The true cost of the 5D

[jon_austin]

"Had a chance to try 5D in a store, looks sweet, smaller than 10D - which I use ..."

 

Actually, the 5D is slightly larger (H and W) and heavier than the 10D.

[phil vaughan - yorkshire u]

<i>"I've been using the 5D with a pair of fairly inexpensive consumer zooms: the Canon 24-85mm and 70-210mm f/3.5-4.5s. Both lenses have performed just fine with 35mm film so why would they suddenly fall apart with a 24x36mm CMOS sensor? In fact they don't fall apart. Stop 'em down to their sweet spots and they perform as I figured they would: just fine."</i><p>Although most consumer lenses perform reasonably at f8 unfortunately I rarely shoot with that small an aperture. I'm not a pixel peeper, but I wouldn't swap my lens collection for a bunch of consumer zooms and a heap of money. They just wouldn't do what I want. I need low light capability and small DoF, just need it, that's all.

[ray robertson]

The 5D plus 24-105L IS looks ideal. My 28-135 IS is the weak link. My copy of the Sigma 15-30 really is quite sharp. The 100-400 IS is not too good at 400mm and f5.6 but much improved at f11.

 

So instead of A$5000, I pay A$7000 and get a nice system with no gaps from 15mm to 640mm, which includes just 3 lenses and 2 camera bodies (Canon 5D, 20D, 15-30, 24-105, 100-400). Not bad, eh!

[Former P.N Member]

<i>

...I'm puzzled by the apparent sharper results in the centre at f/18. What's your explanation for this? My experience...</i>

<p>

As I said, it was an informal, and certainly unscientific, test.

<p>

But to answer your question I would suspect what you're seeing is the use of both the 'sharpen' and 'unsharp mask' filters - as well as the amount of USM that was applied. I was attempting to process each image the same for comparison purposes and I'm sure the settings were not appropriate for the image.

<p>

I would not be surprised if I took the same scans and applied the filters differently or with different settings the results might very well be the opposite.

<p>

...and that analysis is as unscientific as the original test. ;-)

<p>

Dick

[ray robertson]

Dick,

When doing tests like this, you either don't use sharpening at all or you use the same amount of sharpening in all cases for images of the same size. Why would you use different sharpening settings for same size centre crops at different aperures? What were you trying to achieve here?

[Former P.N Member]

<i>Why would you use different sharpening settings for same size centre crops at different aperures?</i>

<p>

I did use the same processing for all images (I believe I listed them) however it is entirely likely that <u>they were not the optimal settings</u> for the specific images. USM parameters are dependent on the content of an image as well as the output resolution. My experience has been that the more detail an image contains generally less USM is needed, otherwise haloing and other artifacts start appearing.

<p>

Since the 15-30 testing was originally done as a matter of satisfying my curiosity rather than a true "scientific" effort. I would suspect that the USM settings were whatever was set in Photoshop when I last used it rather than being optimized for these specific test images.

<p>

As I said, it was an informal test and would not have received any praise from my former university professors. ;-)

<p>

Dick

[david_kieltyka1]

>> Although most consumer lenses perform reasonably at f8 unfortunately I rarely shoot with that small an aperture. I'm not a pixel peeper, but I wouldn't swap my lens collection for a bunch of consumer zooms and a heap of money. They just wouldn't do what I want. I need low light capability and small DoF, just need it, that's all. <<

 

I'm not advocating downgrading anyone's lens kit. :-) I own some of those fast lenses too, and use 'em when desired. My point is that the 5D (or 1Ds series) by itself doesn't obsolete any EF lens. If you've been getting first-rate results with your lenses on 35mm film you'll get first-rate results with the same lenses on the 5D.

 

Now if you're printing larger than in the past *this* may prompt you to upgrade your lens kit. But this is a different issue...

 

-Dave-

[ray robertson]

..."If you've been getting first-rate results with your lenses on 35mm film you'll get first-rate results with the same lenses on the 5D."...

 

David,

Shouldn't one get BETTER results with first rate lenses on a 5D. (I'm assuming here that in order to get first rate results on 35mm you really must have first rate lenses because we all know how severely film can degrade the image).

[andrew robertson]

Film doesn't degrade an image. It records it with its own peculiarities, as does a sensor.

 

High resolution film can still beat digital.

 

But to make a long story short, only you will be able to judge if your images are

unacceptable due to lens problems.

