gdanmitchell

I would definitely get a remote release. I probably would not bother with the more

expensive model. There is a button lock on the non-timer model and when you are

dealing with very long exposures (minutes rather than seconds), split second timing is

totally unnecessary.

(The long exposure equivalent of one stop exposure time difference is often quite long. If

you are at 5 minutes and you want to go one stop either way you'll go to 2 1/2 minutes or

10 minutes. Being off by some seconds is completely inconsequential.)

Dan

There's street and then there's street. There are multiple ways to shoot the subjects lumped together under the heading of "street photography."

When people think of classic street photography they are often thinking of stuff done with high quality rangefinder cameras and one or a very few

primes, and probably not very large aperture primes at that. The object was to work unencumbered by a ton of gear and to be able to react quickly to

the rapid events going on around you.

Some people trying to replicate that experience using a modern DSLR might think about using something like a Digital Rebel (e.g. XT, XTi, or the

upcoming XSi) with some wide-ish primes - somewhere in the 20mm to 35mm range for core lenses, and perhaps including something as long as a

50mm or 85mm prime. Others might even think about working with a good P&S, such as the G9.

Other folks (and I'm in this category) think "street" is more about the subject matter than the gear and will use zooms. I love to wander around with a

17-40 L on my 5D (in interior areas) or my 24-105 (outside) and react quickly and instinctively to what I see going on around me.

And I'm sure that these descriptions don't define limits around what constitutes "street" photography. Personally, I wouldn't have any need or desire for

lenses like the 24mm f/1.4, the 35mm f/1.5, the 50mm f/1.2 or the 85mm f/1.2 for street - but that doesn't mean that such an approach wouldn't be

valid for some photographers. (If I were doing the prime approach, I'd probably want the smaller non-L versions of these lenses.)

Dan

<p>Harry, it is in my <a href="http://www.photo.net/photodb/folder?folder_id=691980">night

photography gallery</a> here at photo.net.<p>

<p>Dan

<p>Mendel wrote:

<blockquote><i>1. The difference between 16mm and 17mm is significant, much more than (say) the difference between 34mm and 35mm. Near the top of the Luminous

Landscape review I linked above, there is a mouse-over demonstration of the difference.</i></blockquote>

<p>Maybe. Maybe not. 17mm of FF is already pretty darn wide. In any case, there is also a more significant difference between 35mm and 40mm.

<blockquote><i>2. Going from an f4.0 to f2.8 does theoretically double the light coming through the viewfinder, but the apparent diffence is not that great. You can see for

yourself by setting an f2.8 lens manually to f4.0 and depressing depth-of-focus preview.</i></blockquote>

<p>Good point. Yes, it is brighter. No, most of us won't really notice a huge difference. (I shoot with lenses whose maximum aperture ranges between f/1.4 and f/4.)

<blockquote><i>3. Just my 2 cents: these 2 lens and their differences seem comparable to the 24-70 vs 24-105 debacle. There is more compromise in the 17-40, both in

spec's and build. The net result, weighing size/weight, image quality and price, makes for a difficult choice.</i></blockquote>

<p>If by "specs" you mean the difference in maximum aperture and focal length range, OK. However, I strongly disagree about there being any difference in build quality. Both

have the same typical high quality L lens build quality. You don't get more "quality" from the 16-35 - you get a larger maximum aperture and a 16-35mm focal length range

instead of 17-40mm.

<p>Dan

Hard to say for sure. I like to alternate between my 17-40mm f/4 L and my 24-105mm f/4 IS

L for street photography. But you say you want f/2.8 or wider.

Have you considered a few primes?

Dan

<p>I would <i>not</i> have Canon do it. Sensor cleaning is part of the job of owning a

DSLR, and it really isn't tremendously difficult.

<p>I've outlined my general approach <a

href="http://www.gdanmitchell.com/2007/10/20/my-approach-to-dslr-sensor-

cleaning/">here</a>. Basically, I don't worry about small amounts of dust on the sensor

since it is an easy matter to remove them in post-processing. When the dust becomes a

bit too thick, I first try a blower, then a static-charged brush, and finally move on to

PecPads and Eclipse fluid.

