would much rather have spatial filtering that can be switched OFF by the user.

Mark

-- hide signature --

Takahashi Epsilon 180ED
H-alpha modified Sony A7S
http://www.markshelley.co.uk/Astronomy/

 sharkmelley's gear list:sharkmelley's gear list
Sony a7S +1 more
elgol20
elgol20 Contributing Member • Posts: 637
Re: Nikon D750 spatial filtering causes magenta stars

sharkmelley wrote:

rnclark wrote:

1llusive wrote:

elgol20 wrote:

1llusive wrote:

Did you make a decision?

The D750 has moved into my sights as well due to its excellent combination of low cost and high performance.

But I do worry about one thing: large pixels and the anti-aliasing filter. For sure, image detail and sharpness won't hold a candle to the D810 or later cameras. You can also buy two or three of them for the price of those.

There doesn't seem to be any talk of this camera eating stars or turning them odd colors, which is good.

you might see here:

a simple comparison between an unmodded D750(left) with the D810a, shot at the same time with same settings, both 135mm, one with Sigma Art the other Zeiss Apo, no post, RAW in LR, screenshots.

not sure if this helps. Pics may be not both in exact focus. It was a test that I did last October, decided to use my older D800E after that.

Thanks, but unfortunately the comparison has limited usefulness when using different lenses.

Both lenses were used at the same f-ratio. In my view, the 5.9 micron pixels (D750) versus 4.9 micron pixels (D800) are not that different.

The largest difference between the two images is the much brighter sky background in the D750 image.

But more important is the filtering seen in the images from both cameras is quite significant and includes mazing and other unnatural patterns. Is that from the raw converter or in the raw file?

Roger

I agree with you that there is a filtering issue but only on the Nikon D750. If you look at the D750 star colours in the images above, there is strong tendency for the stars to turn magenta. The reason for this is that there is something really weird about the Nikon D750 spatial filtering that I haven't seen before. It's something I'm still analysing in an attempt to work out what's going on. I see the same effect happening in D750 dark frames.

The headline summary is that the Nikon D750 spatial filtering algorithm preferentially attacks the green pixels in a star, thus altering the colour balance towards magenta. But why it does this, I'm not yet sure. It could potentially be a firmware bug.

The same magenta effect on the Nikon D750 can be seen in Alan Dyer's comparison - look at the "De-bayering Star Artifacts" section at the bottom of this article:

https://www.diyphotography.net/in-depth-testing-of-the-sony-a7iii-for-astrophotography/

If you are worried about the colour of small tightly focused stars then the Nikon D750 certainly looks like a camera to avoid.

Mark

The pics we loaded in LR. What I learn here is that many things can be crucial in the workflow even though it might seem not. I like simple approaches like APP, LR only for minor post. I do not know rawtherapee, but here just these pics in raw loaded into it:

D750

they were done within like 5 min.

looks better to me.

-- hide signature --

Best regards
_____
Stefan

 elgol20's gear list:elgol20's gear list
Nikon D800E Nikon D750 Nikon D810A Carl Zeiss Apo Sonnar T* 2/135 Nikkor AF-S 300mm f/4E PF ED VR +4 more
sharkmelley
sharkmelley Senior Member • Posts: 2,116
Re: Nikon D750 spatial filtering causes magenta stars
1

elgol20 wrote:

sharkmelley wrote:

I agree with you that there is a filtering issue but only on the Nikon D750. If you look at the D750 star colours in the images above, there is strong tendency for the stars to turn magenta. The reason for this is that there is something really weird about the Nikon D750 spatial filtering that I haven't seen before. It's something I'm still analysing in an attempt to work out what's going on. I see the same effect happening in D750 dark frames.

The headline summary is that the Nikon D750 spatial filtering algorithm preferentially attacks the green pixels in a star, thus altering the colour balance towards magenta. But why it does this, I'm not yet sure. It could potentially be a firmware bug.

The same magenta effect on the Nikon D750 can be seen in Alan Dyer's comparison - look at the "De-bayering Star Artifacts" section at the bottom of this article:

https://www.diyphotography.net/in-depth-testing-of-the-sony-a7iii-for-astrophotography/

If you are worried about the colour of small tightly focused stars then the Nikon D750 certainly looks like a camera to avoid.

