Odds & Ends: 70D, 7D2, New Full Frame, D600 & Photoshop CS6

[neuroanatomist]

As a practical matter, I suspect the crop debate, at least in part, boils down to the fact that FF sensors are not likely to emerge with sufficient pixel density to produce IQ equivalent to a high quality crop sensor in distance constrained situations using the same lens, i.e. the same subject distance. 

Surprisingly, when you shoot the same subject at the same distance and crop the FF image to the FoV of the APS-C sensor, the IQ isn't too different, at least based on my comparison of the 5DII with the 7D.  The 7D image was slightly sharper, the 5DII image was slightly less noisy (and you can trade noise for sharpness, meaning effectively a wash).  But...the cropped 5DII image is only 8 MP, compared to the full 18 MP of the 7D.  So...if you need to crop further (or if 8 MP is not enough resolution for your desired output), then the APS-C sensor is the better choice.

Practically speaking, I often need to crop my 7D images of birds shot at 400mm, therefore I'm keeping my 7D after getting the 1D X.

[BobSanderson]

"IF" Canon can still change their plans to compete more effectively with Nikon's FFs I think what would make most sense would be:

The forthcoming 70D would become the premiere Canon APS-C with many of the bells, whistles of and structure of the past 7D with faster fps and a newer AF but stay at 18 MP. Priced somewhere between $1,400 to 2,000...a better value (and much better than the new Rebel) with an improved sensor (dynamic range and IQ upped a little)

The 7D II would go FF with many of the same features of the 5D III tempered down (less AF points, not as wonderful weather sealing...) with faster FPS but also at 22 MP. Similar but less robust than the 5D III. Call it the Jr. at about $500 more than the 7D II. This still leave a large pricing gap to reach the 5D III. It would be the prosumer FF.

The 5D II would be retired after great service to Canon and photographers.

The 5D III would begin to be offered, after catching up with demand, with some tactical rebates and lens offers to compete more effectively with D800 on price - updates would tweek some of the complaints and offer new features to make this great camera even better and the price would settle down at around $3,000 eventually.

Canon would start looking to deliver a high MP camera with a latest sensor for landscapers and others worried about MP wars....

[dilbert]

A particular format of sensor has no IQ. There's nothing inherently better about APS-H than APS-C or FF.

True when you're talking about IQ on the pixel level, untrue when you're talking about IQ on the image level.  The 20D and 5DII have the same size pixels.  Yes, the 5DII pixels are 'better' but not better enough to account for the IQ differences between the sensors.  The FF sensor has better IQ because as a whole, the larger sensor gathers more total light.

I'll beg to differ with you there. I'm pretty sure that if you took the center 8MP from a 5DII image that it would be better than the 20D's image when using the same lens at the same setting. e.g. from 20m away, use a 50/1.4 and photograph the same subject with both cameras. Yes, the images will not be the same but the center 8MP of both images should be. The center 8MP don't somehow magically benefit from the pixels around the edge of the sensor.

Yes, cropping throws away the benefit of the FF sensor gathering more total light.  I'm saying that if you move closer with the FF camera (or zoom in with a zoom lens) so you're getting the same framing covering the sensor, you'll get better IQ from the FF sensor.  That means there is an inherent advantage to FF, unless you go around shooting everything wider than you need and planning to crop away 60% of all your images.

But when you move or otherwise change the framing then you're undoing the "reach" aspect of crop sensors.

I also wonder whether the FF sensor does capture more light.

What determines how much light lands on the sensor is the lens.

Imagine, for instance, that you've got a light bulb on that's 3m away from the camera. The amount of light that will be collected by the lens is fixed, regardless of the camera/sensor, by the size of the front element of the lens. So the 50/1.4 will provide the same amount of illumination in the lightbox on a crop sensor as it will for a full frame. Now if both sensors have the same pixel density then the amount of light that is captured that represents the light is the same (assuming that the image of the light does not exceed the size of the sensor when projected into the camera by the lens.) So at this point, the full frame sensor does not capture more light over the same area of the sensor as the crop sensor. Overall the full frame sensor does gather more light but not from the subject. Replacing a crop sensor with a full frame sensor does not somehow magically cause there to be more light present.

