First Light D810A, dead of the CCD?

[ollypenrice]

Oh no, now Maurice is on the case as well! Yves and Maurice, this is like taking on two sumo wrestlers when you're five foot two in your high heels (which I am... Well, nearly.)  OK guys, I'm ready. An Old Testament CCD imager, unreconstructed! Simeis 142. Gimme gimme!!! (And Maurice, put that F2.8 thing back in your closet. International fighting rules specify a minimum F ratio of 5.)

By the way, here is a CCD image of the NAM using a monochrome CCD, 8 seconds per colour and 17 seconds in Ha. (Would I lie to you?)

lly

[Davey-T]

what about the Canon 60da i don't think Canon are dragging there heals. i don't have one but i have the 1100d i know its not a top end dslr but it works.and the Canon 60Da is £700 ok it is a lower mgpl but wont that mean less noise (joke) just though id stir it up a bit.

Thought Canon had ceased production of the 60Da and haven't produced a 70Da.

Dave

[vdb]

Ah Sir, that is hard to beat indeed!

On the Sony A7 series ... please read this and then decide if that sensor is really golden or if it's high pass filtering, if you thought the old nikon star eating problem was real ...

http://www.dpreview.com/forums/post/55841466

/Yves

[mftoet]

As explained and demonstrated several times by Mr. 5 ft 2, luminance makes the difference in quality/depth of mono CCD images vs. DSLR images (or OSC CCD images) and, in this particular case, an Ha filter in front of a mono sensor boosts the signal and contrast of the emission nebulas. With a DSLR/OSC CCD an UHC filter helps boosting the contrast, but it cannot match a (3 nm) pure Ha signal. True RGB with a mono CCD (I.e. without luminance) vs. CFA (Bayer) RGB doesn't make much of a difference if you ask me.

[laser_jock99]

Why do people always tend to compare cameras budget-wise instead of sensor-size-wise in the DSLR – CCD debat? I totally agree with Yves; a 36 MP mono CCD + filters truly costs a fortune.

Although I recently acquired a CCD-camera to extend my current imaging gear, I’ll stick to the DSLR for mobile use. Mainly because of the ease of use (quick setup, no need for a laptop).

I held a presentation about the potentials of the current high-end DSLRs as astrocameras at the CEDIC convention in Linz last March. The powerpoint can be downloaded from my website. It includes several DSLR/CCD comparison shots of the same objects. http://www.dutchdeepsky.com/maurice_toet_cedic_2015.pdf

Interesting and largely in agreement with my own findings- in regard to fast optics and dark skies being good for DSLR work.

Can I just check the notation was correct on this page of the presentation?

[ollypenrice]

As explained and demonstrated several times by Mr. 5 ft 2, luminance makes the difference in quality/depth of mono CCD images vs. DSLR images (or OSC CCD images) and, in this particular case, an Ha filter in front of a mono sensor boosts the signal and contrast of the emission nebulas. With a DSLR/OSC CCD an UHC filter helps boosting the contrast, but it cannot match a (3 nm) pure Ha signal. True RGB with a mono CCD (I.e. without luminance) vs. CFA (Bayer) RGB doesn't make much of a difference if you ask me.

I have a new name (and have grown by two inches just by lying on the internet.  :grin:  ) Yes I think the key mono advantages are the ability to shoot L and/or the high contrasts sometimes found in narrowband.

We've talked about Ha here. What about O111? How would these chips cope with the Squid Nebula? Genuine question, not a wind-up.

Olly

[sharkmelley]

Ah Sir, that is hard to beat indeed!

On the Sony A7 series ... please read this and then decide if that sensor is really golden or if it's high pass filtering, if you thought the old nikon star eating problem was real ...

http://www.dpreview.com/forums/post/55841466

/Yves

Slightly off topic but I'll respond because you are the original poster! 

The Sony star eating problem (which appears in Bulb mode only) is very real and very destructive.  It's identical to the infamous star eating algorithm on the old Nikon cameras.  It's very obvious in single exposures but stacking (combined with dithered acquisition) does allow the truncated stars to be re-emerge from the noise as I demonstrated in this synthetic experiment here: http://www.dpreview.com/forums/post/55734616

So the stars in a stacked image are very attenuated but they are still there.  A workaround is to use non-Bulb mode which unfortunately limits you to a maximum 30sec exposure length.

It's not actually a problem with the sensor itself but with the camera's manipulation of the sensor data.    As I said earlier, if Nikon had taken the same sensor they would have produced a much better camera than Sony.

Mark

[laser_jock99]

How would these chips cope with the Squid Nebula? Genuine question, not a wind-up.

Olly

Now that sounds like a challenge to us DSLR imagers!

[ollypenrice]

Now that sounds like a challenge to us DSLR imagers!

Attaboy!

