50mm lens comparisons

[rdenney]

I've spent a bit of time searching around the archives here, but didn't find anything specific and thought I'd introduce myself with a first post. If you want to skip the introduction, go to the last paragraph, heh, heh.

I'm an experienced photographer (I didn't say good, heh, heh), with years of large format, medium format, and 35mm photography under my belt. I still do medium format with my crazy Soviet stuff, and with a view camera using a roll-film adaptor, for those times when I just want to slow down and be traditional. I do mostly landscape and architectural work, but I don't limit myself just to those topics.

For documentation work and when I've needed something light and fast, I've used an EOS Elan II with a range of cheap lenses that work acceptably for that application. But I have too large a backlog of scanning work (I have a Minolta Multi II scanner, among others), and have finally succombed to the digital temptation and bought a 10D.

Right away, my lenses were problems. Three of the four were Sigmas. The 14 and the 28-70 f/2.8 are on their way to be rechipped, but the 18-35 was a rebadged Quantaray that I'll flog on ebay. I replaced it with the 20-35 USM, which has thus far produced excellent results. I also found the ultra-cheap 35-80 that came with the Elan that I had replaced with the faster Sigma. The remaining lens is a Cosina 100-400, which is marginal at best but okay for the once a year I need something that long to take a picture of distant test equipment readings or the like. I can also adapt my East German and Soviet lenses for manual-focus and manual stop-down operation, which makes my longer primes-- 180/2.8, 300/4, and 500/5.6, and EL-Nikkor 105 mounted on bellows--available. Of course, they are not convenient, but having them keeps me from really needing a good long zoom if I need something long for my landscapes. (I love the way the EOS cameras handle such adapted lenses, by the way, giving me aperture priority evaluative metering in stop-down mode--I used to do the same thing with an F-1 and old FL lenses.)

Thus, I've been toying with the idea of a 50mm prime. It seems to me that this lens is an ideal portrait length for the 10D, providing an 80mm equivalent field of view, and it would be much sharper than the Sigma zoom at large apertures. It would work about like the 180mm f/2.8 Sonnar on medium format (okay, it's a little shorter), which has become a favorite lens in my bag. I looked at the 1.8 II, but was scared I would break it if I breathed on it. The 1.4 seems pricey for a single-purpose lens.

So, I find myself considering the 2.5 macro. It seems to be slightly sharper, which would only be useful in copy work, but I do plan to do copy work and the 105 EL-Nikkor is a bit long for my copy stand. I figure I can do copy work for some of my scanning backlog, which is archiving a family picture album for my wife. It will get the same resolution as a scanner for most things, and the workflow is much faster because post-processing can be deferred until I need to make prints. I call this application a bonus if the lens I buy can do it well.

So, here's the question: Has anyone compared the 2.5 macro against the 1.4 50mm lens for portrait applications? Let's say that the extra two stops is not important to me, and I'd stop the 1.4 down to 2.8 in any case. And let's say that I don't care that the 2.5 will focus slower--it's not a problem for the sort of portraiture I do. I'm mostly interested in smoothness, color rendition, and quality of background blur. Has anybody with such experience survived to the end of this long post?

Rick "grateful" Denney

[justme_dc]

I wouldn't discount the 50mm f1.8 mkII just yet. I've had mine for close to 10 years and it has yet to break. I treat this lens horribly, I throw it in the camera bag without end caps on, I leave it laying on the ground when switching lenses, hell I've even thrown it across the room to an assistant so he could put it back in my bag. It is a tough little lens.

As far as background blur and Bokeh the f1.4 is an 8 blade design and the f1.8 is a 5 blade design. More blades=smoother "creamier" bokeh. I am not sure of the number of blades on the macro but would guess it is an 8 blade design as well. (anyone else know?)

My suggestion is this.......Buy the 50mm f1.8, see if you find the focal length useful for your purposes. If you do then spend the extra cash and get the f1.4 if you think you still need it. You can always sell the f1.8 or just use it as a handy camera body cap. The f1.8 costs about as much as a good filter or dinner and a movie. It won't break the bank. Give it a go you'll be suprized by it's optical quality.

