Well... I could do so, but in doing so, I would only get one copy and so I would not have a statistically valid sample, whether positive or negative. I read only the Alternative forum on FM, so I don't know what people in the Canon forum there are saying. I don't own Canon equipment any more, just Leica M and Contax 645/Sinar MFDB (and a couple of old Hasselblads).

Then how about pointing to a test of a bad copy? Or something where someone complains of poor performance?
You say 1 of 3 reviews you've read are unsatisfied users, I'd like to see this as I haven't seen any bad reviews or complaints. Not being able to use a filter isn't an example of a bad copy.

I picked up a 17mm TS about 10 days ago and shot this project last week. Immediately, it became my favorite lens! Before the 17, I would use my 28mm Mamiya more often than not with a P45. I also bought a 5DMk2 and am equally impressed with it. Attached are some of the images taken of this new building project, they've all been retouched but I think you'll get an idea of what the lens is capable of. Jim

Click to view attachment
Click to view attachment
Click to view attachment
Click to view attachment

Well, perhaps you just got as lucky as everyone else and got a good copy ........

Sarcasm aside, it's hard not to be impressed with this lens.

I may sound heretic but first and second pic have MF quality (webwise) Wow! A dozen wow's! Definetely it's my next lens.
Eduardo

[quote name='haefnerphoto' date='Nov 15 2009, 09:50 PM' post='325135']
I picked up a 17mm TS about 10 days ago and shot this project last week. Immediately, it became my favorite lens! Before the 17, I would use my 28mm Mamiya more often than not with a P45. I also bought a 5DMk2 and am equally impressed with it. Attached are some of the images taken of this new building project, they've all been retouched but I think you'll get an idea of what the lens is capable of. Jim

No doubt the 17 TS-E is an impressive piece of glass. Call me crazy but I just don't see much appeal to a 17mm FOV on full-frame, unless shooting tight interiors. Backing up will almost always give you a more pleasing and natural looking perspective, especialling for things like architecture where the 100+ degree FOV's will create a distorted and even misleading impression of the building's proportions.

Granted, sometimes backing up is not an option, but I get the impression some photographers are using these really wide FOV's just because they can. (There are a lot of truly awful pictures taken with the Nikon 14-24m to back this up. IMHO).

Personally I'd much rather have something in the 30-32mm range to slot in-between the 24mm and 45mm options. On the wide end 20mm is about is wide as I would want to go.

I've been shooting with the 17 tse since early August and it's a stunning lens. One needs to be careful using it as things can get wacky pretty quickly if you are too much off angle. The lens has a wonderful 3D quality to it, even more so than the new 24 tse II. In the right conditions (e.g. you can't back up, there's nothing in the foreground getting too distorted, etc.) it works wonders.

Images #1, 2, 3, 5 & 6 in this gallery were shot with it. #1 is a composite of three horiz. frames shot with the 17 tse.

Every image in this gallery except #3 was shot with it.

The three images in this blog post were also shot with it. You can see in the first image that the desk is getting too stretched out but in the last image there's enough other detail that it doesn't matter.

To me the 24mm tse II is the real workhorse on architecture jobs. It gets the most use. If an image can be shot with the 45 tse then great but more often than not it's the 24. When that doesn't work the 17 is perfect and a welcome change from having to use the 16-35 II @ 19mm or the 14mm II.

I'll try to find some of the posts again, but please understand that 1) since my personal observations are worth exactly zero to anyone with an opinion of their own, and 2) it is really dull to re-read threads trying to find posts that you read casually a few days or weeks previously, I may not find them again. If I do, I will post links here, promised. Maybe the few posts that I have seen on the matter really weren't statistically accurate, it is quite possible, I just don't know. It just struck me from the posts I have seen so far that while many are raving and happy, a few are not, and sold it again.

One thing about this lens that I find interesting, and which must have Nikon throwing a hissy fit, is that just when Nikon released a bunch of T/S lenses to match or exceed Canon's stable at the time, Canon responded with a dramatically updated lens (24 TS-E II) and a game changer (17 TS-E). People who would have switched to Nikon are now staying with Canon again. Quite a coup.

The 17mm looks like a nice piece of glass, but I feel that the investment is not worth it. First of all this lens is approaching the price of MF lenses, and Schneider and Rodenstock are not going to let Canon or Nikon ever reach them in quality (a specialist will always out preform a generalist). Would it not be better just to make the investment into a MF system and get the equivalent lenses?

