Typical noob question. What camera?

[Pigpen313]

hey everybody!
Typical noob post.
I REALLY want to get into photography, I have been for awhile. Now is my time to finally be able to purchase a camera, but I need some advise after narrowing it down.
Things I want to use it for are:
Astrophotography.
Landscape (mountains, etc)
Motorcycle/car racing at the track.
Cityscape
Time lapses

Keep in mind, this is a beginner camera for a couple years. I would like one that is able to do all of the above, if possible! With the addition of lenses of course.
Also, if you can recommend affordable lenses for the above things I would like to use it for (for the specific camera you recommend), that would be great.

My budget is about $700, with the addition of lenses needed.
My choices I've narrowed it down to are:
Sony A6000
Nikon d5500 / d3300
Canon t6i / t6 / 60d
Any other cameras you recommend are greatly appreciated!

 

[Derrel]

I woukld go for the Nikon or Sony options, just for their fantastic sensor performance. The A6000 is now 'older', so the price has a lot of value in it at newer, discounted prices. The D5500 from Nikon has the flippy screen which coulkd help with creative, high- or low-angle shots.

Lenses for the Nikon would be the 35mm f/1.8 AF-S DX-Nikkor, or the 50mm f/1.8 AF-S G Nikkor lens.

 

[beagle100]

I agree, go for the Canon 800D or 850D for the super fantastic sensor performance
- and the larger selection of lens, cheaper lens, better IQ image quality, etc)
.
www.flickr.com/photos/mmirrorless

 

[goodguy]

I woukld go for the Nikon or Sony options, just for their fantastic sensor performance. The A6000 is now 'older', so the price has a lot of value in it at newer, discounted prices. The D5500 from Nikon has the flippy screen which coulkd help with creative, high- or low-angle shots.

Lenses for the Nikon would be the 35mm f/1.8 AF-S DX-Nikkor, or the 50mm f/1.8 AF-S G Nikkor lens.

What Derrel said!
Nikon cameras has advantage in dynamic range and with right lenses can give you potentially sharper images

 

[Solarflare]

Btw, the most important part of a camera is always the lens. Nothing else is so important for image quality.

Thats ultimately why Leica is still in business. Not as good sensors as others, and damn expensive, oh and repair service is otherwise good but very slow, but - excellent lenses. Ok, the same logic didnt worked for Pentax. If there was an offer that I would consider superior in lens performance, but inferior in sensor performance, I would go for it. Well, IMHO Fujifilm and Nikon are excellent with both sensors and lenses though anyway, so that question doesnt really come up.

Also I would like to point out that sensor performance is also simply really close nowadays. You can do excellent pictures with an 8 year old D700. The last really big step in sensor performance with Nikon was the step from D90 to D7000 - a full 2 stops of dynamic range thanks to the new Sony EXMOR technology.

hey everybody!
Typical noob post.
I REALLY want to get into photography, I have been for awhile. Now is my time to finally be able to purchase a camera, but I need some advise after narrowing it down.
Things I want to use it for are:
Astrophotography.
Landscape (mountains, etc)
Motorcycle/car racing at the track.
Cityscape
Time lapses

Well thats:

Astro:
Well, you probably need a tripod and possibly a star tracker. Preferably a bright wide angle lens and a full frame sensor with good High ISO performance, but with a star tracker one can use any focal length and aperture nicely. Either way, theres lots of money you can sink into this. Millions, of course, if telescopes count.

Landscape typically needs:
- wide angle to long angle, but not much in respect to aperture - you'll stop down to f/8 or f/11 most of the time anyway
- polarizer filter is pretty close to absolutely mandatory
- ideally a tripod for a 100% stable basis
- often a gradient ND filter and holder
- ND filter for long time exposure effects

Sports need:
- Good autofocus system
- Fast telephoto lenses - which aint cheap
- FPS (frames per second) and Buffer - especially for professionals who need to get the shot

Time lapses need:
Hmm OK I dont know that one. Never wanted to do a timelapse. Anything video doesnt really bother me.

I would recomment to you getting a Nikon D7100 (used market they go for about $500) now, because its an excellent all around performer (The autofocus performance is very good, but it lacks in respect to FPS and buffer compared to the best options, but you dont have the budget for a D500 or the like - for a hobbyist it should be enough) and then an AF-S 18-105mm f3.5-5.6 VR. That plus a polarizer should start you nicely into landscape. For sports you'll find the lens to be very slow - unfortunately lenses with good autofocus performance are costy. The cheapest option I know is the AF-S 70-300mm f4-5.6 VR, which is I think at like $300 on the used market now ?

I agree, go for the Canon 800D or 850D for the super fantastic sensor performance
- and the larger selection of lens, cheaper lens, better IQ image quality, etc)

Hmm ? Canon doesnt have cheaper lenses than Nikon. I dont think they have better lenses, either.

The best company in respect to average lens quality right now is IMHO Fujifilm. Fuji doesnt have the selection, the used market, and the third party offers yet though.

And Canon insists on making their own sensors - which means dualpixel tech, yes, but also inferior sheer sensor performance. Not that that would be too important, mind. Especially after Canon recently started their new fab with now much finer sensor structures.

 

[TCampbell]

I woukld go for the Nikon or Sony options, just for their fantastic sensor performance. The A6000 is now 'older', so the price has a lot of value in it at newer, discounted prices. The D5500 from Nikon has the flippy screen which coulkd help with creative, high- or low-angle shots.

Lenses for the Nikon would be the 35mm f/1.8 AF-S DX-Nikkor, or the 50mm f/1.8 AF-S G Nikkor lens.

What Derrel said!
Nikon cameras has advantage in dynamic range and with right lenses can give you potentially sharper images

This is not entirely true. It's an area that is often over-simplified. The specific point about lenses (the right lenses can give you potentially sharper images" is specifically generally not true. It's one of the reasons why experienced photographer who are aware of the DR differences still choose Canon... because of the lens differences.)

I can show you images from a Nikon D810 that look garbage next to images taken by a Canon 5D IV. Even though... the Nikon D810 allegedly has better dynamic range (it's a very subtle difference.)

Why?

We often tell beginner's that the ISO adjustment allows you to adjust the sensitivity of the sensor ... as if it makes the sensor "more sensitive" to light and thus can take a shorter exposure. While that explanation gets the point across, it's inaccurate to the point of being unhelpful. What the ISO really does is determine how much "gain" or "amplification" will be applied to the amount of light that the sensor collected at base ISO.

Why is this a big deal? Because cameras have two major ways to apply amplification but this varies by model. Nikon cameras seem to largely only amplify image data digitally. In other words, pretend you took the shot at base ISO, saved the file to the card, then went back and "amplified" all the values in the digital file by doubling them all (or increasing them proportionally based on the amount of amplification you need.)

Since an image may have had values for each pixel / photo-site ranging from very dark (near 0) all the way up to something near the max that the camera can capture (for 8-bit color it's 0-255 per channel, for 14-bit RAW that's 0-16,383 per channel) then amplifying those already bright pixels will cause them to clip. You can attempt to protect the bright data by compressing the histogram, but that's still a sacrifice to the dynamic range of the camera. This digital amplification is referred to as "downstream" amplification.

Canon cameras (nearly all models) apply both "upstream" and "downstream" amplification. "Upstream" amplification is handled by the sensor itself while the data is still analog (before it got through the analog to digital conversion process). Because the image data is not yet digital, it's not subject to loss of dynamic range. As you amplify data by "upstream" method, you give up very little in the way of dynamic range. You might think... why not just use upstream for everything? Upstream has it's limits so at some point it's no longer useful to attempt upstream amplification but you can convert to downstream and keep amplifying.

Upstream is amplification that happens before A2D conversion (information is still analog) and downstream is amplification which happens after A2D conversion (information is digital).

