http://www.pbase.com/r_p/image/93551021/original

The rest of the post is here:
http://forums.dpreview.com/forums/readflat...thread=26985923

Looks like the 1Ds3 has some focus issues of its own. Must the new cameras be manually calibrated for every lens used?!?

This seems a major advantage to the camera for me. Lenses can differ, so if I can tune focus on a lens by lens basis, this is well worthwhile.

Actually I was undecided about the 1DsIII, but this pretty much clinches it for me.

I'm thinking that with live view and this mans method it should not take too long to do.

http://openphotographyforums.com/forums/sh...read.php?t=4708

i've been shooting with the 1ds3 for 3 months now, and i have not experienced any focus issues at all. i have also not needed to use the micro adjustment feature with the lenses i have been shooting with.

Exactly the same here

No issues at all with mine [touch wood]

I think having the micro adjustment is a great idea, but let's say you use rental lenses a lot - you'd be having to spend part of each shoot day tinkering with these adjustments. And what if you rent camear bodies? Now it's time to calibrate with all your lenses each time a rental body shows up?
Not cool unless you have vast amounts of free time on your shoot days. I know I don't.
In my book your *professional* camera should come out of the box with the AF working with most lenses alreay, just like the rest of the Canon/Nikon/Contax/etc. cameras. This should not feel like a science project; I just want the thing to work for my $350/day rental fee - so I can take some photos!

In most cases, lens and bodies work fine together. But the micro-adjustment feature is nice to have if you do have a rental lens or body that's a bit off calibration.

Have you used this camera Jonathan?
I have had contact with several photographers who have used this camera, who are now now buying it because of the same AF issues I had...
I'm sure some bodies work as they should, but this is nonsense, IMO.

The results of my own lens calibrations were interesting. All the numbers below are to the right of the adjustment scale (ie: tendency was to front focus, all these represent pushing the focus back). I used the moire target and live view. Very quick and accurate. I think its possible to tweak a lens in 5 minutes once you get the hang of it, so quite reasonable with a rental lens.

All my long focal lengths needed 10-20 in adjustment.

Some normal lenses were fine with 0

Zooms needed 0-5 at the wide end and 10-20 at the long end.

I discovered that AF accuracy depends on whether the lens starts focused close or near. Depending on 'which end' one is coming from the AF settles at a slightly different position. As a result I discovered that if I always focus by rotating the lens barrel anti-clockwise with my left hand (I'm right handed), I could more consistently hit the exact focus point (listening to the beep and not even looking through the viewfinder).

With the 70-200IS L, I found that at the 70 end it needed about 5, and 20 at the long end.

Without a doubt this accuracy is now greater than I've achieved with any of the Canon cameras. I'll be fascinated to see if it improves my accuracy in portrait photography.

I've been working with a couple of DsIII's for 2 months now and have no issues with the AF.

The lens calibration is great - the adjustments are very fine and have made shooting with lenses like the 85 1.2 wide open even better than the DsII.
This lens is my favorite and on the DsII was more hit & miss wide open, on the MkIII after a calibration it's very accurate.

Yes, this all takes a couple of minutes, but it's worth spending the time to get absolute best focus with each lens you use.

I think with this level of rez you don't get away with much, but get it right and the image is amazing for 35mm..

I don't have a -III body yet, but I've shot over 120,000 frames with a 10D, 1Ds, and 1D-MkII and 8 different lenses with no front or back focus problems. And none of those bodies have lens calibration options.

Thanks for the info David. It's encouraging. (Still, it doesn't help those who rent and can't afford the time to calibrate rental body to all one's lenses...)
David, just out of curiosity, did yours focus reasonably well before all the calibration?

Best

I shot with the first body for a few days right out of the box and didn't have any serious issues with the exception that a new 50 1.2 I got at the same time wasn't sharp enough near wide open.
It was also very slightly soft at F8 when compared to images from the 85.

When I tried to calibrate it there wasn't enough adjustment in the calibration tool to bring it into focus wide open.
Canon had a look at the lens and it was found to be off center ( how much I don't know) and adjusted it.
I then checked it on my chart and it's now very sharp and the focus seems bang on.
It's now very sharp at f8 or wide open.

My feeling with the new bodies is that the more the up the resolution the less room they make for any error.
IMO it's worth the extra stuffing around to get shots of such high quality from a camera that's so easy to use.

Rental gear comes in all conditions in my experience and I think with cameras and lenses it would be wise to always do a quick check or have an assistant do it in case the last renter dropped something and didn't fess up.

