And what does Canon do?

[Loswr]

My Rebel T2i is good enough at low ISO.

Clearly, you are neither dramatically underexposing nor pushing your shadows several stops. Get with the program!!

Actually, I do sometimes. This is an example I use for something else, but this was shot in raw, many of the raw pixels in the ceiling were blown so it isn't underexposed, I just had to push the shadows a lot. This was shot on the T2i.

Then you're obviously doing something wrong if you can produce images like that, which aren't totally destroyed by shadow noise and banding. You have to try harder...lift exposures more, use no NR, crop away 90% of your image. If all else fails, try shooting with the lens cap on.

;D

 

[jrista]

And at times it truly does seem as though no one here realizes the differences, or just want to put their heads in the sand as far as the differences go.

I respect you a lot, and have found your input on this forum very useful. But sensor performance at low ISO really isn't a priority for some/many of us. It's not burying our heads in the sand, it's prioritising what's important to us.

I would point out, that it is no longer just low ISO. I was poking around with 1D X and A7s numbers the other day. The A7s is trouncing the 1D X. The former has over a 30% lead on total light gathering capacity at ISO 51200:

1DX: 5202px*3533px*163e-/px = 2,995,722,558e-
A7s: 4288px*2848px*322e-/px = 3,932,336,128e-

The A7s gathers a million more photons at ISO 51200 than the 1D X. Tht is a LOT. On top of the greater total light gathered, it also tops the 1DX in terms of dynamic range. The A7s 8.8 stops of DR at that ISO, vs. the 1D X's 6.6 stops. That is a 2.2 stop lead.

It used to only be ISO 100 where that kind of lead was held by the Exmor camp. Now, it enjoys a lead at both low and high ISO.

Things keep changing. They won't stop changing, and the changes are only going to accelerate.

I've dabbled in astro work. But light pollution is by far the biggest factor holding me back - and I suspect, most UK-based folk. I'd be much better off sinking money into learning to drive and going to a remote spot, or some other way of getting my gear out to darker skies, than buying a Nikon and hacking it. A better sensor isn't gonna help if I'm limited by other factors.

LP need not be the primary issue these days. I image under light polluted skies. My trick is to use an LPR filter, in my case the Astronomik CLS. I regularly get compliments about the depth of my exposures, which is primarily due to the use of the filter.

Software tools have also largely negated the consequences of shooting with LP. PixInsight has an amazing background extraction tool that can both flatten the field and neutralize skyfog gradients from LP (and other sources).

People shooting with Nikon DSLRs are under the same constraints as those shooting Canon DSLRs. There is no difference as far as conditions go. However, when you DO get out to dark skies, the significantly lower read noise of the Nikon cameras is vastly superior to the read noise of Canon cameras. One of the single biggest limitations with DSLRs is the saturation point at higher ISOs...that causes star clipping. With a "hacked" (it's easy...its like installing ML) Nikon, you can get less than 6e- RN at ISO 100. Most Canon cameras barely get that at ISO 400. You can shoot at ISO 100, get BETTER exposures and SNR with a Nikon, and never have to worry about clipped stars. And that's at a dark site.

And, that's just the beginning of the benefits.

Ditto wildlife. What limits me is disposition (I'm not a getting-up-before-dawn person, to get out to the best spots), and ability to travel. After that, autofocus, focal length, and high ISO quality are all far more important than anything else. A D8xx would get me more reach through cropping, but that's about it.

I would't say the D800 is the best camera for wildlife. However, what happens when Nikon, Pentax, and others start reusing the same sensor from the A7s? What happens when the competition gets 8+ stops of DR (TWO MORE STOPS) at ISO 51200 to the 1D X's 6.6 stops? (Let alone any other Canon camera, all of which get considerably less than that, and at even lower ISO settings.)

Canon had an edge for a short while...but their technology is holding them back now even on that front. It is not going to be long before Canon is trounced across the board, high and low ISO, as the competition keeps progressing, and Canon stands still (assuming they are...I'm happy to admit they haven't revealed their next sensor...things could have changed, and I hope they have.)

