Switched to Sony...

beforeEos Camaras

love to take photos.
Sep 8, 2014
248
34
#43
+1 these forums are less about photography and more about fighting...
no sir you have it all wrong your zeal on bringing the so called truth is mind boggling remember sony dat tec? or Betamax and the list is so long of failed items its scary.and digital cameras I got a box point and shoots plus the majic memory sticks don't even say sony is best its not.

I want the use of something that I can have serviced years later plus a proven file compatibly so you won the war go back the sony forums and stop trolling this one
 

Keith_Reeder

No apologies for not suffering fools gladly...
Feb 8, 2014
676
105
58
Blyth, NE England
#44
I'll tell ya. The best thing I ever did was leave this nutsy forum and just DO photography. .
So feel free to continue staying away.

Oh - and the best thing you could ever have done was learn how to convert and post-process your files properly: because all I ever saw, when you posted up images to "prove" this problem or that failing, was someone who had no idea what to do with his files once they were off the camera.

Presumably you'll have found the time to do that, now you're not distracted by CR?
 
Dec 3, 2011
5,341
21
jonrista.com
#49
Interesting that your website is still full of 7D wildlife images then, don't you think? Where are all the superlative Sony shots that blow Canon out of the water?

I suppose we'll be waiting a while for them, eh?
Good to see the old vitriol and hate is alive and well here on CR. *sigh*

I have been doing astrophotography for years now, and it dominates what I do. I work with quite a few different kinds of cameras, and have processed data from dozens of different kinds of cameras, including DSLRs as well as scientific CCDs, astronomy CCD and CMOS cameras, etc. Sensors from Canon, Sony, Panasonic, Kodak, OnSemi, Aptina, and others. I routinely work with signals orders of magnitude fainter than anyone doing terrestrial photography works with (fainter than starlight, and starlight is darn faint!!).

I know how to process just fine. I am intimately familiar with faint signals and how to extract the most from them, and how different cameras compare in terms of noise and signal quality.

Here is one of my latest works, in this case with a Sony IMX183 sensor, 2.4 micron pixels, at about 1.9e- read noise and gain of 1e-/ADU:



This is some of the most beautiful raw data I've ever worked with, and acquired with a relatively modest aperture...and with very tiny pixels. Canon data doesn't come close to this kind of quality. I've processed very faint signals from every major Canon DSLR on the market (I now understand that Canon's sensors suffer from relatively high FPN, which is more of a problem for them than other noise terms, and even with dark calibration, their FPN is often unstable and does not always correct properly). I also use a Panasonic sensor regularly, and it performs very well as well...and in some respects, is actually superior to most of the Sony sensors I use. The only thing that can top the Sony IMX183 and Panasonic M IQ is data from a KAF-16803 (originally Kodak, now an OnSemi CCD sensor with very large pixels and immense dynamic range) when paired with large telescopes...however the camera alone costs at a minimum ten grand, and telescopes that can deliver comparable results with this KAF camera tend to cost eight to twenty grand or more. The KAF-16803 also requires some fairly extensive total exposure time with long individual exposures when operated at similar image scales (here, ~0.82"/px for the image above).
 
Last edited:
Dec 3, 2011
5,341
21
jonrista.com
#51
Thanks, Don! AP has really consumed my imaging, just about entirely. I am actually trying to get back into some terrestrial photography, as I haven't actually done much of any the last couple of years (some floral and fungal stuff mostly). While I am not a huge fan of Canon sensors, I'm a big fan of their glass, and it is well past time for me to upgrade some of my EF lenses (which thankfully, due to their decent backfocus, work on pretty much every camera I own regardless of brand. ;P)
 

AlanF

EOS 5DS R
Aug 16, 2012
4,135
335
#52
Jon, great shot! Is the Sony sensor much different from the one they use in the latest RX100 and RX10 series?
 

Maximilian

The dark side - I've been there
Nov 7, 2013
2,172
104
Germany
#53
Good to see the old vitriol and hate is alive and well here on CR. *sigh*
I have been doing astrophotography for years now, and it dominates what I do. ... Here is one of my latest works...
Hi Jon!

Great picture, compliments

... I work with quite a few different kinds of cameras,
... Sony IMX183 sensor ... some of the most beautiful raw data I've ever worked with
... I now understand that Canon's sensors suffer from relatively high FPN ...
Thank you for sharing your experiences and knowledge. I sometimes really missed that.
I think it is clear to all (even the Canon fans) that Sony sensors are better than Canon sensors, esp. when it comes to low light and/or low ISO.
Some are not willing to state that and that's sad.

I did a quick search for that IMX183 and only found some dedicated astro cams and some rumors from 2017 that it could be in RX cameras.
Can you tell me if it is used in any "consumer" camera?


