A Biased Evaluation of The Differences...

[Hank]

Way back up there to the origins of this thread........

Thanks for the review Michael.  I've frequently said that emergence of a full-frame sensor would not be enough to make me dump my D2X bodies, and that's still true.  But your comments and highlights of other features do in fact make the D3 attractive.  

I may not go so far as destroying a body in the way Toyota depicts owners destroying their vehicles so they can buy a new one, but I frankly won't mourn the passing of a D2X in order to help justify picking up a D3.  It's just that Nikon builds them so darned well, I may be waiting longer than I'd like!

[Graeme Nattress]

The signal frequency is in line pairs as well.
[a href=\"index.php?act=findpost&pid=165893\"][{POST_SNAPBACK}][/a]

The signal should be a frequency, and lp/mm is appropriate. The samples should be in samples/mm.

If you think in audio terms, to get 22khz (or 22,000 cycles / second), the sample rate is 44.1kss (kilo samples per second). You need twice as many samples as hz, or twice as many samples as line pairs.

Graeme

[Rob C]

Wow, a lot of anger in that response.  I am admittedly ignorant of the world you describe, and I am probably happier for it.  While I wouldn't be surprised by those attitudes in fashion, high-end event photography, or commercial product work, I am surprised that you feel it is the prevalent attitude in Fine Art Landscape Photography, which is the subject of the Luminous Landscape website. 

My landscape photo customers rarely see my camera.  99.9% of all our customers couldn't tell the difference between a print made with a 20D or a D3.  99.9% of our customers couldn't tell the difference between a print made with loving precision using a R2400 or an R11880. Certainly the opposite is true and many cr@ppy prints are made with great equipment every day.

I was disappointed because people who know better than to equate equipment with artistic quality (which includes the forum members and all of the excellent contributors to this site) would be the last ones who I would expect to propagate the myth that the gear makes the artist.

Put another way, if I went to one of Michael's expeditions or seminars, would I be considered inferior if I showed up with a two year old camera? A 300D?  We have all been to places where that sort of elitism has excluded people.  I never previously got that vibe from Michael or the contributors to this site. 

I buy new equipment because it allows me to do something new creatively or to grow as an artist, not to impress people.  If there was a purple-plastic digicam that did the trick, I would be there in a heartbeat.
[a href=\"index.php?act=findpost&pid=165685\"][{POST_SNAPBACK}][/a]

Fike

Yes, of course, some anger; it is somewhat annoying to hear people speak with  conviction of things they know little about.

You might or might not choose to spend a lot on equipment, and that´s just you and your personal relationship with your money. But you should try not to extrapolate your own rationale in that way. Like Mr Simon said: one man´s ceiling is another man´s floor.

Further, I don´t think you will find it cast in stone that the LuLa has stated that ´Fine art Landscape´is in any way the exclusive subject of the site; in fact, I do believe that Michael has addressed this in the past. I do not believe that I wrote anywhere that I was making reference to "the prevalent attitude in Fine Art Landscape Photography" -  I wrote of the commercial world of which I do have a long understanding and all too personal experiences. So, I feel quite justified in making the case for the importance of appearances in business, which I do think you accepted.

By the same token, I´d be surprised if the same ethic, if for different motivation, were not to exist within amateur ranks too.

To tell you the truth, there is something about the self-applied title of the genre, Fine Art Landscape Photography, that comes over as unbearably pretentious to some professional photographers, as if its practitioners were unable to say they just liked photographing landscape, that it wasn´t quite enough, there had to be this implied elevation to a higher level ; added value, as the taxman might say.

Anyway, it´s all academic and not really worth either of us getting over-pressured about it - just opinions, and we will both go on with our own set ideas as will the rest of the world.

Cheers - Rob C

[Nemo]

The signal should be a frequency, and lp/mm is appropriate. The samples should be in samples/mm.

If you think in audio terms, to get 22khz (or 22,000 cycles / second), the sample rate is 44.1kss (kilo samples per second). You need twice as many samples as hz, or twice as many samples as line pairs.