Image quality is massively important - but a little less noise isn't that big a deal to some of us. That's not belittling your position, nor is it wilful ignorance of reality. It's an assessment based on needs and desires. (And for example, I expose to the right as much as possible - so higher ISO quality is more important than shadow raising, because I prefer to lower the exposure in post, not the other way round).

It's not just the noise...it's the total dynamic range. If Sony keeps progressing Exmor, both at high and low ISO, as they have been for the last two years...we aren't just talking a small difference in noise. The A7s has TWICE the DR and TWICE the SNR at ISO 51200 as the 1D X. The 1DX...the crowning achievement of Canon two years ago. The A7s realizes a very significant gain at high ISO...it will (and does) have not just a little less noise...it's significantly less noise. People have already demonstrated that visually.

Canon is falling behind on all fronts. Their domain, the high ISO domain, is already being attacked, and the competition has produced superior results there as well. At the moment, the A7s has 1 2/3rds stops more DR at ISO 100, and 2 1/4 stops more DR at ISO 51200. As a wildlife and bird photographer who likes to do landscapes on occasion myself, I would LOVE to have that kind of sensor performance.

 

[Mt Spokane Photography]

Then you're obviously doing something wrong if you can produce images like that, which aren't totally destroyed by shadow noise and banding. You have to try harder...lift exposures more, use no NR, crop away 90% of your image. If all else fails, try shooting with the lens cap on.

;D

Don't underrate shooting with the lens cap on, so many seem to do it, it must produce good results. Or underexposing by 5 stops so that you can see the noise.

Speaking of poor exposures, I well remember the film days in the 1950's or 1940's when the box cameras with fixed f/8 lenses and fixed 1/125 shutter speeds had exposure corrected by the processing lab. In our case, the processing lab was Wrights Service Station, Grocery Store, Radio Repair, and Photo Processing. (Elmer was a great guy, I was fascinated by his technical gadgets when I was a kid).

I have several hundred negatives taken by my parents (and me) on 620 film that I've been scanning. Trying to pull a image from film that is nearly clear is a challenge. Bring a image to life from a underexposed negative that looks like a 10 stop ND filter is equally difficult. I miss all that beautiful grain

I still have a few boxes of those press 25 flash bulbs too. You could see spots for 15 minutes or more after one of them went off in a dimly lit room.

 

[Lee Jay]

At the moment, the A7s has 1 2/3rds stops more DR at ISO 100, and 2 1/4 stops more DR at ISO 51200. As a wildlife and bird photographer who likes to do landscapes on occasion myself, I would LOVE to have that kind of sensor performance.

Too bad that sensor is in such a horrible camera.

 

[scyrene]

And at times it truly does seem as though no one here realizes the differences, or just want to put their heads in the sand as far as the differences go.

I respect you a lot, and have found your input on this forum very useful. But sensor performance at low ISO really isn't a priority for some/many of us. It's not burying our heads in the sand, it's prioritising what's important to us.

I would point out, that it is no longer just low ISO. I was poking around with 1D X and A7s numbers the other day. The A7s is trouncing the 1D X. The former has over a 30% lead on total light gathering capacity at ISO 51200:

1DX: 5202px*3533px*163e-/px = 2,995,722,558e-
A7s: 4288px*2848px*322e-/px = 3,932,336,128e-

The A7s gathers a million more photons at ISO 51200 than the 1D X. Tht is a LOT. On top of the greater total light gathered, it also tops the 1DX in terms of dynamic range. The A7s 8.8 stops of DR at that ISO, vs. the 1D X's 6.6 stops. That is a 2.2 stop lead.

It used to only be ISO 100 where that kind of lead was held by the Exmor camp. Now, it enjoys a lead at both low and high ISO.

Things keep changing. They won't stop changing, and the changes are only going to accelerate.