Second, 17 mm lenses on a full framed DSLR have a pretty big stretch factor, and as you get towards the edge of the image circle, this is only going to get worse. You can get the same view with little stretching by just stitching on a 24 (or 35 in MF). Now mind you do not get the same amount of flexibility with stitching on a t/s lens as opposed to a view camera, but I still feel that it is to pricey for the limited amount of time it would be used.

I right now have a 12 to 24 mm zoom and insofar as professional use, I have only ever used it once because of the stretch factor, and I know I could have gotten that shot with stitching, but it was not necessary. Now some of you are probably saying, it saved you time having that lens. Maybe 5 minutes, but when you put 1 to 4 hours into a single interior, an extra 5 minutes is really no time at all.

Last, I have to agree with Jeff Kohn in that some photographers just get things to get them, even if the investment is not worth it and do not think it through. If you really need this lens for your work, get it. Remember though that Weston took almost all of his images with the same lens he bought for $15 in a junk store and his work is studied by many.

I would rather just save, wait, and make an investment into a MF system when the time is right.

Please explain your concept of "stretch factor" and how stitching a 24 will eliminate this.

Post deleted.

I agree with Jeff Koon. The 17 will excel in tight interiors. The 24 is more suitable for exteriors. If you already own any MFDB then investing in Rodenstock or Schneider is a whole lot better. But in my case and in several thousands more cases around the globe the new glass from Canon is heavenly. If I could justify a digital medium format, no doubt I'd go for it. But my clients can't afford it and I'm sure they will be delighted with the Canon results. On counterpoint, too bad for me that Canon is coming with this fantastic glass because my DMF piggy bank is not getting any fatter. As someone said here before, Hasselblad and PhaseOne worst enemies are Canon and Nikon. Let's hope they don't become their nemesis.
Eduardo

I'm just amused that there is all this debate on DSLR vs MFDB on a forum that by title confines itself to MFDB. We're such dorks... we take our gear and our practices so personally.

Now on to more important matters... I want to get myself a dirtbike.... 2 stroke or 4 stroke?

-Chris <--- has been drinking

Well, it's likely that there are more people in this subforum than others, who can really give a more balanced view of the pros/cons of each system due to having used both. And that's doubly true given the original poster's interest in architecture and interiors, as there are a lot of knowledgeable pros from that genre who hang out here (yourself included).

I don't shoot DMF and probably won't anytime soon, but I hang out here because there's a lot of great discussions about photography, especially more advanced topics (just look at a few of the recent architecture photography threads).

Oh, don't mind me, Jeff. I was just being snarky.

(This is being resubmitted to the thread I originally intended it to be)

For each of the four images I posted earlier the 17mm T/S allowed me to make a better image. In each case backing up was not an option, either walls, trees or water would have come into play (and not beneficially). The images were successful enough, not great, but worked for my purposes (and the client's). Here's another image shot Saturday with the 17 that I'm very happy with. I've printed it to 18" across and it looks great but we know that these files will go even larger! I prefer my P45 files but really like the ability to shift both horizontally and vertically (haven't used the tilt yet, at least not on purpose). I'd also love to have a small technical camera but it's hard to justify the expense (keep in mind I live in Detroit) so the 17mm/5Dmk2 combo made a lot of sense. Jim

Click to view attachment

The equiv MFDB lens is the rodenstock 23HR @ $6500 for an unmounted version.

For the same $ one can buy the 17 tse, 24 tse II, 45 tse and 1.4x and still have change left to put toward a 5d2.

Then the roughly equiv lens to the 24 tse II is the rod 28 HR, which will set you back nearly $5400 for an unmounted one.

So I wouldn't say they're approaching the price of MF lenses.


I have both the Canon and Rodenstock lenses we're talking about here, as well as the rest of the systems to go with the lenses (Alpa and Canon).
MF is better, but unless you can really afford it it's not worth 3-4x the cost for a basic system.

LOL! 4 strokes are where it is at now. You are from Chicago...you need to take a drive to Redbud this spring The pro motocross is awsome.

http://redbudmx.com/

omg what happened here; left for a day and the thread nearly doubled in size!

