My most wanted list, anything more to add?

[Daniel Browning]

But image quality suffers when you try to put a high number of pixels into a smaller space

I disagree. In ample light, a small sensor with more pixels has higher resolution than a large sensor with fewer megapixels. The 50D puts out images with more detail than the 5D1, for example. Noise and dynamic range, though, are of course better with the larger sensor (number of megapixels has nothing to do with it).

less light hits an APS-C rectangle than an APS-H rectangle than an APS rectangle

Not necessarily. If you're using the fastest, longest lens you can afford (or carry), then all the systems get the same amount of light falling on the sensor. That's why I brought up sports and wildlife photographers, because they will benefit the most from a 1.6X pro system.

I suspect that's one of the main reasons that the 1Ds3 has one less stop of ISO than the 1D3.

If that were the case then it would be the other way around, as the 1Ds3 with the same exposure has less noise.

[peety3]

But image quality suffers when you try to put a high number of pixels into a smaller space

I disagree. In ample light, a small sensor with more pixels has higher resolution than a large sensor with fewer megapixels. The 50D puts out images with more detail than the 5D1, for example. Noise and dynamic range, though, are of course better with the larger sensor (number of megapixels has nothing to do with it).

Image quality != resolution, and you seem to have missed that point.

less light hits an APS-C rectangle than an APS-H rectangle than an APS rectangle

Not necessarily. If you're using the fastest, longest lens you can afford (or carry), then all the systems get the same amount of light falling on the sensor. That's why I brought up sports and wildlife photographers, because they will benefit the most from a 1.6X pro system.

No, if the lens is the same, aperture/shutter is the same, and the camera is positioned the same (since pro sports shooters will all be in the photog pit, etc.), the APS-C sensor gets less light. Some of the photos passing through the lens land on black plastic on all sides of the sensor, so less photons (less light) is left for the sensor.

I suspect that's one of the main reasons that the 1Ds3 has one less stop of ISO than the 1D3.

If that were the case then it would be the other way around, as the 1Ds3 with the same exposure has less noise.
<div style="clear: both;"]</div>

Under a given light source, at a given ISO, aperture, and shutter combo, a particular pixel should be a particular numeric (digital) value. That comes from an analog device, the sensor, and its companion amplifier(s) and the resulting processing by the DiGiC chip(s) inside the camera. There's less light on a per-pixel basis hitting a 1Ds3 sensor than a 1D3 sensor, so the amplifiers/processors have to recover more data to bring the image up to the same digital value.


I can't find an online reference to point you towards, but if you were to put a particular lens on a 1Ds3 and a 1D3, take the same picture through both cameras, then crop away the pixels in the 1Ds3 picture until you had the same framing in the 1D3 picture, you'd find that the 1Ds3 has about 15MP inside an APS-H frame. The 1D3 has 10MP in that space. Those 5 extra megapickles have to get light somehow; they're sharing it with the other 10MP. The DiGiC chips in the 1Ds3 are having to boost the image more to match the same on-screen brightness you see. If it has less noise, that's great; it's probably why the camera has a slower frame rate.

[Jon Ruyle]

Dan has an interesting list- in particular, the long slow lenses... I've often wondered why they aren't more of them made, *especially*macro, where the reach is very useful and (very expensive) wide aperture isn't. Part of the problem with a long macro, though, is the huge amount of travel required... to be able to focus on infinity and at 1x, a 300mm macro would require a foot (300mm) of travel. So it might be a heavy tube. On the other hand, it would have barely more aperture-- and probably less glass-- than lightweights like the 100mm f/2 and 85mm f/1.8.


I've noticed that slow lenses tend to be more expensive than faster ones of the same aperture. (For example, 400mm f/5.6 costs more than 200mm f/2.8, 800mm f/5.6 costs more than 400mm f/2.8). Faster optics of a given aperture should be much less expensive to make, so I'm not sure why this is. Anyone out there have a clue?