What I find curious is that all of my AF lenses, which all worked fine on the 1DsMK1, MK2 and 5D, with no perceptual front or back focus issues, are all over the map with the MK3, with some needing front adjustment and some needing back, and one lens that seems to need different compensations for different distances, and that lens, the 200mm 1.8 was absolutely spot on with every other camera at all distances. Well, at least I have a camera back and working for the most part.

Of course this is not right. I don't know why anyone would say this is what to expect just because it is a a high resolution camera?? Hopefully they will have a firmware solution soon...

DUH!!! The higher the resolution of the sensor, the narrower the tolerance of acceptable focus becomes. It's an inevitable consequence of increasing resoluion; small focus errors and lens aberrations become more noticeable in direct proportion to the increase in resolution. 35mm film equates to 3-5 megapixels of digital capture, depending on the film type and ISO. Most of Canon's current lenses were designed prior to digital capture being the dominant form of photography, and it's not surprising that a fourfold increase in resolution behind the lens strains the limitations of the design a bit.



They do. It's called the lens calibration feature, and it's already available.

While it's true that higher resolution does tend to fine tune that point of critical focus if the lens is capable of resolving the difference, it's also true that the real world differences between the 1DsMK2 and 3 are extremely small, and the best lenses can easily resolve to the limits of the sensor. I don't know where you came up with 3-5 mp for a film equivalent, but that's a fairly laughable figure. Film today, even 35mm, can actually resolve at least as much as these new digital cameras, but extractly all that's on the film is quite difficult. Even scanning at a true 8000 spi on my Howtek at 3 microns does not get everything on the film, but it does provide a hell of lot more information than a three to five megapixel digital capture. Just curious, how did you arrive at your conclusion?

BTW, have you seen just how sharp a 200 1.8 is when it's properly focused?

I spoke with Canon CPS Irvine yesterday and am taking the lens in later today. Yes, they will still work on it as long as it doesn't need parts, and there are both near and far focus adjustments internally.

My observation that a 10D could match or beat 35mm scans and optical prints shot with the same lens, and that the 1Ds-I blows away both scans and optical prints in head-to-head comparisons using the same lens. I have yet to see any credible evidence that 35mm film can match the original 1Ds in any way, shape, or form. The 1Ds-III is a significant improvement over the 1Ds-I, and your assertion that 35mm film can compete with the 1Ds-III is not very credible. Do you have any head-to-head comparisons to back that up?

I think 35mm film is very close to 1Ds2 territory, but due to the noise (grain) the digitals print better.
But anyway, that's not the discussion here. What we are talking about is a new camera going from WAY front to WAY back focused moment to moment. And this it not all due to increased resolution.

Indeed the focus issue has got fu++ all to do with greater resolution of the 1ds3 sensor!

I'm on my 3rd 1DS3 and every one has had a focus issue, none of which the micro adjustment has any hope of sorting out. The same lenses on both a 5D and 1ds mk2 were perfect. The focus issue is with longer distances and out toward infinity where some lenses will not make it past about 15mtrs using AF centre spot, both my 85f1.2, 50f2.5 and 16-35mk2 have problems with the same issue, however the zoom works fine on the 35 end but not wider. Incidently you dont need to check the image as it's so obvious on the lens scale, and even more obvious on the screen. Strangely my 135f2 has no problems at all.

So i'm off to canon to see why all 3 1DS3's have had this problem, and the 1DS2 and 5D had no problems with the same lenses. Again nothing to do with extra resolution, that's just bollocks i'm afraid. Focus is focus, and out of focus isn't.

Jonathon,

Have you ever actually made an 8000 spi scan on a hi-end drum scanner? Your comparisons are reallly just comparing your digital cameras to whatever scanner you had available and really just showing what a low or medium end scanner is. If you take a piece of Velvia 50 or T-Max 100, properly exposed and shot with somethiing like the 200 1.8 @f/4 or 5.6 and look at that film with a low powered microscope, say 50-100X or perhaps through an Omega grain focuser with the enlarger at the top of the column, you will see what I'm talking about. Even scanning at 3.17 microns (8000 spi) on the drum will not quite record everything on those two film stocks. Now the problem, of course is that even if you make an analog projected print, you are limited by the quality of the enlarger lens and you can never print everything that is on the film. When you take into account the optical losses of analog printing or digital scanning, 35mm film seems to be just shy of a 1DsMK2 in real world terms. The 1DsMK3 is only marginally sharper than the MK2. I know. I have both and have compared them head to head. What I have observed seems to be right in line with what others have reported in terms of film's resolving power vs. the resolving power of a specific pixel pitch/lens combination.