Mark

The pics we loaded in LR. What I learn here is that many things can be crucial in the workflow even though it might seem not. I like simple approaches like APP, LR only for minor post. I do not know rawtherapee, but here just these pics in raw loaded into it:

D750

they were done within like 5 min.

looks better to me.

Sure - you can process an image in many different ways, either to saturate or de-saturate star colour. Different profiles in different raw convertors will also have quite a large effect.

For this particular D750 image, examination of the raw file would quickly determine if the cause of the pink stars is spatial filtering or not i.e. has the pink been caused by capping the values of the green pixels.  I would be very interested to take a look.

Mark

-- hide signature --

Takahashi Epsilon 180ED
H-alpha modified Sony A7S
http://www.markshelley.co.uk/Astronomy/

 sharkmelley's gear list:sharkmelley's gear list
Sony a7S +1 more
sharkmelley
sharkmelley Senior Member • Posts: 2,116
Re: Nikon D750 spatial filtering causes magenta stars
3

sharkmelley wrote:

For this particular D750 image, examination of the raw file would quickly determine if the cause of the pink stars is spatial filtering or not i.e. has the pink been caused by capping the values of the green pixels. I would be very interested to take a look.

Stefan sent me the raw NEF which I opened using default settings in Photoshop/ACR. Here is a crop of that image at 200% scale:

Nikon D750 crop at 200% scale

A large number of pink and purple stars can be seen.

So I examined the raw data for 10 of the more obvious pink stars in the above crop. In every case the star contained a pair of green pixels with identical values. The reason for this is that a bright green pixel at the core of the star had its value capped to the value of a dimmer neighbouring green pixel, by the spatial filtering algorithm.

There can be no doubt that the spatial filtering is causing the pink stars. Unfortunately stacking together many dithered images (i.e. images with random positional offsets) will still not fix this pink bias in star colour.

Thank you Stefan for your help.

Mark

-- hide signature --

Takahashi Epsilon 180ED
H-alpha modified Sony A7S
http://www.markshelley.co.uk/Astronomy/

 sharkmelley's gear list:sharkmelley's gear list
Sony a7S +1 more
elgol20
elgol20 Contributing Member • Posts: 637
Re: Nikon D750 spatial filtering causes magenta stars

sharkmelley wrote:

sharkmelley wrote:

For this particular D750 image, examination of the raw file would quickly determine if the cause of the pink stars is spatial filtering or not i.e. has the pink been caused by capping the values of the green pixels. I would be very interested to take a look.

Stefan sent me the raw NEF which I opened using default settings in Photoshop/ACR. Here is a crop of that image at 200% scale:

Nikon D750 crop at 200% scale

A large number of pink and purple stars can be seen.

So I examined the raw data for 10 of the more obvious pink stars in the above crop. In every case the star contained a pair of green pixels with identical values. The reason for this is that a bright green pixel at the core of the star had its value capped to the value of a dimmer neighbouring green pixel, by the spatial filtering algorithm.

There can be no doubt that the spatial filtering is causing the pink stars. Unfortunately stacking together many dithered images (i.e. images with random positional offsets) will still not fix this pink bias in star colour.

Thank you Stefan for your help.

Mark

sounds strange to me, D750 no good for AP?

-- hide signature --

Best regards
_____
Stefan

 elgol20's gear list:elgol20's gear list
Nikon D800E Nikon D750 Nikon D810A Carl Zeiss Apo Sonnar T* 2/135 Nikkor AF-S 300mm f/4E PF ED VR +4 more
sharkmelley
sharkmelley Senior Member • Posts: 2,116
Re: Nikon D750 spatial filtering causes magenta stars

elgol20 wrote:

sharkmelley wrote:

sharkmelley wrote:

For this particular D750 image, examination of the raw file would quickly determine if the cause of the pink stars is spatial filtering or not i.e. has the pink been caused by capping the values of the green pixels. I would be very interested to take a look.

Stefan sent me the raw NEF which I opened using default settings in Photoshop/ACR. Here is a crop of that image at 200% scale:

Nikon D750 crop at 200% scale

A large number of pink and purple stars can be seen.