Good Luck to you.

[rdenney]

dwerner wrote:
The narrower depth of field is one of the things that makes fast 80mms the magical portrait lenses that they are, and the 50s (not even the 1.4) really have that background blur that you need.... And, add to that the lower anguler distrotion and the more "natural" perspective that "real" 80's have and there's no contest.



Yes, you are correct that depth of field is tied only to the focal length of the lens. I hadn't thought of that.

But I don't get your second remark. The perspective on the subject entirely depends on the distance from the camera to the subject. If I use an 80mm lens on a 35mm camera, then I have to stand, say, 10 feet away to get a particular head-and-shoulders shot (just to pull some numbers out of the air for discussion's sake). To get the same image content on the 10D, I'd have to use the 50...from exactly the same position. So the perspective would be the same. Sure, a 50 is a 50 is a 50, and the 10D just crops the 35mm frame, but 35mm is just a crop of 120, and 120 is just a crop of 4x5, and so on. More importantly, an 80 is just a crop of the field shown by a 50. What matters is how far the camera is from the subject. So, if I have a particular field of view in mind and I want to work 10 feet from the subject, and to get that field of view I'd need an 80 on the 35mm camera, then I'd need a 50 on the 10D...and a 180 on 6x9, and a 300 on a 4x5.

But the depth-of-field issue could be significant. I'll experiment a bit to see how blurry I want the background to be. I can always adapt a Zeiss medium-format 80 to the EOS for experimentation--it's a 2.8 and has reasonable bokeh. In that case, I'll probably end up with the 50 macro (for copy work) and a fast 80, though it is pushing too telephoto in that format for my favorite working distances.

Rick "appreciative" Denney

[soumya63]

Hi Rick,

The cropping part is correct but perspective distortion will be different for a 80mm and a 50mm lens. So even you will get same magnification of the subject with a 50mm on 10D as a 80mm on 35mm body, but the perspective distortion will be different. Just draw a line diagram of a 50mm angle of vision and a 80mm angle of vision and you will immediately see the difference. So even though you will get a cropped view of 50mm with 80mm equivalent of magnification, but the perspective will look different.

Thanks

Soumya Mitra

www.mitraphoto.com

[rdenney]

soumya63 wrote:
Hi Rick,

The cropping part is correct but perspective distortion will be different for a 80mm and a 50mm lens. So even you will get same magnification of the subject with a 50mm on 10D as a 80mm on 35mm body, but the perspective distortion will be different. Just draw a line diagram of a 50mm angle of vision and a 80mm angle of vision and you will immediately see the difference. So even though you will get a cropped view of 50mm with 80mm equivalent of magnification, but the perspective will look different.

Thanks

Soumya Mitra

www.mitraphoto.com

With respect, I must argue the point. If you draw your line diagram, be sure you draw it from the edges of the frame. I've taken you up on your challenge--here is a drawing:

http://www.rickdenney.com/scratch/focal_length_and_format.gif

The closer the subject is to the camera, the greater the cosine error in the angles. In my drawing, which would be a very close focus distance, the effect focal length changes, and even though my frame heights are in scale, the focal lengths are drawn 48 and 68 respectively. For subjects this close, the 50 on the 10D would provide more like the perspective and field of view of a 70 on a full frame. At subject distances much greater than focal length, say, by a factor of 10 or more, the relationship between focal length and format is close enough to linear.

Note that the small object in the background covers the same percentage of the frame in both upper diagrams. That shows that the perspective is the same between a 10D with a 50 and a full-frame camera with an 80.

Only when I move the camera closer to the subject does the perspective change, as shown by the bottom diagram, where the more distant object projects a smaller image compared to the full frame. The relative sizes of projected near and far objects is, in fact, the definition of perspective.

For your lovely work in the macro range, focal length is such a high percentage of subject distance that just changing the focus setting will alter the perspective. But for normal distances, this effect is negligible.