So far as I'm aware, no vendor discloses how much upstream or downstream amplification is applied, but via testing it's possible to back into the numbers. In general, most entry-level Canon cameras will apply "upstream" amplification up to around ISO 800 without resorting to "downstream" amplification. Some mid-range and high end models will take it as far as ISO 1600 or even ISO 3200. Usually it's not actually an even ISO level and they may actually even provide a slight mix of both.

But for this reason, many Canon models actually have higher dynamic range than Nikon models IF you're shooting at high ISO values. When you do downstream amplification you lose 1 stop of DR for every stop of ISO boost (that's by the math, it's not a matter of opinion).

Also, while I see the "dynamic range" thing frequently being cited by Nikon shooters... there's a lot more to consider about a camera than just dynamic range (and usually dynamic range isn't a problem for any camera).

The claim about dynamic range usually comes from results of DxO labs testing. I don't hold the opinions of DxO in very high regard as I doubt their understanding of science and they use methods that they do not disclose. It's their secret sauce. But it flies in the face of other's who do testing and end up getting different results. If "my" science is correct then I should be able to explain my hypothesis, testing methods, and results and then "you" should be able to repeat those tests and get the same results. If we are getting very different results, it means one of us is doing something wrong or my hypothesis is wrong.

DxO doesn't explain their methods, they don't share their data, they don't show us their sample images. They just provide us with "scores". During an interview (probably a couple of years back) they were asked how they come up with these "scores" and they declined to share that publicly. They claimed that needed to be kept secret. However, the interviewer was shown some of the behind-the-scenes testing being performed and one thing he stumbled onto... was that DxO doesn't think it's fair to compare cameras if they have different sensor resolutions (but this is something you'd have to do since different camera models have different sensor resolutions). SO... DxO believes the "fair" thing to do is to resample the images to make them the same resolution.

That mans when they compared the Nikon D810 (36MP camera) to say, a Canon 5D III (a 22MP camera) or a 5D IV (a 30MP camera) DxO is resampling the Nikon images down to the lower resolution... but they are not resampling the Canon images!

You CAN NOT RESAMPLE an image without cleaning up much of the noise; that's mathematically not possible. The whole point of "resampling" is that you're averaging groups of pixels at high resolution to determine what the pixel value should be of the combined group when displayed at a lower resolution. Of course it averages out and substantially reduces noise. This skews the noise scores in Nikon's favor.

So the interviewer points this out to DxO - who as I recall from the article, seemed to confess that this could be an issue but how could they otherwise make a comparison since the cameras don't have identical resolution. But the interviewer's issue was that while it's one thing to do this... it's quite another to not disclose it to the readers. That's unethical.

It used to drive me nuts that I'd see the "scores" from DxO that flew in the face of reports that I could find on DPreview or PopPhoto. But DPreview and PopPhoto both showed their sample image data so you could get a real idea of what to expect and DxO doesn't do that. In science, if you don't disclose your testing methods and show your work and your data, then what you are doing isn't really science at all. It must be possible to allow others to vet your work ... either to confirm your accuracy and your findings... or to find flaws in your methods that contradict your accuracy. Otherwise your results can't really be considered reliable.

Scores are worthless. It's sort of like attending a class with a teacher who won't divulge the how they come up with your final grade.... only find out that how you combed your hair and whether you tied your shoe-laces was scored very highly and whether you answered questions correctly on tests contributed less to your final grade then you may have thought.

A while after Canon released the 1DX Mark II, a ran across an article pointing out some early testing. A lab showed that the DR on that camera was so good that it beat all Nikon models they'd ever tested and come in 2nd to one particular Sony model camera (and was close that they reported it was nearly a tie). Meanwhile... DxO's testing showed the camera's DR was considerably worse. <sigh>

THAT is my beef with DxO labs. They seem to suffer from delusions of adequacy.


Meanwhile... back to those lenses.

The lens lineup from Canon generally outperforms the lens lineup from Nikon... while that's an accurate "generalization", the challenge is that generalizations aren't helpful because YOU don't plan to buy the entire lens lineup... you only intend to buy perhaps a few lenses. What would be more helpful is to understand the differences in the lenses you actually care about.

Canon's 50mm "nifty fifty" is more affordable than Nikon's "nifty fifty". If you're on a budget, that's helpful to know.

Canon recently reworked their entry-level consumer lenses with several features that make them huge upgrades over previous lenses and in some cases nearly on par with Canon's "L" series glass.

Two very popular Canon lenses are the Canon EF 24-70mm f/2.8L and the Canon EF 70-200mm f/2.8L IS - these two lenses are often paired by full-frame shooters. Nikon has lenses in this same category but Nikon has a THIRD lens... the 12-24mm and Canon didn't offer a lens in that range (Nikon owners often refer to it as the "holy trinity" and would point out that Canon is missing the third lens.). Canon has a 16-35mm f/2.8 (Nikon has one too, but it's an f/4 lens). Nikon shooters usually don't call attention to this. Canon now offers an 11-24mm f/4 which to cover the ultra-wide range. BTW, Nikon's 12-24 is an f/2.8 (Canon's is an f/4).

Canon's 16-35mm f/2.8 is basically unrivaled (nobody makes a better lens in that range). They also make an f/4 version which is considerably less expensive. Technically not quite a sharp... but very close.

Canon's 24-70 f/2.8 Mk II is also basically the best 24-70 in the industry.

Canon's 70-200 f/2.8 Mk II also beats Nikon's lens -- not just in areas such as sharpness and chromatic aberration (Canon seems to be around 30% sharper and suffers from about half the chromatic aberration issues as Nikon's glass) but Nikon's 70-200 is a VERY HEAVY breather. Zoom to 200mm but focus on a close subject (e.g. for a portrait shot) and Nikon's glass is only offering something near 140mm focal length (even though the zoom ring would say it's 200). All lenses "breath" (focal length shifts as you adjust focus) but when you do this with the Canon (which also breathes), the focal length drops to about 190mm (So Canon's glass is within 5% of what you'd expect).

BTW, earlier I mentioned DxO resampling images when comparing things like ISO performance & noise. But DxO apparently does _not_ resample IF they are comparing lenses. So when a Nikon lens which is technically less sharp than a competing Canon lens are compared, DxO will use a camera like a D810 (with high resolution) and show a lens "score" that is higher. Swap camera bodies to something like a Canon 5Dsr (50mp camera) and things trade places very quickly. But even when you get to bodies that aren't so far apart but give Nikon the edge (like a D810 at 36mp vs. a Canon 5D IV at 30mp) the Canon camera+lens combo outperforms the Nikon camera+lens combo (the extra 6mp resolution isn't enough to save the Nikon because the glass isn't as good).

This is why I caution people from making a comparison based on one thing.

Yes, most (but not all) Nikon bodies have more dynamic range than most Canon bodies if you shoot exclusively at base ISO. Come off base ISO and things aren't so clear cut.

This is NOT to suggest that you shouldn't buy a Nikon. Don't make a camera purchase decision for an entry level camera body based on the performance of high-end camera body because that's not the camera body you are going to get.

In general, the only way you'll see some of these differences is if you're pixel peeping (nobody else but you is going to notice these differences). You're also unlikely to use the full resolution that your camera offers. Sure you'll take the shot at max resolution... but when you display it on the web, you'll resample the image down to a more web-friendly size. At that point, most of these differences won't matter.

It's hard to buy a "bad" camera these days... from almost any vendor. My point in responding to this thread is to call attention to the fact that some of these generalizations used to justify one model vs. another are tending to over-simplify things. You don't get the full information and you don't realize that the generalization isn't actually always true.

No camera outperforms in every area. It's always a game of trade-offs.

 

[Derrel]

Yeah...Canon users are always trying to negate the company's second-tier sensor performance. Meanwhile, Fuji digital medium format, Hasselblad medium format, Pentax's entire line of smaller d-slr and medium format cameras--all these companies have elected to source their sensors from Sony, and of course Nikon has mightated to sensors made by Sony on steppers designed and built by Nikon itself. And of course, Sony is using its own sensors in its brand of cameras too.