So I examined the raw data for 10 of the more obvious pink stars in the above crop. In every case the star contained a pair of green pixels with identical values. The reason for this is that a bright green pixel at the core of the star had its value capped to the value of a dimmer neighbouring green pixel, by the spatial filtering algorithm.

There can be no doubt that the spatial filtering is causing the pink stars. Unfortunately stacking together many dithered images (i.e. images with random positional offsets) will still not fix this pink bias in star colour.

Thank you Stefan for your help.

Mark

sounds strange to me, D750 no good for AP?

Decide for yourself.

In response to Illusive's comment "There doesn't seem to be any talk of this camera eating stars or turning them odd colors, which is good" you immediately posted a D750 image with odd star colours - the pink stars. I assumed you were answering his question.

There are plenty of good reviews of the D750 for astrophotography and I'm sure many folk are very happily using it.

I'm not saying it's no good for AP but given a choice between a model that produces pink stars and a similar model that doesn't then for me the choice is clear.

Mark

-- hide signature --

Takahashi Epsilon 180ED
H-alpha modified Sony A7S
http://www.markshelley.co.uk/Astronomy/

 sharkmelley's gear list:sharkmelley's gear list
Sony a7S +1 more
elgol20
elgol20 Contributing Member • Posts: 637
Re: Nikon D750 spatial filtering causes magenta stars

sharkmelley wrote:

elgol20 wrote:

sharkmelley wrote:

sharkmelley wrote:

For this particular D750 image, examination of the raw file would quickly determine if the cause of the pink stars is spatial filtering or not i.e. has the pink been caused by capping the values of the green pixels. I would be very interested to take a look.

Stefan sent me the raw NEF which I opened using default settings in Photoshop/ACR. Here is a crop of that image at 200% scale:

Nikon D750 crop at 200% scale

A large number of pink and purple stars can be seen.

So I examined the raw data for 10 of the more obvious pink stars in the above crop. In every case the star contained a pair of green pixels with identical values. The reason for this is that a bright green pixel at the core of the star had its value capped to the value of a dimmer neighbouring green pixel, by the spatial filtering algorithm.

There can be no doubt that the spatial filtering is causing the pink stars. Unfortunately stacking together many dithered images (i.e. images with random positional offsets) will still not fix this pink bias in star colour.

Thank you Stefan for your help.

Mark

sounds strange to me, D750 no good for AP?

Decide for yourself.

In response to Illusive's comment "There doesn't seem to be any talk of this camera eating stars or turning them odd colors, which is good" you immediately posted a D750 image with odd star colours - the pink stars. I assumed you were answering his question.

There are plenty of good reviews of the D750 for astrophotography and I'm sure many folk are very happily using it.

I'm not saying it's no good for AP but given a choice between a model that produces pink stars and a similar model that doesn't then for me the choice is clear.

Mark

yes, that's what I meant...

-- hide signature --

Best regards
_____
Stefan

 elgol20's gear list:elgol20's gear list
Nikon D800E Nikon D750 Nikon D810A Carl Zeiss Apo Sonnar T* 2/135 Nikkor AF-S 300mm f/4E PF ED VR +4 more
rnclark Senior Member • Posts: 3,713
Re: Nikon D750 spatial filtering causes magenta stars
2

sharkmelley wrote:

sharkmelley wrote:

I agree with you that there is a filtering issue but only on the Nikon D750. If you look at the D750 star colours in the images above, there is strong tendency for the stars to turn magenta. The reason for this is that there is something really weird about the Nikon D750 spatial filtering that I haven't seen before. It's something I'm still analysing in an attempt to work out what's going on. I see the same effect happening in D750 dark frames.

The headline summary is that the Nikon D750 spatial filtering algorithm preferentially attacks the green pixels in a star, thus altering the colour balance towards magenta. But why it does this, I'm not yet sure. It could potentially be a firmware bug.

The same magenta effect on the Nikon D750 can be seen in Alan Dyer's comparison - look at the "De-bayering Star Artifacts" section at the bottom of this article:

https://www.diyphotography.net/in-depth-testing-of-the-sony-a7iii-for-astrophotography/

If you are worried about the colour of small tightly focused stars then the Nikon D750 certainly looks like a camera to avoid.