I've dabbled in astro work. But light pollution is by far the biggest factor holding me back - and I suspect, most UK-based folk. I'd be much better off sinking money into learning to drive and going to a remote spot, or some other way of getting my gear out to darker skies, than buying a Nikon and hacking it. A better sensor isn't gonna help if I'm limited by other factors.

LP need not be the primary issue these days. I image under light polluted skies. My trick is to use an LPR filter, in my case the Astronomik CLS. I regularly get compliments about the depth of my exposures, which is primarily due to the use of the filter.

Software tools have also largely negated the consequences of shooting with LP. PixInsight has an amazing background extraction tool that can both flatten the field and neutralize skyfog gradients from LP (and other sources).

People shooting with Nikon DSLRs are under the same constraints as those shooting Canon DSLRs. There is no difference as far as conditions go. However, when you DO get out to dark skies, the significantly lower read noise of the Nikon cameras is vastly superior to the read noise of Canon cameras. One of the single biggest limitations with DSLRs is the saturation point at higher ISOs...that causes star clipping. With a "hacked" (it's easy...its like installing ML) Nikon, you can get less than 6e- RN at ISO 100. Most Canon cameras barely get that at ISO 400. You can shoot at ISO 100, get BETTER exposures and SNR with a Nikon, and never have to worry about clipped stars. And that's at a dark site.

And, that's just the beginning of the benefits.

Ditto wildlife. What limits me is disposition (I'm not a getting-up-before-dawn person, to get out to the best spots), and ability to travel. After that, autofocus, focal length, and high ISO quality are all far more important than anything else. A D8xx would get me more reach through cropping, but that's about it.

I would't say the D800 is the best camera for wildlife. However, what happens when Nikon, Pentax, and others start reusing the same sensor from the A7s? What happens when the competition gets 8+ stops of DR (TWO MORE STOPS) at ISO 51200 to the 1D X's 6.6 stops? (Let alone any other Canon camera, all of which get considerably less than that, and at even lower ISO settings.)

Canon had an edge for a short while...but their technology is holding them back now even on that front. It is not going to be long before Canon is trounced across the board, high and low ISO, as the competition keeps progressing, and Canon stands still (assuming they are...I'm happy to admit they haven't revealed their next sensor...things could have changed, and I hope they have.)

Image quality is massively important - but a little less noise isn't that big a deal to some of us. That's not belittling your position, nor is it wilful ignorance of reality. It's an assessment based on needs and desires. (And for example, I expose to the right as much as possible - so higher ISO quality is more important than shadow raising, because I prefer to lower the exposure in post, not the other way round).

It's not just the noise...it's the total dynamic range. If Sony keeps progressing Exmor, both at high and low ISO, as they have been for the last two years...we aren't just talking a small difference in noise. The A7s has TWICE the DR and TWICE the SNR at ISO 51200 as the 1D X. The 1DX...the crowning achievement of Canon two years ago. The A7s realizes a very significant gain at high ISO...it will (and does) have not just a little less noise...it's significantly less noise. People have already demonstrated that visually.

Canon is falling behind on all fronts. Their domain, the high ISO domain, is already being attacked, and the competition has produced superior results there as well. At the moment, the A7s has 1 2/3rds stops more DR at ISO 100, and 2 1/4 stops more DR at ISO 51200. As a wildlife and bird photographer who likes to do landscapes on occasion myself, I would LOVE to have that kind of sensor performance.

I appreciate you taking the time to respond in depth I do use an LP filter when I do night sky work, and it has been transformational. Ditto a tracking mount. When I return to astro work, I'm gonna get a full frame filter so I can use the 5D3 - but the 50D has done sterling work in my very amateur opinion.

As far as every competitor getting an A7s sensor, so what? I like that camera, and if I had a spare couple of grand, I'd get one. But the low megapixel count means far less cropping, which is relevant to my work. And the advantages are largely at the highest ISO settings (as far as the comparisons I've seen posted online are concerned). It would be a great low-light party camera, or a good dusk/dawn video option. But autofocus capability usually trounces other image quality aspects in my experience, for birds especially. And unless Sony comes out with native super tele lenses, that rules this body out for that work. (If Nikon used the sensor, it would be more feasible).