If not I suppose that you can also admit, that this is a dedicated tool for a really special purpose. Isn't it?
And before you think I want to start that "... more than just the sensor ..." debate, no I won't.
Of course it stands as example for other Sony sensors that are built in "consumer" cameras and I want to understand your input in a better way.

I would be also interested in how you use it. I suppose you you do quite long exposures and read out the raw/base data.
Just in general: How do you read out and proceed that data?
As I am not into astro that much I don't want to get too much into detail.

Thank you in advance.
 
Last edited:

JBSF

EOS M50
Dec 19, 2014
94
37
#54
Good to see the old vitriol and hate is alive and well here on CR. *sigh*

I have been doing astrophotography for years now, and it dominates what I do. I work with quite a few different kinds of cameras, and have processed data from dozens of different kinds of cameras, including DSLRs as well as scientific CCDs, astronomy CCD and CMOS cameras, etc. Sensors from Canon, Sony, Panasonic, Kodak, OnSemi, Aptina, and others. I routinely work with signals orders of magnitude fainter than anyone doing terrestrial photography works with (fainter than starlight, and starlight is darn faint!!).

I know how to process just fine. I am intimately familiar with faint signals and how to extract the most from them, and how different cameras compare in terms of noise and signal quality.

Here is one of my latest works, in this case with a Sony IMX183 sensor, 2.4 micron pixels, at about 1.9e- read noise and gain of 1e-/ADU:



This is some of the most beautiful raw data I've ever worked with, and acquired with a relatively modest aperture...and with very tiny pixels. Canon data doesn't come close to this kind of quality. I've processed very faint signals from every major Canon DSLR on the market (I now understand that Canon's sensors suffer from relatively high FPN, which is more of a problem for them than other noise terms, and even with dark calibration, their FPN is often unstable and does not always correct properly). I also use a Panasonic sensor regularly, and it performs very well as well...and in some respects, is actually superior to most of the Sony sensors I use. The only thing that can top the Sony IMX183 and Panasonic M IQ is data from a KAF-16803 (originally Kodak, now an OnSemi CCD sensor with very large pixels and immense dynamic range) when paired with large telescopes...however the camera alone costs at a minimum ten grand, and telescopes that can deliver comparable results with this KAF camera tend to cost eight to twenty grand or more. The KAF-16803 also requires some fairly extensive total exposure time with long individual exposures when operated at similar image scales (here, ~0.82"/px for the image above).

Very beautiful, and quite accomplished.
 
Dec 3, 2011
5,341
21
jonrista.com
#56
The IMX183 is actually a different sensor than the one used in the RX100. The RX100 uses an Exmor RS sensor, which is actually arguably BETTER than the IMX183 sensor I am using right now. The IMX183 is a machine vision sensor, and it has some amp glow, a kind you would never see in a mirrorless photography sensor. There are some astro cameras built on Sony Exmor sensors, including the IMX128 and IMX094, which are used in familiar cameras like the D810 and D750. Sony Exmor sensors are, IMO, the best there is. The IMX183 is a good sensor, with significantly lower noise, vanishingly low FPN, and overall a much cleaner noise profile in general than any Canon sensor, but it is still not quite as good as an Exmor sensor.

The key issue with amp glow is, you really cannot use the data without calibrating it properly. Calibration requires subtraction of a master dark frame. For normal photography, which is usually composed of a single frame, calibration is never done and usually not an option with the commonly used raw processing software. When calibrated, the IMX183 has amazing characteristics, but that glow does put it a step behind the Exmor sensors when it comes to cutting edge strait out of camera IQ. The read noise of the IMX183 is very low, and in fact may even be lower than a lot of the Exmor sensors, and it has ultra low dark current (around 0.002e-/s), and its noise profile is just wonderful. For the tiny 2.4 micron pixels, it actually has a rather larger 15,000e- full well capacity, which is a bit of a technical feat in and of itself. The IMX183 also comes in monochrome versions which work better with astrophotography imaging filters.

If I had the funds, I'd get an astro camera with an IMX128 or IMX094 in it. These are the full frame sensors used in the D810, D750, A7r, etc. They are amazing sensors. Sadly they also only come in color versions, there are no mono versions, which limits their versatility for astrophotography. Still...two of the best sensors on the planet at the moment.
 