Mark

As evidence, here's one of my plots of the raw values in a Nikon D750 dark frame:

Every pixel in the raw data has been plotted against the maximum value found in the 24 neighbouring pixels in the 5x5 surrounding block.

You can immediately see that no green pixel is brighter than its neighbours. On the other hand there are a significant number of red and blue pixels that are significantly brighter than their neighbours. This is purely a result of the spatial filtering algorithm.

In the context of a small tightly focused star this means that on average the red and blue pixels will tend to dominate, leading to a magenta/pink star colour.

Mark

Hi Mark,

Amazing analysis and insight.  Thank you for the dedication.  You've done quite a few cameras and found filtering in many.  This raises multiple questions.

Sensor characteristics (gain, read noise, dark current) are derived from noise analyses.  If the raw data are filtered, it will impact all of these and make derived sensor specs look better.  Eric Fossum (inventor of CMOS) continues to drop little hints of skepticism about some of the very low read noise derivations, especially those around 1 electron.  You said your analysis showing filtering is not seen in FFT data.  What other filtering could have escaped detecting so far and what effect on sensor analyses could that have?  The reason I ask is by looking at short exposures of light frames, the raw data on some cameras has texture that looks odd--either spatial filtering or odd sensor response, e.g. sometimes looking wormy.  The bottom line might be that with filtering, the sensor noise analyses may result in 1) sensor gain is off, 2) read noise is lower, 3) dark current is lower.  I am also skeptical because some of the derived gains do not follow expected trends with newer generation sensors.  For example, the derived specs on the Nikon D850 are lower than earlier cameras.  Is that true, or due to less/no spatial filtering so derived sensor characteristics are more realistic?  This is particular of interest in light of the fact that you found that the D850 has less aggressive filtering in the long exposure dark frames than earlier cameras.

What impact does the filtering you have detected have on derived dark current rates, assuming derived gains are accurate?

In the D750 plot above, and in some other cameras, a lot of pixels in your plot show lines parallel to the horixontal and vertical axes and in your cloudynights thread you said that was fine.  But construct an image of random noise and subject it to your analysis.  I did that a while back and the result is a symmetric cloud, not with any data parallel to the axes.  So perhaps the parallel to the axes is yet another filtering algorithm at play.

If there are other algorithms affecting derived sensors gains, read noise and dark current, it sure biases sensor reviews.

My theory on the filtering is that 1) camera manufacturers are trying to deliver the best overall image from their perspective and testing, and 2) they vary the algorithms to see their effects in real-world use (Nikon seems to be responding to users; Sony not).  Note there is no law/rule that says raw data must be raw.

Roger

sharkmelley
sharkmelley Senior Member • Posts: 2,116
Re: Nikon D750 spatial filtering causes magenta stars
2

rnclark wrote:

Hi Mark,

Amazing analysis and insight. Thank you for the dedication. You've done quite a few cameras and found filtering in many. This raises multiple questions.

Sensor characteristics (gain, read noise, dark current) are derived from noise analyses. If the raw data are filtered, it will impact all of these and make derived sensor specs look better. Eric Fossum (inventor of CMOS) continues to drop little hints of skepticism about some of the very low read noise derivations, especially those around 1 electron. You said your analysis showing filtering is not seen in FFT data. What other filtering could have escaped detecting so far and what effect on sensor analyses could that have? The reason I ask is by looking at short exposures of light frames, the raw data on some cameras has texture that looks odd--either spatial filtering or odd sensor response, e.g. sometimes looking wormy. The bottom line might be that with filtering, the sensor noise analyses may result in 1) sensor gain is off, 2) read noise is lower, 3) dark current is lower. I am also skeptical because some of the derived gains do not follow expected trends with newer generation sensors. For example, the derived specs on the Nikon D850 are lower than earlier cameras. Is that true, or due to less/no spatial filtering so derived sensor characteristics are more realistic? This is particular of interest in light of the fact that you found that the D850 has less aggressive filtering in the long exposure dark frames than earlier cameras.

What impact does the filtering you have detected have on derived dark current rates, assuming derived gains are accurate?