<p>I probably don't have to do the wet cleaning more than a few times per year.

<p>Dan

<p>Hop Phan, you have listed the reasons that the f/4 version of this lens is

<i>preferred</i> by quite a few photographers <i>today</i>.

<p>f.2.8 will always be one stop faster and will always provide a slightly narrower DOF

for those who need this,so I don't believe that the f/2.8 lens is likely to go away. But the

f/4 lens is a great lens for many, many photographers who don't need the extra stop and

would like a bit less size/bulk.

<p>Oh, and who don't mind paying a bit less.

<p>Dan

On my full frame body the sharpest apertures with the 70-200mm f4 L are in the f/8-f/11

range. f/16 is next sharpest and almost indistinguishable from f/8 and f/11. Things are not

quite as sharp at f/5.6 and f/4, though still quite good enough that I would not hesitate to

use those apertures when necessary due to low light or need for narrower DOF.

<p>Some time ago I made up a small graphic to illustrate the actual difference in size (at

same dpi) of sensors with different numbers of photosites. You can find it <a

href="http://www.gdanmitchell.com/2005/09/10/dslr-megapixel-count-

comparison/">here</a>.

<p>It is hard to say when or if moving to a sensor with "more megapixels" will make a

difference. With a qualification I'll mention below, I feel pretty certain that doubling the

number of photosites can make a significant difference. Increasing the number of

photosites by 50% may or may not make a difference in your photographs.

<p>One qualification has to do with how you reproduce your photographs. If you mostly

print at letter size or smaller and/or share electronic versions of your images on the web

or in email, going from 8MP to 12MP will almost certainly <i>not</i> result in any visible

improvement in your images.

<p>If you are making fairly large prints (at least 12 x 18 inches and perhaps larger) there

could be a difference that you'll notice. But even here I think that the difference is only

likely to be significant if you are using excellent lenses, and careful technique - perhaps

shooting from a tripod, for example.

<p>Other factors can also explain better image quality that you might feel you get from

the G9. For example, the in camera processing (sharpness, saturation, levels) may be

different and responsible for the difference. I think that is more likely, given the small

sensor in the G9, than that the higher photosite density is causing the perceived

improvement.

<p>I think the XSi sounds like a great camera. Whether you'll see improvements in your

photographic results with it is another question.

<p>I agree with Jos that the most significant difference in image quality could come from

moving to full frame - but again, only if you use your photos in a way where the

improvement will be visible.

<p>Dan

I do not delete many photos at all, and I probably have a collection that rivals yours in

size. I delete photos that are technically very bad or that I know I would never use because

the person photographed looks awful in the particular shot. (Its an ethical thing.)

However, If I make 30 shots of a subject, I do not cull out some percentage of the weakest

shots. I have several reasons:

First and foremost, too many times I've gone back through my old photos and found really

wonderful images that I simply overlooked the first time through. I think that when I shot

them I had some particular thing in mind, the photos were not that particular thing so I

flew past them, but with some distance I can look at them for what they are. (Once a year I

actually go back through ALL of my images from the past year to look for these.)

Storing images on a hard drive is very cheap. No fuss, no muss - why delete.

I've had people contact me for images of particular subjects. I found such photos among

those that I might have deleted otherwise, since they were not personally interesting to

me.

Dan

Fred, "confusion" or wrong-headedness on my part is not the only reason why I might keep saying the same things about diffraction.

I responded to you over in the 17-40 thread and tried to offer a more detailed explanation of why the effect of diffraction on photographs made with a given lens/aperture

will not be the same when sensor size is different. If that explanation is too wordy, take a look at the simpler one posted by someone else right below your message.

In fact, you are essentially half right in your assumptions about diffraction. However the other half that you neglect is absolutely critical to understanding the effects of

diffraction on photographs made with different sizes of sensors.

If I wrote that diffraction is "worse" on crop sensor bodies I apologize for my imprecise use of words. More accurately, as you stop down your aperture diffraction blur

becomes a significant issue sooner on crop sensor bodies than on full frame bodies. By extension the same could be said relative to FF DSLRs and MF bodies - you can stop

down further on the MF body before diffraction becomes significant.