The worse dynamic range, the narrower range of colors, the lworse noise performance at HIGH ISO, and the MUCH worse shadow recovery that Canon sensors are plagued with is something Canon fans hate to read about or hear others made aware of. it's become an ugly little fact that Sony-sensored camwera makers and users have become aware of since 2009.

Some people love to say that Canon and Nikon leapfrog one another. That might have been true, on a decade-by-decade basis in the 70's,80's,and 1990's in the film era, but Canon has not leapfrogged ANYBODY in years...Since 2009 when SONY invented EXMOR sensor technology, Canon was left in the dust.

But yeah...Canon has some lenses. Thery even have a handful of lenses that other makers do not have. But they also LACK some lenses that other makers have in their lineup. Where is Canon's 105mm f/1.4? Where is Canon's 300mm f/4 phase fresnel lens? Where is Canon's affordable and sharp 200-500mm f/5.6 zoom? And so on, and so on. Canon has nothing really on Nikon, lens-wise. Lenses are mostly a wash between Canon and Nikon, so what really counts is the part of the camera that MAKES the images: the sensor, and the associated signal-processing software.

Good writing and careful wording goes a long way toward minimizing Caon's shortcomings against Sony sensors and the many brands and models of cameras that use Sony sensors. But a quick trip to Dxo Mark, or to dPreview will show the facts: a Sony sensor can often be deliberately UNDER-exposed by five full EV, and the resulting almost-black raw file can be recovered 100% fully in software. Meanwhile, a Canon file treated the same way will be filled with hideous pattern banding and ugly,blotchy,horrible noise.

Yes, it's hard to buy a bad camera these days. But it is easy to buy a GREAT camera as well. If good is the goal, Canon has some good cameras, for sure. But Nikon, Pentax, Sony, Fuji, and Hasselblad have some great cameras. It's like in auto racing, where the experts often say, "Ehhhh...the engine...who needs a great engine! A good one is plenty." Oh wait, nobody tries to minimize the importance of the engine in race cars, but we have people talking about the windshield ...AKS, the glass quality....

A second-rate sensor is still a second-rate sensor. When Canon HAD the lead over others, all we heard were Canon fanboys touting their company's sensor superiority. Well, it's a different era now, and Canon is decidedly BEHIND, and has been since 2009. The sensor in the 5D Mark III is second-rate compared to the full-frame sensors made by Sony in 24-MP, 36-MP, and 42-MP sizes. Period.

Canon 5D Mark III:

Canon EOS 5D Mark III Review

Nikon D610
Nikon D610 Review

NO Contest, Period. the SOny-made sensor blows the Canon sensor out of the water.

 

[table1349]

Pigpen313 The best camera is the one you like and will use. Go to a camera store, get your hands on the bodies you are interested in. See which one(s) you like. Check out the various systems. See if they have the lenses and accessories you will want in the future and make you decision from there.

Any one on your list will do what you want. It is the photographer that takes the shots, the camera is just the tool. To be able to take the kinds of shots you want you need to understand photography. Once you buy a camera read the Owners manual more than once. Then read this: Digital Photography Tutorials All of it. As you do practice the various tutorials with your camera in hand.

 

[beagle100]

I woukld go fd prices. which coulkd help with creative, high- or low-angle shots.

Lenses for thotentially sharper images

This is not entirely true. It's an area that is often over-simplified. The specific point about lenses (the right lenses can give you potentially sharper images" is specifically generally not true. It's one of the reasons why experienced photographer who are aware of the DR differences still choose Canon... because of the lens differences.)

I can show you images from a Nikon D810 that look garbage next to images taken by a Canon 5D IV. Even though... the Nikon D810 allegedly has better dynamic range (it's a very subtle difference.)

Why?

We often tell beginner's that the ISO adjustment allows you to adjust the sensitivity of the sensor ... as if it makes the sensor "more sensitive" to light and thus can take a shorter exposure. While that explanation gets the point across, it's inaccurate to the point of being unhelpful. What the ISO really does is determine how much "gain" or "amplification" will be applied to the amount of light that the sensor collected at base ISO.

Why is this a big deal? Because cameras have two major ways to apply amplification but this varies by model. Nikon cameras seem to largely only amplify image data digitally. In other words, pretend you took the shot at base ISO, saved the file to the card, then went back and "amplified" all the values in the digital file by doubling them all (or increasing them proportionally based on the amount of amplification you need.)

Since an image may have had values for each pixel / photo-site ranging from very dark (near 0) all the way up to something near the max that the camera can capture (for 8-bit color it's 0-255 per channel, for 14-bit RAW that's 0-16,383 per channel) then amplifying those already bright pixels will cause them to clip. You can attempt to protect the bright data by compressing the histogram, but that's still a sacrifice to the dynamic range of the camera. This digital amplification is referred to as "downstream" amplification.

Canon cameras (nearly all models) apply both "upstream" and "downstream" amplification. "Upstream" amplification is handled by the sensor itself while the data is still analog (before it got through the analog to digital conversion process). Because the image data is not yet digital, it's not subject to loss of dynamic range. As you amplify data by "upstream" method, you give up very little in the way of dynamic range. You might think... why not just use upstream for everything? Upstream has it's limits so at some point it's no longer useful to attempt upstream amplification but you can convert to downstream and keep amplifying.

Upstream is amplification that happens before A2D conversion (information is still analog) and downstream is amplification which happens after A2D conversion (information is digital).

So far as I'm aware, no vendor discloses how much upstream or downstream amplification is applied, but via testing it's possible to back into the numbers. In general, most entry-level Canon cameras will apply "upstream" amplification up to around ISO 800 without resorting to "downstream" amplification. Some mid-range and high end models will take it as far as ISO 1600 or even ISO 3200. Usually it's not actually an even ISO level and they may actually even provide a slight mix of both.

But for this reason, many Canon models actually have higher dynamic range than Nikon models IF you're shooting at high ISO values. When you do downstream amplification you lose 1 stop of DR for every stop of ISO boost (that's by the math, it's not a matter of opinion).

Also, while I see the "dynamic range" thing frequently being cited by Nikon shooters... there's a lot more to consider about a camera than just dynamic range (and usually dynamic range isn't a problem for any camera).

The claim about dynamic range usually comes from results of DxO labs testing. I don't hold the opinions of DxO in very high regard as I doubt their understanding of science and they use methods that they do not disclose. It's their secret sauce. But it flies in the face of other's who do testing and end up getting different results. If "my" science is correct then I should be able to explain my hypothesis, testing methods, and results and then "you" should be able to repeat those tests and get the same results. If we are getting very different results, it means one of us is doing something wrong or my hypothesis is wrong.

DxO doesn't explain their methods, they don't share their data, they don't show us their sample images. They just provide us with "scores". During an interview (probably a couple of years back) they were asked how they come up with these "scores" and they declined to share that publicly. They claimed that needed to be kept secret. However, the interviewer was shown some of the behind-the-scenes testing being performed and one thing he stumbled onto... was that DxO doesn't think it's fair to compare cameras if they have different sensor resolutions (but this is something you'd have to do since different camera models have different sensor resolutions). SO... DxO believes the "fair" thing to do is to resample the images to make them the same resolution.

That mans when they compared the Nikon D810 (36MP camera) to say, a Canon 5D III (a 22MP camera) or a 5D IV (a 30MP camera) DxO is resampling the Nikon images down to the lower resolution... but they are not resampling the Canon images!

You CAN NOT RESAMPLE an image without cleaning up much of the noise; that's mathematically not possible. The whole point of "resampling" is that you're averaging groups of pixels at high resolution to determine what the pixel value should be of the combined group when displayed at a lower resolution. Of course it averages out and substantially reduces noise. This skews the noise scores in Nikon's favor.

So the interviewer points this out to DxO - who as I recall from the article, seemed to confess that this could be an issue but how could they otherwise make a comparison since the cameras don't have identical resolution. But the interviewer's issue was that while it's one thing to do this... it's quite another to not disclose it to the readers. That's unethical.