In the D750 plot above, and in some other cameras, a lot of pixels in your plot show lines parallel to the horixontal and vertical axes and in your cloudynights thread you said that was fine. But construct an image of random noise and subject it to your analysis. I did that a while back and the result is a symmetric cloud, not with any data parallel to the axes. So perhaps the parallel to the axes is yet another filtering algorithm at play.

If there are other algorithms affecting derived sensors gains, read noise and dark current, it sure biases sensor reviews.

My theory on the filtering is that 1) camera manufacturers are trying to deliver the best overall image from their perspective and testing, and 2) they vary the algorithms to see their effects in real-world use (Nikon seems to be responding to users; Sony not). Note there is no law/rule that says raw data must be raw.

Roger

Hi Roger,

Thanks for your comments.

I think the question you are asking is this: is there some (hidden) data manipulation being applied to all exposures that will bias the statistics either deliberately or accidentally.

I suspect we'll never know the answer for sure but I really doubt this is happening. Here's my argument.

There are a couple of things we do know for sure:

  • Nikon cameras apply spatial filtering for exposures longer than a certain threshold (typically 0.25sec or so)
  • Sony mirrorless cameras apply spatial filtering for longer exposures (either bulb mode or 4 sec and greater)

These effects are very obvious from the change in the shape of the histogram when the exposure threshold is reached i.e. the long tail of the histogram becomes obviously curtailed.

The Sony spatial filtering (both the original version and the more recent replacement) is very crude and easily shows up in a 2D fast Fourier transform (FFT) as an attenuation of high frequencies.
The (various) Nikon spatial filtering algorithms are more sophisticated and rarely show up in an FFT.

The analysis I have been doing recently plots a pixel value against the maximum of its immediate neighbours. It is far more sensitive than an FFT but it only works because we have prior information (or guesswork) about the type of spatial filtering algorithm being used.

When we calculate gain, read noise and QE we are using very short exposures, unaffected by this spatial filtering that we know exists. The question is whether these very short exposures are affected by some other data manipulation.

If this data manipulation does exist then it doesn't show up in an FFT. We already know of the existence of filtering that doesn't show up in an FFT so it's certainly possible that it can be done. But would it have a material effect on gain, read noise and QE calculations?

Let's take the original Sony star eater as an example. Every pixel brighter than its 8 neighbours of the same colour has its value capped and every pixel dimmer than its 8 neighbours has its value floored. As a side effect this reduces the apparent noise (i.e. the standard deviation) by approximately 20%. The Nikon algorithms only cap pixel values (they don't floor them) so the worst algorithm results in a 10% reduction in noise. The better Nikon algorithms have a much lower effect.

So if Nikon is using some sophisticated data manipulation on short exposures that doesn't show up in an FFT but has a material reduction in noise then why is there a need for the additional unsophisticated algorithm used for exposures 0.25sec and longer? That wouldn't make sense to me.

Measuring dark current is a difficult problem because consumer cameras subtract the average level of the accumulated dark current. We can only measure the thermal noise and we know this is reduced by the spatial filtering so it requires some kind of adjustment to obtain a reasonable dark current estimate. This is fraught with problems.

You also mentioned that the derived specs on the Nikon D850 are lower than earlier cameras. Are you referring to the DXO derived estimate QE estimate of 44% at PhotonsToPhotos:
http://www.photonstophotos.net/Charts/Sensor_Characteristics.htm#Nikon%20D850_14

I don't know exactly how that is derived but Bill Claff has come up with a significantly higher estimate of 52% here:
https://www.dpreview.com/forums/post/61988234

This second estimate makes a lot more sense to me.

In summary, we don't know for sure that the raw files are not being "cooked" but I don't think a (so far) undetected algorithm can really be having a material effect on gain, read noise and QE estimates. But dark current estimates are certainly problematic because they rely on spatially filtered data.

