paulmessmer

<p>Does anyone who has the Sigma 50mm f/1.4 on a Canon body notice a conspicuously loud noise when the lens stops down? After returning two wretchedly-focusing copies of Canon's 1.2L, I decided to try the Sigma to replace my partially-broken Canon 1.4. The lens came today, and appears to be focusing well and producing promising images. I noticed though when I press the DOF preview button to stop the lens down, it's surprisingly loud and odd sounding (compared to Canon lenses), but I don't have other Sigma lenses to compare to. Just wondered if this was normal?, or whether this one is headed for some sort of premature failure.</p>

<p> </p>

From the same position, if you take the "same" picture... A) Using

a full-sized sensor 1Ds with an 80mm lens, and then B) Using an APS sized sensor 10D with a 50mm lens, both at the same aperture, the results (on say an 8x12 print from each camera) are _almost_ the same. The only difference is that B) has more depth of field. The field of view is the same, the perspective compression is the same.

The answer is that it depends on the f/4 lens you are connecting it to. I have both the

new pro Tamron 1.4x and 2x teleconverters, because I wanted to use them with a

Tokina 300mm f/2.8 as well as with Canon lenses. They have the extra contacts on

them and they are able to report the change in maximum aperture when they are

connected to a lens, just like the Canon teleconverters do. So, for example, if I stick

the 2x on my 70-200 IS, the camera knows the maximum aperture has gone from f/

2.8 to f/5.6.

 

However, unlike the Canon ones, these Tamron teleconverters can physically mate to

(almost?) any lens. It appears that, when they mate to a lens not originally designed

to take a Canon teleconverter (a lens that doesn't have the extra contacts on it), the

camera does not know the modified maximum aperture. My Tokina 300mm f/2.8 is

an example... it does not have the extra pins on it, and the camera still sees it as f/2.8

even with a teleconverter on it.

 

So... if you connect the Tamron Pro 2x to a lens like the EF 300/4L, which I assume

has the extra contacts on it, the camera will see it as f/8 and not autofocus. If

you connect it to a Canon lens not originally designed to take a TC (I tried my 50/1.4),

or a third party lens, without the extra contacts on it, the camera wouldn't know

and would still try to autofocus.

Back when I shot film with Minolta AF, I had this lens in the Maxxum mount. When I

switched to Canon digital with a 10D and later a 300D as well, I bought this lens in

the EF mount. It has never disappointed me. While I own a bunch of L lenses, I never

hesitated to get the Tamron 90mm. The fact that it can be found used cheaper than

the Canon 100mm didn't hurt either. I played with the Canon 100mm in the store

once, but have never owned it, so can't really compare the lenses directly.

 

For myself, Pros:

- Has always been fabulous optically.

- I like the manual focusing feel and big focusing ring.

- Small/Lightweight

 

Cons:

- Auto focus is noisier and slower than Canon USM. When used as a short tele lens,

it's fine, particularly for something like portraits. When used in macro applications,

focus may hunt but manual focus is typical in such applications.

- Front filter is 55mm, which shares nothing with my other lenses.

- No tripod collar

 

I got my EF mount one used/mint in an auction for US $300 and would do it again if I

needed to replace it.

I updated the firmware of both my 300D yesterday (and my 10D as well). The 300D

reports the current version now as 1.1.1 in the camera.

"With the 1.6X magnification of the 300D sensor rendering an effective focal range of

80mm, I thought that f2.5 would be enough to throw the background out of focus,

especially in tighter headshots."

 

Something to consider - while it may have an effective focal length of 80mm, it will

not give you the same out of focus background as a real 80mm lens @ f2.5 on a full

frame camera will. There will be less blur. As you use smaller sensors/film, at the

same "effective focal length" the depth of field will increase... you will not get as much

background blur. If you use a consumer digicam with a small sensor for example,

you'll notice that even at f/2.8 you don't get much background blur. The opposite is

also true... e.g. large format cameras have less depth of field than 35mm when taking

the same picture.

"Personally, I don't find head 'n shoulder portraits flattering with 50mm lenses--the

face looks flat (nose appears too wide, etc.). The 1.6x cropping factor simply crops

the image. The flat face perspective remains the same. "

 

The perspective _will not_ remain the same when using a 50mm lens on a full frame

camera verses on a 1.6 crop camera if you make the subject the same "size" in the

pictures (i.e. percent of frame), because you had to change the camera to subject

distance between the two cases to achieve similar results. You had to move away

from the subject with the 1.6 crop camera to get a similar picture to the full frame

camera, and this changed the perspective, right?

 

Or, think about not moving: Suppose you mount a 50mm lens on a camera with a 1.6

crop factor, select an aperture, and shoot a frame of a subject. Without changing

camera to subject distance (sensor plane to subject distance, if you like), mount an

80mm lens on a full frame camera and take a picture of the same subject at the same

aperture. What is the difference between these two images? As far as I know, the

only difference is that the first one has slightly more depth of field than the second.

 

Unless this is mistaken, then except for depth of field differences, using a 50mm lens

on a 300D is going to yield the same result as using an 80mm lens on a full frame

camera at the same distance to subject would.

Yes, I could have used film, but that wasn't really the idea. :-) Film would involve

processing and scanning - too much work. Film comes with its own problems as

well, instead of single pixel artifacts we'd have dust and grain. I never meant to

complain about digital, merely to make an observation.

 

If anyone is curious what the output looks like, you can see an example from a world

map circa 1646 that I own. The file is scaled up a bit beyond 6 megapixels for text

legibility, and is a 2 megabyte download:

 

http://www.spootopia.com/MapFinal.jpg

 

This map is framed and is behind glass, lit with available light. Canon 10D, Tamron

90mm f/2.8 macro at f/8, 1.5 sec, ISO 100, RAW w/ special post-processing.