When it changes is when we move closer to the object to replicate the same field of view that we had with the crop sensor. I'll note that in doing so, it is the increase in size of the image on the sensor that allows more light to be captured - the exposure parameters will also change as you get closer. Then because the distance relative to the object changes, we must therefore capture more light. That also goes to say that using a zoom lens in a manner that does not substantially change the distance to the subject will also not deliver more light.

Consider also that all sensors are the same distance from the lens' focal point.

What full frame sensors give us is better separation of detail. Of course, having said that, I feel like I'm wrong but I just can't see it.

For those that don't follow, the amount of light that reaches a given point from an object is an inverse-squared relationship, meaning that at 4m from an object you receive 1/4 of the light that you do when you're 2m from it.

[dilbert]

As a practical matter, I suspect the crop debate, at least in part, boils down to the fact that FF sensors are not likely to emerge with sufficient pixel density to produce IQ equivalent to a high quality crop sensor in distance constrained situations using the same lens, i.e. the same subject distance. 

Surprisingly, when you shoot the same subject at the same distance and crop the FF image to the FoV of the APS-C sensor, the IQ isn't too different, at least based on my comparison of the 5DII with the 7D.  The 7D image was slightly sharper, the 5DII image was slightly less noisy (and you can trade noise for sharpness, meaning effectively a wash).  But...the cropped 5DII image is only 8 MP, compared to the full 18 MP of the 7D.  So...if you need to crop further (or if 8 MP is not enough resolution for your desired output), then the APS-C sensor is the better choice.

Practically speaking, I often need to crop my 7D images of birds shot at 400mm, therefore I'm keeping my 7D after getting the 1D X.

Right, so if the 1DX was 46MP and had the same pixel density as the 7D then you could dispense with the crop camera

[KyleSTL]

TTMartin and unfocused are singing my tune.  If Canon released a 60D body with a FF sensor in it, I would be all over it.  Furthermore, it would probably push down the selling price of the 5DI, due to lower market demand.  A $500 5DI sounds awesome.  It would also probably affect the 5DII market since it would almost definitely have 9 cross-type AF sensors (from the 40D, 50D, 60D, and T4i) and a newer sensor and Digic 5, etc of the latest generation.  I hope the D600 comes out soon, since we're likely to see a reply from Canon shortly thereafter.

[dilbert]

Yes, cropping throws away the benefit of the FF sensor gathering more total light.  I'm saying that if you move closer with the FF camera (or zoom in with a zoom lens) so you're getting the same framing covering the sensor, you'll get better IQ from the FF sensor.  That means there is an inherent advantage to FF, unless you go around shooting everything wider than you need and planning to crop away 60% of all your images.

As a practical matter, I suspect the crop debate, at least in part, boils down to the fact that FF sensors are not likely to emerge with sufficient pixel density to produce IQ equivalent to a high quality crop sensor in distance constrained situations using the same lens, i.e. the same subject distance.  At least not at a cost attractive to those who would be the most likely to benefit from such a camera.   Perhaps the rumored high-MP FF would achieve this performance level I dont' know, but I note that the motivations for FF bodies are generally different from that of the cop bodies, such that to date we've not seen a FF  body that can do this.

I would be surprised if there weren't already FF sensors in existence with the same pixel density as crop sensors today. The problem is manufacturing. With a larger sensor there is also a similar increase in the chance of a flaw in the fabrication process to render the sensor useless. With fewer sensors per slice of silicon than with crop sensors, the requirements for a good yield are much more stringent with full frame sensors. The chance of a flaw during production goes up as the size of the pixels go down and the megapixel count goes up.