It used to drive me nuts that I'd see the "scores" from DxO that flew in the face of reports that I could find on DPreview or PopPhoto. But DPreview and PopPhoto both showed their sample image data so you could get a real idea of what to expect and DxO doesn't do that. In science, if you don't disclose your testing methods and show your work and your data, then what you are doing isn't really science at all. It must be possible to allow others to vet your work ... either to confirm your accuracy and your findings... or to find flaws in your methods that contradict your accuracy. Otherwise your results can't really be considered reliable.

Scores are worthless. It's sort of like attending a class with a teacher who won't divulge the how they come up with your final grade.... only find out that how you combed your hair and whether you tied your shoe-laces was scored very highly and whether you answered questions correctly on tests contributed less to your final grade then you may have thought.

A while after Canon released the 1DX Mark II, a ran across an article pointing out some early testing. A lab showed that the DR on that camera was so good that it beat all Nikon models they'd ever tested and come in 2nd to one particular Sony model camera (and was close that they reported it was nearly a tie). Meanwhile... DxO's testing showed the camera's DR was considerably worse. <sigh>

THAT is my beef with DxO labs. They seem to suffer from delusions of adequacy.


Meanwhile... back to those lenses.

The lens lineup from Canon generally outperforms the lens lineup from Nikon... while that's an accurate "generalization", the challenge is that generalizations aren't helpful because YOU don't plan to buy the entire lens lineup... you only intend to buy perhaps a few lenses. What would be more helpful is to understand the differences in the lenses you actually care about.

Canon's 50mm "nifty fifty" is more affordable than Nikon's "nifty fifty". If you're on a budget, that's helpful to know.

Canon recently reworked their entry-level consumer lenses with several features that make them huge upgrades over previous lenses and in some cases nearly on par with Canon's "L" series glass.

Two very popular Canon lenses are the Canon EF 24-70mm f/2.8L and the Canon EF 70-200mm f/2.8L IS - these two lenses are often paired by full-frame shooters. Nikon has lenses in this same category but Nikon has a THIRD lens... the 12-24mm and Canon didn't offer a lens in that range (Nikon owners often refer to it as the "holy trinity" and would point out that Canon is missing the third lens.). Canon has a 16-35mm f/2.8 (Nikon has one too, but it's an f/4 lens). Nikon shooters usually don't call attention to this. Canon now offers an 11-24mm f/4 which to cover the ultra-wide range. BTW, Nikon's 12-24 is an f/2.8 (Canon's is an f/4).

Canon's 16-35mm f/2.8 is basically unrivaled (nobody makes a better lens in that range). They also make an f/4 version which is considerably less expensive. Technically not quite a sharp... but very close.

Canon's 24-70 f/2.8 Mk II is also basically the best 24-70 in the industry.

Canon's 70-200 f/2.8 Mk II also beats Nikon's lens -- not just in areas such as sharpness and chromatic aberration (Canon seems to be around 30% sharper and suffers from about half the chromatic aberration issues as Nikon's glass) but Nikon's 70-200 is a VERY HEAVY breather. Zoom to 200mm but focus on a close subject (e.g. for a portrait shot) and Nikon's glass is only offering something near 140mm focal length (even though the zoom ring would say it's 200). All lenses "breath" (focal length shifts as you adjust focus) but when you do this with the Canon (which also breathes), the focal length drops to about 190mm (So Canon's glass is within 5% of what you'd expect).

BTW, earlier I mentioned DxO resampling images when comparing things like ISO performance & noise. But DxO apparently does _not_ resample IF they are comparing lenses. So when a Nikon lens which is technically less sharp than a competing Canon lens are compared, DxO will use a camera like a D810 (with high resolution) and show a lens "score" that is higher. Swap camera bodies to something like a Canon 5Dsr (50mp camera) and things trade places very quickly. But even when you get to bodies that aren't so far apart but give Nikon the edge (like a D810 at 36mp vs. a Canon 5D IV at 30mp) the Canon camera+lens combo outperforms the Nikon camera+lens combo (the extra 6mp resolution isn't enough to save the Nikon because the glass isn't as good).

This is why I caution people from making a comparison based on one thing.

Yes, most (but not all) Nikon bodies have more dynamic range than most Canon bodies if you shoot exclusively at base ISO. Come off base ISO and things aren't so clear cut.

This is NOT to suggest that you shouldn't buy a Nikon. Don't make a camera purchase decision for an entry level camera body based on the performance of high-end camera body because that's not the camera body you are going to get.

In general, the only way you'll see some of these differences is if you're pixel peeping (nobody else but you is going to notice these differences). You're also unlikely to use the full resolution that your camera offers. Sure you'll take the shot at max resolution... but when you display it on the web, you'll resample the image down to a more web-friendly size. At that point, most of these differences won't matter.
It's hard to buy a "bad" camera these days... from almost any vendor. My point in responding to this thread is to call attention to the fact that some of these generalizations used to justify one model vs. another are tending to over-simplify things. You don't get the full information and you don't realize that the generalization isn't actually always true.
No camera outperforms in every area. It's always a game of trade-offs.

true, join a local photography club and go on monthly photowalks where shooting the same subjects the same time and it's easy to see the pics from Canon cameras are usually better. Of course having a 600 f/4 could make any wildlife photo look better!
And most pros shoot with Canon gear so it's easier to upgrade.

But now Canon is seriously getting into the mirrorless game, watch out Sony !

www.flickr.com/photos/mmirrorless[/QUOTE]

 

[table1349]

I woukld go fd prices. which coulkd help with creative, high- or low-angle shots.

Lenses for thotentially sharper images

This is not entirely true. It's an area that is often over-simplified. The specific point about lenses (the right lenses can give you potentially sharper images" is specifically generally not true. It's one of the reasons why experienced photographer who are aware of the DR differences still choose Canon... because of the lens differences.)

I can show you images from a Nikon D810 that look garbage next to images taken by a Canon 5D IV. Even though... the Nikon D810 allegedly has better dynamic range (it's a very subtle difference.)

Why?

We often tell beginner's that the ISO adjustment allows you to adjust the sensitivity of the sensor ... as if it makes the sensor "more sensitive" to light and thus can take a shorter exposure. While that explanation gets the point across, it's inaccurate to the point of being unhelpful. What the ISO really does is determine how much "gain" or "amplification" will be applied to the amount of light that the sensor collected at base ISO.

Why is this a big deal? Because cameras have two major ways to apply amplification but this varies by model. Nikon cameras seem to largely only amplify image data digitally. In other words, pretend you took the shot at base ISO, saved the file to the card, then went back and "amplified" all the values in the digital file by doubling them all (or increasing them proportionally based on the amount of amplification you need.)

Since an image may have had values for each pixel / photo-site ranging from very dark (near 0) all the way up to something near the max that the camera can capture (for 8-bit color it's 0-255 per channel, for 14-bit RAW that's 0-16,383 per channel) then amplifying those already bright pixels will cause them to clip. You can attempt to protect the bright data by compressing the histogram, but that's still a sacrifice to the dynamic range of the camera. This digital amplification is referred to as "downstream" amplification.

Canon cameras (nearly all models) apply both "upstream" and "downstream" amplification. "Upstream" amplification is handled by the sensor itself while the data is still analog (before it got through the analog to digital conversion process). Because the image data is not yet digital, it's not subject to loss of dynamic range. As you amplify data by "upstream" method, you give up very little in the way of dynamic range. You might think... why not just use upstream for everything? Upstream has it's limits so at some point it's no longer useful to attempt upstream amplification but you can convert to downstream and keep amplifying.

Upstream is amplification that happens before A2D conversion (information is still analog) and downstream is amplification which happens after A2D conversion (information is digital).

So far as I'm aware, no vendor discloses how much upstream or downstream amplification is applied, but via testing it's possible to back into the numbers. In general, most entry-level Canon cameras will apply "upstream" amplification up to around ISO 800 without resorting to "downstream" amplification. Some mid-range and high end models will take it as far as ISO 1600 or even ISO 3200. Usually it's not actually an even ISO level and they may actually even provide a slight mix of both.

