Canon EOS 1Ds
A Preliminary Field Report
This is the third part of a multi-part field test of a pre-production Canon EOS 1Ds. If you have not yet read Part One, the introduction to the piece, please do so first as the entire review will only make sense once you've seen all of the parts in the correct order.
There will be a segment in The Luminous Landscape Video Journal's next issue (Issue #6, scheduled for release in November) showing the camera being used on a landscape and nature shoot and an analysis of prints produced with the 1Ds.
Day 3 & 4 — Friday, September 27th & Saturday September 28th
Though I have split this report into multiple parts to avoid slow loading pages, nevertheless you're going to find that this page is slow unless you have a broadband connection. But, since I need to display many test examples a short wait is the price you'll have to pay.
Yes Virginia, dust does collect on the 1Ds' sensor. Since the camera that I'm testing is Canon's pre-production sample I have no idea of where it's been or the use its been subjected to prior to coming into my hands. But in the past 36 hours I've probably changed the lens 30 times, and I'm seeing some dust images in places that I hadn't seen them when I first got the camera. This is a more frequent occurrence of dust than I've seen before from the D30 or the D60, though not as bad as with the original 1D.
What do I make of this? Not much. Dust is a fact of life in imaging, traditional, scanning or from digital cameras. Compared to the amount of "spotting" usually needed from film scans, it's no big deal.
Canon EOS 1Ds with Canon 70-200mm f/2.8L IS. 185mm @ ISO 100
This photograph is about a 90% crop from full frame. A larger version may be viewed by clicking on it.
Taken just after dawn, the warm early morning light gives a very pleasing skin tone rendition. I particularly am taken with the 1Ds' hybrid auto-white-balance approach which places an incident white balancer sensor on the front of the camera. The 1Ds (and the 1D — where we first saw this) measures the colour balance of the ambient light and integrates it with the usual white balance determination made from the image. Just using white balance based on what the lens sees can be problematic, if for example you are shooting a green car on a grassy field, with no sky or other colour clues available. I'll have more on this topic in the days ahead. In the meantime I'm pleased to see that the 1Ds does a very clean job with skin tones.
Wide Angle Lenses and The Full-Frame DSLR
There has been much discussion on various online Boards about how wide angle lenses for 35mm cameras will not work properly with full-frame digital sensors. There has been discussion about how because the photosites (pixels) are in "wells", the light hitting the edges and corners of the frame will be attenuated, especially at small apertures where diffraction effects will also come into play.
When the Contax N Digital 6MP full-frame camera started shipping this past spring (2002) I was curious to see if the pundits and self proclaimed experts would be right. I wasn't able to test this for myself because Contax, though they're promised several times that they'd provide me with a camera for testing, has not been able to get me one for the past 6 months. Their loss. Nevertheless, reviews out of the UK where some magazines have run test reports have not indicated this to be an issue.
I was eager to see if it would be a problem with the 1Ds. There had been some discussion out of Photokina about how the 1Ds uses "micro lenses" to deal with this issue, and that the Kodak DCS 14n didn't need to. In any event, the test below is just one of several that I did, and they all produced similar results.
The first three frames show a shot taken with the Canon 16-35mm f/2.8L lens, at 16mm. This is about as wide a lens as most people will use. The first was taken wide open at f/2.8, the second at the lens' optimum aperture of f/8 and the third stopped all the way down at f/22.
16mm full-frame @ f/2.8
16mm full-frame @ f/8
16mm full-frame @ f/22
What you'll notice is that the shot taken wide-open at f/2.8 shows vignetting, something that is usual with any ultra-wide angle lens, even when using film. The other two shots look essentially the same as each other, even when viewed at maximum resolution.
To show this, below are crops from the upper right-hand corner of the frame.
16mm corner @ f/2.8
16mm corner @ f/8
16mm corner @ f/22
The f/2.8 frame is a bit soft, again as one would expect from the corners of an ultra-wide-angle lens used at maximum aperture. The f/8 and the f/22 frames are sharper, and are essentially identical to each other. I think that this issue is now laid to rest. Full frame digital cameras (at least as exemplified by the Canon 1Ds) do not suffer from the forecast small aperture diffraction effects that some people were forecasting. So much for the nattering nabobs of negativity.
Canon EOS 1Ds with Canon 16-35mm f/2.8L lens at 16mm @ ISO 100
After a couple of years working with the D30 and D60, and living with the 1.6X magnification factor, it has been a joy to shoot with a wide angle lens that once again gives wide coverage. It's going to be hard going back.
I'm not in a position to comment on battery life yet. The battery that came with the pre-production camera I received was a Canon demo battery and I have no idea about its provenance. Canon claims 600 frames on a charge, but my experience with such claims is that one typically gets about half that. This is the same battery that is provided with the much more power-hungry 1D, so I would expect pretty decent life. Regardless, a second battery is a must, as is a car inverter so that batteries can be recharged on the road. I also am interested in the Digital Camera Battery, an outboard battery pack that was recently reviewed very positively by Rob Galbraith and which could be a viable solution for something like a week-long backpacking trip.
