DJ - yes I agree - the 1Ds3 really is a fine camera. I'm enoying using mine alot. With its many new features and technical advances it is much more use-friendly than the two previous models and produces great image quality. While Canon has difficult issues of corporate culture and quality consistency to deal with, they do know how to make very good cameras.

As for the near-by post from Dinarius generalizing about the superiority of Nikon lenses - I'll ask the same question I keep asking when I see these off-the-cuff generalizations - what is the body of evidence to confirm this contention? Both companies make lense that range in quality from excellent to so-so. Until I see convincing evidence one way or another I would refrain from generalizations about which company makes better lenses.

I've asked this question too, asked for links to the studies or specific lenses that beat Canon's, and I have yet to find any evidence. I asked specifically what Nikon lens is superior to the Canon 70-200IS L in that focal range?

What is the relevance of this discussion in a thread headed "Canon 1ds3 alignment problems, Viewfinder vs image alignment"?

I'm sure another thread, maybe called "Lens masturbation for the uninitiated" could be started with just a few keystrokes... and we can get back on topic here. I notice others have tried to suggest this with such subtlety it has obviously been overlooked - please forgive my own lack of...!

Has anyone else had a MkIII back from Canon Service yet? - Any comments/thoughts good or bad?

Following my earlier disappointment I'm still on the fence about buying this camera.

Carl,

When people have teething problems with an 8000 dollar camera they naturally ask questions about whether they should look elsewhere and it becomes a slippery slope once the discussion meanders into related issues such as the qualities of the options. But you are correct, this thread has slid down the slope a bit more than it should have. But if only all of life's problems were so.

Now as for the camera, let me put it this way: I have now made about 2000 captures with this camera. I just returned from a wonderful 9 day field shoot in Yellowstone National Park and the Tetons (with "Travel Images" led by John Baker). It was really cold outside and the camera performed flawlessly. Focusing is fine and mine does not appear to have an alignment problem. The battery performance was astounding. A colleague with a 1Ds2 (Ni-cad battery) was having frozen battery problems while mine (L-I battery) just kept performing with tons of capacity to spare - which I can read to the accuracy of 1 percentage point on this camera - a new feature. My fingers froze (even wearing gloves) long before this battery ever will. The camera's useability is simply tremendously improved over the previous models. Everything is so much more accessible and easy to adjust. I am persuaded that the image quality is superb. For example, I've used it outdoors at night during a snowstorm at high ISO here in Toronto with low shutter speed and IS lenses and I think the results are fine Toronto Winter With My !ds3 . I haven't had a chance to seriously review the Yellowstone shoot yet, but based on testing I've done with a range of other more usual subject matter here in Toronto, I think the image quality is fine.

Now whether you should buy this camera or not, at this point in time I would say is hard to advise about. There are a number of questions which only you can answer to your own satisfaction, so the only help I can legitimately offer is to suggest what those questions are about: (1) How many MP of resolution do you need for the output you will be producing, remembering that you can get commercially acceptable inkjet print quality for large prints at output PPI in the minimum range of 180-240 PPI. For many people 12 MP is more than enough, and there is a growing number of good choices in this size range. I like the 21 MP idea because it gives me much leeway to crop and still print big, but not everyone would find that important. (2) What are you using now and how urgent is it to up-grade? (3) What's your investment in Canon lenses and how much budget space do you have to contemplate a system change, considering that if you were to sell Canon and move to Nikon you will get about 50 cents on the dollar for the lenses. (4) If you want full-frame at the "bleeding edge", you have a choice between the 1Ds3, the Nikon D3 (still in even shorter supply than the Canon), and perhaps later in the year, a speculative Sony 24 MP camera, and a Nikon D3 successor speculated at 24 MP. Who knows when these last 2 cameras will appear, what they will cost and how they will perform compared with the Canon, but if you have time to wait and are prepared to switch systems, why not wait and have an expanded range from which to select once enough comparative performance data emerges. The Nikon D3, from what I've seen, seems to be a fabulous product, but 12MP.