Dec 3, 2011
5,341
21
jonrista.com
#58
Regarding usage, there is actually a lot of flexibility with it. I use moderately long exposures of 10 minutes with narrow band imaging (narrow band filters cut out all but about (actually less than) 1% of the visible spectrum, which GREATLY reduces the photon flux levels to far below even normal starlight levels...the image above is actually a multi-channel narrow band image). However the camera has variable gain, and can be used at pretty high gain settings. At higher gains I can use 2-3 minute narrow band exposures, which are actually very short in the grand scheme of astrophotography exposures.. Higher gains have less dynamic range though...so, it is often a balancing act. Most CCD cameras require exposures at least 20-30 minutes for narrow band, and in some cases exposures of 60-90 minutes are required for CCDs with read noise in the range of 8-11e-. I know of some CCD users who push their individual sub exposures to 2 hours... (FTR, the Canon 5D III, which I've used extensively for both birds, wildlife and astrophotography, has 6-8e- read noise when used at viable ISO settings...and it, too, requires pretty long exposuers. Canon FPN is horrible, though...)

For LRGB imaging (luminance and RGB), exposures are usually in the range of 30-90 seconds. This can actually present a bit of a challenge, as with astrophotography we work with signals so faint that and yet objects with such immense dynamic range that a single exposure is insufficient to capture enough of the faint signal without clipping too much of the bright signals (i.e. stars, galaxy cores, etc.) We have to compensate for that by stacking lots of individual sub exposure frames together. When exposures are only 30-60 seconds, but we need several hours of data in total to produce an image with good SNR (again, we are working with signals many orders of magnitude fainter than starlight, and orders of magnitude fainter still than the darkest shadows anyone might pull out of a terrestrial photo), which can mean stacking hundreds of frames. That can get a bit tedious once you get beyond about 200 frames, as disk space, cpu and memory requirements start to skyrocket when you need to handle 300, 500, 1000 frames.

With narrow band, where signals are again orders of magnitude fainter STILL, we sometimes get an object photon once every minute or so! It can take several minutes to pick up enough photons to produce a usable photoelectric object signal in each sub in the faintest areas of the object. So even with 10 minute subs, we may only have a few electrons worth of signal. I usually stack around 60 10m subs, which is 10 hours of exposure. Such an integrated exposure may, finally, have a signal reminiscent of the shadows in a normal terrestrial photo. :p

The value of calibration then becomes significant here, as dark signal introduces FPN, and FPN can actually be a fairly significant noise term in each individual exposure. In fact, having a lot more experience with a wide range of cameras and sensors these days, I think one of Canon's bigger issues is high FPN...the 7D was a good example of this. Even with strong signals, the 7D images could still look noisy...and this was because of FPN, not read noise. Subtracting a master dark calibration frame removes most of the FPN, producing a cleaner signal. The IMX183 has one of the cleanest dark signals and lowest FPN of any sensor I have used so far, save some of the Exmor sensors (which are about as low) and a couple of the Sony ICX CCD sensors (there are two ICX sensors on the market that are definitely better than the IMX sensors I use...however Sony has sadly ended production of CCDs and these amazing sensors, the ICX814 and ICX834, will no longer be an option once existing supplies run out...plus, cameras built on these sensors tend to cost about $4500 base, and with all necessary accessories for astrophotography, usually closer to $6-7k.)
 
Last edited:

Don Haines

posting cat pictures on the internet since 1986
Jun 4, 2012
7,317
311
Canada
#59
I do not, and never will, understand brand fanaticism. If you are threatened by someone else’s choice, you are the one with the problem.....

Personally, I have been getting into astrophography as well. My understanding is that the most important gear is a good tracking mount and astroguider, and knowing how to process images and what type of images to take is far more important than camera choice. Once I get better at that, I intend to revisit camera bodies, and if Sony is the answer, so be it! My choice of an astrophotography camera has zero impact on what I choose to use in the rest of my life....

Jon has been an inspiration to me. His images show what is possible and he is generous with advice. I am glad to see him back on the forum and look forward to see what the future brings.
 
Jul 6, 2017
834
62
Davidson, NC
#60
Personally, I have been getting into astrophography as well. My understanding is that the most important gear is a good tracking mount and astroguider, and knowing how to process images and what type of images to take is far more important than camera choice.
I have not tried taking any pictures through my telescope since I used slide film over 15 years ago. I imagine that the old Canon mount for the telescope might work with my current digital bodies, but I haven't tried. Light pollution around here is too great to photograph much of anything. Even back then, I don't think the clock mechanism was that accurate. I got some decent shots of Jupiter, but didn't try Saturn, though it was relatively close at the time and looked good through the telescope. An exposure long enough was too likely to be a blur, so not worth wasting film on. Also eyeballing the angle to Polaris with my equipment that is rather primitive by modern standards would have given even more of a tracking error.

So I would be interested to know what you come up with in method, equipment, subject matter, even though I'm not gung ho enough to travel to dark places and try my hand.