I happen to own a very sharp 800mm f/8 lens manual focus lens (otherwise known as a 4" refracting telescope), and it cost under $2000. One can get them cheaper than that, and I would say that in general, telescopes (refractors, anyway) can make decent long slow mf lenses if you are aware of certain drawbacks. Below is a link to a sample pic. (I'm not holding it up as an example of good photography, mind you , or even as a good use of the 800mm f/8. It's just what I got when I told my daughter to stand still 20 feet from the lens). The sharp vignetting is caused by a star diagonal, which I didn't have to use. The pic is much sharper than it looks in this scaled down version (a 100% crop reveals more detail, like tiny wrinkles on my 6-year-old daughter's perfect skin).


http://picasaweb.google.com/jonruyle/January2009#5301770473802868274

[Daniel Browning]

Image quality != resolution, and you seem to have missed that point.

"Image quality" is nebulous. I broke it down into resolution, noise, and dynamic range. What other important factors did you have in mind?

No, if the lens is the same, aperture/shutter is the same, and the camera is positioned the same (since pro sports shooters will all be in the photog pit, etc.), the APS-C sensor gets less light. Some of the photos passing through the lens land on black plastic on all sides of the sensor, so less photons (less light) is left for the sensor.

Those photons might as well be falling on black plastic, because the part of the sensor they are falling on will have to be cropped to get the same magnification as the APS-C. (If the bigger sensor had a longer lens, then it would have the same magnification, but we already established that they're using the longest lens they can afford or want to carry).

There's less light on a per-pixel basis hitting a 1Ds3 sensor than a 1D3 sensor

Less light per pixel combined with many more pixels results in more total light.

then crop away the pixels in the 1Ds3 picture until you had the same framing in the 1D3 picture

In that case, the total amount of noise will be the same at similar output sizes (spatial resolution).


<span class="post"]99% of photographers, reviewers, magazines, etc.
focus on pixel size when total sensor area is what matters to light
gathering ability and noise. If pixel size were the most important
factor, then a small sensor with big pixels would outperform a larger
sensor with small pixels, but that is the opposite of reality.

The
common mental model for "small pixels are noisier" goes like this: a
single pixel, in isolation, when reduced in size, has less sensitivity,
more noise, and lower full well capacity. (So far, so good.) Therefore,
a given sensor full of small pixels is worse than the same sensor full
of large pixels. (Incorrect.)

The correct mental model is to
forget about pixels and think about light. The amount of light falling
on a given sensor area does not change, no matter the size of the
pixel. Large and small pixels alike record that light falling in
certain positions. Both reproduce the same total amount of light when
displayed; one just does it with greater accuracy.

The principle reasons that the myth persists are the following mistakes made when comparing data from pixels of different sizes:


<ul type="square"]
<span class="post"][*]Unequal spatial frequencies.
[*]Unequal sensor sizes.
[*]Unequal processing.
[*]Unequal expectations.
[*]Unequal technology.

[/list]


<span class="post"]
One
good example is a chart of noise power and spatial frequency. Many
claim that the 50D is noisier than the 40D because it has much smaller
pixels. The following chart by Emil Martinec shows why reality doesn't match the claims:


<span class="post"]

[Daniel Browning]

I've noticed that slow lenses tend to be more expensive than faster ones of the same aperture. (For example, 400mm f/5.6 costs more than 200mm f/2.8, 800mm f/5.6 costs more than 400mm f/2.8). Faster optics of a given aperture should be much less expensive to make, so I'm not sure why this is. Anyone out there have a clue?

I've wondered about this a lot as well. In the case of the 200 f/2.8, I think the reason is aberration correction. The 400 f/5.6 is corrected to a much higher standard than the 200 f/2.8. In fact, the 400mm f/5.6 has even less aberrations than the 400mm f/2.8 (when both are wide open).


The 800mm f/5.6 is definitely an outlier. Some have postulated opportunism on Canon's part, but who knows what the reason is.

[Colin]

While It's fair to say that closer pixels mean more noise per pixel, it's not more noise, necessarily, if you consider that more pixels means finer filtering of noise (and less loss of detail in the process).