Mark

-- hide signature --

Takahashi Epsilon 180ED
H-alpha modified Sony A7S
http://www.markshelley.co.uk/Astronomy/

 sharkmelley's gear list:sharkmelley's gear list
Sony a7S +1 more
DMcGarvey
OP DMcGarvey Forum Member • Posts: 71
Re: Nikon D750 for Astrophotography

Hello all.  Sorry, but I have not checked this thread in a while.  I did end up getting the D750 and I'm pleased with it so far.  It performs much better than my D3200.  There is zero amp glow and a 10-minute dark frame taken at room temperature and stretched with curves revealed virtually no pattern noise...leaps and bounds beyond my D3200.  I've read through this thread and admittedly the discussion has gone a bit over my head with all of the detailed sensor analysis, etc.  I can't say that I've noticed an unusual level of magenta fringing around stars.  Perhaps I am used to some level of it  since I had this with data captured with my D3200 as well.  I find that the defringing correction in Lightroom (Adobe Camera Raw) corrects this quite well.  With small adjustments to the magenta/orange/green sliders, I can get rid of most of the fringing.  I've also noticed differences in the fringing depending on the lens used - which is why I never connected it with the camera.  If I can provide raw files or any other data that would be helpful, please let me know.

David

sharkmelley
sharkmelley Senior Member • Posts: 2,116
Re: Nikon D750 for Astrophotography

DMcGarvey wrote:

Hello all. Sorry, but I have not checked this thread in a while. I did end up getting the D750 and I'm pleased with it so far. It performs much better than my D3200. There is zero amp glow and a 10-minute dark frame taken at room temperature and stretched with curves revealed virtually no pattern noise...leaps and bounds beyond my D3200. I've read through this thread and admittedly the discussion has gone a bit over my head with all of the detailed sensor analysis, etc. I can't say that I've noticed an unusual level of magenta fringing around stars. Perhaps I am used to some level of it since I had this with data captured with my D3200 as well. I find that the defringing correction in Lightroom (Adobe Camera Raw) corrects this quite well. With small adjustments to the magenta/orange/green sliders, I can get rid of most of the fringing. I've also noticed differences in the fringing depending on the lens used - which is why I never connected it with the camera. If I can provide raw files or any other data that would be helpful, please let me know.

David

You have just discovered the most important test of all: is it a camera that you as an individual are happy with!

If the answer is yes then you can safely ignore all the technical issues that are sometimes discussed.  It's always good to be aware though, just in case you experience the issue at a later date.

Mark

-- hide signature --

Takahashi Epsilon 180ED
H-alpha modified Sony A7S
http://www.markshelley.co.uk/Astronomy/

 sharkmelley's gear list:sharkmelley's gear list
Sony a7S +1 more
Trollmannx Senior Member • Posts: 5,345
Re: Nikon D750 for Astrophotography

DMcGarvey wrote:

Hello all. Sorry, but I have not checked this thread in a while. I did end up getting the D750 and I'm pleased with it so far. It performs much better than my D3200. There is zero amp glow and a 10-minute dark frame taken at room temperature and stretched with curves revealed virtually no pattern noise...leaps and bounds beyond my D3200. I've read through this thread and admittedly the discussion has gone a bit over my head with all of the detailed sensor analysis, etc. I can't say that I've noticed an unusual level of magenta fringing around stars. Perhaps I am used to some level of it since I had this with data captured with my D3200 as well. I find that the defringing correction in Lightroom (Adobe Camera Raw) corrects this quite well. With small adjustments to the magenta/orange/green sliders, I can get rid of most of the fringing. I've also noticed differences in the fringing depending on the lens used - which is why I never connected it with the camera. If I can provide raw files or any other data that would be helpful, please let me know.

David

The purple fringing is a lens issue - even my CZ lenses show this purple fringe (that is the Achilles heel of lots of lenses - also the last generation Zeiss lenses).

We can defocus the lenses ever so slightly to hide some of the purple fringing but then the tiny slightly out of focus stars easily wind up with a cyan color cast.

So no simple solution here.

My old Minolta 2.8/100 mm macro lens was corrected for photographic film, so this lens show strong blue halos around stars.

Well - as Mark so wisely say, enjoy your new camera!

And in astrophotography - whatever we do - there are always issues. The astonishing fact is that consumer grade cameras and lenses work pretty well for this demanding field. That is actually pretty amazing.