Nikon/Sony delivering a 36MP sensor camera for $500 less than the 5D3 also says a lot about the manufacturing processes being used. It may be that Canon have got more catching up to do than we realise.

[neuroanatomist]

Same light per unit area, true (thus, same exposure).  More area = more total light.  If you frame the subject identically (which, IMO, is the most relevant comparison since composition determines the picture), the subject is occupying a larger fractional area of the image circle, meaning more total light from the subject will fall on a FF sensor.

Right, so if the 1DX was 46MP and had the same pixel density as the 7D then you could dispense with the crop camera

Heck, if it was 36 MP and could still deliver that high frame rate, I'd dispense with the 7D...

[dlleno]

Right, so if the 1DX was 46MP and had the same pixel density as the 7D then you could dispense with the crop camera

There is little question that one could produce a FF sensor with the same pixel density as a crop sensor and therefore preclude the need to crop externally,  but I'm not convinced that today's state-of-the-art is capable of producing such a camera in a way that makes it attractive to those who want to use it for its FF capabilties, i.e. without throwing away the advantages relating to low pixel pitch. 

There is, certainly,  some FF pixel count which will produce the optimimum "end to end resolution IQ" competition for an 18MP crop sensor, of the same technology cycle,  for the distance-constrained situations where one must  crop the FF image to obtain the same FOV.  I suspect that pixel count is less than 46, and that such a camera would produce images inferior  to the otherwise equivalent lower density FF in those situations that benefit from that lower density. 

THere is no question that the FF is superior when the subject can be framed optimially.  The only area in which the crop debate rightfully exists, imho,  is in those situations where is not NOT possible to frame the subject optimally-- for example, wildlife situations where the subject distance cannot be reduced, or one does not have a longer lens

[TTMartin]

I would be surprised if there weren't already FF sensors in existence with the same pixel density as crop sensors today. The problem is manufacturing. With a larger sensor there is also a similar increase in the chance of a flaw in the fabrication process to render the sensor useless. With fewer sensors per slice of silicon than with crop sensors, the requirements for a good yield are much more stringent with full frame sensors. The chance of a flaw during production goes up as the size of the pixels go down and the megapixel count goes up.

Nikon/Sony delivering a 36MP sensor camera for $500 less than the 5D3 also says a lot about the manufacturing processes being used. It may be that Canon have got more catching up to do than we realise.

Well, Canon has built a 120mp APS-H sensor.
http://www.dpreview.com/news/2010/8/24/canon120mpsensor

And a huge CMOS Sensor.
http://www.dpreview.com/news/2010/08/31/canonlargestsensor

Everybody talks like Sony/Nikon have some sort of amazing technological advantage over Canon. Except for DXOMark scores there is nothing that shows this. Dispite what DXOMark says tests at DPReview show that the 5D Mk III has more dynamic range than the D800 http://www.dpreview.com/reviews/nikon-d800-d800e/19

So either Nikon has some really crappy processors that mess up the 2.7 stop advantage that DXOMark says the D800 has, or we should stop quoting, worrying about, or looking at what DXOMark says.

If we're going to base or view of Canon's sensors on DXOMark scores, than Canon might as well not do anything. Because no matter how good a sensor Canon produces, DXOMark will still rate it the same.

[x-vision]

APS-C does not give the same IQ as APS-H
APS-H has more reach than ff

... but by that logic, APS-H does not give the same IQ as FF, no?

APS-H is the compromise solution that has no major weakness

Nope.

Even the kids know that bigger is better  .
So, from a marketing perspective, APS-H has one really major flaw: it's smaller than FF.
And if your camera has a smaller thingy than the competition, you can't charge same/more than the competition.
It's that simple. 

Nikon switched from DX to FX in their pro Dx line because Canon was winning with a bigger thingy in their cameras.
And Canon in turn switched from APS-H to FF in the 1-series because now Nikon had a bigger thingy.

So, despite all the cheerleading, APS-H is out forever. Get over it .