I'm really not confused about this at all.

Take care,

Dan

<p>Jim Larson wrote:

<blockquote><i>"Every time Canon increases the MP count, there are those who say "you don't need those extra MP" "Noise will increase" and other

stuff.</i></blockquote>

<blockquote><i>Honestly, this happens every time a newer, higher density, sensor is released."</i></blockquote>

<p>I'm with you, Jim.

<p>Now the messages often say "10MP is plenty. There is no advantage in going to 12MP. There will be too much noise at 12 MP; 10MP is the sweet spot."

<p>The funny thing is that the last time around they said "8MP is plenty. There is no advantage in going to 10MP. There will be too much noise at 10MP; 8MP is the

sweet spot."

<p>And at the time of the previous update it was "6MP is plenty. There is no advantage in going to 8MP. There will be too much noise at 8MP; 6MP is the sweet spot."

<p>:-)

<p>Dan

If you just need a cpu to use with existing peripherals, for many purposes a Mac mini is

just fine. However, it is not a particularly powerful system. As a Mac user I'd hesitate to go

that route for any sort of significant PS work.

If you are buying a complete system, however, Intel iMac systems are very price

competitive with PC alternatives and make great PS workstations for almost everyone. (I

use a 20" iMac and am very happy with it.)

I run mine as a plugin. I am very pleased with the program and recommend it highly.

Dan

<p>Some hot spots are going to be unavoidable in night shots including extensive dark areas and direct artificial light sources. Sometimes I handle

this by combining multiple exposures in post. I might make a main shot for optimum exposure of the bright areas, and then make another shot or two

exposed for shadow detail. In PS I'll most likely start with the darker image (the one exposed for the bright lights) and then use layers/masks to "paint

in" some detail from the other exposure(s).

<p>Here is an example that used this technique:

<p><img src="http://d6d2h4gfvy8t8.cloudfront.net/5552972-md.jpg">

<p>(More <a href="http://www.photo.net/photodb/folder?folder_id=691980">here</a>.

<p>BTW, you perhaps don't need the electronic timer remote release avoid motion blur. It is true that you <i>could</i> produce some blur if you are

using a really flimsy tripod (and/or not putting it on a solid surface) and/or being rely careless with the remote. However, you can also slip the release

lock over the button and let go of it.

<p>BTW #3: I wonder if putting the tripod on the grassy surface could be a problem. If you have a solid tripod with metal points for feet you can get

good purchase on a lawn, but if you have a light tripod and the usual tripod feet this may not be the case. Adding some weight to the center column

hook could also help a bit here.

<p>Dan

<p>Common question. The answer depends on how/what you shoot.

<p>A first question might be, are you finding the apertures of the 10-22 to be a serious limitation in your photography? If not, your lens is <i>almost</i> certainly the f/4 17-40.

<p>The 16-35mm f/2.8 II L gets you better image quality wide open. If shooting an ultra-wide at f/2.8 and f/4 is a big part of your photography the extra cost of this lens could

well be worth it for you. Reportedly, the "betterness" of the new version of this lens is found only at the f/2.8 aperture - and it doesn't offer image quality improvements at smaller

apertures compared to the previous version of the lens.

<p>The doesn't seem to be any real evidence that it is better than the 17-40mm f/4 L when stopped down. There seems to be some general sense that, in fact, the 17-40

<i>might</i> be slightly (but probably insignificantly) sharper at smaller aperture, especially in the center. If you mainly shoot ultra wide at smaller apertures (e.g. - landscape

going for big DOF) then the 17-40mm f/4 L could be the "better" choice.

<p>Dan

<p>(Disclaimer: I own and extensively use the 17-40mm lens. I considered but chose not to purchase the 16-35mm lens, so my knowledge of that lens is based on extensive

research but I haven't used it. My main interest in ultra-wide lenses on FF is primarily landscape, though I have used the 17-40 for street photography and some interior

photography as well.)