As for Canon's quality control issues, this alignment problem seems to be on the course of resolution. The evidence suggests that they are repairing them and returning them to their owners in fine condition. Perhaps they will have now tackled the problem at source, so for future purchasers this will be an irrelevant "early adaptor" issue with no longe-term consequences. It would be good if Canon came out of its shell and told people the status of the problem, but others have given-up trying to tell Canon things that are in their own interest, so I have no illusions about any announcements. Nonetheless, they aren't dummies and if they've solved the problem silently under the hood, in the final analysis that is what matters.

Hope this helps.

Mark

Mark,

Thanks for taking the trouble to respond in such depth.

My only issue with the camera was the alignment problem and then subsequently the slow trickle of focus problems and dodgy sensors... but I am realistic about the value of on-line forums and I do appreciate that people tend to sway to them when they themselves have or foresee an issue. For me, the problem seems very widespread and that makes me question the whole assembly/quality-control process Canon have chosen to adopt for this particular model.

I spent a few days with a 1DsIII in my studio before returning it with the alignment issue - I agree, the usability is a major step forward from the MkII I'm working with at the moment.

With regard to image quality, i did a few tests which I repeated on both bodies as a comparison with controlled lighting in my studio. (Here's when I reach for my flack jacket!) I interpolated the MkII images up to the same size as the MkIII (just with Photoshop) did a little subtle sharpening on both...... and could see only the slightest perceptible visual difference at pixel resolution on a nice Lacie CRT. I think we are both experienced enough to realise this would make sweet-fanny-adams difference on a double page spread spewed out of a four+ colour press. I have to stress this was in a studio with over 15KJ of flash on tap, so maximum usable ISO or Nikons 25,000 ASA D3 don't keep me awake at night!

I could see no difference in the colour with regard to the increased bit depth (I used the same RAW calibration settings I produced for the MkII using a hybrid of the Thomas Fors script) I think the advantage would only be evident if you're the type to get the exposure badly wrong or simply like crunching the colour around in Photoshop.

When the water settles I will upgrade to a MkIII later this year for two reasons.

1. My MkII is 3 years old and at that age needs to be relegated to a back-up
2. I don't want to give all my profits to the tax man!

For the time being I've started to experiment with more exotic lenses and adapters to see the effect on both image quality and feel....


Cheers,

Carl

Carl,

Yes, a four-colour printing press would be the clinch-leveler of quality differences between sensors. There is only a 14 percent difference in resolution between a Mk2 and a Mk3, so I can readily see why it would be hard to detect for many kinds of images where liberal cropping is not being undertaken. My output world is inkjet prints, where quality differences in terms of tonal gradations and detail would be more visible. I think this is one of those areas where one really needs to print the comparisons and look at them closely to see the differences - using imaging situations which create stress between the previous technology and the current one, to see just how much that extra resolution and bit depth would make a tangible difference under a variety of capture conditions. I up-graded from an 11MP Canon 1Ds, giving me a 38% increase of resolution and the jump from a DIGIC 1 to a DIGIC III processor, so in my case the I.Q. gap may be more noticeable - I have yet to test this, because the weather here has been inimical to working with my favorite test site for so long a time now. I fully intend to do this, however, as it may be interesting to more than me.

Reading the forums, it seems to me that sensor and focus problems have been very few with this model, the alignment issue being the majority concern - and even then we don't know how many of their total production so far have experienced this flaw. It is inevitable save for the most rigorous testing - and even then - that several cameras out of thousands in a production line will turn-up defects that escape the factory processes. I don't think that represents a major risk factor as the camera is under warrenty - it is at most an annoying nuissance which Canon should do their utmost to prevent or repair quickly in the interest of themselves and their clients.

err, dont wish to be rude, but this is not remotely the right way to compare high end digital cameras files

1: a CRT is not at all sharp compared to a digital flat panel
2: the idea that the arbiter of it all is a 4 color litho dps is plainly not a runner.

litho presses run on incredibly small gamut CMYK inks, through a very rough 'screen'. photographic prints from e.g. an Epson are far finer and have a dramatically bigger gamut.