I am note sure I understand that part. I made it in bold for your reference. Are you saying your 17mm can be shifted in both horizontal and vertical at the same time just like a view camera?

Okay, the stretch factor. When using a lens that is as wide as a 17 mm lens, you end up with stretching of objects on the edge of the image circle to make up for the wide view of the lens and this only gets worse as you increase the angle your are viewing the subject at. So if you are taking a shot of a subject which has no depth dead on with the camera parallel to the image plan (like the front of a house or building where you are not looking into the architecture) you are fine and you will receive an accurate description of what the subject looks like. But as soon as you start to turn the camera and look at the subject on an angle, the edges of the image will start to be stretched. For instance you could take an image of a house and have it look three times longer than it really is. I have examples of this and will try to post them tomorrow; I am too tired to find them right now. I will also try tp post an explanation of why this happens in terms of lens engineering, but am much too tire to post that.

Insofar as prices are concerned, when I wrote the post I was thinking of the cost of the new 24mm II compared to the 35 mm Schneider, 2000 vs 3500, only 75% more. Schneider lenses are the lenses I most look at and it just so happens that they are symmetrical focus lenses which are cheeper to make than retro focal lenses. Also, 17 mm on a DSLR is somewhere between a 23 and 28 mm on a MF system, so comparing the 17 to the Rod 23mm is not really a fare comparison.

I would also like to add that I also fell in love with photography when I was introduced to large format and 4x5 work and would not be a pro if I did not get into working in that type of style, so I may have a bias when it comes to this topic.

Ok, this has been covered many times, so I won't go into it (there's a thread on it, about perspective and stitching and wide angle lenses) but lenses (rectilinear) have no power to stretch as you describe. It's a function of perspective. If you stitch 2 24mm shots to get the same angle the 17 gives you the "stretch" will be the same as long as you don't change camera position.

Maybe Bernard L can repost the test shots he did showing this.

My experience proves the same. You may not want to believe it, people love to hold on to the idea a "wide angle" lens "stretches" or "distorts" at the corners, and the great thing about stitching is using "longer" lenses to avoid that distortion, but try it and see. I can even do it on my Alpa with my Schneider 35xl and get the exact framing in 3 or 4 frames I'd get with my 28HR, only a larger files. But the corners wouldn't be any more or less "stretched".

Also, according to Alpa's calculations and my testing the 17 on 35ff dslr format is equiv to a 23-24 in 36x48 mfdb format. I attached a screenshot, you can download the Alpa calculator for free, but if you disagree test it and see. I'll agree the 35 and 24 are closer than the 28 and the 24. But 75% more isn't nothing, it's roughly half the cost of a 5d2 body, or a 45tse plus 1.4x, plus the all the added costs of MFDB.


Btw, I've been shooting LF professionally for 20 years (almost exclusively until a few years ago), the Alpa allowed me to switch to digital, and 35 digital (for what I do) has only been acceptable to me since this spring when I got the 24 tse II, followed shortly by the 17, used in combination with the 5d2.

asf it's true that a rectilinear wide-angle lens and a stitch of multiple shots with a longer lens will have the same perspective or 'stretching' if you use a rectilinear projection in the pano software. But with pano stitching you have other projection options that don't 'stretch' as much as you get away from the center/vanishing point. Personally I don't much care for rectilinear projections beyond about 90-100 degrees, regardless of whether it's from a single image or a stitch. Even at 90 degrees I'll usually pick something other than a rectilinear projection when stitching, precisely because I don't like what happens to objects away from the center.

Thanks for the clarification Jeff

Yes, there's lots of variables possible when it comes to stitching. As far as I'm aware stitching arch and interiors is usually done with rectilinear projection (always for me as I use Alpa shift adapter, stitching by shifting a tse lens or on the artec is basically the same thing). There may be instances where other forms of projection would be more pleasing. I'll try to test that out.

While agreeing in general about the principle of the angle and the shooting location, there are different tradeoffs lens makers can make in the corners of superwides. For example, allowing a little barrel distortion can make the image look more relaxed in the corners, but of course is no good for stitching.

I recall reading about some of these tradeoffs at one point in one of my theory books, but don't recall where right now (my memory is really going).