One more small point to hopefully avoid confusion or misinterpretation of the above. I don'

mean to create the impression that "diffraction is worse" on smaller formats.

It would be more accurate to say that diffraction becomes significant sooner as you stop

down on a smaller format.

<p>Fred C wrote:

<blockquote><i>"Dan, I'll make it clear where diffraction comes from: the lens. Given the same lens at the same aperture, the amount of diffraction is the same. Sensor or format size has nothing to do

with it, no matter how often your assert to the contrary. See my detailed explanation here: http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00O8Q2"</i></blockquote>

<p>You are only considering part of the issue, and failing to consider the most important part in regard to sensor size. To build on your statement: Yes, "diffraction comes from... the lens." But it is also

significant to look at "where the diffraction ends up."

<p>The absolute dimension of the diffraction <i>at the sensor</i> (airy circle and all that) will indeed be the same regardless of sensor/film size.

<p>But that is <i>precisely why sensor size is relevant</i> when discussing diffraction in photographs.

<p>Let's take a hypothetical lens/aperture combination that produces diffraction blur that is (to use a convenient though not precise description) 1/10 of a millimeter wide. You are indeed correct that

the diffraction will be the same .1mm wide "blur" on crop, FF, MF, or 8x10 film or even 8" x 10" film.

<p>Do you see the issue yet?

<p>Point: A specific absolute amount of diffraction covers a larger portion of the frame on a smaller sensor.

<p>Our ".1mm diffraction blur" comprises a much larger portion of the image on the very small crop sensor image than it does on the larger image. The "blur" would "cover" only 1/1000 of the width of

the image on sensor/film that is 100mm wide but it would be 1/100 the width of the image on a sensor that is 10mm wide. Thus it would be "10 times worse" - to use approximate language - on a

print made from the smaller sensor.

<p>Another way to think about it is the following: If diffraction begins to "spill across" a <i>significant portion of the image at a particular aperture</i> on a small format, it would not yet "spill

across" the same portion of the image in a larger format - in other words, it would not be "significant" at this same aperture. One could continue to stop down the lens a bit more on the larger format

before the diffraction increase to the point where it would "spill across" the same portion of the image. Between crop and FF digital one should get roughly the same amount of "blur" expressed in terms

of relationship to the overall frame size at about f/8 on crop and perhaps about f/14 on FF.

<p>No, I'm not terribly "confused" about this issue. Part of your understanding is exactly correct, but only as far as it goes - and it doesn't go quite far enough until you consider the effect of a given

amount of diffraction relative to the overall dimensions of the capture media.

<p>Take care,

<p>Dan

<p>(An analogy: It is true that a difference in speed of 1 mph is the "same difference" whether you are comparing the speed of two airplanes or the speed of two people walking. But while a 1 mph

velocity difference is pretty insignificant when comparing the speeds of two supersonic aircraft, it could be very significant when comparing how long it might take two people to walk one mile.)

I strongly agree with Larry Cooper.

You certainly have some fine pieces of equipment in that list. If you just want to know if those are high quality lenses, the answer is "yes."

If you want to know if they are the best choices for you and your photography, it isn't possible to say based on just seeing a list.

Depending upon the sort of photography you do and the product you produce, it can be valuable to have high quality lenses and other

equipment. But which _specific_ equipment within the framework of "high quality" is best for your work is an entirely different question.

(On the off chance that you are just starting out in DSLR photography and trying to buy "the whole kit" right off the bat, I think that doing it

that way can be a very poor idea.)

There is no universal "right lens," much less a universal "right set of lenses." To give you one example, some folks who shoot, for example,

weddings and portraits justifiably regard the 24-70mm f/2.8 L lens highly. On the other hand, some who (like me) favor other subject such as

landscape or perhaps street photography may much prefer the 24-105mm f/4 IS. Who is right? We both are.

You'll find a similar situation among the four Canon 70-200mm L zooms and between the 17-40mm f/4 L and the 16-25mm f/2.8 L. There

are excellent reason why any one of the four 70-200 lenses could be the best choice for one photographer, while a different one would be the

right choice for another.