In my workflow where I use a CRT and have used a number of CRT's from Lacie consecutively over the last nine years, it is relevant. The only flat panel screen I've used and has compared favorably to my Lacie CRT is an Eizo CG221 22".... Yes, an AdobeRGB(1998) gamut and very sharp - but until my Lacie gives up the ghost I'm not inclined to throw almost £3000 at a monitor for no reason. With the exception of the Eizo, I think alot of people have been seduced by the idea of flat screens and then convinced themselves they are better than high-end CRT's... bit of emperors new clothes!!

My camera comparison is based on real-world output in my real-world workflow - as opposed to pixel peaking and comparing irrelevant graphs and specifications.

My work is printed predominantly on a litho press, so isn't that where my comparison should be?

I don't think you are being rude, you simply didn't understand my point.

On the point of mis-alignment (where we should really be) I spoke to a Canon 'Tech' at the UK Focus exhibition on Tuesday. Apparently Canon were aware of the issue back in December...

Just remember that color gamuts are improving across all devices (new LED backlit displays, 16-bit printer drivers etc) so what may not be apparent today, may be in a year or two. I noticed lots of differences in my images when I switched to an LCD (from a LaCie CRT - loved that monitor) and again when I plugged my computer into my hdtv (plasma) which has a much wider gamut (it would seem).

Well, today my 1Ds3 is expected via overnight FedEx from Canon Irvine. Yesterday I spoke directly with Chris Canada of Canon Professional Services and he assured me that while the problem is not 'widespread,' they are aware of it and that the camera had been repaired correctly. I'll report back on that when it arrives. I'm crossing my fingers as repair reports of this defect at Irvine haven't been particularly good. BTW: mine had a .8% tilt, serial #614---.

I just received mine back from Canon yesterday, and I'm impressed with the service. They sent me a shipping label and expedited the return shipping. I sent it on Wednesday and received it Monday, which is faster than I expected. It seems to be in perfect alignment now. While it was annoying to have to send it in, it was resolved fairly quickly.

Not to nitpick, but...

1Ds Mark II = 16.7 megapixels
1Ds Mark III = 21.1 megapixels

A little math seems to indicate that's a 26.3% increase in pixels... what am I missing here?

What you are missing is a bit of geometry. There may be approximately a 26.3% increase of total pixels, but not of resolution, when resolution is measured as pixels per inch or more conveniently total pixels on one linear dimension (be it the long one or the short one, if we are comparing two sensors of the same linear dimensions). Remember the total pixel count is contained in an area (A) of rows and columns which has a height (h) and width (w), so A = h*w in square inches, let us say. Therefore, the formula for calculating the linear resolution in total pixels let us say on the long dimension is "(the square root of the number of megapixels multipled by the square root of the aspect ratio)*1000". Do this calculation twice - one for the 16.6 MP sensor (not 16.7) and the other for the 21.1 MP sensor, subtract the smaller result from the larger result, take this difference as a percentage of the smaller one and you will get the increase of resolution.

So here we go: the aspect ratio is 1.5 for both of these cameras, the square root of which is 1.225 (confining to 3 decimal places). The square rooots of 16.6 and 21.1 are 4.074 and 4.593 respectively. 1.225*4.593*1000=5625. 1.225*4.074*1000=4971, 5625-4971=654. 654/4971=13.2% increase of resolution between a 1Ds Mk2 and a 1Ds Mk3. I hope that clarifies the matter for you.

1ds II = 3328x4992 => 12x18 print @ 277ppi
1dsIII = 3744x5616 => 12x18 print @ 312ppi

277ppi -> 312ppi = +12.6%

Edit: too slow!

i guess we can all use a little geometry now and then!!

as for canon irvine's repair, it did arrive today and appears to be pretty close to horizontal. without putting up a sheet of grid paper, my informal testing puts it at either completely level or at most about .1% tilted. my gut is that it's off that small amount and i'll continue to try and confirm this, but for now, it's probably good enough.