One thing I do remember is that a strictly accurate superwide lens will often look artificial and stretched if you don't stand at the equivalent point when viewing the print. In other words, using a super-wide, and then standing far back will make the corners look stretched. You need to stand closer, so that the angle of viewing equals the angle of the capture.

It's true that rectilinear stitching may often be needed for architecture to keep straight lines actually straight. If you only have to worry about keeping verticals straight (and horizontals that pass through the center/vanishing point) then a cylindrical projection can give a very pleasing result. Additionally, a good stitcher like PTGui will give you some control over the projection parameters such as horizontal and/or vertical compression. For landscapes, I most often use PTGui's Vedutismo projection, which gives very natural looking results IMHO. I haven't tested to see how well it would work for architecture though.

From reading the recent post, it appears like we have come to the conclusion about how and when optics with wide views will stretch the image and give inaccurate renderings. Lets get back to the point of price; I feel what I said earlier is still right, you would be better off putting the money into a MF system then getting these lenses. Lets look at two options for a system.

For about $8500 you can get a:
Canon 1D, 21 mp recording at 14 bits per channel, 10.5 stop range
the new Canon 24 ts

For around $16,000 you can get a:
Alpa 12max
Schneider 35mm
Leaf Aptus II 5, 22 mp recoding at 16 bits per channel, 12 stop range (and unlike Phase One's entry level backs, this one still has a large sensor)

Now as far as I see it, aside from saving you 8000, the DSLR system offers you a lighter camera easier to transport, automatic settings, and the ability to use very long lenses. Aside from the last one (and the saving of 8k does sound good), the other two I do not care about. You could also make the claim that it would be easier to find someone to look at your canon if need be.

The MF system over the DSLR offers greater optical resolution, a much larger range and color gamut (remember bit and stop growth grows exponentially) giving you more freedom in PS and more lush transitions, the ability to shift on both axes, sturdier construction, levels on the camera, etc. Also, when the sensor becomes obsolete, you do not have to replace the whole body, only the back. Last, the pixels on this back are larger in size then on the Canon, decreasing the chance of getting apo-chromatic aberrations (it is Schneider glass though so you would not have to worry about it anyway). Although you do increase you chances of getting morie by going with an entry level back, but that would not be any different then the Canon.

You could even lower that price by about 2500 and get a Cambo Wide DS.

Now 8000 more is a large sum, but if you are spending 8500 anyway, why not just put out the extra to get a system better equipped to shooting the subject?

Only problem with your comparison is that a Canon 1D offers no real advantages over a 5D11 for shooting architecture. So the competitive price for the Canon alternative would actually be 4900.00 not 8500.00.

Good point. That does make the DSLR more attractive, although I still would go with the other system.

the first point is that the aptus 22 works with 14 bits too. the two additional bits are EMPTY/ kind of noise floor.

and second is that you will have to go wider if you want to work seriously in architecture photography. so count in the 28 or 23Hr as well, and if you do that you already get an entire system with the canon with a 17/24/45/90 1,4 extender and 100-400 L is lens. all together will cost app: $ 10.000,--

and if going with the XY most likely you will have to buy some canon or nikon equipment too ( backup and at least long lenses ) . i dont know people who dotn own an additional 35mm digital aside their mf setup.

but all this does not say it does not make sense to go with mf systems AND a canon/nikon,-
but certainly i wouldnt advise that to anybody who wants to start to come in architecture photograhy or who is studying photography. and less to lease all the stuff ...

Kirk and Rainer beat me to it, but there is no reason for the 1ds3, the 5d2 is actually preferable. 14bit vs 16bit is another fallacy that won't die.

And if you go Alpa (I use the Max) there are many expensive accs you will need, compounding the price difference. The adapter for the digital back alone is $1100, add in a dongle for $1300, stitch adapter for $350, a minimum of a 2nd lens for another $3500. I have 4 lenses for my Alpa system, it's not really enough.

Have you used these systems? If you want to get an MFDB you should go right ahead, but unless money is not a concern for you it's better to know what you're getting into before dropping a minimum of $20k on a basic one lens Alpa/Leaf setup as opposed to less than half that for a 4 lens Canon setup. Or recommending others do the same.

I started using the 5d2 as backup/long lens/above 100asa addition to the Alpa/Leaf and have come to realize just how good and versatile it is.