And so forth...

I completely understand why one particular photographer might be justified in believing that the best kit would be: 16-35mm f/2.8 L, 70-

200mm f/4 L, 70-200mm f/2.8 IS. Yet, for another photographer the "best" kit might be 17-40mm f/4 L, 24-105mm f/4 IS L, 70-200mm

f/4 IS. For someone else the "best" might be a 40D with the 10-22mm EFS and the 24-105mm f/4 IS L. It all depends one what you are trying

to do.

Dan

<blockquote><i>"Hi everybody! I like to buy a 70-200mm L IS lens, but F2.8 or F4 which one I am not sure. Is it F4 wide

open sharper than F2.8 at F4 from 70mm to 200mm? Thank you for sharing your experience."</i></blockquote>

<p>Choosing among the uniformly excellent Canon 70-200mm L lenses (f/2.8 or f/4, IS or non-IS) on the basis of

"sharpness" is like <i>choosing between a dollar bill and a stack of four quarters on the basis of "value."</i> There

could be reasons to prefer a dollar bill or a stack of quarters (need change for a parking meter?) but "value" isn't one of

them.

<p>Sharpness is <i>essentially equivalent among all four of these lenses</i>. Any differences can largely be attributed

to sample variations, and besides being unpredictable are generally completely insignificant in actual photographs.

<p>The <i>significant</i> differences that might cause photographers to select any of the four as his/her choice

include:

<ul>

<li>need for f/2.8 aperture or not.

<li>need for IS or not.

<li>whether or not weight/bulk are important issues.

<li>cost.

</ul>

<p>Note that <i>any one of these lenses could be the best choice</i> for a given photographer depending upon the

intended use, and that all four can produce excellent photographic results.

<p>Some examples of situations in which one or the other might be the best choice - and the list is not comprehensive:

<ul>

<li><b>70-200mm f/4 non-IS/</b> - could be the best choice for a photographer who wants a less bulky and lighter

lens and who generally shoots from a tripod, say for landscape work.

<li><b>70-200mm f/4 IS</b> - could be the best choice for a photographer who doesn't need the somewhat narrower

DOF of the f/2.8 lenses, and who often shoots handheld in decent light, especially if camera shake is likely to be more of

a problem than subject motion.

<li><b>70-200mm f/2.8 non-IS</b> - could be the best choice for a photographer wanting the narrower DOF of f/2.8

but working from a tripod, or for one shooting action sports where shutter speed is the limiting factor in stopping the

action.

<li><b>70-200mm f/2.8 IS</b> - could be the best choice for someone who needs to cover all of the possible options,

isn't concerned about cost, weight, bulk.

</ul>

<p>The question isn't "which is sharpest." It is "which matches best to the types of photography I do?"

<p>Dan

This brings up a question/point that I think about a lot as I read forum posts. I _seems_

that a good number of folks buy a particular lens because a) it is supposed to be a really

good lens, b) someone tells them that it is a lens that they should own, or c) it looks like a

really cool lens.

The 70-200mm f/4 qualifies on all counts. I use mine a lot. It is a really fine lens.

But why did _you_ get it?

Lens purchases should usually be based on identifying a role for that particular lens in

ones photography. For example, if one owns, for example, the kit lens and often is

frustrated by being unable to shoot some particular subject because 55mm is not long

enough, it is time to think carefully about what specific lens features might help with this.

Obviously, a longer lens. But also how long, what apertures, etc.? At that point, the lens

that "fills the gap" you have identified can be found, and you make a purchase.

But it just strikes me as somewhat odd that a person would buy a lens without

understanding why the need it our how they'll use it.

Take care,

Dan

A few thoughts:

1. f/8 should be fine. I'd tend to keep it at f/8 and perhaps increase ISO to 200 rather than

going to f/5.6. Odds are that the edges will be sharper at f/8 than at f/5.6 - though you

know this particular lens better than I.

2. If there is a focal length sweet spot for this lens (perhaps not at 10mm?) try to position

yourself further back from the building such that the longer focal length can be used.