Graeme
[a href=\"index.php?act=findpost&pid=165934\"][{POST_SNAPBACK}][/a]

Then, 300 "samples" is ok. The lp/mm consistent unit gives an intuitive idea of how many "pixels" you must to put into a millimeter for meeting the requirements of the Nyquist Theorem. For instance, 100 lp/mm leads to 5 micron spaced pixels, quite small for actual reflex cameras (10MP FourThirds Olympus cameras, however, reach that number)... and we are talking of at least 300 lp/mm for a correct reproduction of 150 lp/mm (as transmitted by the lens). This is three times the maximum pixel density in actual reflex cameras. That is the idea.

Thanks for the correction Graeme.

[Rob C]

Way back up there to the origins of this thread........

Thanks for the review Michael.  I've frequently said that emergence of a full-frame sensor would not be enough to make me dump my D2X bodies, and that's still true.  But your comments and highlights of other features do in fact make the D3 attractive. 

I may not go so far as destroying a body in the way Toyota depicts owners destroying their vehicles so they can buy a new one, but I frankly won't mourn the passing of a D2X in order to help justify picking up a D3.  It's just that Nikon builds them so darned well, I may be waiting longer than I'd like!
[a href=\"index.php?act=findpost&pid=165932\"][{POST_SNAPBACK}][/a]

Hank -

Please forgive the hijacking that follows, but as I consider you my man in the country, some more clarification of US folklore, please!

Chevrolet. Years ago, it used a blue, slanted cross in its logo. Today, it has a golden slanted cross. Do you remember when this change happened, why it happened or whether it has something to do with GM buying the Daewoo brand?

This matters to me, because I find, as I age, that things I took for granted as being the way I thought them to be, have lately begun to break the security of that fond feeling. Maybe you can shed a light on the car thing, at least!

Rob C

[Quentin]

An unfortunate side effect of the exciting progress made with digital cameras is that they might just have become consumer goods like cars, to be changed as a matter of routine, not cherished in to old age as film cameras might have been.  The Nikon's look great, but how about the D4 or D5?   Will the Canon 1Ds VI have 50mp?  Its enough to make you yearn for film.

So I'll buy new kit of course when it offers a real advantage, but I don't feel any peer pressure to ensure I have the latest from Japan adorning my kneck any more than I'll ever buy a new car to impress the neighbors.  Its just not what floats my boat.  I have no interest in what the Jones' are doing, much less trying to keep up with them.

Quentin

[Hank]

Nothing was aimed at you or anyone else Rob.  Rather I wanted to sidestep the frey and respond directly to Michael.  Sorry for the interruption.  It's what I was trying to avoid.

Edit-

I just saw Quentin's post.  My sincere hope is that just as I'm ready to replace our D2X bodies, the D4 will be on the scene.  With such a jump from the D2 to the D3, I can only hope for a similar jump from the D3 to the D4!

[John Sheehy]

But Photoshop uses bicubic for downsampling, which is hardly an optimum downsampling filter.
[a href=\"index.php?act=findpost&pid=165881\"][{POST_SNAPBACK}][/a]

You can change PS' default interpolation method in Preferences (and it ships as bicubic, IIRC), but this is not used for image display in Photoshop.  Photoshop uses some bizarre hybrid of nearest neighbor and some kind of averaging.  You see the averaging in some  ratios, and in some you don't, and it looks like pure NN.

Make a single-pixel checkerboard image in photoshop, and run it through different percentages in Navigator.  You will see a bunch of things that look totally unlike the original, as you would imagine it resampled.

If you don't have an easy way to make a checkerboard, just make a new canvas, and open filter factory and enter:
R: (x+y)%2?0:255
G: (x+y)%2?0:255
B: (x+y)%2?0:255

It is really, really bad.  This decision was probably made back when computer had 33 MHz CPUs, and 2 Megabytes of RAM, and no one has given it a thought since.  There is no excuse for such nasty resizing in a state-of-the-art image processor.

And again, my point is that this kind of viewing increases image-level noise, and gives an unfair advantage to images with less, but bigger pixels, because the noise of individual pixels is exaggerated.

[John Sheehy]

The signal should be a frequency, and lp/mm is appropriate. The samples should be in samples/mm.

If you think in audio terms, to get 22khz (or 22,000 cycles / second), the sample rate is 44.1kss (kilo samples per second). You need twice as many samples as hz, or twice as many samples as line pairs.
[a href=\"index.php?act=findpost&pid=165934\"][{POST_SNAPBACK}][/a]

More is still better, though.  I think the "nyquist" theory is a bit too optimistic.  I think you really need to sample at 4x to 5x the highest frequency.