[poias]

Dispite what DXOMark says tests at DPReview show that the 5D Mk III has more dynamic range than the D800 http://www.dpreview.com/reviews/nikon-d800-d800e/19

You are joking, right? Turn the ADL to Extra High and see D800 range go from the end to end. Let's face it, 5D3's DR is even worse than its 4 year old predecessor's.

[dilbert]

Everybody talks like Sony/Nikon have some sort of amazing technological advantage over Canon. Except for DXOMark scores there is nothing that shows this. Dispite what DXOMark says tests at DPReview show that the 5D Mk III has more dynamic range than the D800 http://www.dpreview.com/reviews/nikon-d800-d800e/19

So either Nikon has some really crappy processors that mess up the 2.7 stop advantage that DXOMark says the D800 has, or we should stop quoting, worrying about, or looking at what DXOMark says.

Or dpreview is using JPEGs and not RAW data.

While you're on that dpreview page, make these changes:
Set the "Canon EOS 5D Mark III" to "HTP On"
Set the "Nikon D800" to "ADL Extra High"

This should give you:
Canon EOS 5D Mark II - ~33 to ~2
Nikon D800 - ~38 to ~2

[TTMartin]

APS-C does not give the same IQ as APS-H
APS-H has more reach than ff

... but by that logic, APS-H does not give the same IQ as FF, no?

APS-H is the compromise solution that has no major weakness

Nope.

Even the kids know that bigger is better  .
So, from a marketing perspective, APS-H has one really major flaw: it's smaller than FF.
And if your camera has a smaller thingy than the competition, you can't charge same/more than the competition.
It's that simple. 

Nikon switched from DX to FX in their pro Dx line because Canon was winning with a bigger thingy in their cameras.
And Canon in turn switched from APS-H to FF in the 1-series because now Nikon had a bigger thingy.

So, despite all the cheerleading, APS-H is out forever. Get it over with .

Well, remember the original APS-H spec (Kodak Film) was for a 16X9 aspect ratio. With the same height as APS-C. So, Canon could have a 16X9 APS-H sensor with a mirror that was the same height as an APS-C camera, only wider. So, it would clear EF-S lenses. If they did that, Canon could have an APS-H sensor camera that was compatible with EF-S lenses.

[Albi86]

Everybody talks like Sony/Nikon have some sort of amazing technological advantage over Canon. Except for DXOMark scores there is nothing that shows this. Dispite what DXOMark says tests at DPReview show that the 5D Mk III has more dynamic range than the D800 http://www.dpreview.com/reviews/nikon-d800-d800e/19

Well, it's obvious that Sony is a much bigger company with a much bigger wallet, and thus can afford to invest a lot more in R&D than Canikon. Also, Sony is into many more things than just photography.

A sensor is a piece of tech, by itself has little to do with photography. I think Nikon was very smart to buy them from someone who knows better in this respect, while keeping processor coding in-house.

[dilbert]

APS-C does not give the same IQ as APS-H
APS-H has more reach than ff

... but by that logic, APS-H does not give the same IQ as FF, no?

APS-H is the compromise solution that has no major weakness

Nope.

Even the kids know that bigger is better  .
So, from a marketing perspective, APS-H has one really major flaw: it's smaller than FF.
And if your camera has a smaller thingy than the competition, you can't charge same/more than the competition.
It's that simple. 

Nikon switched from DX to FX in their pro Dx line because Canon was winning with a bigger thingy in their cameras.
And Canon in turn switched from APS-H to FF in the 1-series because now Nikon had a bigger thingy.

So, despite all the cheerleading, APS-H is out forever. Get it over with .

Well, remember the original APS-H spec (Kodak Film) was for a 16X9 aspect ratio. With the same height as APS-C. So, Canon could have a 16X9 APS-H sensor with a mirror that was the same height as an APS-C camera, only wider. So, it would clear EF-S lenses. If they did that, Canon could have an APS-H sensor camera that was compatible with EF-S lenses.

Quit it already, will you?