But for this reason, many Canon models actually have higher dynamic range than Nikon models IF you're shooting at high ISO values. When you do downstream amplification you lose 1 stop of DR for every stop of ISO boost (that's by the math, it's not a matter of opinion).

Also, while I see the "dynamic range" thing frequently being cited by Nikon shooters... there's a lot more to consider about a camera than just dynamic range (and usually dynamic range isn't a problem for any camera).

The claim about dynamic range usually comes from results of DxO labs testing. I don't hold the opinions of DxO in very high regard as I doubt their understanding of science and they use methods that they do not disclose. It's their secret sauce. But it flies in the face of other's who do testing and end up getting different results. If "my" science is correct then I should be able to explain my hypothesis, testing methods, and results and then "you" should be able to repeat those tests and get the same results. If we are getting very different results, it means one of us is doing something wrong or my hypothesis is wrong.

DxO doesn't explain their methods, they don't share their data, they don't show us their sample images. They just provide us with "scores". During an interview (probably a couple of years back) they were asked how they come up with these "scores" and they declined to share that publicly. They claimed that needed to be kept secret. However, the interviewer was shown some of the behind-the-scenes testing being performed and one thing he stumbled onto... was that DxO doesn't think it's fair to compare cameras if they have different sensor resolutions (but this is something you'd have to do since different camera models have different sensor resolutions). SO... DxO believes the "fair" thing to do is to resample the images to make them the same resolution.

That mans when they compared the Nikon D810 (36MP camera) to say, a Canon 5D III (a 22MP camera) or a 5D IV (a 30MP camera) DxO is resampling the Nikon images down to the lower resolution... but they are not resampling the Canon images!

You CAN NOT RESAMPLE an image without cleaning up much of the noise; that's mathematically not possible. The whole point of "resampling" is that you're averaging groups of pixels at high resolution to determine what the pixel value should be of the combined group when displayed at a lower resolution. Of course it averages out and substantially reduces noise. This skews the noise scores in Nikon's favor.

So the interviewer points this out to DxO - who as I recall from the article, seemed to confess that this could be an issue but how could they otherwise make a comparison since the cameras don't have identical resolution. But the interviewer's issue was that while it's one thing to do this... it's quite another to not disclose it to the readers. That's unethical.

It used to drive me nuts that I'd see the "scores" from DxO that flew in the face of reports that I could find on DPreview or PopPhoto. But DPreview and PopPhoto both showed their sample image data so you could get a real idea of what to expect and DxO doesn't do that. In science, if you don't disclose your testing methods and show your work and your data, then what you are doing isn't really science at all. It must be possible to allow others to vet your work ... either to confirm your accuracy and your findings... or to find flaws in your methods that contradict your accuracy. Otherwise your results can't really be considered reliable.

Scores are worthless. It's sort of like attending a class with a teacher who won't divulge the how they come up with your final grade.... only find out that how you combed your hair and whether you tied your shoe-laces was scored very highly and whether you answered questions correctly on tests contributed less to your final grade then you may have thought.

A while after Canon released the 1DX Mark II, a ran across an article pointing out some early testing. A lab showed that the DR on that camera was so good that it beat all Nikon models they'd ever tested and come in 2nd to one particular Sony model camera (and was close that they reported it was nearly a tie). Meanwhile... DxO's testing showed the camera's DR was considerably worse. <sigh>

THAT is my beef with DxO labs. They seem to suffer from delusions of adequacy.


Meanwhile... back to those lenses.

The lens lineup from Canon generally outperforms the lens lineup from Nikon... while that's an accurate "generalization", the challenge is that generalizations aren't helpful because YOU don't plan to buy the entire lens lineup... you only intend to buy perhaps a few lenses. What would be more helpful is to understand the differences in the lenses you actually care about.

Canon's 50mm "nifty fifty" is more affordable than Nikon's "nifty fifty". If you're on a budget, that's helpful to know.

Canon recently reworked their entry-level consumer lenses with several features that make them huge upgrades over previous lenses and in some cases nearly on par with Canon's "L" series glass.

Two very popular Canon lenses are the Canon EF 24-70mm f/2.8L and the Canon EF 70-200mm f/2.8L IS - these two lenses are often paired by full-frame shooters. Nikon has lenses in this same category but Nikon has a THIRD lens... the 12-24mm and Canon didn't offer a lens in that range (Nikon owners often refer to it as the "holy trinity" and would point out that Canon is missing the third lens.). Canon has a 16-35mm f/2.8 (Nikon has one too, but it's an f/4 lens). Nikon shooters usually don't call attention to this. Canon now offers an 11-24mm f/4 which to cover the ultra-wide range. BTW, Nikon's 12-24 is an f/2.8 (Canon's is an f/4).

Canon's 16-35mm f/2.8 is basically unrivaled (nobody makes a better lens in that range). They also make an f/4 version which is considerably less expensive. Technically not quite a sharp... but very close.

Canon's 24-70 f/2.8 Mk II is also basically the best 24-70 in the industry.

Canon's 70-200 f/2.8 Mk II also beats Nikon's lens -- not just in areas such as sharpness and chromatic aberration (Canon seems to be around 30% sharper and suffers from about half the chromatic aberration issues as Nikon's glass) but Nikon's 70-200 is a VERY HEAVY breather. Zoom to 200mm but focus on a close subject (e.g. for a portrait shot) and Nikon's glass is only offering something near 140mm focal length (even though the zoom ring would say it's 200). All lenses "breath" (focal length shifts as you adjust focus) but when you do this with the Canon (which also breathes), the focal length drops to about 190mm (So Canon's glass is within 5% of what you'd expect).

BTW, earlier I mentioned DxO resampling images when comparing things like ISO performance & noise. But DxO apparently does _not_ resample IF they are comparing lenses. So when a Nikon lens which is technically less sharp than a competing Canon lens are compared, DxO will use a camera like a D810 (with high resolution) and show a lens "score" that is higher. Swap camera bodies to something like a Canon 5Dsr (50mp camera) and things trade places very quickly. But even when you get to bodies that aren't so far apart but give Nikon the edge (like a D810 at 36mp vs. a Canon 5D IV at 30mp) the Canon camera+lens combo outperforms the Nikon camera+lens combo (the extra 6mp resolution isn't enough to save the Nikon because the glass isn't as good).

This is why I caution people from making a comparison based on one thing.

Yes, most (but not all) Nikon bodies have more dynamic range than most Canon bodies if you shoot exclusively at base ISO. Come off base ISO and things aren't so clear cut.

This is NOT to suggest that you shouldn't buy a Nikon. Don't make a camera purchase decision for an entry level camera body based on the performance of high-end camera body because that's not the camera body you are going to get.

In general, the only way you'll see some of these differences is if you're pixel peeping (nobody else but you is going to notice these differences). You're also unlikely to use the full resolution that your camera offers. Sure you'll take the shot at max resolution... but when you display it on the web, you'll resample the image down to a more web-friendly size. At that point, most of these differences won't matter.
It's hard to buy a "bad" camera these days... from almost any vendor. My point in responding to this thread is to call attention to the fact that some of these generalizations used to justify one model vs. another are tending to over-simplify things. You don't get the full information and you don't realize that the generalization isn't actually always true.
No camera outperforms in every area. It's always a game of trade-offs.

true, join a local photography club and go on monthly photowalks where shooting the same subjects the same time and it's easy to see the pics from Canon cameras are usually better. Of course having a 600 f/4 could make any wildlife photo look better!
And most pros shoot with Canon gear so it's easier to upgrade.

But now Canon is seriously getting into the mirrorless game, watch out Sony !

www.flickr.com/photos/mmirrorless

 

[TCampbell]

The worse dynamic range, the narrower range of colors

I'm not sure if you understand how "color" works in cameras. There's basically no such thing as a "color" sensor... only monochrome. But the monochrome sensor has a color filter array in front of it so that different photo sites only detect different wavelength ranges. Hand me a completely monochrome camera and I can build a color image for you (true color - not colorized like Turner movies.)