On the subject of astro, while I've got you - I tend to shoot at high ISO, partly because I can't trust my mount to do exposures longer than, say, 2mins at 100mm. Is Nikon/Sony in the lead enough at say ISO 3200 to be worth swapping? I mean, I use an old camera as it is, and get results that please me (I'm not trying to compete with the serious astro nuts).

 

[Loswr]

Canon is falling behind on all fronts.

Since you previously praised Canon's AF performance, a conclusion like that seems unwarranted. What you mean is 'falling behind' on sensor performance, but even there, the 3rd best golfer on the PGA tour, or even the last place golfer, is still really damn good. Similarly, Canon's 'fallen behind' sensors are amply good for the vast majority of photographers.

At the moment, the A7s has 1 2/3rds stops more DR at ISO 100, and 2 1/4 stops more DR at ISO 51200. As a wildlife and bird photographer who likes to do landscapes on occasion myself, I would LOVE to have that kind of sensor performance.

So would I. But...let me guess...you also wish for excellent subject tracking AF, and you also wish for native compatibility with excellent long lenses, and you also wish for good ergonomics and UI, and you also wish for true RAW files without compression...and you wish for all of that in a single camera that you can afford.

Well, there's certainly nothing wrong with wishing for those things. Of course, as my dad used to say: "Wish in one hand and sh!t in the other, and see which fills up first."

 

[3kramd5]

Things keep changing. They won't stop changing, and the changes are only going to accelerate.

Can I ask how (without an increase in bit depth)?

 

[raptor3x]

The A7s gathers a million more photons at ISO 51200 than the 1D X. Tht is a LOT. On top of the greater total light gathered, it also tops the 1DX in terms of dynamic range. The A7s 8.8 stops of DR at that ISO, vs. the 1D X's 6.6 stops. That is a 2.2 stop lead.

I find the difference in dynamic range interesting as even at ISO 200K there's only ~2/3 stop difference between the 1DX and A7S in SNR according to DxO and at ISO 50K the advantage is ~1/3 stop.

 

[Orangutan]

Of course, as my dad used to say: "Wish in one hand and..."

I just love those old Irish aphorisms.

 

[Clownbaby]

I'm a long time Canon user and I don't see how people can defend Canon so much. They literally produced nothing interesting for years, Price their gear really high and really don't seem to care about their customers. Perhaps the most annoying thing to me was the change in CPS service. After two months of shelling out $500 for platinum, they change the terms of repair costs, super annoying.

Besides lack of innovation, they cripple their hardware so that they lack features. Thanks to ef mounts become more available on other cameras, I certainly won't be buying any more Canon bodies unless something really blows me away.

 

[Orangutan]

I'm a long time Canon user and I don't see how people can defend Canon so much. They literally produced nothing interesting for years, Price their gear really high and really don't seem to care about their customers. Perhaps the most annoying thing to me was the change in CPS service. After two months of shelling out $500 for platinum, they change the terms of repair costs, super annoying.

Besides lack of innovation, they cripple their hardware so that they lack features. Thanks to ef mounts become more available on other cameras, I certainly won't be buying any more Canon bodies unless something really blows me away.

Could you be specific: which features are lacking in which bodies?

 

[jrista]

At the moment, the A7s has 1 2/3rds stops more DR at ISO 100, and 2 1/4 stops more DR at ISO 51200. As a wildlife and bird photographer who likes to do landscapes on occasion myself, I would LOVE to have that kind of sensor performance.

Too bad that sensor is in such a horrible camera.

Yeah, I agree. I also don't think it will stay horrible forever...at some point, Sony will start putting more effort into their bodies. Even if they don't, you have to figure other manufacturers will start using that sensor as well.

 

[jrista]

Canon is falling behind on all fronts.