Also: The levels on the Alpa are fantastic, but the shoe level I use on the Canon works perfectly well. The larger pixels on my Leaf back are a nightmare of moire in some situations, Canon has smaller pixels and an AA filter and no moire (I don't understand how you can say the Canon is just as susceptible).You have half a point about shifting on both axes, but it's only more convenient on the Alpa, I can and do shift in both axes on the Canon.

Rainer and ASF beat me to it....

I had written (but ran out mid-post):

Neither of those setups is a "system." Both of those setups are kits in my view. The difference being that what you listed above is fine if you are doing work for yourself and only have yourself to please. But neither cuts it if you are doing professional work and have clients to answer to.

As asf pointed out yesterday, the true cost differential is a minimum 3-4x more for a digital medium format based system as compared to a dslr system.

--

What you do is up to you but for people reading the thread having a true sense of costs, trade-offs, and other issues is preferable.

exactly the same hapens to me at this moment.
although my beginning in digital was with 35mm too ( with kodak slr and canon 1ds ), at a time where the 24xl already did not exist and so no real wide angle for mf. at that time i mixed digital 35mm with 4x5" film work. than came a long break, using mf nearly exclusive except for details and super wide shots, but now the 35mm thing again looks very attractive, and for the first time as stand alone system too.


.

Isn't that Leaf back using the same sensor that was in the entry-level Mamiya from a couple years back? That wasn't exactly a highly-regarded sensor. Not only is it not a true 16-bit sensor, but it's far less versatile than today's DSLR's with regard to ISO range, frame rate, not to mention the lack of live-view. Some might consider the lack of AA filter an advantage, but not me. A 22mp 36x48mm sensor is not only going to have issues with moire but also edge aliasing and color aliasing. Even the D3x seems like a bargain in comparison, all things considered.

Sorry, this is another rumor like 14bit vs 16bit that won't die.

The Aptus 22 is nothing like the Zd

There was a thread about that recently too

14 and 16 are the same? This is the first time that I have heard this. Would not 14 bit have 16,384 level and 16 have 65,536? I am sorry, but I am a math man and get side tracked looking at the numbers sometimes.

Now I do have a Canon system and that is how I am shooting right now and I have used a MF system on occasion, and money is a concern, and I am not looking to get just one lens (I through that out there as an example). I kind of miss spoke before with where I was. Now It just happens that the Canon system annoys me, not so much the tech, but the camera itself. I learned photography off of a view camera but when I went into professional photography two years ago I did not have the capital to purchase a MF system, film did make any sense to continue to use (and I so wish I had access to a darkroom), and then the economy dropped out. I would get along with it much better if they have spirit levels in the camera; getting the camera level and completely vertical drives me nuts. Never thought of getting a shoe level; get that at Home Depot? Now as the economy is starting to recover, the amount of work I am getting is picking up and MF systems are looking with in range, especially considering the lower end leaf back. For one just starting to come into the medium, the canon and ts lenses are the best alternative, but for me, I feel that it would be better to put the capital in a MF system instead of the new generation of ts lenses. I have never had an issue with morie in MF with my limited use, so it seems I am wrong there. Will have to look into that more.

Also, concerning the 17 ts, I feel that it is too wide and I never work that wide. There are many photographers who are much more experienced then I and much more popular that do not go that wide either. I just do not like working with the large angle of view for the problems you run into with stretching and limited use. I do not think that investment is worth it for me. Now when going to MF, I do anticipate getting a 28 mmn but not a 23 for the same reason.

P.S., I really need to stop posting to this topic; it appears to pi**ing a lot a people off.

Joe, I wouldn't trust those hot shoe mount bubble levels... there's so much slop in the mount and who's to say the hotshoe is perpendicular to the sensor anyway? I carry a small "torpedo" level and just place that across the front of the lens, whether dslr or 645 to get plumb, for level I count on grid focus screens.

-C

Level:

http://www.bhphotovideo.com/c/product/1123..._Flash_Hot.html

I have 3 of these and use with 2 Canons. No slop and they are aligned with the sensors.

As for 14bit and 16bit, your numbers are right, unless those last 2 levels are empty, which is the case here.

Math exists in its own universe. Ours isn't so perfect. We have salesmen and advertising, and the internet which loves to perpetuate rumors and false info.