What we perceive as "color" is literally in our heads, and everything you perceive w.r.t. color in a camera is a result of post process digital manipulation. You can get the camera to do this if you want. If you have a Canon camera, go play with "picture styles". You can tell it to saturate, de-saturate, manipulate specific colors, do color substitution, etc.

I shoot RAW, which means that the color in my images is created by Adobe Lightroom's RAW demosaicing process. If you don't like it, use a different demosaicing process.

But I should point out that when I take a shot with my Canon camera, white balance it, bring it in to my calibrated display, and then compare it to my color checker... the colors are bang-on accurate.

If you prefer saturated colors... Lightroom has an adjustment for that.

worse noise performance at HIGH ISO, and the MUCH worse shadow recovery that Canon sensors are plagued with is something Canon fans hate to read about or hear others made aware of.

And that's just it... I can go to DPReview or PopPhoto (now going out of business) and see examples that show the opposite of what you claim.

Here's another interesting tidbit... members of one of my astrophotography groups noticed that stars were missing in the RAW data from Nikon cameras. We were looking at Milky Way shots and areas that should have been loaded with stars actually appeared smoother than reality. And these were RAW images. It wasn't a focus problem. The cameras were cooking the RAWs.

Where is Canon's affordable and sharp 200-500mm f/5.6 zoom? And so on, and so on. Canon has nothing really on Nikon, lens-wise. Lenses are mostly a wash between Canon and Nikon, so what really counts is the part of the camera that MAKES the images: the sensor, and the associated signal-processing software.

Come on Derrel, who's in denial now? (All in fun, of course.)

Shall we mention the Nikon 49.7-142mm lens? I think that's the lens in your bag labeled "70-200".

As for a 200-500mm...the leader in that category... is actually (oddly enough) the Canon EF 100-400mm f/4.5-5.6 IS USM II which is so far ahead of the pack that even with a 1.4x teleconverter on the lens (turn it into a 140-560mm) it still outperforms all the other lenses. And this isn't just one lab's study... I see that come up over and over again by different people reviewing it.

I'm not saying you should go buy that lens. Sigma and Tamron have offerings in the category as well (Sigma has a two, one being a bit cheaper than the Nikon and the other being quite a bit more expensive.) So if you want the price point you can certainly get it. But the top performer in the category is the Canon lens.

Good writing and careful wording goes a long way toward minimizing Caon's shortcomings against Sony sensors and the many brands and models of cameras that use Sony sensors. But a quick trip to Dxo Mark, or to dPreview will show the facts: a Sony sensor can often be deliberately UNDER-exposed by five full EV, and the resulting almost-black raw file can be recovered 100% fully in software.

This is because Nikon cameras basically only have one ISO. They have no upstream amplification. But what that means is that as you crank up the ISO or try to recover in post, you're losing dynamic range. That's math. There's no getting around this.

I pulled up the DPreview comparison of a D810 vs. a 5D IV and when you compare shots taken "at ISO" (not underexposed and pulled up) such that the Canon camera engaged it's upstream amplification AND downstream amplification, the Canon images are better than the Nikon. And the Canon camera doesn't start giving up dynamic range until you start to pass ISO 3200. The Nikon loses a full stop of DR as soon as you go to ISO 200.

That's also math. It's not a matter of opinion.

Here... have a look at the "hideous pattern banding and ugly,blotchy,horrible noise" in that Canon photo on the left next to that obviously better Nikon photo, right?

So much for that sensor that just blows away anything Canon makes. Of course, this would involve actually looking at an image sample instead of reading DxO "scores".

I'm not trying to bash Nikon cameras. They're great cameras. I'm trying to put an end to all the FUD about how every Nikon is better than every Canon in every way.

Oh BTW... now that you've seen the above actual sample of data comparing the Nikon D810 and 5D IV... go have a peek at what DxO has to say when they compare the ISO performance "scores" for these two cameras.

This is what I mean about DxO has "delusions of adequacy" ... they won't show us their "data" only their "scores". But when I see actual data provided by someone else... it contradicts DxO.

 

[Derrel]

Love the ISO 25,600 shot....why not reply to the actual substance of my post and the ctualk example I listed?

Hiding behind ISO 25,600... Funny Stuff.

You don't think I understand how a digital sensor works after 17 years of shooting both high-end Nikons and Canon 5D and 20D d-slr cameras? I love the carefully crafted smear technique there! The carefully slected "example" of uselessly-high ISO images from two cameras...kind of a useless example you selected to try and prove a point. When was the last time ANYBODY tried to prove how good a camera is by shooting at ISO 25,600?

Again...try what I said....under-expose a Canon 5D Mark III or IV or a Canon 70D or 6D frame that was shot at a highlight-protecting exposure of Minus 4 to Minus 5 EV for the overall scene (perhaps Minus 8 to 10 EV for the shadow values) so the RAW file that comes out of the camera is almost BLACK....do the same thing with a Nikon D7100 or D7200 frame, or a Nikon D810 frame.

The Canon frame will be a hideous MESS of noise....the SONY-sensored frames, shot by a Fuji or Pentax medium format, or a Pentax small-sensor d-slr, or by a SONY-brand camera, or a NIKON-branded camera...those frames will "lift" up to a nearly normal, perfectly usable frame.

The outdated Canon sensor tech? UGLY patterned noise.

SONY sensor technology is far better than what Canon has developed. And that is why SONY cameras, Pentax cameras, Hasselblad cameras, and Nikon camera are all using Sony-made sensors.

You want to tell us that one company, Canon, has better sensor technology than Fuji,Pentax,Sony,Hasselblad, and Nikon all have elected to use?

Dynamic Range: Benchmark Performance: Nikon D810 review

Nikon D810, as good as Medium Format Digitiual in terms of DR. Canon 5D Mark III...looks like crap....

Headline: "D810 Crushes the 5DS R"

HERE is exactly what I mentioned above: take a shot that is under-exposed, to protect bright highlights, then LIFT the shot by 5 EV in post. Here is the Sony-sensor vs Canon-sensor comparison that actually counts....not some B.S shot done at ISO 25,600 like you posted above.

Wow..look at the NOISE Canon can create, even at ISO 100, in the under-exposed areas!

Impressive.Not.

ISO 100. Not at ISO 25,600...but noise at I-S-O one-hundred.

 

[beagle100]

The worse dynamic range, the narrower range of colors

I'm not sure if you understand how "color" works in cameras. There's basically no such thing as a "color" sensor... only monochrome. But the monochrome sensor has a color filter array in front of it so that different photo sites only detect different wavelength ranges. Hand me a completely monochrome camera and I can build a color image for you (true color - not colorized like Turner movies.)

What we perceive as "color" is literally in our heads, and everything you perceive w.r.t. color in a camera is a result of post process digital manipulation. You can get the camera to do this if you want. If you have a Canon camera, go play with "picture styles". You can tell it to saturate, de-saturate, manipulate specific colors, do color substitution, etc.

I shoot RAW, which means that the color in my images is created by Adobe Lightroom's RAW demosaicing process. If you don't like it, use a different demosaicing process.

But I should point out that when I take a shot with my Canon camera, white balance it, bring it in to my calibrated display, and then compare it to my color checker... the colors are bang-on accurate.

If you prefer saturated colors... Lightroom has an adjustment for that.

worse noise performance at HIGH ISO, and the MUCH worse shadow recovery that Canon sensors are plagued with is something Canon fans hate to read about or hear others made aware of.

And that's just it... I can go to DPReview or PopPhoto (now going out of business) and see examples that show the opposite of what you claim.

Here's another interesting tidbit... members of one of my astrophotography groups noticed that stars were missing in the RAW data from Nikon cameras. We were looking at Milky Way shots and areas that should have been loaded with stars actually appeared smoother than reality. And these were RAW images. It wasn't a focus problem. The cameras were cooking the RAWs.