Is The 1Ds Lens Limited?
There has been much speculation that the 1Ds has such high resolution (higher than any colour film) and that most lenses may not have enough resolution to make the most of what the camera has to offer. (A couple of early reviewers have written that the camera resolves beyond the capability of the standard testing charts that they've been using).
Since I don't photograph test charts I'm not in a position to refute or confirm this, other than to say that some of the photographs I've taken with the 1Ds these past few days are among the sharpest that I've ever seen from 35mm equipment.
The math shows that the 1Ds should have a resolution of about 72 LP/mm , at the high end of what most lenses can produce at optimum aperture. It's worth digressing for a moment to comment on the absurd statement made by some people online (you know by now that I have a bee in my bonnet about these yobos, don't you?), that the best lenses have resolutions of 200, 300 or more line pairs per millimeter. Sure — aerial resolution with high contrast subjects. Get real. The very best lenses (such as a Leica M 90mm f/2 APO) top out at 90-100 LP/mm . Top lenses when measured at the center, at optimum aperture, produce excellent performance at 60-80 LP/mm . The 1Ds is right there to take advantage of this, and does better than most colour films in this regard. (Read my essay on Understanding Sharpness for more on this).
Canon EOS 1Ds with Canon 70-200mm f/2.8L IS. 200mm @ ISO 100
This photograph is about a 50% crop from full frame. A larger version may be viewed by clicking on it.
Some people are also saying that this new generation of digital SLRs will need prime lenses to make the most of what they have to offer. Maybe. But in truth most serious photographers use primes when they need speed, and otherwise shoot with zooms because of these lens' versatility. Inexpensive consumer zoom lenses will under-perform on this camera. But then, how many people are going to put a $200 lens on a $9,000 camera? Canon's L series zooms are most professional photographer's bread and butter lenses and they will perform optimally with this camera.
A Joyous Noise
On film we call it grain. In a digital image it's noise. In either case it's what happens when you run out of information and start to look at the very fabric of the image. Incidentally, as as pointed out succinctly to me by a reader, when you're scanning film you scan the image along with the grain. It is therefore a second generation image. Digital cameras produce a first generation images and are only subject to the noise inherent in their own imaging chip and circuits, which as you'll see is very low on the 1Ds.
For the past couple of years we've been taken with how noise free digital SLRs with large photosites can be. Much less noise (grain) than film. This is one of the reasons why a relatively low resolution (these days) camera like the Canon D30 was so impressive when it first appeared. What it lacked in megapixels it made up for in clean noise-free images. Consequently I stated at the time that in prints up to A4 and slightly larger D30 images were as good as if not superior to scanned film. At the time many scoffed. Now, almost no one disagrees, except for a few Net-know-nothings. Certainly not anyone who has used this camera.
Now with the 1Ds we enter a new realm. Below is a shot taken at the beach in Florida. The first image is full-frame. No sharpening has been applied. This is the way it came out of the camera.
The second frame is enlarged 1,000% (X10 above actual pixels resolution). It's of the surfer's hand above his head. Of course you can see significant pixilation. The next frame is of the sky, again at 1,000% magnification. No tricks, nothing up my sleeve.
Full Frame @ ISO 100
Sky @ 1000% enlargement
This is why those who say that a 3MP, 6MP, 11MP (pick a number) digital SLR can't equal film are basing their belief on either wishful thinking or bad science. Megapixels alone aren't the issue. All that larger imaging chips give you is better resolution, which translates into larger print sizes. Not a bad thing. But it's digital's lack of noise that gives it the edge over film. (This does not necessarily apply to the tiny imaging chips with very small photosites found in digicams. The smaller the individual pixels the more prone they are to noise.)
Canon EOS 1Ds with Canon 16-35mm f/2.8L lens at 35mm @ ISO 800
This photograph is about a 60% crop from full frame. A larger version may be viewed by clicking on it.
This photograph was taken well before dawn. Shot at ISO 800, even a large print is more than acceptable in terms of noise / grain. Combine the 1Ds with a fast lens, and high quality available-darkness colour work is possible.
Progress note: If my comments on the 1Ds thus far have appeared somewhat breathless and enthusiastic, it's because I am enthusiastic. I'm simply knocked out by the quality of the images I'm seeing. Colours are accurate and images are increadably sharp. Noise is essentially nonexistent at low ISO ratings, and very acceptable at high ones. Handling of the 1Ds is essentially the same as the 1D — which means superb. I've been using a Canon 1V since it first came out several years ago and the 1Ds handles almost identically, which means that it becomes an extension of your hands and eye, something that I can't really say for the D30 and D60.
It's going to be very hard to give this camera back to Canon next week.