But it's important to understand this distinction if you do lots of work with large format using cameras that cannot be hand-held, as I have done. You have to be able to picture the image without the ground glass. I can take a mat board with an open rectangle in it, and use it to frame the scene. If I hold it at arm's length, it's like a telephoto. If I hold it up to my face, it's like a wide angle. But the perspective is controlled entirely by where I'm standing. To change the perspective, I have to move.

Thus, changing focal length has exactly the same effect as changing the frame size--it only changes how much of the scene is visible. A change in frame size can be offset by a change in focal length, but perspective will remain the same at normal working distances.

Rick "who always chooses the perspective before selecting the lens, when using cameras that require a tripod" Denney

[Lazyg]

Maybe I can help with a visual example.


http://65.34.52.4/images00/perspective.jpg

[rdenney]

With all due respect, these images are taken from different camera positions. If they were taken from the same camera position, they would appear the same, except that the subject's face would be smaller in the image. If you took them all from the same position, enlarge them variously until the subject's face was the same size in each of them, and the cropped to show the same frame, then they would be identical. I have said all along that perspective goes with camera position, and focal length controls the amount of scene in the image.

Rick "unpersuaded" Denney

[PaulB]

Paul, "who also doesn't like Ricks' tone or style", nevertheless has to agree with him that perspective depends solely upon camera position.
Rick, the pretend laughing with the heh,heh is just as bad AS SHOUTING IN A POST.

[rdenney]

Sorry. Each forum has its own unwritten rules. I have been using the "heh, heh" actively in other forums for a long time with nobody being offended. I'll stop forthwith.

Frankly, I feel the same way about emoticons, but I know that others don't, so I understand the point of view, and will try to comply. If I forget, just remember that "heh, heh" = :)

There is a fine line between debating a point with clarity and focus, and doing so obnoxiously. But I know that I'm new here and am defending my statements vigorously without having built up any good-will account. I apologize--I've not found the line for this forum yet, and should have approached it more from the conservative side.

Rick "who wasn't pretending" Denney

[Neil Thomas]

Your diagrams illustrating perspective make a very good point. It's a pity it wasn't received more graciously.

There are people on the Canon SLR forum at www.dpreview.com who use the Canon 50 mm (and 85 mm) lenses, so it might be worth posting your question there.

[robertwgross]

rdenney wrote:
With all due respect, these images are taken from different camera positions. If they were taken from the same camera position, they would appear the same, except that Tim Allen's face would be smaller in the image. If you took them all from the same position, enlarge them variously until Tim's face was the same size in each of them, and the cropped to show the same frame, then they would be identical. I have said all along that perspective goes with camera position, and focal length controls the amount of scene in the image.

Rick "unpersuaded" Denney

Who is Tim Allen? Does he post here?

---Bob Gross---

[Pekka]

rdenney wrote:
So, here's the question: Has anyone compared the 2.5 macro against the 1.4 50mm lens for portrait applications? Let's say that the extra two stops is not important to me, and I'd stop the 1.4 down to 2.8 in any case. And let's say that I don't care that the 2.5 will focus slower--it's not a problem for the sort of portraiture I do. I'm mostly interested in smoothness, color rendition, and quality of background blur. Has anybody with such experience survived to the end of this long post?

Rick "grateful" Denney

I have used 50/1.4 with great success in all kind of portrait situations from two people and a horse to closeup facials (and this is non-porn terminology mind you) and it is simply excellent. But, about the macro version Canon lens work II tells that ".. and general portraits with pleasing background blur", also MTF there is very good, equal sharpness to 1.4 version.

So it seems 50/2.5 would be a great portrait lens, too.

Note that if total color accuracy is required, 1.4 is the most accurate 50mm.

[rdenney]

My mistake. I thought it was a picture of Tim Allen, the TV comedian. I see now that only the wide-angle shot looked like Tim Allen.

Rick "just noticing the 'edit' button" Denney

[rdenney]

Thank you. That's just what I was looking for.

Rick "appreciative" Denney

[soumya63]

rdenney wrote:
With respect, I must argue the point. If you draw your line diagram, be sure you draw it from the edges of the frame. .....

