M51 Whirlpool - First light with Sony A7S

[sharkmelley]

This is the first light for my new Sony A7S full-frame mirrorless camera on a Celestron C11 with 0.8 reducer/flattener i.e. imaging at F8.  A total of 2 hours of data in 90sec subs at ISO 2000 was taken in SQM 21.0 quality sky at 6C ambient temperature.  The subs were short to overcome a flexure or mirror shift problem I was experiencing but this has not affected the image quality because of the remarkably low read noise at ISO 2000.

This is a crop of the image because I had very squiffy stars, probably due to incorrect spacing of the reducer. 

The full size (i.e. 1:1 scaling) of the crop can be found here:  http://www.markshelley.co.uk/Astronomy/2015/m51_a7s_c11_20150420.jpg

I have deliberately left the image as "raw" as possible. Dithering was done during acquisition using PHD2 and FeloPaul's PHDMax software to control the Sony Remote Camera Control Software. Darks, flats and bias have been applied and then a straight sum of the 80 frames was done after aligning the frames.  I then did a stretch of the luminance channel (an arcsinh transform, to be specific). The RGB balance is the camera's own daylight balance and no colour saturation or sharpening has been applied.  This makes it easier to judge the overall image quality and noise characteristics.

Analysis showed the following statistics for each sub:

Read noise:  1.4e/pixel

Thermal noise:  2e/pixel

Skyglow noise: 4.5e/pixel

These can be combined quadratically to give a total noise of 5.1e/pixel

It is notable that even at such a slow F-ratio (F8) the skyglow (i.e. light pollution) is still the main source of image noise, though this is partly because of the low ambient temperature of 6C.  However even on a warm Summer evening in the UK i.e. 20C the thermal noise would only rise to 4e and be similar to the skyglow.

Note that this image was taken with the camera in its unmodded state.  I'm hoping to DIY mod it next week

I also intend to provide a more in-depth review of the camera in the near future.

Mark

[Zakalwe]

Thats blooming amazing for 2 hours at f8!

[laser_jock99]

Interesting first light- I'm looking at the full frame Sony cameras with an eye on having a light weight full frame option. It will interesting to see if the modded version is an improvement.

You seem to have a darkish sky (SQM 21) and yet LP is still an issue even at F8?

[lensman57]

Very impressive for short data. The very dark sky really does make all the difference. Very well done.

A.G

[Chris]

Crikey!!

[sharkmelley]

You seem to have a darkish sky (SQM 21) and yet LP is still an issue even at F8?

Yes, SQM was 21.0 and naked eye limiting mag somewhere between 5.5 and 6.0 I reckon.  The electron counts per pixel for the 90sec exposure were 30,55,25 respectively for R,G,B.

There are 2 features of this camera that make it attractive:

1) The low read noise means that very short exposures can be made without read noise becoming a problem

2) The low dark current means that it can be used on slower F-ratio scopes without thermal noise being an issue

I have to use my current Canon 350D on a fast Tak Epsilon (F2.8) in order for my (fairly low) skyglow to swamp the thermal noise - this camera can be used on much slower scopes (say F6) and still the skyglow will drown the thermal noise i.e. no cooling required.

At the Kelling Heath Star Party, last weekend, I put a 50mm F1.8 lens on it and a few us took turns looking through the electronic viewfinder (i.e. in liveview) at ISO 102400 (yes, one hundred thousand).  Hundreds of stars became visible where the naked eye could only see tens of stars. 

It certainly opens up a wide range of possibilities.

Mark

[bwana]

This is the first light for my new Sony A7S full-frame mirrorless camera on a Celestron C11 with 0.8 reducer/flattener i.e. imaging at F8.  A total of 2 hours of data in 90sec subs at ISO 2000 was taken in SQM 21.0 quality sky at 6C ambient temperature.  The subs were short to overcome a flexure or mirror shift problem I was experiencing but this has not affected the image quality because of the remarkably low read noise at ISO 2000.

This is a crop of the image because I had very squiffy stars, probably due to incorrect spacing of the reducer. 

m51_a7s_c11_20150420_small.jpg

The full size (i.e. 1:1 scaling) of the crop can be found here:  http://www.markshelley.co.uk/Astronomy/2015/m51_a7s_c11_20150420.jpg

I have deliberately left the image as "raw" as possible. Dithering was done during acquisition using PHD2 and FeloPaul's PHDMax software to control the Sony Remote Camera Control Software. Darks, flats and bias have been applied and then a straight sum of the 80 frames was done after aligning the frames.  I then did a stretch of the luminance channel (an arcsinh transform, to be specific). The RGB balance is the camera's own daylight balance and no colour saturation or sharpening has been applied.  This makes it easier to judge the overall image quality and noise characteristics.