You Canon boys are not alone with viewfinder/sensor alignment issues. Although it's not in the same league as your cameras, my (RIP) D70 had a three degree misalignment. Nikon service claimed that they fixed it, but no. They did reduce it from 3 to 2 degrees. My D200 is perfect.

Much too long winded. Aspect ratio has nothing to do with it, you muliplied and then divided by it.

sqr(21.1/16.6) - 1 = 0.127 = 12.7%

The diff is your rounding error. I hope that clarifies the matter for you.


Your

That's what I said before. It would be tempting to just clobber Canon, alas they all make (some) cameras in exactly the same manner.

Indeed, you don't need to spend 8 kilobucks to have beautifully sloping horizons. The same thing occurs commonly in Rebels XTi, which I checked on my sample (former, since I got rid of it) showing about 0.75 degree slant (quite annoying) and several store cameras. My 20D has only a very very small tilt which I didn't notice before. These Rebels and the Mark 3s are assembled using the same technology, so to fix it they probably need to use shims and a trial and error procedure which requires some skill.

I also said before, the fact they can not reliably fix it proves my earlier assumptions that this is not truly adjustable.

I know there is a small rounding error on the aspect ratio - no big deal - the principle of measuring the resolution using square roots is the basic point I was making. What you show is correct; however, the value-added of considering the aspect ratio in the calculation is that it translates the total MP data into a linear metric which then lends itself to more easily seeing PPI, which is the usual way of defining resolution, hence easier for people to relate to. How's that for a long-winded response

Are you a technical expert on Canon assembly processes, or is this an inference based on some random experience? And what makes you think they can't reliably fix this problem? There is already a track record of successful re-alignments.

I am no expert on how to manufacture cameras either, but I think it more sensible to believe that a company like Canon is fully capable of manufacturing cameras in a manner to avoid this problem reliably, and whatever errors, oversights or shortcuts are now causing this problem are fixable - if not already done. I suspect the real problems are rooted in corporate culture and commercial considerations - not technical ability per se; but in this day and age of the internet, with forums such as this one, bad news circulates real fast and the reputational risk and possible consequences of doing nothing are high. So sooner or later these issues get addressed, at some inconvenience to the early adopters.

I have to agree with Mark here, I've gone to Canon with my fair share of problems, both their fault and (more often) mine, and they have been 100% in fixing them.

They also DON'T have the annoying habit of pretending there's no problem or that the problem is entirely with the photographer unlike some companies I've done business with.
(Profoto, Hasselblad and Bowens are masters of this in Australia)

In fact, service from these guys is so good it makes a joke iMHO out of everyones else in the industry here.

I'd agree with that. I'm also looking forward to the side by side testing of the 3 vs the 1 and 2 not to hijack further

My replacement 1DsMk3 arrived this morning. Note "replacement" not repair.

Haven't had a chance to try it yet. Here's hoping.

D.

I'm certainly not an expert on Canon camera assembly, but I did stay at a Holiday Inn Express... and that's the point - a Canon person may be tempted to explain it better. In this day and age of Internet company representatives should indeed talk to the customers also in forums like this one, and use the same tools to defend their reputation. And believe it or not, some companies do that, even companies bigger than Canon (such as Intel for example).

So of course you are right that it's corporate culture and not technical ability per se - I'm sure the engineers are able to eventually solve this issue -- maybe in the next model. Btw. the robotic assembly of today's point and shoot cameras with their tiny lens elements and other parts is really something to admire and getting quickly better. Of course, these cameras in general can not be repaired by humans after assembly, but since they are cheap, they can be replaced. Not so with Ds MarkIII...

