Canon EOS 1Ds
A Preliminary Field Report
This is the second 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 2 — Thursday, September 26th
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.
Moire and Resolution Concerns
I wasn't looking for it yesterday but this morning a friend pointed out to me the strong moire pattern that can be seen in one of yesterday's test shots.
1Ds Moire Example
If you look at the middle left of the frame below you'll see a grating just above the flag poles. The moire isn't very visible, even in a 13X19" print, but when enlarged on-screen so that the total image size would be 56" X 37" (that's almost 5 feet across), at an on-screen resolution of 72 PPI, the moire pattern is very visible. So too is the astonishingly high resolution.
I've discussed this with a colleague who has far greater familiarity with related technical issue than me, and his comment after examining some 1Ds sample images (at full resolution) is that what Canon appears to have done with this camera is make the trade-off of the risk of moire Vs. resolution, by making the anti-aliasing filter weak. His opinion is that they got the compromise just about right with the D60, but with the 1Ds (and the 1D for that matter) they have decided to go for the highest possible resolution over anti-aliasing.
I for one have no issue with this, as moire is seldom a problem in landscape work, nature and wildlife. But fashion photographers, for example, are going to have serious concerns since certain fabrics at certain magnifications are going to moire like crazy. In the meantime some other testers are complaining that the 1Ds has such high resolution that it's beyond the limits of their standard resolution test charts. You can't win 'em all.
D60 Moire Example
Above you see the lack of moire pattern in the same subject when photographed with the D60, but also the substantially lower resolution. That's why I've reproduced this at 100%, rather than try to match the size of the 1Ds image. Clearly there's more than the absolute number of pixels at work here — the 1Ds simply has higher resolution as well as more photosites. (See below for more on this subject).
Canon 1Ds Vs. Canon D60
This is the test that many were waiting for yesterday, but due to a momentary loss of reason I accidentally erased the D60's Microdrive last night before I could copy the files to my computer. So, this morning I returned to the same spot and redid the test. Fortunately it was a sunny day once again and so the lighting conditions were almost identical.
Immediately below are two test files, each shown in relative size. Naturally a tripod was used, and the aperture was set at f/8. The lens was a Canon 70-200mm f/2.8L IS.
The first thing that catches ones eye is the relative sizes. The D60 produces a file of about 12.8" X 8.5" at a 240 PPI printing resolution while the 1Ds will produce a 17" X 11" file at the same printing resolution.
The D60 file was imported as sRGB (the D60's native mode) while the 1Ds shot was taken in Matrix Mode 4, which is Adobe RGB.
The next thing that one notices is that the colour palette is very similar. In these examples all I've done is to normalize them by setting the gray balance in Photoshop to the white building in the lower right.
Megapixels, like horsepower sometimes make a huge difference. So does the way the chip's anti-aliasing has been configured. The Canon 1Ds is very strong on both counts, but as discussed above, watch for moire problems with certain subject matter.
Though the D60 comes up with the short end of the stick in this comparison, it's no slouch. This is somewhat like comparing the latest Ferrari to a (pick your favourite car). You get the point. Especially since the 1Ds will cost 3 to 4 times what a D60 costs. On the absolute goodness scale, the 1Ds wins, but using the value for money yardstick the D60 is a champ.
The Canon D30 and D60 have developed strong reputations for their low noise files. When I tested the Canon 1D last year I was surprised to find that it had even lower noise than the D30, even though it uses a CCD chip rather than CMOS. Canon seems to have a thing for noise regardless of which chip they use.
The first image below is the full-frame shot taken with the 1Ds and the Canon 70-200mm f/2.8L IS.
Following are 6 frames taken with the 1Ds at ISO ratings from 50 to 1250. ISO 50 is a special setting that needs to be invoked though a Custom Function. To save space I have cropped the files to include an area which has good highlights, shadows and mid-tones as well as fine detail. Have a look at them and then read my impressions below.
If you're interested in how these compare with the Canon D60 I have prepared a page with similar shots taken at the same time with the D60 at various ISO settings. (Please note that I have not bothered colour balancing between the two systems for this test.)
1Ds @ ISO 50
1Ds @ ISO 100
1Ds @ ISO 200
1Ds @ ISO 400
1Ds @ ISO 800
1Ds @ ISO 1250
What I see holds no surprises, and no magic. There appears to be little advantage to the special ISO 50 setting, though I must admit that I haven't explored it fully yet. ISO 100 and 200 are virtually noise free while ISO 400 is very usable — certainly better than any ISO 400 colour film. ISO 800 and 1250 are noisy, but quite usable if needed.
As for the comparison with the D60, I don't find a lot of difference between the two, though ISO 1250 on the 1Ds does appear to be slightly "finer grained" than ISO 1000 on the D60.
Questions and Issues
One question that I've been asked by a couple of people since yesterday's first comments is "Has there been any post-processing of the images"? The answer is, "Yes, of course there has". Strangely, some people think that there's something sacrosanct about digital files — that one can only evaluate them when they're completely "virgin". Nonsense.