[ray robertson]

.."Film doesn't degrade an image."...

 

Who are you kidding! Of course film degrades an image. The best you can say of it is, the degradation is smaller the larger the format. For 8"x10" format, the degradation is insignificant. But for 35mm it's huge, especially color film. First you have degradation caused by the lens, which results in about 50% loss of contrast in detail at 50 lp/mm. Then you have a further loss of contrast as a result of the film (about another 50% loss at 50 lp/mm), and a yet further loss of contrast when (or if) you scan the film.

 

It's no wonder that APS-C digital sensors can produce more appealing results than 35mm film, and 35mm FF sensors equal the quality of medium format film.

[david_kieltyka1]

>> David, Shouldn't one get BETTER results with first rate lenses on a 5D. (I'm assuming here that in order to get first rate results on 35mm you really must have first rate lenses because we all know how severely film can degrade the image). <<

 

Don't assume anything. With proper technique you can get great 16x20" prints from 35mm film. Go check out Galen Rowell's gallery in central/eastern California for first-hand evidence. Some of those photos were taken with consumer-level zooms too. Galen's most well-known photo, of a rainbow over the Potala Palace in Tibet, has been printed at 20x30" and still looks mighty fine. For the photo Galen used a Nikon 35-70mm lens...you can pick one up used at KEH for around US$170.

 

-Dave-

[gareth_harper]

Film might be slower to work with, but it's far more flexible and still capable of better results than digital.

 

Film degrades the image? What nonsense.

[andrew robertson]

You speak of 'the film' as if there's only one kind.

 

There is no image but a temporary one without a recording medium. Good film can record

upwards of a hundred line pairs per mm, regardless of the size of the film. You have a

strange way of looking at the imaging chain.

[ray robertson]

Well, what can I say! I think some of you guys should learn how to read MTF charts. It's not true that I speak of film as if there is only one kind, as Andrew implies. I did mention that color film in particular degrades the image more. If I wanted to demonstrate any superiority that film might have over digital of the same format, I would choose a fine grain B&W film like T-Max 100, if it's still available, and a contrasty target with lots of dark, fine edges set against a light background, power lines suspended in the sky, car number plates, billboards and shop signs etc.

 

In other words, if my target contains high contrast detail at 100 lp/mm (at the focal plane) then I should be able to capture it with a good lens and 35mm T-Max 100 film. I would stand no chance with a 5D or even a 1Ds2, that's true.

 

Unfortunately (for film diehards) color film is used most of the time by most photographers and most of the subject matter we shoot contains little information that resembles high contrast B&W lines on a test chart, or power lines swinging in the breeze, so we are mainly concerned about how well the sensor or film captures low contrast detail with a contrast ratio of 2:1 (or less) as opposed to 1000:1.

 

Here a DSLR does much better than film, which is why even a 3MP D30 can produce better looking prints up to 8x12" than 35mm film, and why a 5D can produce prints equal to or better than medium format at print sizes up to 16x24".

 

How do you guys think that large format film cameras can produce better, sharper and bigger prints than smaller format film cameras? Do you think it's because the larger format lenses are so much better than 35mm lenses? That might be a contributing factor depending on DoF requirements and f stop used, but principally it's because large pieces of film degrade the image less and small pieces of film degrade the image more, comparing equal 'fields of view' of course.

[danny lee]

"even a 3MP D30 can produce better looking prints up to 8x12" than 35mm film"

 

What on earth ?

[david_kieltyka]

Bleep MTF charts. Look at *photographs* for cryin' out loud! Photography is an empirical pursuit. The results you see with your eyes are all that matter. If it looks good it is good whether the recording medium is film, digital, glass plates, polished metal, whatever. Technical wankery is nothing but an impediment to the whole process.

 

-Dave-

[ray robertson]

..."Look at *photographs* for cryin' out loud! Photography is an empirical pursuit. The results you see with your eyes are all that matter."...

 

David,

Where on earth did you get the impression that I don't look at photographs? I've been taking photographs, off and on, for over 40 years. I have literally thousands of slides and negatives stored in acid free sleeves, which I scan now and again on my KM Elite 5400 ll or Nikon 8000ED.

 

I also have even more digital images taken with my Canon D60 and 20D, taken over the past 4 years or so. I read MTF charts as a means of understanding the reasons for the current limitations of image capture as well as for inspiration regarding future possibilities as technology advances.