Say you sample a sine wave with a frequency of 20KHz at 40Ksps, and the sample period is centered on the parts of the sine wave where it crossing the zero point?  What do you get?  Nothing.  In the real world, there is usually some frequency modulation, so you get changing phase relationships, and therefore amplitude modulation.  These occur at other ratios to the theoretical maximum sampled frequency, too.  A slow sine wave sweep drops in volume at several points in the sweep, as you approach the nyquist.  Ratios like 3/5, 4/5, 2/3 the nyquist, etc.

Back to imaging, I know it is a subtle thing, but I think over-sampling more than conventional sampling technique requires leaves a lot more room for quality resampling of the data.  Unfortunately, as you include more samples, you also have to increase the total data, or compromise it to some degree with compression.

[John Sheehy]

With such a jump from the D2 to the D3, I can only hope for a similar jump from the D3 to the D4!
[a href=\"index.php?act=findpost&pid=165957\"][{POST_SNAPBACK}][/a]

Well, that's a pretty big jump.  The D3 is a lot closer to the physical limits of sensors than any previous Nikon.  I doubt you will ever see a jump that big again unless one of the following occurs:

1) A way of guiding light by color into cells instead of wasting it on color filters is
    invented.

2) A way of collecting more photons than current well depths allow is invented (lower
    low ISOs; won't improve higher ones, though).

3) A way of reading out the sensor and just counting photons, adding no read noise, is
    invented (deep shadows would be liberated).

[sojournerphoto]

More is still better, though.  I think the "nyquist" theory is a bit too optimistic.  I think you really need to sample at 4x to 5x the highest frequency.

Say you sample a sine wave with a frequency of 20KHz at 40Ksps, and the sample period is centered on the parts of the sine wave where it crossing the zero point?  What do you get?  Nothing.  In the real world, there is usually some frequency modulation, so you get changing phase relationships, and therefore amplitude modulation.  These occur at other ratios to the theoretical maximum sampled frequency, too.  A slow sine wave sweep drops in volume at several points in the sweep, as you approach the nyquist.  Ratios like 3/5, 4/5, 2/3 the nyquist, etc.

Back to imaging, I know it is a subtle thing, but I think over-sampling more than conventional sampling technique requires leaves a lot more room for quality resampling of the data.  Unfortunately, as you include more samples, you also have to increase the total data, or compromise it to some degree with compression.
[a href=\"index.php?act=findpost&pid=165986\"][{POST_SNAPBACK}][/a]

John

I'm not sure Nyquist theory is optimistic as simply too generalised in use. Your example is one I've considered on many occasions and, like you, reached a view that oversampling is necessary to ensure good data up to the limiting frequency (I believe SACD samples at 96kHz for a ~20kHz bandwidth). Using a Bayer array makes the situation worse as there is less data in each channel of course.

However, positively, fortunately most real audio or visual events that are sampled do not globally fall into the 'difficult' regions for sampling - i.e. most music is not fixed 20kHz sine waves (not that I could hear them anyway in my middle age!) and most images are not of fixed or even slowly sweeping frequency data.  Also, I believe it's the case that the way sight works means that we are mostly much less discerning as to the quality of images than sound. For evidence of what is acceptable consider what happens when you freeze a TV picture - either 625 line or HD! - although I accept that moving pictures are differnt than still images.

Overall I am amazed at how good the images from my 5D look:) Who would have thought that so few image elements and so little data could create such a nice impression! On the other hand I have a 30 by 15 inch black and white print of some deer and trees on the desk next to me and I would be glad of a bit more resolution to clean up the foliage... the only thing is that I'm not convinced that the extra 30% gained from 21Mp is really enough. (he says now!)

Mike

[sojournerphoto]

So I'll buy new kit of course when it offers a real advantage, but I don't feel any peer pressure to ensure I have the latest from Japan adorning my kneck

Quentin
[a href=\"index.php?act=findpost&pid=165956\"][{POST_SNAPBACK}][/a]

I'm picturing the latest Wista 45DX4 it better still the T810 Mk2:)

[Ray]

Basically, you are in the business of taking money from your client´s bank account and putting it into yours. In order to do this, you generally need a client with more money than you have. And that client must be reassured that he isn´t being taken for a ride, that you are the right guy with whom he should be spending his money. And that feeling is very well helped along by the showing of the trimmings of success................