Since you previously praised Canon's AF performance, a conclusion like that seems unwarranted. What you mean is 'falling behind' on sensor performance, but even there, the 3rd best golfer on the PGA tour, or even the last place golfer, is still really damn good. Similarly, Canon's 'fallen behind' sensors are amply good for the vast majority of photographers.

At the moment, the A7s has 1 2/3rds stops more DR at ISO 100, and 2 1/4 stops more DR at ISO 51200. As a wildlife and bird photographer who likes to do landscapes on occasion myself, I would LOVE to have that kind of sensor performance.

So would I. But...let me guess...you also wish for excellent subject tracking AF, and you also wish for native compatibility with excellent long lenses, and you also wish for good ergonomics and UI, and you also wish for true RAW files without compression...and you wish for all of that in a single camera that you can afford.

Well, there's certainly nothing wrong with wishing for those things. Of course, as my dad used to say: "Wish in one hand and sh!t in the other, and see which fills up first."

Sure I do. Aside from the AF thing, I could attach the A7s to my Canon lenses with an adapter. I'm not a big fan of mirrorless, but again...technology in cameras is progressing on all fronts. EVFs won't stand still either. The successor to the A7s could be a great camera all around. Sony has also demonstrated much improved subject tracking...who knows exactly how it compares to Canon's today, although once you realize the difference (and Sony's tracking system DOES have a feature that I don't think Canon or Nikon have), it seems really impressive...but, again...Sony AF technology won't just stand still.

As for wishing vs. shitting...I'm doing neither. I'm watching. I don't need to wish...it won't be all that long before Sony's currently lacking bodies and new AF system become better bodies and better AF system. Canon moves so slowly compared to the rest of the industry...by the time the next high performance Canon DSLR hits the streets, there could have been two more generations of Sony and Nikon cameras.

No need for wishing...just watch, and see improvements occur generation over generation on the competitors cameras. At some point, one of them has to end up good enough that there is absolutely no reason not to pick one up...and it'll likely outperform a Canon camera completely on the sensor IQ front, and perform well enough on other fronts (i.e. AF). If it's mirrorless and you can get an adapter...well...then you have kick-ass lenses paired with a better sensor and a decent AF system.

It's just a matter of time, now. Unless, of course, Canon has actually done something really intriguing on the sensor front with the 7D II, or ends up doing something radical with the 5D IV sensor. Things can still change...I hope they do...but the world ain't gonna wait for Canon.

 

[Clownbaby]

I'm a long time Canon user and I don't see how people can defend Canon so much. They literally produced nothing interesting for years, Price their gear really high and really don't seem to care about their customers. Perhaps the most annoying thing to me was the change in CPS service. After two months of shelling out $500 for platinum, they change the terms of repair costs, super annoying.

Besides lack of innovation, they cripple their hardware so that they lack features. Thanks to ef mounts become more available on other cameras, I certainly won't be buying any more Canon bodies unless something really blows me away.

Could you be specific: which features are lacking in which bodies?

On the 5diii which is still more expensive than better newer cameras but compared to d800
Worse IQ
Less sharp
Worse color
Less dr
True 1080p


My c100 lacks a ton that its competitors have, too much to list

 

[Orangutan]

Could you be specific: which features are lacking in which bodies?

On the 5diii which is still more expensive than better newer cameras but compared to d800
Worse IQ
Less sharp
Worse color
Less dr
True 1080p


My c100 lacks a ton that its competitors have, too much to list

So why did you buy a 5D3 and C100? Or why don't you sell them and buy a competing product?

 

[Old Sarge]

Speaking of poor exposures, I well remember the film days in the 1950's or 1940's when the box cameras with fixed f/8 lenses and fixed 1/125 shutter speeds had exposure corrected by the processing lab. In our case, the processing lab was Wrights Service Station, Grocery Store, Radio Repair, and Photo Processing. (Elmer was a great guy, I was fascinated by his technical gadgets when I was a kid).