•••

"But anyway, that's not the discussion here. What we are talking about is a new camera going from WAY front to WAY back focused moment to moment. And this it not all due to increased resolution."

I agree. I retested my 200 1.8 with zero correction this morning and it was front focusing at closer distances about the same amount on both camera bodies. Manual focus was right on. Curt and I drove to Irvine at lunch today - and dropped off several lenses including my 200. Since there are both near and far adjustments in the lens, I'm looking forward to seeing how much improved the situation is. The other lenses all seem fine so far.

I don't have a drum scanner, and I have no plans to buy one. But Michael did a comparison between the 1Ds and drum scanned 6x7, and the differences were pretty minimal. So unless you have a head-to-head comparison between 35mm film and digital to back up your assertion, I'm calling BS.

Let's give it a rest Jonathan. If you want to discuss film vs. digital, please start another thread, or read some of the other 4 million threads on the web about that topic, OK?

In the mean time, I think it would be great if ONLY those who have some experience shooting with the 1Ds3 would give their opinion/experience here...thanks.

As a result of this thread went out and tested my 24-105 and the 100-400 zoom to see how they faired on my new 1Dslll.

No problem with either, however the ability to trim the focus point, if needed, is useful.

Here here!! Jonathon is very good at stating facts as reported by other photographers without actually having any first hand experience, then reporting like some kind of a guru about how he's shot 200,000 images on whatever camera like that gives him authority status. Let's wait to see what he has to say when/if he buys a 1DS3, and if he gets a properly functioning one. As for the film/ digital debate, don't get me started, i would toss all this digital shit in the bin and go back to shooting 6x7 and 4x5 colour neg if economics would allow, simple as that.

Bullshit. I did do my own comparisons between my 1Ds and scanned 35mm film using the same lens and subject, and the 1Ds was clearly superior. I do have some first-hand experience in this area. And Michael is hardly a neophyte when it comes to photography. If you dispute his conclusions in Michael's comparison I linked to, what are the grounds? What were the flaws in his methodology?

If film is so obviously superior, it should be a simple matter to post your own comparisons similar to the one Michael did and demonstrate the validity of your claims. I haven't ever seen a head-to-head comparison between 35mm film and 35mm digital with real-world subject matter that demonstrated any sort of superiority for for film. If there is one out there, I'd like to see it. Put up or shut up.

Keep your helmet on!

I was talking about your Zero first hand experience of the 1DS3, nothing to do with film so shut up yourself.

"Bullshit. I did do my own comparisons between my 1Ds and scanned 35mm film using the same lens and subject, and the 1Ds was clearly superior. I do have some first-hand experience in this area. And Michael is hardly a neophyte when it comes to photography. If you dispute his conclusions in Michael's comparison I linked to, what are the grounds? What were the flaws in his methodology?"

Yeah, we know you did your own comparisons, but you were using a crappy scanner, so your conclusions were only as good as your hardware. Also don't know what your were using in the way of lenses either. As far as the test you refer to, it was not done on a drum scanner, it was on an Imacon - a scanner that simply does not meet its published specs either in density range or resolution. If it's hitting a real 2400 whiile scanning at 3200, I'd be surprised, plus the focus issues they have to boot. Again, what is your actual first hand experience here. I own two of my own Howtekl drum scanners, one of which can hit real world numbers in the 7000+ resoluition and still can't record everything I can see on a Velvia slide with a lo-power scope.

"If film is so obviously superior, it should be a simple matter to post your own comparisons similar to the one Michael did and demonstrate the validity of your claims. I haven't ever seen a head-to-head comparison between 35mm film and 35mm digital with real-world subject matter that demonstrated any sort of superiority for for film. If there is one out there, I'd like to see it. Put up or shut up."

Oh Jonathon, quit your whining. You've missed the point entirely here. No one ever said film was superior in real world resolution to a 1DsMK3. In fact more than one of us pretty much said the same thing if you care to re-read the posts. You're the one who claimed that film was no better than, what did you say - a 3-5 mp capture. So please, until you get yourself a 1DsMK3 with some known premium lenses and have access to a true high end scanner for whatever comparsons you might want, you might want to sit on the sidelines and listen. If you want to live in your ignorant bliss, fine, but keep it to yourself.

Go back and re-read the page, paying close attention to the last comparison at the bottom of the page. There is a 3-way comparison between the 1Ds, the Imacon scan, and a drum scan made by an Isomet 405 HR drum scanner at 5334 PPI. The 6x7 drum scan matches, but is not significantly better than, the 1Ds capture.