Hi Rick,

Sorry I was busy and have not visited this forum for few days. First let me thank for your diagram to explain what is correct that if you crop a photo you will get the same magnification of a longer lens. I have never questioned that. This is a well know fact.

What I said was the perspective distortion (due to lens aberration) will be very much apparent even from those crops.

For example take a close-up portrait shot with a 20mm lens. Which will behave like a 32mm on 10D. Now shoot the same subject with a 35mm lens fitted on a 35mm camera. Magnification wise both of them will show the face but you can immediately see the difference of the 20mm shot and a 35 mm shot.

Same will happen with 50mm on 10D and 80mm on 35mm camera. The lens characteristics are so different, you will see it immediately even though both of them will fill the frame with same subject.

The line diagram part is a mistake on my part as it is not possible to draw a simplified line diagram for a complex objective to prove the point. Moreover, you are correct that angle of view for the 10D cropped image and that of the 35mm lens on a 35mm body will be the same. But the optical characteristics and lens distortion for 20mm and 32mm will not be the same, which will cause them to look very different.

Lastly, I quite like your style heh heh :D It is a nice change from those dumb smiles.

Thanks

Soumya
www.mitraphoto.com

[rdenney]

Then it must be something other than perspective projection, and now I'm truly curious. I have made these conversions many times when making backup images of a given subject using a different format (and that's what the 10D is--a different format than 35mm), and have never noticed any difference in the results.

Your suggestion is that the perspective distortion is caused by lens aberrations. Geometric distortion? It's possible, I suppose--of all these probably only the 50 macro is truly rectilinear. I frankly can't see the process in my mind as to how it would be noticeable. Perspective seems to me to be the relative magnification of near and far objects, and so I can't dissociate perspective and magnification.

Let's put it to the test. I will conduct an experiment this weekend. On my Elan II, I'll take a still life (something very three-dimensional) using the 80mm setting on a zoom lens, well stopped down to ensure adequate depth of field. Then, I'll take the same picture from the same place with the 50 on the 10D (I went ahead and bought the 50/2.5 macro), and post the results. Of course, it'll take a few days because I'll have to wait for the film to be processed (oops--I almost said 'heh, heh'). Just for fun, I'll take the same picture with my view camera and a new-to-me 6x9 roll-film back (which I need to test anyway), using a 210mm lens (which is about like a 54 on the 10D). Again, I'll use the same camera position. For comparison, I'll take a few at other camera distances with different focal lengths just to eat up the roll. I know--I'll take a picture with a 47mm lens on the 6x9 camera, and a picture from the same position with the 50 on the 10D. We would both predict, I think, that the latter would look just like a 15x23mm rectangle cropped out of the 6x9 frame.

Will this experiment reveal the effect you are describing?

Rick "stay tuned" Denney

[soumya63]

rdenney wrote:
Will this experiment reveal the effect you are describing?

Rick "stay tuned" Denney

Sure. But may I suggest of using something wider like 20mm or 17mm? Longer the focal length, less prominent is the effect.

Moreover please shoot with biggest common aperture.

Rick, you are right it is not the perspective, as in the both cases they will be the same. It is a wrong statement on my part to say perspective will be different. What I meant was the qualitative difference between a wide and normal lens will be very easily noticeable even though magnificationwise they will be same on 10D and 35mm body.

Right now I have no 35mm body lying with me, so that I cannot perform the experiment myself. It will be interesting to see your results. If they appear the same, I will be happy to correct my notion.


Thanks

Soumya

www.mitraphoto.com

[rdenney]

Okay, now I think I see where you are going. The lens with the shorter focal length will have more depth of field, because depth of field is solely related to focal length and not to angle of view or format (unless you enlarge to the limit). Thus, the 80 on a full-frame camera will have a smoother background and crisper selective focus than a 50 on a 10D at the same aperture. That will definitely affect the appearance of the result.

Another responder had mentioned the same thing, but that exchange had gotten heated and both of us have deleted those messages, probably before you reviewed the thread. He was right on that issue, and it was a good call. On that point, I actually favored the 1.4, which might be similar to a 2.8 80 in full frame, but I went with the macro because of my need for a copy lens. And the 85, 1.8 would be even better, of course. (It's too bad--I have an excellent Russian 85mm f/2 lens, but no way to adapt it from its Leica thread mount to the Canon body.)