A digital file is like a film negative (positive). It is the raw stuff (and sometimes the RAW stuff) from which we create prints. Some USM needs to applied or else the image will lack appropriate accutance. Appropriate White Balance needs to be applied and so forth. To coin a phrase, "It's the prints stupid".
I have applied an appropriate amount of USM to each file, both digital and film scans. Sometimes I do it manually using Photoshop, sometimes I use nik Sharpener Pro. I simply use my judgment as to what is needed to make a particular image at a particular display size look its best. Remember, USM is not about sharpness, it's about accutance. It's like choosing a different film developer.
My scanner is profiled. My D60 is profiled and the 1Ds will be soon. On the 1Ds I am using Matrix Mode 4, which is Adobe RGB.
Once more with feeling — trying to evaluate a digital image (a camera original or a scan) in its unprocessed state tells you nothing. Image processing is part of what one does to a file to make it usable, from Bayer interpolation and Linear conversion to White Balancing etc. Some of it is done in-camera by firmware, some after the fact in the RAW conversion stage, and some in Photoshop. Sometimes you can choose where, sometimes not. Please stop listening to the Net-know-nothings who spout this misinformation.
The other issue is that (as is always the case when I do these tests) some people have brought me to task for not having done an A Vs B, or X Vs. Y test, or some other combination which they believe will be a more fair comparison. Here's the thing — I do these tests in an attempt to reproduce what a typical professional user would do. In fact I do these tests primarily for my own edification. I want to know how these products compare when using the tools and techniques that I normally use. Not something theoretical, but real-world. If I'm going to buy a $6,000 to $9,000 camera I want to know what I'm getting and how it compares to what I am currently using. I figure that others do too, and so I publish the results, here and in my magazine reviews.
I do not run a testing lab, and I am not interested in cosmic comparisons. In other words, yes, I could do a 500MB scan of film and then have a Lightjet 5000 print made. (And, who knows — I still might for this series of tests). But the reality is, I've done this before. I've had many drum scans made and I've exhibited and sold literally dozens of Lightjet prints. The reality is that I spent $10,000 several years ago on an Imacon scanner because it could produce scans that were as good and sometimes better than those from a drum scanner. (Part of the reason for this is that I make them myself. No technician in a lab making $15 / hour cares about the scans the way I do.) As for Lightjet Vs. inkjet prints, yes, Lightjets can be better, but not that much better, and not all the time. Again, it depends on the technician.
Since I don't have $30,000+ to spare for a drum scanner, nor $150,000+ for a Lightjet printer, my tests have proven to me that I can usually make as good as if not better prints by scanning and printing myself with the gear that I've chosen. An Imacon scanner and (currently) an Epson 2200 printer. Some other combinations will produce different, and possibly better results. But this is what I use, and its as good as if not better than what most photographers have available.
Canon EOS 1Ds with Canon 70-200mm f/2.8L IS. 200mm @ ISO 400
This photograph is about a 50% crop from full frame. A larger version may be viewed by clicking on it.
File Sizes and Storage
If you shoot in combined RAW and Fine JPG mode (as I do) you end up with about 15MB per shot; 11MB for the RAW file and about 4MB for the JPG. That gives you some something under 70 frames per 1 GB Microdrive — less than the equivalent of 2 rolls of film. Clearly, unless you have a pocket full of 1GB drives you're going to run out of "bits" long before the day is done.
In mid-September I published a review the 30GB Delkin eFilm PicturePad. This is a pocket-sized battery operated hard disk with built-in colour LCD screen, about the size of a paperback book. It is also sold as the Nixvue Vista. With one of these you can shoot about 2,100 frames or 58 rolls of 36 exp film. Don't even consider a large image sensor camera like the Canon 1Ds without having one of these in your camera bag.
Yesterday I dropped my film off at my regular lab for processing. This is one of the largest professional labs in Toronto, which is the largest city in Canada, and the country's center for advertising and publishing. The owner / manager happened to be at the counter, and I asked him in passing what the effect of digital cameras and backs has been on his business. His reply was illuminating.
He said that they had now stopped doing C41 and B&W processing due to lack of demand and were only still processing E6 film. He said that most pros in the city were now shooting digital (studio and location) and that film was used mostly now for special applications. Interesting.
Canon EOS 1Ds with Canon 300mm f/2.8L IS @ ISO 200
This photograph is about a 70% crop from full frame. A larger version may be viewed by clicking on it.
I heard today from a large Canon dealer, (in a country and city that will remain unnamed), that the Canon distributor has asked them to get deposits from anyone wishing to place an order for a 1Ds. No reason is forthcoming, but my surmise is that since cameras are going to be in short supply when they start shipping in November, this is ensure that no dealer over-orders, and that cameras go first to legitimate purchasers who are willing to make a commitment now. Interesting.