You don't have to take my word for it - ask some of the scientists here (Panopeeper where are you?), or ThierryH who was the Sinar rep for a long time, or Rainer who developed the ArTec and the sinar software workflow (with Brumbaer), or call Rick Adshead at Leaf, or do you own research ...

Sounds like you want to go MFDB. Good choice, do it. Go Alpa and Leaf. I have them and they are the best. But right now if I had to give up one system I wouldn't think twice.

I AGREE-the shoe is not square. I carry the bubble level in the shoe but use it by holding it against the viewing screen (with some black masking tape on the side which contacts the view screen to avoid scratches). There are a couple of advantages here. One-the viewing screen has proven to me to be parallel to the sensor (at least on the 5D and 5DII). Two-the bubble level is small, cheap and conveniently carried in the shoe. And three and most important, when shooting at twilight, the image on the screen provides enough illumination to read the bubble.

I'm not sure what thread you're referring to, got a link? Some of the supporting electronics may be different (I never said the backs were the same, just the sensor). But it was my understanding that Kodak stopped making a 22mp sensor quite a while ago, only Dalsa still had one so it pretty much had to be the same as the sensor in the Mamiya (I may have it backwards on who stopped making their 22mp part and who didn't, I can't remember). At the very least, the sensor design is a generation or two old.

Now I'm not saying the Leaf Aptus II 5 is junk but I am questioning just how much better than the D3x or 1Ds3/5DII it really is, especially considering all the compromises that come with these backs as far as ergonomics and versatility. At base ISO, the image quality may be better if you can live with the artifacts that comes with not having an AA sensor with such large photosites.

I am glad to see more DB's breaking the 10K barrier though. Maybe eventually we'll have a semi-affordable back with good pixel density and live-view.

I'm not sure what's more amazing about those hot-shoe bubble levels: that companies like Hama have the nerve to charge over $30 for them, or that photographers actually pay over $30 for them. Talk about the most over-priced photographic accessory that doesn't have 'Leica' in the brand name. A while back there was a link on another site to a place where you could order those same levels for 4 or 5 bucks (sorry, I don't have the link anymore or know if they're still available). But like others mentioned, I found them not to be 100% reliable.

That said there are other ways of getting the camera level without having to upgrade to a view camera. Some Nikon DSLR's have a leveling function built in that is pretty accurate. And lots of ballhead and quick-release clamps have bubble levels in them.

I don't think I have ever paid more than like 8 bucks for a bubble level and if you use them the way I suggest they work fine (that is not relying on the level foot and camera shoe).

If you want a higher tech solution than the bubble level (not necessary IMO KISS) see this short discussion on the ZigView from my blog:
Assignment Notes-leveling the camera

Yeah it seems kinda overkill; the calibration is nice, but having yet another device to worry about batteries for is not.

Now if there were a device that could be used to help get the camera parallel to a subject that might be something to consider, as that can often be more difficult than getting the camera level.

If Moire is a concern, you may consider Hasselblad see my post (# 32) on:

Moire & Phocus

For accurate fast leveling I really like mine Multiflex head which seems to be a clone of the much more expensive Arca Swiss C1 Cube.
More expensive but very fine to use.

That's what I use.
The Multiflex is excellent.
http://www.photoclam.com/
About $950 USD

N.B My Rolex is real!

Thanks all for the level suggestions, and I will not pay that much for a level Kirk. Those levels supplied by camera shops are more over priced then the old lens board, a piece of metal with a hole drilled in it. And I would love to see the math and fractals on how they work the empty bit space on MF backs; any sites that explains that in detail. Not that I do not believe you, just as a person who studied math in college, I would be interested in seeing the fractals.

Also, I did research the Aptus II 5 more and it appears that it does use a sensor at least one generation old. I looked at a couple of test images Leaf has on there website and it appears to have good image quality in the are that are in focus, but when you get into the areas out of focus, some noise starts to appear; I do not have leaf software and ran it through Adobe Raw. With that said, I still prefer the view camera, not because of the tech, but because of the camera itself.

I've used a Bogen 410 geared head for years. Does everything I need and handles my 4x5 too=but costs just $219. 410

When you get your Alpa you can have a nice discussion with them about how overpriced the back adapters are, and why their mounted lenses are more expensive than anyone else's. Or with Leaf about the price of the dongle, the batteries, the charger, their FW cables and sync cords ... Or with parent company Phase about their Value Added packages ...