Are you sure the shutter speed was that fast? Just eyeballing my first camera (a Brownie Hawkeye) and my mother's camera that I used before that (forget which model that was...it was sort of neat, had two "f/ stops", not marked as such....one for sunny and one for cloudy) I would have pegged the shutter at close to 1/50th. But I could be wrong. If I remember the film we used was about ASA 80. The first rolls I used were orthochromatic then later they were panchromatic. Does that sound right to you?

 

[Clownbaby]

Could you be specific: which features are lacking in which bodies?

On the 5diii which is still more expensive than better newer cameras but compared to d800
Worse IQ
Less sharp
Worse color
Less dr
True 1080p


My c100 lacks a ton that its competitors have, too much to list

So why did you buy a 5D3 and C100? Or why don't you sell them and buy a competing product?

I admit going canon was a mistake but I have a lot of canon glass. I've thought about it but that would just end up costing money to sell my bodies at a loss before I planned to replace them but going forward is different. It's not like they're terrible but I just don't feel like I'm getting a good value out of being a Canon customer.

 

[Chuck Alaimo]

It's the differences in IQ that I'm referring to. Features are a dime a dozen. ...

We use cameras to take pictures. When it comes to taking pictures, most Canon cameras have not changed much in YEARS as far as IQ is concerned.

We use cameras to take pictures, not just sensors (where have I heard that before? : ). Of all the pictures taken with a dSLR, what fraction would you say we're shot without autofocus? Less than 2%? Less than 0.2%? The 40D got a big AF boost, in the T4i that was improvement reached the entry-level line. The 7D had the best APS-C AF available, now in the xxD line, and if the rumored spec is true, the 7DII will once again have the best APS-C AF system. The 1D X and 5DIII have top AF systems, and both are significant improvements over their predecessors.

Better AF – more cross-type points, spread further across the frame, f/2.8 accuracy in the center, those 'dime-a-dozen' features translate directly to a higher rate of in-focus images. So...does a blurry, misfocused image with 13 stops of DR have better IQ than a properly focused image with 2 stops less DR? Are you going to take a blurry, misfocused image and lament over noise when you push the shadows...or just delete the image?

For the small minority (there's that word again) who's shooting style depends primarily on the sensor...tripod users shooting static scenes, Canon sensor IQ has improved only slightly...and as the many award-winning, awe-inspiring landscape scenes shot with Canon cameras will attest, the sensor IQ was excellent already.

I think people are clamoring for it. However they are clamoring for other products. The sheer number of new Nikon DSLRs being used with the black point hack in astrophotography...

I didn't say people aren't clamoring for it. I said not many people, as in a minority of dSLR buyers, are clamoring for it. "The sheer number of ... DSLRs being used ... in astrophotography...," LOL. I know you later acknowledged it's a niche, but thanks for the laugh on the way there...

oh come now neuro, we all know what clients and buyers want...that is of course to pixel peep! that is why I am gonna change what I show to clients from now on. Instead of showing them an image, I will take an extremely small crop of a corner of the image.

I mean, the evidence is clear, see below...the first 2 are tiny crops of the bottom left and right corner...I mean who would want to buy that---it's so shadowy...I know...if only i had an exmor......don't even look at the 3rd, i mean, the 2 super crops should be enough to tell the story right?

 

[jrista]

As far as every competitor getting an A7s sensor, so what? I like that camera, and if I had a spare couple of grand, I'd get one. But the low megapixel count means far less cropping, which is relevant to my work. And the advantages are largely at the highest ISO settings (as far as the comparisons I've seen posted online are concerned). It would be a great low-light party camera, or a good dusk/dawn video option. But autofocus capability usually trounces other image quality aspects in my experience, for birds especially. And unless Sony comes out with native super tele lenses, that rules this body out for that work. (If Nikon used the sensor, it would be more feasible).

Who's to say A7s level performance won't be the norm with the next round of APS-C or high res FF releases? I am still a firm believer that total sensor area and Q.E. are the primary factors that affect noise. In the case of the A7s...it's clear that they are doing some rather intense noise reduction in the BionzX chip as well (much like the DIGIC 6 does). All three of those factors, IMO, are more important than pixel size in the long run. So I see no reason you couldn't have that level of ISO performance in a sensor with smaller pixels.