 

Given that this comes from a 6 megapixel camera and not a scanner, I'm pretty

stoked. The idea of affordable 11 or 14 megapixel sensors is even more exciting.

I don�t know if this is common knowledge or not, but I ran across something recently

which I thought would be of interest. I have come to the conclusion that the CMOS

sensor on my Canon 10D has thousands of bad sensor sites which are mapped away

by the camera firmware for each picture.

 

A friend of mine who has an enormous collection of antique maps has been

interested in capturing images of maps digitally for quick reference / study. While

the proper tool for the job is of course a large format flatbed scanner, he doesn�t feel

like purchasing one and doesn�t need excessive resolution. We have from time to

time considered the possibility of accomplishing this with a digital camera. As

camera resolution has gone up, this has started to become feasible.

 

While we achieved reasonable results shooting 10D RAW images and post-processing

the images with standard software, many of the maps are black and white, or have

very limited coloration. Because of this fact, we considered that we might get better

results using custom processing of the RAW sensor data. Instead of performing

standard interpolation of the Bayer pattern data to get a full color image, I used a

custom scheme where the luminance information of the image comes from

uninterpolated sensor data (to which chroma information from traditional

interpolation is later added). For any map areas that have coloration this yields a

patterning that looks something like dithering, but it yields the most resolution for

legible B&W text and art.

 

The interesting thing that I found during this work, looking at greyscale images of

uninterpolated sensor data, is that there were single pixel artifacts in the images.

Had they all been dark spots I might have attributed them to sensor dust or

something, but many were light. The only reasonable explanation we have for this

behaviour is that these are bad sensor sites and the camera firmware is mapping

these away by stealing some nearby value from a sensor site with the same color and

substituting. Attached to this posting is a picture that shows two images of the same

area of the map, 300% zoom from actual pixels, uninterpolated sensor data. I moved

the camera slightly between the frames so that the identical image would fall on and

be recorded by two different areas of the CMOS sensor, so you can see the

differences which are these artifacts (e.g. the M in Meer).

 

Given the number of these pixel artifacts I can count in an area, I estimate that my

sensor has at least 2000 defective sites. In normal photographic situations these are,

of course, invisible and meaningless. In very critical work (with standard

interpolation) you can see these if you look very hard, but they are much more

difficult to detect because you aren�t missing (or stealing) an entire pixel from

somewhere else � more like just an R, G, or B component for a pixel. Given the

realities of semiconductor manufacture and economics, I guess this shouldn�t be

surprising, but I thought it was a rather interesting find.

 

If anyone would like more information, please email me or ask, but otherwise I won�t

bore you all with more details. Certainly though, anyone who gets a few bad pixels

and has Canon service map them away shouldn�t fret, as this already seems to be

going on in a big way unless someone can come up with another explanation?

 

-- Paul<div></div>

I just snagged one of these in "9" condition from Midwest Photo Exchange for $249. I haven't had a chance to shoot with it yet, but if the image quality is as good as rumored then at this price - and more importantly at this small size and weight (compared to other L glass) - I'll be very pleased.

I wish I could find something similar as well... I haven't looked at the EP-EX15, but I doubt it does what we want. The description says it extends the eyepiece 15mm from the camera body (useful for eyeglass wearers) but _reduces_ viewfinder magnification by 30%. Sound like it's going in the wrong direction. I did play with the Canon's Angle Finder on the 10D for a while at a local store. It definitely allows you to magnify the viewfinder image for critical focusing (1.25x or 2.5x), but seemed a cumbersome arrangement to use constantly for normal shooting.

I own both the 24-70 L f2.8 and the 17-40 L f4. Also, I had had the 16-35 L f2.8 for a short time, but exchanged it for the 17-40. The 24-70 is great optically, and the hood is very functional because of the cute design where the lens changes length while zooming, but the hood + lens doesn't change length. On a 10D though the effective focal length at the short end is ~38mm. I would use that as a primary point to decide between the two. Do you ever want to go wider?? If not, and you don't mind the size, weight, and extra cost, the 24-70 is a great choice. If I could only keep one of these lenses for a full frame camera it would be the 24-70. For my 10D though, I would keep the 17-40 if forced to choose because I would not be happy without real wide-angle coverage. I returned my 16-35 because it did not have as much corner sharpness as the 17-40 on the short end (where I would use it instead of the 24-70), and I also found the flare characterists of the 16-35 more objectionable than the 17-40. Wide open, however, my 16-35 @ f2.8 or @f4 seemed sharper than the 17-40 @f4 - this was the only thing I felt I was giving up really. The 17-40 also allegedly vignettes more than the 16-35, but this didn't bother me since I am almost always shooting on a 10D and not on a full-frame camera where it could be more bothersome.

I'm glad you're enjoying your 10D... I also bought one recently and think it's

fabulous. These sure look like hot/stuck pixels to me. It is not dirt on the sensor...

dirt on the sensor creates darkened blobby shapes. If you are seeing these in most of

your pictures (at reasonable exposure times), this isn't normal. Constantly editing

these out will get old... so I would either have Canon service the camera or get a

replacement.

Very coincidental... I purchased a 135/2 from B&H last week and experienced this same problem, and just sent the lens back for an exchange. In the range where it scraped, it was also harder to turn and would occasionally stick during MF. It did not have AF problems. I did notice that by continuing to turn the focus ring past infinity or past the nearest focusing point I could change where the sticky range was. Hopefully it is exchanged for a lens without this issue, as this was somewhat annoying during MF.