Olly

[toxic]

Ooooo a squid challenge 

if its a sick squid i cant afford it.

all sead and done it is a lovely image.

that must be what a dark site can do Olly 17 seconds     oh and a ccd

[Alien 13]

What us poor sub 400mm FL imagers realy want is much smaller pixels be it CCD or CMOS 1-2 μm would be nice start.

Alan

[lensman57]

I have a new name (and have grown by two inches just by lying on the internet.  :grin:  ) Yes I think the key mono advantages are the ability to shoot L and/or the high contrasts sometimes found in narrowband.

We've talked about Ha here. What about O111? How would these chips cope with the Squid Nebula? Genuine question, not a wind-up.

Olly

It will be a squid covered in black ink. But joking aside have a look at this, 45 hours exposure in Ha, Oiii and RGB @ F3.5, that is a years imaging time for me but the result is unbelievable and he has brought the squid out with a 6nm Oiii Astronomik clip filter.http://www.astronomersdoitinthedark.com/index.php?c=110&p=525, my only observation is that he has an extremely dark sky with SQM figures in the 21 and 22s.

Regards,

A.G

[Tom OD]

Oh no, now Maurice is on the case as well! Yves and Maurice, this is like taking on two sumo wrestlers when you're five foot two in your high heels (which I am... Well, nearly.)  OK guys, I'm ready. An Old Testament CCD imager, unreconstructed! Simeis 142. Gimme gimme!!! (And Maurice, put that F2.8 thing back in your closet. International fighting rules specify a minimum F ratio of 5.)

By the way, here is a CCD image of the NAM using a monochrome CCD, 8 seconds per colour and 17 seconds in Ha. (Would I lie to you?)

lly

This is getting interesting. In the Red Corner we have the DSLR, and in the even Redder Corner we have the CCD. Well I nailed my colours to the mast years ago.

I'm off to France soon, and in the hot Summer months, the cooler of the CCD was essential. I even have the data from a CCD camera without its cooling on verses it a few minutes later with the cooling on and the difference is huge. I have a big chip, the 11000 in the Atik camera, which I choose for cost, FOV and not efficiency, which is pretty low really.  I've been waiting for years for a full frame Sony chip, that has just not happened, and never will from what I've heard.

While my only DSLR imaging was the Canon 350D, and some 7D, I'll wait for more data before I 'm swayed, as they have not come anywhere close to the CCD. However I'm ready, and a little hoping to be wowed with a jump in the chip quality or in other areas which help us gather the photons more easily / efficiently.

Tom

[sharkmelley]

I held a presentation about the potentials of the current high-end DSLRs as astrocameras at the CEDIC convention in Linz last March. The powerpoint can be downloaded from my website. It includes several DSLR/CCD comparison shots of the same objects. http://www.dutchdeepsky.com/maurice_toet_cedic_2015.pdf

Very interesting presentation.  The DSLR vs Atik comparisons are quite informative.

Mark

[vdb]

It will be a squid covered in black ink. But joking aside have a look at this, 45 hours exposure in Ha, Oiii and RGB @ F3.5, that is a years imaging time for me but the result is unbelievable and he has brought the squid out with a 6nm Oiii Astronomik clip filter.http://www.astronomersdoitinthedark.com/index.php?c=110&p=525, my only observation is that he has an extremely dark sky with SQM figures in the 21 and 22s.

Regards,

A.G

I was going to post that same link, anyway if the camera was not cooled then a new generation could do it in less then halve the time ... that Samyang 135 f2 seems all of a sudden very tempting ...

/Yves

[vdb]

Tom I'm interested on how to camera will perform in "hot" summer nightsas well,

though at an altitude it's not that bad ... but none the less, I will be curious and it could very well be that a cooled ccd pulls away out of sight ...

/Yves

[vdb]

Some more comparison on "hot" summer night simulation;

http://www.astrobin.com/185845/

(unfortunately it is not mentioned if these where first images taken or after a series a dslr warm's up ... but it's clear the D810A is something amazing)

And reviews of the camera;

http://blogg.astrofotografen.se/2015/05/nikon-d810a-review.html

(the dark test do +4EV which is nice as you really see the noise pattern)

and;

http://blog.adamwoodworth.com/2015/06/nikon-d810a-review.html

/Yves

[wxsatuser]

It's very tempting to be a Nikon owner.

[ollypenrice]

Ooooo a squid challenge 

if its a sick squid i cant afford it.

all sead and done it is a lovely image.

that must be what a dark site can do Olly 17 seconds     oh and a ccd

No, the times ware a joke. (Would I lie to you?) The image had 3x15 per colour and 4x30 in Ha. No lumiminance. So about 5.5 hours. Not a lot by our usual standards but a lot more than Yves' and from an incomparably better site. I'd be hard pressed to believe there'd be much in it in a straight shootout. The Nikon might even win on this one.