Hi Dottore - yes, well I've stayed at a Holiday Inn Express too, so that makes both of us geniae. In whatever way they are assembling the 1Ds3, the re-assuring element is that people who've had this problem are getting them back repaired or replaced satisfactorily - that is what most of the evidence so far indicates, so I guess I'm a bit more optimistic about the high-end of the product line than you seem to be. I'm sure Canon has learned a lot from recent episodes with both the 1D3 and the 1Ds3, but it seems unlike some other companies, they simply don't have the corporate culture of candour with their customers. In fact one Canon rep told me over the phone that they DO NOT come to any conclusions about performance issues or design their support policies around what is said on web forums - (but it's also clear that they do read them). Perhaps that attitude will be be cured eventually with increasing competition in this market niche.

Just to say that I have only just now found time to install the grid focusing screen and check this replacement (not repair) camera.

It's totally flawless!

And not only does it have a perfectly aligned viewfinder, it also came with zero exposures clocked on the counter (my first one didn't) and the sensor is totally spotless (again, the first camera's sensor was a little spotty).

All in all, this camera is exactly as my 5D was and is how my first 1DsM3 should have been a couple of months back.

Just goes to show that it pays to throw your toys out of the pram occasionally.

I am (finally) a very happy camper.

D.

Soon you will be contributing to the 'in praise of the Ds3 thread':)

Mike

Just over a year ago I bought a 30D and have experienced the following-

- Sloping horizons (the subject of this thread)
- AF tracking totally worthless
- Inconsistent exposures

This latter one I have not seen discussed at all but take a look at these two shots- www.tipphorse.com/subpage7.html

They were taken within one minute of each other with absolutely no change in camera settings other than a slight adjustment to focal length on the Sigma EX 80-400 zoom. I was pointing the camera in more or less the same direction. Now, as photographs they would normally be deleted but I've kept them as one further reason why I no longer have any faith in Canon.

Justin.

Repost them without stripping the EXIF data, and isolating the cause of the problem will be a lot easier.

Back to the thread - after running micro adjustment tests tonight on my lenses (only the 85 f1.8 needed an adjustment of +5 and the 25-105L was a bit funny, being spot on at 24mm for close things, but front focusing slightly at longer distance, and at 105 spot on regardless of distance) I thought I'd have ago at the viewfinder alignment. It seems pretty straight, but part of the image in the viewfinder is cropped off the bottom of the frame and the top gains a bit. Less than the gaps on the 5D, but cause for canon to fix it none the less.

Mike

Its getting late so I have not time to dig out the metadata on the two shots but I am pretty sure from memory the figures were the same except for the focal length.

But that is beside the point. A camera should not produce this sort of variance in picture quality when no alteration has been made to the settings and the lighting has hardly changed, if at all. If you are looking to rely upon a camera to produce consistent images then a problem is a problem irrespective of the cause.

Justin.

Not true. You may have invertently activated bracketing, which would alter the exposure even if you were in manual mode. Or you may have had the camera in Av or Tv mode, in which case changing the framing could have altered the camera's meter reading. Or you may have bumped the mode dial to an adjacent mode. Or the aperture in the lens could be sticky and not always stopping down to the correct aperture. Or even if you were in full manual mode, and no camera settings changed, if it was a partly cloudy day the lighting could have changed that much between shots. There's many possibilities to consider other than a defective camera.

Post links to the original files from the camera.

Just to let people in Melbourne know that Burwood now does the vf allignment on site (they used to have to send them to Sydney). When I dropped my camera off this morning they told me that they now have "the tool" to do it.
I'll be picking the camera up on Tuesday ... I'll let you know how it goes.

Cropping, in addition to mis-alignment, was also a problem with my first Mk3. The replacement is perfect on all counts.

D.

Interestingly, Canon suggested I initially contact the retailer for a replacement. I'll report how it goes.

Mike

I didn't deal with Canon at all. I dealt only with my dealer who is a professional reseller (i.e. they have no high-street presence) who insisted, on my behalf, on a replacement and not a repair.

On the issue of cropping, the simplest way to test it is to frame something close and tight such as some text from a newspaper. Then switch on Live View and compare what you are seeing on the screen with what's visible in the viewfinder. You may be surprised!

D.

It is interesting to note that one correspondent upon this subject has sloping horizons on some images presented within a post on another thread.