If you want honesty and photographic good times without obligations,  go be an amateur and please only yourself; but do understand that the two - pro and am - are worlds apart, hardly the same species even.

Rob C
[a href=\"index.php?act=findpost&pid=165677\"][{POST_SNAPBACK}][/a]

Rob,
That's sounds like a rather cynical point of view. You remind me of a professional photographer I once met who was working in a hardware shop in Australia, presumably because he didn't make enough money as a photographer, but I recall it might also have been because it was his own hardware shop.

We naturally got talking about photography and I remember being somewhat amazed when this guy claimed that he never did any shooting unless a client paid him to produce photos.

As you say, the professional and the amateur can be worlds apart. Some professionals won't lift their camera unless someone pays them to. Isn't that a bit sad?

[Jeffrey Friedl]

A very enjoyable read, thank you.

About the 12-bit vs. 14-bit NEFs, as one commenter has noted, it could be about file size, but it's more likely related to backwards compatibility. Updating a raw converter for a new Nikon model normally involves just adding a color profile for the new camera, and perhaps a few other camera-related settings, but the move from 12-bit to 14-bit requires a bigger programming change that might take longer than someone wants to wait. Thus, using the older format provides a quicker path to getting stuff done.

It's likely for this same reason that the lossy NEF compression is included along side the new lossless NEF compression.

About using DX lenses on the D3, you might consider turning off the auto-masking, and doing the crop yourself later on. I tried a 17-55/2.8 on a D3 this way, and found that for all but the most extreme wide angles, the sensor had full image coverage. I was limited to viewing the image on the back of the camera, so I don't know whether there was reasonable quality toward the edges, but it seems to me that it's better to keep the extra information until you know you don't need it, just in case. At least, for some shooting situations.

One thing I don't care for about the D3 is that all the AF points are clustered in the DX area. I understand why they did this, but I prefer a wider range.... it seemed cramped.

The Auto ISO is indeed nice, but many find the maximum of 1/250th for the lower threshold to be useless. Sports shooters, for example. I can't conceive of a reason that Nikon wouldn't add a few more rows to the AutoISO menu to allow for greater shutter-speed floors. Like the inability to use MUP and timer together, it just boggles the mind, especially when viewed in relief of so many brilliant design decisions.

-------------------------------------------------------------------------------------------
Jeffrey Friedl   http://regex.info/blog/     Kyoto Japan

[Rob C]

Ray

Cynical or not, I think the point is that a professional HAS to make the business pay its own way and also the keep of the photographer, his wife, his children and, perhaps, even an ex-wife or two; the richer the photog, perhaps the more of the latter.

Your buddy in the hardware store can´t have been much of a pro photographer, because had he been so, he wouldn´t have had the time to devote to anything else. Possibly, he was an "advanced" amateur?

However, his claim that he doesn´t get out of bed unless somebody wants to pay him would certainly fit some busy pros: do you really think that it stays fun doing photography when you have also to do it for real? For survival? My take on this, having had both experiences, the full-time pro and now the retired one, is that I do it now, in retired mode, because if I did NOT do it, then there would be one huge void looking for something to fill it. Life and age bring their own tribulations - believe me - you are damn lucky if you escape that; you need something to lift you out of the grind, which is NOT about financial survival but IS about staying sane in a world you see going to hell in a rickshaw.

I got into this business because of love for it; I stuck with the good and the bad,  concatenated and taking their turns to come and play for no obvious cause; I paid my dues and collected the reward, but somewhere along the line it became clear that what had started as love had become its own monster with appetite greater than my own. Now, as I said, I take the devil for walkies every so often.

Ciao - Rob C

[Rob C]

Nothing was aimed at you or anyone else Rob.  Rather I wanted to sidestep the frey and respond directly to Michael.  Sorry for the interruption.  It's what I was trying to avoid.