3. For a long and wide subject like this, have you thought about stitching two horizontal

images into a panorama format? The building would be less dwarfed but the dark sky and

foreground.

4. MF set at infinity may not be exactly the best focus. Since there are some bright lights

on that building I'd try to AF one of them (perhaps using only the center AF point), turn of

AF so that focus point doesn't change, and the recompose. (Focus and recompose

shouldn't cause any problems with this subject.) Other methods may work if you can't get

AF that way: Put a very bright light in the scene so that you can AF, remove it after focus

and proceed as above. Or, focus during day time or twilight when you can get good AF and

use that setting.

Dan

<p>Fred C asked:

<blockquote><i>Why is the lens' performance at f/4 weaker with an APS-C sized sensor?</i></blockquote>

<p>That wasn't my point. What I was pointing out is that (following on what I write below about diffraction) if you

accept that diffraction can become an issue beyond about f/8 on crop sensor bodies you'll have a relatively narrower

range of "usable" apertures with this lens on a crop body than on a full frame body.

<p>(And, yes, I understand that f/8 is not exactly a diffraction "hard stop" on crop, and that f/16 is not a "hard stop"

on FF.)

<blockquote><i>And at f/8, why is diffraction a problem on APS-C but not on 24x36mm

sensors?</i></blockquote>

<p>I won't go into the full explanation here, but <i>diffraction</i> (at least in terms that are meaningful in a print)

is a function of sensor/film size and aperture. At some point as you stop down diffraction begins to produce a

softening of the image that can be significant. If we use f/8 as a point beyond which diffraction softness can begin to

increase on 1.6x crop sensor bodies, the onset point would be somewhere around f/14 on FF. In my tests (see <a

href="http://www.gdanmitchell.com/2007/04/12/sharpness-and-aperture-selection-on-full-frame-

dslrs/">http://www.gdanmitchell.com/2007/04/12/sharpness-and-aperture-selection-on-full-frame-dslrs/</a>)

there is a very tiny amount of diffraction that may be visible to pixel peepers on a 12 MP full frame body starting at

about f/16 - but it is not enough to be significant in a print.

<p>BTW, Fred, you posted a response to me in the "17-40" lens thread that I don't understand. I left a follow-up

there. Perhaps you could reply?

<p>Thanks

<p>Dan

<p>Fred C quoted me:</p>

<blockquote><i>"On a crop sensor body... I found that there were some sharpness issues in the far corners. They could be reduce significantly by stopping down, which works well for landscape and

similar subject in most cases. On a crop body you are somewhat limited though in terms of how far you'll want to stop down by the onset of diffraction. This can start to diminish overall sharpness beyond

about f/8 or so.<i></blockquote>

<p>And then Fred C wrote:

<blockquote></i>Just can't stop repeating this piece of creative writing, huh? http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00O8Q2</i></blockquote>

<p>Sorry, I'm not getting your point...

<p>In the thread you link to I wrote:

<blockquote><i>If I were shooting a crop sensor body and looking for lenses in that category, I'm pretty certain that I would go with the EFS 17-55mm f/2.8 IS. It really has a great feature set: f/2.8 max

aperture, IS, good optical quality, reasonably price for what it provides.

<br><br>If you are planning a move to full frame soon then one of the L wides could make sense. Otherwise they may not really gain you that much.

<br><br>Between the two of them, the 17-40 is quite decent if you do mostly landscape type stuff, though on a crop sensor you could find yourself kind of caught between the slightly weaker wide

aperture performance at f/4 and the onset of possible diffraction after f/8. (I love this lens for landscape on full frame.)

<br><br>The 16-35 does give you an extra stop, and it holds its image quality much better at wide open apertures. However, if that isn't how you shoot it (wide open) most of the time, the advantage

over the 17-40 sort of goes away.</i></blockquote>

<p>(The question in that thread was slightly different and the L lenses had been mentioned IIRC.)

<p>Are you saying that I made this up? Or maybe something else. I don't get the meaning of your comment.

<p>I stand by what I wrote in both places. Well, except for a typo or two.

<p>Dan