As to my real-world experience, I have looked at hundreds of prints from many photographers who shoot 35mm film, ranging from full-time professional to rank amateur, printed both digitally (scanned with a variety of equipment) and optically in a variety of sizes up to 24x36 inches. And over the last 10 years or so, I have never seen a print from 35mm film that can match or beat the output from the 1Ds with regard to resolution. In every case it has been immediately obvious which print came from film and which was digital, and the digital was superior.

If you have any comparisons of your own to support your claims, I'll be happy to look at them and offer a retraction if warranted, but so far I have never seen one shred of credible evidence that film can match the image quality of digital of the same format.

"203", just because a couple of reviews indicate a specific owner has a specific problem, it doesn't necessary point to a generic defect of the camera model. If you are looking for a canvas of experience from 1Ds3 owners, I have made about 2000 captures with mine since I received it at the end of November and I am fully satisfied with its focus accuracy using all three high quality Canon zooms I own. It can focus almost in the dark; it is accurate at both short and long distance focusing tasks.

Given the very high resolution of this sensor it will be very unforgiving of inferior quality lenses. One also needs to be mindful of the conditions in which people observe "defects". For example, from things I read, it could be that there is an Airy disk issue affecting the resolution of images made at excessively small apertures relative to the sensor resolution (the Cambridgeincolour website has considerable material on this subject).

We'll see I guess. Hopefully the one I just ordered ($6,884.99!) will work better than the first one I tried.

"Go back and re-read the page, paying close attention to the last comparison at the bottom of the page. There is a 3-way comparison between the 1Ds, the Imacon scan, and a drum scan made by an Isomet 405 HR drum scanner at 5334 PPI. The 6x7 drum scan matches, but is not significantly better than, the 1Ds capture."

Yeah, you're right, there was a drum scan involved and it was better than the Imacon, but there are drum scanner and then there are drum scanners. There's more to drum scanning resolution than just a claimed dpi. You need to know the aperture used in the scan. You know, for instance, that the much revered Heidelberg Primescan has a minimum aperture of 10 microns, giving a hardware max of 2540, no matter what is claimed. The Isomet is not a scanner I've ever seen or used and have no idea what it's true specs are. There are only two scanners I know of that use a true 3 micron scanning aperture - the ICG and the Howtek/Aztek. As with most of Michael's tests, there are too many areas to criticize, and this is no different. There are much sharper lenses for 6X7 than the Pentax 200. Take any Mamiya 7 Sekors and you'll see what sharp really is - and having compared directly a 1DsMK2, which is significantly sharper than the MK1 to an 80mm M7 lens, the Mamiya still is quite a bit better especially in very fine detail.

"As to my real-world experience, I have looked at hundreds of prints from many photographers who shoot 35mm film, ranging from full-time professional to rank amateur, printed both digitally (scanned with a variety of equipment) and optically in a variety of sizes up to 24x36 inches. And over the last 10 years or so, I have never seen a print from 35mm film that can match or beat the output from the 1Ds with regard to resolution. In every case it has been immediately obvious which print came from film and which was digital, and the digital was superior."

Now there's a controlled study Jonathon. Up medium size prints like the ones you mention, the 1Ds prints certainly look good, but I still prefer the look of a drum scanned Velvia or Kodachrome at 24 X 36 or larger. Go to 32 X 48 or larger and the film looks even better and never looks digital, but let's get back to your original premise - that it only takes 3-5 mp to equal scanned film. You keep avoiding dealing with that, but maybe that's your experience, and if it is, that's fine. It's just not mine or so many others.

"If you have any comparisons of your own to support your claims, I'll be happy to look at them and offer a retraction if warranted, but so far I have never seen one shred of credible evidence that film can match the image quality of digital of the same format."

It's funny how you change the rules for your comparison part way through your argument. I'll repeat myself one more time and then I'm out of here, as I'm completely slammed with real work. Let's see your 3-5 mp digipix that equal scanned film. That's what you claimed, but it only proves how little experience you have with really high end scanners and cameras.

This thread has somehow gotten so far off topic. It's time to move it back.

pfigen, I guess this is a fundamental part of the whole issue isn't it: what you "prefer" and what "is" are not the same thing. And Iwonder about fluid concepts such as looking "film" vs looking "digital"? I would have thought that prints have objective qualities which are not defined in terms of their respective technologies but are enabled by them.

Roger Clark has published a considerable amount of rigorous analytics comparing resolution between various kinds of sensors and films Clarkvision; but quite apart from those findings, the maketplace has spoken - most notably the professional studios which converted from their Hasselblads to the Canon 1Ds-1 starting back in 2002 - because in their judgment they could at the very least maintain the quality of their output much more efficiently that way.

But I won't add anything more to this debate because your final observation above is correct.