For now, if you need cropping ability, then pixel size is probably a key factor. Doesn't mean that A7s level performance will stay only in sensors with bigger pixels forever. (Or even necessarily for all that long.)

I appreciate you taking the time to respond in depth I do use an LP filter when I do night sky work, and it has been transformational. Ditto a tracking mount. When I return to astro work, I'm gonna get a full frame filter so I can use the 5D3 - but the 50D has done sterling work in my very amateur opinion.

On the subject of astro, while I've got you - I tend to shoot at high ISO, partly because I can't trust my mount to do exposures longer than, say, 2mins at 100mm. Is Nikon/Sony in the lead enough at say ISO 3200 to be worth swapping? I mean, I use an old camera as it is, and get results that please me (I'm not trying to compete with the serious astro nuts).

Astro is a very different beast. SNR, DR, and RN levels are all important. In astro, final signal strength or signal power is critical. For best results, you need to find the sweet spot that maximizes them all. I'd say before getting a different camera, start guiding your mount. More reliable tracking will probably do more for you than a better sensor, at least to start. I think that the A7s could certainly do better at high ISO than any current Canon sensor. It would certainly allow for the use of higher ISO settings...but I'm not sure that will really buy you much in the end. The key benefit of the increased DR of the A7s is simply a lower chance of clipping stars. In the long run, even with an A7s, guiding is going to allow you to expose for twice as long or longer than your current 2 minute exposures. (And, at 100mm, I would say even with a mount's periodic error, you should be able to expose for four minutes easily...unless the mount has some seriously wicked PE!)

In general ISO 3200, IMO, is rather high for astro unless your doing ultra wide field work (i.e. 16mm milky way imaging). For anything DSO related, you should be using the lowest ISO you can get away with and still have the minimal or nearly minimal read noise. That will maximize your DR and SNR, limiting the chance that stars will clip, maximizing your ability to increase the total signal power, and usually at no cost to your actual REAL exposure.

First, exposure. Exposure really has nothing to do with ISO. Exposure is purely a matter of total light quantity gathered. That's light volume over time...aperture and shutter. Once the camera's read noise reaches a floor, then increasing ISO any further is not actually going to change anything. Well, it will change something...it will put a ceiling on how long you can expose for...which will directly impact your signal power and SNR, which will limit your ability to expose dimmer details above the noise floor. (And when detail is buried in the noise floor, ultimately it doesn't matter how many subs you take...you can use averaging to improve SNR, but no amount of averaging can improve signal power, and THAT is really what brings out the fainter details.)

In the case of the 50D, you reach a read noise floor around 3.3e- by ISO 800. Technically speaking, the sensor is "ISO-less" from that point on. At ISO 400 RN is 5.5e-, and at ISO 200 it's 9.5e-. You could probably get away with ISO 400...even high end CCD cameras rarely get down to the 5e- RN level, many are at 7-8e- and some are over 10e-.

So, for best astro results with a Canon like the 50D (or pretty much any current Canon APS-C), you want to shoot at ISO 400. You might take a small hit to read noise, however you'll gain considerably on the dynamic range front. That will allow for much longer exposures before you clip your stars...which means it shouldn't be difficult to shift the whole histogram more to the right a little bit, and separate the lowest color channel away from the left-hand edge of the histogram (which is all you really need to do to get the signal above the read noise floor.)

When it comes to the 5D III, I shoot at ISO 800 and 400. ISO 800 has good DR (more than any Canon APS-C at ISO 400), and 6.1e- RN which is still pretty good...but the larger pixels have a tendency to undersample shorter lenses/telescopes and therefor it's rather easy to clip stars. ISO 1600 is where it drops to the RN floor, but at half the dynamic range...so it's even worse for clipping stars...I usually have to considerably shorten my exposure times, which completely negates any benefit the lower read noise might have offered. I rarely shoot at ISO 1600 unless it's a REALLY faint object (and those are usually too small to really be resolved usefully by such a large sensor with large pixels...so I basically never use ISO 1600). ISO 400 has a ton of dynamic range (relatively speaking), but 10.1e- RN. I usually expose for a little bit longer than one stop more than at ISO 800 to get ISO 400 shots above the read noise floor, and I am usually safe from clipping my stars.