Where is Canon's affordable and sharp 200-500mm f/5.6 zoom? And so on, and so on. Canon has nothing really on Nikon, lens-wise. Lenses are mostly a wash between Canon and Nikon, so what really counts is the part of the camera that MAKES the images: the sensor, and the associated signal-processing software.

Come on Derrel, who's in denial now? (All in fun, of course.)

Shall we mention the Nikon 49.7-142mm lens? I think that's the lens in your bag labeled "70-200".

As for a 200-500mm...the leader in that category... is actually (oddly enough) the Canon EF 100-400mm f/4.5-5.6 IS USM II which is so far ahead of the pack that even with a 1.4x teleconverter on the lens (turn it into a 140-560mm) it still outperforms all the other lenses. And this isn't just one lab's study... I see that come up over and over again by different people reviewing it.

I'm not saying you should go buy that lens. Sigma and Tamron have offerings in the category as well (Sigma has a two, one being a bit cheaper than the Nikon and the other being quite a bit more expensive.) So if you want the price point you can certainly get it. But the top performer in the category is the Canon lens.

Good writing and careful wording goes a long way toward minimizing Caon's shortcomings against Sony sensors and the many brands and models of cameras that use Sony sensors. But a quick trip to Dxo Mark, or to dPreview will show the facts: a Sony sensor can often be deliberately UNDER-exposed by five full EV, and the resulting almost-black raw file can be recovered 100% fully in software.

This is because Nikon cameras basically only have one ISO. They have no upstream amplification. But what that means is that as you crank up the ISO or try to recover in post, you're losing dynamic range. That's math. There's no getting around this.

I pulled up the DPreview comparison of a D810 vs. a 5D IV and when you compare shots taken "at ISO" (not underexposed and pulled up) such that the Canon camera engaged it's upstream amplification AND downstream amplification, the Canon images are better than the Nikon. And the Canon camera doesn't start giving up dynamic range until you start to pass ISO 3200. The Nikon loses a full stop of DR as soon as you go to ISO 200.

That's also math. It's not a matter of opinion.

Here... have a look at the "hideous pattern banding and ugly,blotchy,horrible noise" in that Canon photo on the left next to that obviously better Nikon photo, right?



So much for that sensor that just blows away anything Canon makes. Of course, this would involve actually looking at an image sample instead of reading DxO "scores".
I'm not trying to bash Nikon cameras. They're great cameras. I'm trying to put an end to all the FUD about how every Nikon is better than every Canon in every way.
Oh BTW... now that you've seen the above actual sample of data comparing the Nikon D810 and 5D IV... go have a peek at what DxO has to say when they compare the ISO performance "scores" for these two cameras.
This is what I mean about DxO has "delusions of adequacy" ... they won't show us their "data" only their "scores". But when I see actual data provided by someone else... it contradicts DxO.

yeah, it seems obvious but I would recommend anyone buying a new camera to bring an SD card to a store and check the differnent ca out. And join a local photo club where members

The worse dynamic range, the narrower range of colors

I'm not sure if you understand how "color" works in cameras. There's basically no such thing as a "color" sensor... only monochrome. But the monochrome sensor has a color filter array in front of it so that different photo sites only detect different wavelength ranges. Hand me a completely monochrome camera and I can build a color image for you (true color - not colorized like Turner movies.)

What we perceive as "color" is literally in our heads, and everything you perceive w.r.t. color in a camera is a result of post process digital manipulation. You can get the camera to do this if you want. If you have a Canon camera, go play with "picture styles". You can tell it to saturate, de-saturate, manipulate specific colors, do color substitution, etc.

I shoot RAW, which means that the color in my images is created by Adobe Lightroom's RAW demosaicing process. If you don't like it, use a different demosaicing process.

But I should point out that when I take a shot with my Canon camera, white balance it, bring it in to my calibrated display, and then compare it to my color checker... the colors are bang-on accurate.

If you prefer saturated colors... Lightroom has an adjustment for that.

worse noise performance at HIGH ISO, and the MUCH worse shadow recovery that Canon sensors are plagued with is something Canon fans hate to read about or hear others made aware of.

And that's just it... I can go to DPReview or PopPhoto (now going out of business) and see examples that show the opposite of what you claim.

Here's another interesting tidbit... members of one of my astrophotography groups noticed that stars were missing in the RAW data from Nikon cameras. We were looking at Milky Way shots and areas that should have been loaded with stars actually appeared smoother than reality. And these were RAW images. It wasn't a focus problem. The cameras were cooking the RAWs.

Where is Canon's affordable and sharp 200-500mm f/5.6 zoom? And so on, and so on. Canon has nothing really on Nikon, lens-wise. Lenses are mostly a wash between Canon and Nikon, so what really counts is the part of the camera that MAKES the images: the sensor, and the associated signal-processing software.

Come on Derrel, who's in denial now? (All in fun, of course.)

Shall we mention the Nikon 49.7-142mm lens? I think that's the lens in your bag labeled "70-200".

As for a 200-500mm...the leader in that category... is actually (oddly enough) the Canon EF 100-400mm f/4.5-5.6 IS USM II which is so far ahead of the pack that even with a 1.4x teleconverter on the lens (turn it into a 140-560mm) it still outperforms all the other lenses. And this isn't just one lab's study... I see that come up over and over again by different people reviewing it.

I'm not saying you should go buy that lens. Sigma and Tamron have offerings in the category as well (Sigma has a two, one being a bit cheaper than the Nikon and the other being quite a bit more expensive.) So if you want the price point you can certainly get it. But the top performer in the category is the Canon lens.

Good writing and careful wording goes a long way toward minimizing Caon's shortcomings against Sony sensors and the many brands and models of cameras that use Sony sensors. But a quick trip to Dxo Mark, or to dPreview will show the facts: a Sony sensor can often be deliberately UNDER-exposed by five full EV, and the resulting almost-black raw file can be recovered 100% fully in software.

This is because Nikon cameras basically only have one ISO. They have no upstream amplification. But what that means is that as you crank up the ISO or try to recover in post, you're losing dynamic range. That's math. There's no getting around this.

I pulled up the DPreview comparison of a D810 vs. a 5D IV and when you compare shots taken "at ISO" (not underexposed and pulled up) such that the Canon camera engaged it's upstream amplification AND downstream amplification, the Canon images are better than the Nikon. And the Canon camera doesn't start giving up dynamic range until you start to pass ISO 3200. The Nikon loses a full stop of DR as soon as you go to ISO 200.

That's also math. It's not a matter of opinion.

Here... have a look at the "hideous pattern banding and ugly,blotchy,horrible noise" in that Canon photo on the left next to that obviously better Nikon photo, right?



So much for that sensor that just blows away anything Canon makes. Of course, this would involve actually looking at an image sample instead of reading DxO "scores".
I'm not trying to bash Nikon cameras. They're great cameras. I'm trying to put an end to all the FUD about how every Nikon is better than every Canon in every way.

Oh BTW... now that you've seen the above actual sample of data comparing the Nikon D810 and 5D IV... go have a peek at what DxO has to say when they compare the ISO performance "scores" for these two cameras.
This is what I mean about DxO has "delusions of adequacy" ... they won't show us their "data" only their "scores". But when I see actual data provided by someone else... it contradicts DxO.

right, some of the Canon models "crush" the Nikon cameras for ISO "noise' but it depends on the model

true, Nikon owners may have "delusions of inadequacy" but what's funny on monthly photo walks in my local photography club they insist their photos look good even though you can see the difference in quality .... they blame the photo "degradation" on Flickr and Facebook or the lens .... !

www.flickr.com/photos/mmirrorless

 

[table1349]

 

[TCampbell]

Love the ISO 25,600 shot....why not reply to the actual substance of my post and the ctualk example I listed?

Hiding behind ISO 25,600... Funny Stuff.

You don't think I understand how a digital sensor works after 17 years of shooting both high-end Nikons and Canon 5D and 20D d-slr cameras?