Analysis showed the following statistics for each sub:

Read noise:  1.4e/pixel

Thermal noise:  2e/pixel

Skyglow noise: 4.5e/pixel

These can be combined quadratically to give a total noise of 5.1e/pixel

It is notable that even at such a slow F-ratio (F8) the skyglow (i.e. light pollution) is still the main source of image noise, though this is partly because of the low ambient temperature of 6C.  However even on a warm Summer evening in the UK i.e. 20C the thermal noise would only rise to 4e and be similar to the skyglow.

Note that this image was taken with the camera in its unmodded state.  I'm hoping to DIY mod it next week

I also intend to provide a more in-depth review of the camera in the near future.

Mark

Great Whirlpool galaxy!  The A7S is truly a wonderful imaging camera!?  It is pretty good unmodded but excellent full spectrum modded!!

I've been using a A7R and A7S (both now full spectrum modified) since they were released with excellent results.  However, I image with a considerably different technique and parameters:

- I don't shoot or use darks, flats or bias subs with either camera.  Sony sensors don't really require darks or bias subs.  I do gradient removal in PixInsight or use GradientXTerminator in Photoshop as a replacement for flats, and with a clean sensor, dust spots can easily be removed manually in Lightroom.

- I normally image at ISO 3200 with the A7R and ISO 3200-12800 with the A7S.  I've gone as high as ISO 25600 with the A7S with reasonable results, although I have to contend with amp glow above ISO 6400 (darks would probably eliminate this limitation?).

- I shoot most astro-images at 30 sec. exposures and normally shoot 50-60 subs.  I've gone as long as 120 sec. on occasion but it takes a real dark sky to allow this long an exposure.

- Since acquiring the A7R in Nov. 2013 I haven't fired up guiding for any astrophotography, and any problems with system flexure have disappeared compared to shooting 5-20 min. subs.

- When I purchased the A7R I had three OSC (one shot color) CCD's.  By the time I bought the A7S all OSC's were sold.  The A7S is simply better than any OSC CCD I've ever owned!

- I used to require a computer for imaging and guiding.  I now simply use a wired remote timer, a Bahtinov mask for focusing, do a reasonable polar alignment and image...  I do use StarSense on the Celestron CGEM DX and AVX mounts for auto star alignment and fast polar alignment; my only failing for high tech .

- In other words, the Sony A7R and A7S cameras have totally changed the way I undertake astrophotography!

- I have retained a mono CCD (ATIK 428Ex); however, I haven't used it since I started shooting with the A7S which I now use for both OSC and narrowband astro-imaging.

The following image of the Soul & Heart nebulas is a mosaic of two sessions with an Astro-Tech 65 EDQ scope & Hutech HEUIB-II filter, 30x60 sec @ ISO 5000 (on a Celestron AVX mount):

Misty Valley Ranch Observatory  In case you peruse the Misty Valley albums, please note they also contains images off an unmodded Sony A7 II and, if you go back a ways, several OSC and Canon DSLR images.

Enjoy!  It is a brave new world of astrophotography!

bwa

[sharkmelley]

I agree with you about OSC cameras.  As long as the optical train can take good advantage of a full frame camera with large pixels then I think this camera beats any of the current generation of cooled OSC cameras in terms of sheer image quality. 

I do have a number of niggles though:

1) The narrow diameter of the E-mount opening means that very severe vignetting can occur at the corners of the full-frame sensor depending on what correctors and adapters are in the optical train.

2) Sony have stayed with their 12 bit image format for this camera and they also continue to compress the raw file.  To be fair though, this appears to have little (if any) impact on final image quality - at least when deep sky imaging at ISO 2000-4000. 

3) There is very little software support in the astro-community as yet.  For accurate focusing I wrote a PixInsight script that auto-opened images that that the Sony software dumped in  a chosen monitored folder, at 1:1 scaling - this allowed me to use Bahtinov Grabber.

4) Sony appears to apply some kind of hot pixel and dark pixel suppression - the evidence for this is almost everywhere in the raw files where pairs of identical warm pixels and pairs of identical cool pixels appear all over the place.  I'm quite concerned that this algorithm could have nefarious effects similar to the infamous Nikon "star eater" algorithm on undersampled stars.  I'm still investigating this issue.   It could certainly rule out photometry at shorter focal lengths.

Mark

[ollypenrice]

This is impressive. Since it seems that manufacturers are losing interest in CCD technology I hope that something effective for astronomy will come along and replace it. I'd be very interested to see what a monochrome version of this chip could do.

Olly

[sharkmelley]

Yes, a mono cooled version of this sensor would be extremely interesting.  For one thing, the low read noise would make lengthy exposures for narrowband a thing of the past.

Mark