I've found that ISO 400 is generally a pretty good ISO setting to use, on all Canon cameras, despite the higher read noise...especially when I need to expose for a long time (over four or five minutes), or am imaging a really bright target (Orion Nebula, Andromeda). ISO 800 is great for areas of fewer stars and dimmer targets...or areas where there may be lots of stars, but smaller ones, and dimmer nebula.

When it comes to a Nikon camera using the black-point hack, the game has totally changed. Read noise is ridiculously low at ISO 100, it tops out at around 6e- at most (it's usually less than that). At ISO 200 it's still closer to 3e-. There is little reason not to consider the D800 with the hack as a true linear sensor, and just use ISO 100 for everything unless you need to use really short exposures or are imaging a particularly dim target. Then, you probably only need to bump up to ISO 200. With all the dynamic range these sensors have, you can practically expose forever. Most people who haven't spent $8000 or more on a mount are unlikely to be exposing for much longer than 10 minutes anyway, and at ISO 100 on a hacked D800, you can easily expose for 10 minutes and still not even risk clip your stars. The linearity of a hacked D800 is much better than any Canon camera for astro work. The read noise is lower. The dark current is considerably lower (you could do away with dark frames if you use dithering). It would be the closest thing to having something like a QSI690 CCD camera ($3500-4200), with the primary drawback being that the D800 isn't cooled (which reduces dark current to ludicrously low levels).

 

[DominoDude]

...
Astro is a very different beast. SNR, DR, and RN levels are all important. In astro, final signal strength or signal power is critical. For best results, you need to find the sweet spot that maximizes them all.
...


In the case of the 50D, you reach a read noise floor around 3.3e- by ISO 800. Technically speaking, the sensor is "ISO-less" from that point on. At ISO 400 RN is 5.5e-, and at ISO 200 it's 9.5e-. You could probably get away with ISO 400...even high end CCD cameras rarely get down to the 5e- RN level, many are at 7-8e- and some are over 10e-.

So, for best astro results with a Canon like the 50D (or pretty much any current Canon APS-C), you want to shoot at ISO 400. You might take a small hit to read noise, however you'll gain considerably on the dynamic range front.
...

I've found that ISO 400 is generally a pretty good ISO setting to use, on all Canon cameras, despite the higher read noise...especially when I need to expose for a long time (over four or five minutes), or am imaging a really bright target (Orion Nebula, Andromeda). ISO 800 is great for areas of fewer stars and dimmer targets...or areas where there may be lots of stars, but smaller ones, and dimmer nebula.
...

Jumping into the debate here, after seeing something that tickled my interest. I've cut wildly in your response, Jon, but it's mostly since what I want to ask about is the usage of "full-stop ISO". Perhaps I should've started a new topic about it, but lets see if it gets to be interesting enough to deserve a thread of its own.

When I first started to use a digital camera, it was the EOS 50D, and rather soon it showed some ugly noise while I shot at all the regular ISO-stops. Looked for answers in magazines all over the place, and googled like crazy, and after a while the answers seemed to say: If you want low amounts of noise from a Canon sensor stay away from the regular ISOs.
So, have you noticed anything similar while shooting deep into the star-filled skyes? Or is your specialty so much different than regular shooting?

In my experience what gives me the cleanest results out of any Canon dSLR (I can always mess things up later in post... ) is to shoot at ISO 160, then ISO 320, followed by ISO 800.

I leave it at this for now - I'm in a hurry, it's early morning over on this continent, and I have to leave for the day to drive a shitload of miles. Will peek in tonight to see if this leads anywhere, and if anyone else has been curious about the odd-ISO usage.