At the same time, if you have a lens that can support a full frame sensor, a full frame sensor will in fact get more light. It gets all the light that would hit a 'crop' sensor, and on top of that it gets the light around it. More light, more signal to noise, in an analog fashion.


Now, where a crop sensor REALLY would be handy, from a weight perspective, is if you can get an EF-S lens, which doesn't have to have all the glass necessary for supporting a full frame sensor. Otherwise, if you've got the same pixel density, you can just crop a full frame for the same result.


Want lighter or longer? CROP!


of course, you do lose image detail in the optics if you crop, regardless of your pixel density, but that's true on any lens, and why larger format cameras have an easier time, everything else being equal, getting more detail.

[peety3]

There's less light on a per-pixel basis hitting a 1Ds3 sensor than a 1D3 sensor

Less light per pixel combined with many more pixels results in more total light.

Nope, you can't invent more light here.

<span class="post"]
One
good example is a chart of noise power and spatial frequency. Many
claim that the 50D is noisier than the 40D because it has much smaller
pixels. The following chart by Emil Martinec shows why reality doesn't match the claims:


<span class="post"]
[url="http://luminous-landscape.com/forum/index.php?s=bbb03fa9c87055adf6da0c7e89c4b46c&amp;s howtopic=29801&amp;st=20&amp;p=241562&amp;#entry24 1562][/url]






<div style="clear: both;"]</div>

You'd have a point, if we were comparing apples to apples. Instead, we're talking a newer generation sensor and a newer generation processor. Assuming it was possible to replace the 50D's DiGiC IV with a DiGiC III, you'd have more noise in the 50D.

[Daniel Browning]

Nope, you can't invent more light here.

Right, it's only more light if it's a larger sensor. (Once you crop the sensor down to the same size, it's the same light.)

You'd have a point, if we were comparing apples to apples. Instead, we're talking a newer generation sensor and a newer generation processor. Assuming it was possible to replace the 50D's DiGiC IV with a DiGiC III, you'd have more noise in the 50D.

The processor only applies to JPEG, not the capabilities of the sensor. I wasn't trying to make an apples-to-apples comparison, I was just pointing out the incorrectness of the 40D superiority. In any case, there is no such thing as apples-to-apples because no matter how similar something is, it will never be the same. Even unit-to-unit variation can cause differences.

[Sean Setters]

The Nikon CLS system is superior to Canon's wireless ETTL flash system. How about building a 580EX with RadioPopper capabilities built in (and possibly a more friendly user interface)?
<div style="clear: both;"]</div>

I'll bite: what makes Nikon CLS superior?
<div style="clear: both;"]</div>

Damn those wonderful forward/backward buttons on my mouse. I love those buttons, but I wrote a big long reply and it was gone in an instant because I accidentally hit the "BACK" button on my mouse. *sigh*


Ok, to sum it up:


1.) Nikon has been building optical slave capabilities and/or PC sync ports into their flashes for years. Canon only recently incorporated a PC port into its latest model, the 580EX II. In order to get the same functionality, a Canon user (not using a 580EX II) must purchase a hotshoe with a PC port.


2.) A Nikon body can use its pop-up flash to control off-camera flashes. A Canon user must purchase additional equipment (either a master flash, or an ST-E2), both of which are pricey. Also, because a Nikon body can use its pop-up as a commander, you can adjust your flash settings from the camera body rather than using a camera-mounted flash (which, I'm sure, is somewhat easier).


3.) This is trivial, but worth a mention. Once the flash goes into power-saving sleep mode (when off camera), you can wake a Nikon flash up by hitting the test fire button on your triggering system. A Canon flash requires that you touch a button on the flash itself. This is important if you happen to have your flash on a tall lightstand and don't shoot for a little while (which happened to me while shooting a 50th wedding anniversary...I had to disable the power-saving auto-off feature through a custom menu option).





There's my take on it.

[Groundfault]

A Volkswagen will always be a Volkswagen. It will never be a BMW...