Edit-

I just saw Quentin's post.  My sincere hope is that just as I'm ready to replace our D2X bodies, the D4 will be on the scene.  With such a jump from the D2 to the D3, I can only hope for a similar jump from the D3 to the D4!
[a href=\"index.php?act=findpost&pid=165957\"][{POST_SNAPBACK}][/a]

Hank

I think you misunderstood me: there was NO hidden agenda in my last post about Chevvy - it really was all about the marque and nothing deeper had crossed my mind. I really am interested to know about the change from blue to golden logos.

Ciao - Rob C

[Paul Kay]

Wow, is there a lot to read through here!

Interesting article.

My own take is slightly different as I have no bias to either manufacturer (although I now use Canons I miss the ergonomics of my Nikons - another personal point) but I'll post it anyway.

When I switched from Nikon to Canon when the 1DS was released (I am still using 1Dss) I bought Canon's 24/1.4 (and now have 2 for obscure reasons - but essentially because one is used underwater mostly) and would be lost without it. To tempt me back to Nikon (the D3 sounds like it has an appropriate spec for my work) they'd have to bring out a similar lens because I wouldn't want to lose it! The 28/1.4 sounds good but isn't wide enough, and is as pointed out very expensive.

[Ray]

Ray

Cynical or not, I think the point is that a professional HAS to make the business pay its own way and also the keep of the photographer, his wife, his children and, perhaps, even an ex-wife or two; the richer the photog, perhaps the more of the latter.

[a href=\"index.php?act=findpost&pid=166138\"][{POST_SNAPBACK}][/a]

Indeed! I think it's a difficult profession and even more difficult for the photographer who tries to run the business himself. A business is a business, photographic or otherwise, and one needs a partner with some business acumen to run it. I would never attempt to be a one-man band.

I sometimes think maybe I should get my own website and advertise a few photos for sale. Who Knows! Someone might like a few. Then I have second thoughts. Do I really want the responsibility of producing prints on demand, filling out GST (VAT) forms, being available to reply to emails, queries, complaints etc?

I don't think so. However, if some nice, attractive, well-connected and competent young lady would like to apply for the job of handling the non-picture-taking aspects for me, be an agent, oranizer and accountant, I'd certainly reconsider   .

[Ray]

Interestingly, it's been my experience that when professional photographers, writers, reviewers, and photographic educators get together they rarely discuss equipment's image quality, but rather prefer chatting and bitching about features and functionality. Food for thought.

As pros know all to well, great image quality, while important, does not trump handling deficiencies. Sure, if you're shooting snaps of your cat on the living room couch, or on your summer vacation, poor camera design may not be too big an issue. But if you shoot 100,000+ frames a year and live with a camera in your hand 10 hours a day, every poor design aspect becomes like a thorn and can be hard to ignore.

If the above statement from Michael's review is true, I wonder why it's true? I'm having trouble reconciling the following statement..As pros know all to well, great image quality, while important, does not trump handling deficiencies.

If great image quality does not trump handling deficiencies, why would anyone bother with large format cameras? Don't tell me an 8x10 field camera, a Mamiya RB67 or even a Mamiya ZD handles as well as any reasonably well designed 35mm camera. The point has been made more than once in this forum's MFDB section that those who spend $40,000 on a P45 do so because they are concerned primarily with absolute image quality; the maximum quality they can get.

However, I can understand that having got the image quality one desires or can afford, then the field of dicussion can be narrowed down to ergonomic issues and additional features which are useful, such as the auto ISO feature of the D3. I can't understand why Canon has not long since introduced this feature, as well as ISO bracketing, considering they have, till now, been the leader in low noise at high ISO.

[Graeme Nattress]

Say you sample a sine wave with a frequency of 20KHz at 40Ksps, and the sample period is centered on the parts of the sine wave where it crossing the zero point?  What do you get?  Nothing.
[a href=\"index.php?act=findpost&pid=165986\"][{POST_SNAPBACK}][/a]

Correct, in your example you get nothing because you violated sampling theorem. It states "less than", not "less than or equal to" and in your case, it's "equal to". As long as it's band limited, ie a pure sine wave at a frequency less than half the sample rate, the reconstruction filter (which we don't really have in image processing - it was the gaussian spot on the CRT, but that's gone now...) will reproduce that sine wave, correctly. I used to think as you do that the sine wave could not be reproduced, but then I learned more and more until I finally got it, playing with digital audio samples and waves I'd created, watching the results of the reconstruction filter on an analogue oscilloscope so I could see what was going on.

Graeme