 

However good something looks, it's only good in relation to something worse. It's not good in relation to something better. If Galen Rowell had had the fortune to take that rainbow shot in Tibet with a

Canon 1Ds2, it would have been better. Better still! Okay?

[david_kieltyka1]

>> However good something looks, it's only good in relation to something worse. It's not good in relation to something better. If Galen Rowell had had the fortune to take that rainbow shot in Tibet with a Canon 1Ds2, it would have been better. Better still! Okay? <<

 

I think we're talking past each other here. You're defining "better" in a strictly technical sense whereas I'm defining it in aesthetic terms. I don't care much about the technical nuts & bolts of photography. I care about how my prints look. Obviously I need to know a certain amount about how my gear works in order to have control over the final printed result. But beyond that the technical stuff is just a distraction.

 

If I have questions about how the 5D performs with various lenses I don't spend time theorizing about it...I go get my hands on a 5D and take a bunch of photos. This is in fact what I did. A friend of mine recently loaned me his new 5D for a couple weeks. I shot with just about every lens I own, in low & high contrast conditions. Not test charts, not brick walls but actual real-world photos. My initial comments in this thread were based on what I've seen.

 

My 12x18" prints from the 5D are superior in terms of tonality to my 12x18 prints from scanned 35mm film. If I were using a drum scanner the gap would narrow, but film grain is also a limiting factor. With low-ISO digital files you can enlarge a great deal and still retain tonal smoothness whereas with film the grain degrades tonality beyond a certain point. Of course some folks (me included at times) like grain effects...aesthetics overriding technical purity again. Resolution isn't so important here. A Velvia slide contains at least as much spatial detail as a 5D file and both are sufficient for crisp 12x18 or 16x20 prints. Smooth tonal gradation (or the increasing lack thereof as you enlarge more) is the limiting factor with 35mm film. Contrast is something you can control in Photoshop or similar apps provided there's enough of it in the film scan or RAW file to record detail.

 

What I meant to counter in my initial post was the notion that using a 5D will reveal lens flaws that in turn will lead to poorer print quality than can be had by using 35mm film or a D-SLR with a smaller-than-24x36mm sensor. The photos I've taken with the 5D tell me this is not the case. Reports of poor corner performance are IMO either exaggerated or not based on real-world photos. Even a relatively inexpensive lens like the Canon 24-85mm performs well on the 5D. But people need to be realistic in their expectations. Stop down to f/8 or f/11 at the wide end and you'll be hard-pressed to see any difference in print between a 24-85 photo and one taken at the same focal length & aperture with an L lens. Canon's pre-USM 35-135mm f/3.5-4.5 is another inexpensive lens I really like on the 5D. Wide open at the long end it makes a darn good portrait lens.

 

Galen's 35mm photos look great even at 20x30" because of his technique, his use of low-speed, fine-grained film and most importantly his aesthetic sensibilities. If he'd been using a 5D instead of Kodachrome or Velvia the larger prints would be smoother & cleaner technically. But better? Don't know, and I don't think it matters much. The best of his photos have soul...and the prints convey this.

 

-Dave-

[ray robertson]

Dave,

Of course esthetic considerations stray from technical considerations. When I make a statement that film degrades an image, I have perhaps fallen into a trap of not distinguishing between the esthetic and the technical. These points are often raised by photographers; it's the person behind the camera that counts, not the sophistication of the equipment; yet we still strive to own better and more sophisticated equipment.

 

I've been mulling over this upgrade to a 5D for some time. It was only a year ago I bought a 20D. Well, I've taken the plunge and bought a 5D and 24-105L IS package at a relatively attractive price. I've yet to test everything.

 

Part of the motivation is due to the fact I already own an Epson 7600 printer which produces prints 24" wide and as long as you like.

 

I regularly produce 23x15.5" prints of 5400 dpi scanned 35mm color film which look quite good. I feel I need that size to appreciate the detail that 35mm film is capable of producing. Nevertheless, I'm expecting that the 5D will produce more impressive results at that size.

[fourfa]

"For the photo Galen used a Nikon 35-70mm lens...you can pick one up used at KEH for around US$170."

 

not to nitpick, but it was the Series E 75-150 Nikkor. similar price though

 

read an interview with Rowell about it here: http://www.outdoorphotographer.com/content/2003/apr/mountain.html