My idea was to eliminate that variable to isolate the perspective projection in the test, but in fact that is the effect that interests you. I'll do it both ways so we can see just how much effect there is. I have a fixed-focal 80mm Zeiss lens for medium format that I can adapt to my Elan, and I'll use that in the test instead of my much-slower standard zoom. That way, we can compare at f/2.8 instead of f/4. And both lenses will be in their sweet center.

I'll try to find the aperture on the longer lens that provides the same background blur as the 50 does wide open. That would be good to know, too.

Rick "seeing progress in this discussion" Denney

[Tom W]

Rich, just to make sure I'm thinking kind of straight (or to correct my frame of mind here), let's say I decide to take a picture with my Elan II, looking down a narrow street. I'm ignoring depth-of-field here (which you explain well in the last post), but instead viewing the perspective of physical depth of buildings which appear "shallow" when I look at them with telephoto lenses.

So, with that in mind, I take a certain picture while standing in a cetain spot with the Elan II with my 70-200 zoom and I find that the frame I want occurs at 160 mm of focal length.

Next, I borrow a digital camera, in this case a 10D and stand in the same spot. However, due to the smaller sensor size, I set the focal length to 100 mm to frame the same picture.

Now in this case, I am expecting the same perception of building depth - that is, that same "telephoto" look to all the buildings in the picture. They don't look "deeper" with the shorter focal length, but look the same due to the relative field of view being the same. (again, I'm ignoring depth-of-field here - my focus is on the appearance or illusion of how distance from the camera appears shorter when telephoto is used).

I would think that this is so, just as it seems to be so when I use my little 22 mm telephoto lens on my S-400 digital. At 22 mm, it gives the same apparent FOV of approximately a 105 mm lens in 35 mm terms.

[rdenney]

Tom W wrote:
Rich, just to make sure I'm thinking kind of straight (or to correct my frame of mind here), let's say I decide to take a picture with my Elan II, looking down a narrow street. I'm ignoring depth-of-field here (which you explain well in the last post), but instead viewing the perspective of physical depth of buildings which appear "shallow" when I look at them with telephoto lenses.

So, with that in mind, I take a certain picture while standing in a cetain spot with the Elan II with my 70-200 zoom and I find that the frame I want occurs at 160 mm of focal length.

Next, I borrow a digital camera, in this case a 10D and stand in the same spot. However, due to the smaller sensor size, I set the focal length to 100 mm to frame the same picture.

Now in this case, I am expecting the same perception of building depth - that is, that same "telephoto" look to all the buildings in the picture. They don't look "deeper" with the shorter focal length, but look the same due to the relative field of view being the same. (again, I'm ignoring depth-of-field here - my focus is on the appearance or illusion of how distance from the camera appears shorter when telephoto is used).

I would think that this is so, just as it seems to be so when I use my little 22 mm telephoto lens on my S-400 digital. At 22 mm, it gives the same apparent FOV of approximately a 105 mm lens in 35 mm terms.

Your expectation is correct as long as you are standing in the same place.

To summarize, focal length does not control perspective. Where you are standing controls perspective. Focal length just controls how much of the scene is in the picture. Format also controls how much scene is in the picture. A 100 on the 10D includes the same amount of scene as a 160 on an Elan II, and if you don't move the perspective will be the same. The focal length and format offset each other.

Rick "whose experiment is done but the film hasn't made it to the lab yet--but the viewfinder tells the story" Denney

[bluebomberx]

Here is one photographer's opinion on the normal lenses offered by Canon.

http://emedia.leeward.hawaii.edu/frary/toolbox5.htm

[rdenney]

bluebomberx wrote:
Here is one photographer's opinion on the normal lenses offered by Canon.

http://emedia.leeward.hawaii.edu/frary/toolbox5.htm

Thanks. I had seen this site before, but I'd lost the link and couldn't find it.

Rick "appreciative" Denney