Yes, that is precisely what I think. There is no reason a photographer would accidentally gain this knowledge without having to specifically go dig to learn it. You are unlikely to come by this information accidentally. In other words, this is an area where years of experience don't mean much. The fact that you keep driving home the ISO 100 thing... pretty much solidifies the case.

I got my first 35mm film SLR back in 1977 timeframe and I've been shooting ever since. Years later, I was literally the first person I knew of to own any digital camera (I was a computer geek and in those years of my career I was writing CAD/CAM graphics software ... so I've picked up a trick or two on how to manipulate digital data including both rasterized and vector graphics (and, as a software developer). When digital camera's came out... I had to have one. The digital tech wasn't very good back then so I continued to use film cameras for anything serious but the digital camera was a bit of a fun novelty (and a lot of people wanted to check it out because very few people had ever seen one.) It was a while before I noticed the technology getting good enough to where I thought I should maybe try to make the switch to DSLRs and stop using film. Regardless, I didn't have any low-level understanding of how the sensors worked (even though I did understand how the files and data were stored.)

But what really caused me to gain more of an understanding into specifically what the camera is doing under the covers... didn't happen until I started doing astrophotography.

In astrophotography, the image acquisition process is very heavily based on learning all about signal to noise rations (SNR) and how to boost the "S" and decrease the "N" (thus improving the ratio). Doing that required learning a bit about where the noise is actually coming from. There are lots of ideas about how to improve the SNR and lots of experiment has been done... is it better to shoot a greater quantity of short duration images vs. a shorter quantity of longer duration images. Is it better to shoot at high ISO for shorter duration vs. low ISO for higher duration. You can find piles and piles of articles on this stuff.

The other aspect of astrophotography is that you're trying to capture as much detail as you can on incredibly faint objects in the night sky... while simultaneously trying to not over-expose the stars (which would result in a loss of color). So you're trying to protect the dynamic range of the image.

It turns out if you dig hard enough... there are some answers. But as you dig, you'll find some of these answers require a somewhat low-level understanding about various sensors work.

Every camera (model) has some optimal ISO setting. This is the setting which maximizes signal, while minimizing noise, reduces the overall exposure times (and there are huge incentives to do that because exposure times can go so long that it can take many nights to collect enough data).

For my Canon 60Da, that magic value is ISO 800. For my 5D III it's ISO 1600... but really it's a bit above that so even ISO 3200 works. And for the 5D IV, it seems to be closer to ISO 3200.

For most Nikon models... it's ISO 100 (if you boost ISO to reduce exposure duration then you give up dynamic range.) Most Nikon's do actually start out with more DR so you could afford to sacrifice a bit. But when you boost it enough to compet with what you can do with some of these Canon models, the Canon cameras actually end up with lower noise and more DR. BTW, one notable exception was the D3100... that camera does seem to apply real upstream amplification (prior to the A2D conversion) and it looks like it makes it up to around ISO 1600 before it starts switching to downstream amplification. But the D3100 doesn't have as much DR as some of the other models if compared at ISO 100.

I love the carefully crafted smear technique there! The carefully slected "example" of uselessly-high ISO images from two cameras...kind of a useless example you selected to try and prove a point. When was the last time ANYBODY tried to prove how good a camera is by shooting at ISO 25,600?

If by "carefully crafted" you mean actually showing samples? So people can judge with their own two eyes? Rather than looking at DxO "scores"? ...guilty as charged.

Of course, I don't need to use 25,600... it shows up at 12,800, it shows at 6400, 3200, even 1600. It shows up whether I use a 5D IV as a comparison body or a 5D III as a comparison body.

Here... have a look ... here's ISO 3200 and 12,800 w/ Canon 5D III & Nikon D810. In both of these examples, the lowly 5D III you hate so much has less noise than the D810 (and the D810 is an impressive camera, I'm not bashing it.) I didn't bother to put the ISO 1600 test (the Canon appears to have less noise but the difference is very subtle) nor did I bother with the ISO 6400 version (the difference is roughly half-way between the noise level of ISO 3200 and 12800 just as you'd expect) so I figure the two examples should make the point that we don't just see a difference at ISO 25,600.

So why is it that you can show data of a D810 smoking a 5D III... and then I can turn around and show the 5D III smoking the D810?

The answer lies in understanding how digital sensors work and, more specifically, how gain is applied.

The "ISO invariance" tests that you can find are actually poor and unrealistic examples of how someone would shoot (and, in fact, betray that the people conducting the test do not actually understand how digital sensors work. Yes, you read that right... I just said that many of the people who are conducting these tests are not actually qualified to test equipment.

If you run around shooting everything at -5EV and then try to review your images or check the histogram on the camera, you're going to be one frustrated photographer. You're going to shoot at the suggested exposure and maybe you'll shoot above or below by perhaps a stop... maybe 2. But 5 stops? Never (not unless you were shooting it as part of HDR bracketing sequence.)

Again...try what I said....under-expose a Canon 5D Mark III or IV or a Canon 70D or 6D frame that was shot at a highlight-protecting exposure of Minus 4 to Minus 5 EV for the overall scene (perhaps Minus 8 to 10 EV for the shadow values) so the RAW file that comes out of the camera is almost BLACK....do the same thing with a Nikon D7100 or D7200 frame, or a Nikon D810 frame.

Why would you want to do that?

For astro-imaging, I'll tell you why you definitely do not want to do that... ever.

I'll shoot 5-minute sub-exposures using my 60Da at ISO 800 (ISO 800 is the "magic" ISO point where the camera starts trading "upstream" amplification for "downstream" amplification. In other words, if I don't shoot above ISO 800 then I'm actually protecting nearly all of my dynamic range. If I exceed ISO 800 then EVERY STOP of ISO that I increase costs me a stop of DR.)

Recall earlier than I mentioned the magic ISO for nearly all Nikon bodies with Sony sensors is... ISO 100.

At ISO 200, your camera has 1 stop less DR than it did at 100.

That means if YOU shoot at ISO 800, you're giving up 3 stops of DR. Since you don't want to give up DR in astrophotography you have to extend the exposure time by 3 stops. So my 5 minute sub works out to requiring a (I hope you're sitting down) 40 MINUTES on a Nikon.

The 6D you dislike so much... it can shoot at ISO 1600 while protecting nearly all of it's DR. You'd need 1 hour and 20 minutes on a Nikon to compete with that. The 5D III's recommended ISO for astrophotography is ISO 3200. That's 2 hours 40 minutes on a Nikon to capture what that 5D III can get in just 5 minutes. I don't have any data for the 5D IV but it generally outperforms the 5D III rather noticeably.

I used to subscribe to PopPhoto (too bad they seem to be going out of business) and their camera reviews (especially in the digital editions) used to let you play with their graphs interactively and they showed the DR of the various cameras overlaid on the same graph and you could adjust the ISO and watch the graph update. You could clearly see that the Nikon's had more DR at base ISO, but there was generally always a trade-off as you boosted ISO ... where Canon had more DR at high ISO.

There are a lot of situations where shooting at higher ISO makes a lot of sense (action photography in poor lighting comes to mind).

Hopefully you're catching on to why upstream signal amplification is a good thing and the ISO invariance test of using ISO 100 and try to boost ISO 5 stops is an unrealistic and unhelpful test when comparing cameras.

If you use the Canon camera the way it was intended to be used... you get noise comparisons like the ones I posted above. Those aren't doctored-up images... I just had to pick any example with ISO levels above ... say 1600... and that pattern starts to emerge (and becomes a stronger difference the higher you push it.)




I'll repeat one more time...

I'm not bashing Nikon cameras. They're great cameras. You've often likely seen me reply to the "which camera should I get" posts and suggest Nikons as often as Canons.

I'm just trying to get you (and others) to stop bantering about the Nikon is so much better and and it smokes the Canon models blah blah blah ... and in every category and every day of the week and twice on Sundays... because the data shows it's just not true.