Justin.

Update:

Today was the first day I've really had a chance to use the replacement Mk3 in earnest. I've been shooting MFDB the last two weeks.

All is not as well as I'd hoped, but I'm going to put up with it.

1. While alignment is flawless, there is some discrepancy between what I am seeing in the viewfinder and what is on Live View/Captured File, namely, in a tightly framed scene there is some clipping across the bottom of the frame. So, with critical compositions it is necessary to use the Live View to be sure you're getting everything you want in shot.

2. When shooting a uniformly lit subject against a white background I noticed that there was a faint shadow line across the bottom of the frame. This had all the hallmarks of an incorrect sync speed, so I set this shutter to 1/200 (instead of 1/250). Still the same problem, though the shadow line was fainter. At 1/160 it was fainter still, though still visible.

Only at 1/125 sec. and longer was there a clean image. So, unless I have set something incorrectly in the camera (suggestions on a postcard please! ) my camera does not sync as claimed. Annoying, but not the end of the world. I guess they'll never get this camera quite right.

By the way, anyone wishing to try to test for this will easily see the shadowing on the camera's screen (make sure the the white wall, or whatever, is evenly lit) but, you can confirm it using ACR's colour picker.

D.

i think you camera is syncing properly: if you are shooting with studio strobes the recommended sync speed is 1/125 or slower...the higher sync speeds generally apply to compact ie. speedlite and speedlite compatible units...

I wasn't aware of that. Thank you for clearing that up.

D.

Yep. The strong strobes take a bit more time for all that light to come out .

Sixth time lucky...

It's now about three months since my first disappointing brush with the 1DsMkIII.

Following my return of FOUR bodies in January with mis-alignment issues me dealer called earlier this week to let me know he'd received four fresh new bodies, oh what jolly excitement.

Having tentatively offered my credit card I took the camera to my studio for a closer look...and surprise, surprise... discovered they still haven't figured out how to put the sensor in straight!! - OK, it's not as bad as the first four, but still not where it should be.

So, I make a cup of coffee, sit down and start to wonder if I simply have to accept this shoddy workmanship as the new norm for Canon.... or see if another camera manufacturer can do a better job.

I decided to hang onto the camera and take a print of my findings to the dealer. He agreed it wasn't good and asked if I'd like to give yet another one a last try... OK says I.

I nearly fell over when I discovered the sensor in this SIXTH body was in EXACTLY the right place. Wow, was it a fluke? Who knows, but it seems (to me anyway, from my own hands-on experience) the odds are stacked well against getting a good copy of this camera.


It has been a frustrating and time consuming experience from which I've now made three observations..

1. When it's put together correctly, the 1DsMkIII is truly excellent.

2. Canon needs to pull it's socks up when it comes to quality control and customer care.

3. It's important to develop a good working relationship with your local camera dealer.


...and the cherry on the cake is that the camera has dropped in price by £1200 since January - so thanks be to Canon for an unexpected week in the Greek Islands in return for all the hassle!!

Interesting that you had problems with the sensor. I never did, or at least I don't think I did. With the three Mk3s I used, what I was seeing on Live View was exactly what I was getting. Of course, live view can only see what the sensor is seeing but my point is that it was perfectly aligned with the grid.

On the other hand, two of the bodies had alignment problems, while all three (including the one I now own) had/have problems with clipping across the bottom of the viewfinder.

I guess you could argue that the viewfinder frame is correctly installed and that the sensor is not installed to match it! But, my money is on the viewfinder frame not being the 100% viewable area that it should be.

Bangs for my buck, the 5D is still the best camera I have ever owned and that includes all my film cameras, with the possible exception of my Sinar P.

D.

I don't think your experience alone is sufficient to conclude that in general the odds are stakced against getting a correct copy, but your three observations have real merit. One hopes #2 will happen as a result of increasing competition.

Third time's the charm.

Dropped the camera off at Irvine a second time with a print-out of the issue and received it back, fixed, after four business days.