Olly

[mftoet]

Can I just check the notation was correct on this page of the presentation?

Yes, the notation is correct. I understand why you’re confused. We know from our daily practice with a DSLR that increasing ISO increases the noise. However, what normally happens (i.e. what the camera automatically does) is that increasing ISO by one stop, cuts the exposure time in half. 50% less light (= less signal!) is hitting the sensor, so the camera needs to amplify the signal twice as much in order the get the same brightness of the scene.

While increasing ISO from 100 tot 1600, I kept the exposure time and aperture the same in my example. I used linear debayering and matched the histograms in PixInsight in order to get the same brightness of the scene. This will, of course, blow out the highlights (notice the white wall). Though not as much as you might expect. At least not with a modern midrange to high-end DSLR. Of the 14 bits of dynamic range those DSLRs in theory yields, approximately only 12 bits (12 photographic stops) are useable. Increasing ISO brings down the read noise (in e-1), so you gain more useable dynamic range at the low end of the histogram, where our nebulas hide , at the cost of loss of dynamic range at the high end of the histogram, where our stars shine brightly. I always shoot at ISO 1600 with my 5D Mark II (Baader modded). I guess it’s the sweet spot considering read noise versus dynamic range. Please refer the graph on slide 36 of my presentation to see what I mean. According to me, it doesn’t completely blow out the star cores (star colours) in my images. And if it did, I would just shoot several images at a lower ISO for preserving star colour and blend it in the stacked high ISO image.

I even have the data from a CCD camera without its cooling on verses it a few minutes later with the cooling on and the difference is huge.

Another important difference between DSLRs and CCDs. A CCD-camera needs to be cooled in order to get proper data. A DSLR doesn’t. It will only perform better when cooled ;-)

I've been waiting for years for a full frame Sony chip, that has just not happened, and never will from what I've heard.

I’m not quite sure about that. Rumours have it that Sony will launce a fullframe monochrome RX1 (mirrorless camera) this year.

[ollypenrice]

It seems as if this might not be a time to buy a camera so much as a time to wait...

On the summer heat, surely a simple cool box would solve the problem? It wouldn't need to cool massively, just enough to replicate a winter night. I have to say that I take Yves' camera very seriously. It could take OSC data while a CCD was collecting NB. The resolution at 36mp would be nice to have. Watch that focus, though!

Olly

[laser_jock99]

It will be a squid covered in black ink. But joking aside have a look at this, 45 hours exposure in Ha, Oiii and RGB @ F3.5, that is a years imaging time for me but the result is unbelievable and he has brought the squid out with a 6nm Oiii Astronomik clip filter.http://www.astronomersdoitinthedark.com/index.php?c=110&p=525, my only observation is that he has an extremely dark sky with SQM figures in the 21 and 22s.

Regards,

A.G

So it has already been done -(20.8 hrs) using an Astronomik 12nm OIII clip filter/SQM in the low 21's. A long time but still 'do able' with a DSLR then.

With a darker site and faster optics (and a 'narrower' narrowband filter) it could be imaged in less time methinks.........  

[ollypenrice]

So it has already been done -(20.8 hrs) using an Astronomik 12nm OIII clip filter/SQM in the low 21's. A long time but still 'do able' with a DSLR then.

With a darker site and faster optics (and a 'narrower' narrowband filter) it could be imaged in less time methinks.........  

Very good effort there by Scott Rosen. He did use an O111 filter though. I was wondering if the Squid would ever show simply in OSC eventually. Once you want to start using NB filters it would strike me as preferable to use all the pixels, though.

Olly

[lensman57]

So it has already been done -(20.8 hrs) using an Astronomik 12nm OIII clip filter/SQM in the low 21's. A long time but still 'do able' with a DSLR then.

With a darker site and faster optics (and a 'narrower' narrowband filter) it could be imaged in less time methinks.........  

I am not sure if his site can get any darker in practical terms. SQM 21~22 is pretty dark already. He is also a master of processing faint stuff. The truth is that given a site as dark as his and a 6nm Oiii on top of a cooled CCD the OU4 can probably be imaged at around 8 hours of exposure with 30 minute subs. Still pretty good for a DSLR I'd say.

A.G

[laser_jock99]

Very good effort there by Scott Rosen. He did use an O111 filter though. I was wondering if the Squid would ever show simply in OSC eventually. Once you want to start using NB filters it would strike me as preferable to use all the pixels, though.

Olly

I have a feeling the brighter Ha background would swamp The Squid with OSC. I think an Astronomik 6nm OIII filter would help a lot. With OSC there are at least twice as many green pixels as red & blue - so that might aid data gatthering too?

My 12nm Baader OIII probably wouldn't work too well- I've noticed it seems to pass way more light than than the 7nm Ha. I can't really justify a 3nm Astrodon filter for this experiment!!