DMcGarvey
OP DMcGarvey Forum Member • Posts: 71
Re: Nikon D750 for Astrophotography

Thank you Mark.  I will keep all of this in mind.  And if you are interested in seeing any more raw data or would like me to perform a certain test with my D750 that may shed some light on something, I would be willing to do so.  Here is a recent Horsehead Nebula image consisting of data captured with this camera if anyone is interested in checking it out!

David

DMcGarvey
OP DMcGarvey Forum Member • Posts: 71
Re: Nikon D750 for Astrophotography

I agree - there are always issues, but the challenges are part of the fun!  And they offer a person like me the opportunity to learn a lot.  I agree with regard to lenses.  The first set of lenses I used were "kit" lenses that were basically lower-quality zoom lenses.  Shooting at low f-ratios really brought out horrible aberrations.  I have an 18-300mm that I used to shoot the Andromeda Galaxy and ended up tossing the data due to how badly the stars were distorted in the corners.  Within the past year I pretty much decided to only use prime lenses of better quality.  I picked up a used Nikkor 300mm f/2.8 and a used Rokinon 135mm f/2 and was floored at the difference.  So I agree that the lens can really make a difference.  And yes - I do like the camera!  For me it's a huge step up.

Rob Farmiloe Junior Member • Posts: 33
Re: Nikon D750 for Astrophotography

I have a D750, not claiming its lab colour perfect but I have not noticed any  issues with purple stars. I pick up a good range of star colours.

On the other hand, dark /bias frames can be stretched to crazy levels and remain practically black.

I cannot fault it as a camera.

1llusive
1llusive Senior Member • Posts: 1,556
Re: Nikon D750 for Astrophotography

Rob Farmiloe wrote:

I have a D750, not claiming its lab colour perfect but I have not noticed any issues with purple stars. I pick up a good range of star colours.

On the other hand, dark /bias frames can be stretched to crazy levels and remain practically black.

I cannot fault it as a camera.

What kind of tracking and optics do you use with it?

Rob Farmiloe Junior Member • Posts: 33
Re: Nikon D750 for Astrophotography

I have an Astrotrac but have also recently bought an LX2 mini track (clockwork tracker).

Lens i have tried for Astro

Nikon 400mm f3.5 ai - good at f5.6

Nikon 180mm ED 2.8ai good at f2.8, marginal improvement at f4

Nikon 300mm f4 E PF good at f4

Nikon 50mm F2 not great

Samyang 85mm f1.4 good at f2.8 marginal improvement at f4

Irix 15mm f2.4 good at f2.4 except vignetting in corners is strong. Much better at f2.8

I also have scopes but use a CCD with those.

The D750 is my daytime camera so have no desire to modify it. There are photos on astrobins though that demonstrate they give reasonable ha data as is.

In terms of colour i read an interesting article a nightscape photographer posted.

He tested the D750 at various ISO settings and compared noise and colour. Each setting did produce a slighty different result.

For images around 60s or below i have stopped taking darks. Appears to offer no benefit.

I have read the D750 is ISO tolerant. For general photography this is probably fair. For say wildlife, i still aim to stay below 2000 or the image does suffer.

For Astro I struggle to use above Iso800 simply because of clipping the histogram or dropping exposure times. ISO400 is where I have settled for most.

I saw a surprisingly excellent none tracked milkyway image taken at ISO8000. I intend on trying this come summer at a dark site.

I am no expert and have not tested the D750 against other newer Nikons. However i have no desire to replace it which says something.

Rob

Rob Farmiloe Junior Member • Posts: 33
Re: Nikon D750 for Astrophotography

Just to clarify the above.

For 300mm and above I prefer to use a mono CCD and filters.

I have used the Astrtrac up to 420mm FL. I use heads on top of the Astrtrac to give setting circles otherwise locating objects takes up additional time.

I use the D750 for mobile Astro or for when there us a short window of time. Otherwise I have a full go to on a pier.

The Samyang 85mm I purchased to go with the Irix 15mm for use on the LX2 mini tracker.

Both are dual purpose lens and perform very well for daytime use. The Samyang was really excellent value.

If you want a camera for purely astro work there would be other options to consider.

Rob

Keyboard shortcuts:
FForum MMy threads