Nikon D70 vs D2Xs

[NikonUser]

D70 = 6MP; D2Xs=12MP
As close to identical a set-up as I could get; simply switched camera bodies and slight re-aligment.

4T close-up lens + MF 105mm Micro Nikkor @ full extension +16 cm bellows; 1/250s @ f8 on lens, ISO 200, single SB800 flash; stack of 18 images @0.1mm, Helicon Focus 4.1.
(The D2Xs will go down to ISO 100 which may give a better resolution but the D70 goes only to ISO 200).

Crops from the full-sized images (inserts). Full frame width=6mm.

Post processing with Photoshop to get the best final images.

Trying to convince myself that I like the D2Xs image the best ( I'll let viewers guess which is which).
Bottom Line: Little, if any, difference between 6 (relatively cheap) and 12 MP (expensive) for this type of image.

Bug is a female Crane Fly (Fam: Tipulidae: Nephrotoma ferruginea)

[DaveW]

Even if there were a great deal of difference in resolution, as I understand it we would not see it as the extra megapixels are lost on a computer monitor or digital projection screen, since it only resolves about the equivalent of a 4 megapixel camera I believe? You will only really notice any difference on a print.

Another 10 megapixel Nikon user

DaveW

[rjlittlefield]

Great comparison!

I'm going with the top one as from the higher resolution camera, based on sharper rendition of the finest textures such as on the palps and the fine hairs of the antennae.

I agree that the difference is minimal at the resolution that's shown here.

However, what's shown here seems to be well below full resolution even for the 6 MP camera. As I do the calculation, working from the full-frame insets, the crops correspond to full-frame images that would be only about 1025 x 1538 pixels. In other words, the crops are showing us only about 1 out of every 4 pixels for the 6 MP camera, and only 1 out of every 8 pixels for the 12 MP.

Given this amount of reduction from camera resolution, it's an interesting question why we're seeing any detectable difference at all.

I think there may be a clue about that in the upper antenna. Notice that in the region 3-5 segments from the end, there is clear doubling of some of the large bristles in the lower image. In addition, there is a general sense of "smearing" in the fine texture all along the antennae in the lower picture, but not the upper.

Both of these effects indicate that the software had trouble properly registering multiple frames in the lower picture. In theory, registration is more robust when more pixels are used to cover the same details, which again would give the edge to the 12 MP camera.

There's no guarantee, of course, that whatever registration algorithm is used by HF conforms to that theory. I have no idea what approach HF uses. Several versions back, it used to make this sort of error fairly often, but I haven't seen it recently. These images are interesting for that reason also.

In any event, those are the features that I notice. I will be intrigued to hear if in fact the top image comes from the 6 MP camera.

--Rik

[NikonUser]

Dave:
What little printing I do is on 13x19" paper. And yes, there is a noticeable difference between a D70 image and a D2Xs image when printed at, what is usually, a great enlargement of the actual subject.

Rik:
Correct, top is from the D2Xs.
I took 800px (wide) crops from the full frame images and one can clearly see problems with the D70 (lower image). (I guess the 'problem' is with HF). Perhaps I should upgrade to HF 4.5 from 4.1.
As shown by Photoshop, the D70 full frame is 3,008 x 2,000 pixels; the D2Xs is 4,288 x 2,848

[DaveW]

There are other differences between the two cameras than just megapixels. They are both withdrawn now, but the D70 was first announced in January 2004 whereas the D2Xs was announced in June 2006 30 months later, and in digital camera terms regarding improvements in circuitry surrounding the sensor that is quite significant. Also the D70 is a CCD sensor and the D2Xs a CMOS one. See:-

http://www.dpreview.com/reviews/compare ... 0&show=all

As the D2Xs has a high speed 6.8 megapixel cropped mode it would have been interesting to have seen that compared to the 6 megapixel D70 to see what improvements had occurred in sensor technology in the 30 months between cameras.

I am not saying the extra pixeled Professional D2Xs does not produce better images, since I use a 10 megapixel D200 myself, only that due to limited computer monitor screen resolution it would be hard for us to see the difference. In fact the new D300 prosumer camera, at a far smaller price, now beats the pants off the D2Xs although having a similar sensor size from all accounts, which is what another 14 months of sensor development and it's surrounding circuitry does.

DaveW

[NikonUser]

Thanks Dave.
Fair criticism. I have edited the title to better reflect the actual comparison; i.e., Nikon D70 vs D2Xs rather than Nikon 6 vs 12 MP.
Presumably the D700 is better than the D300.
How do you feel about the D90 and D3?

[g4lab]

Anybody guess why the top one shows so much more chromatic disturbance than the bottom one?

[rjlittlefield]

Anybody guess why the top one shows so much more chromatic disturbance than the bottom one?

That's a good question.

Some registration & stacking algorithms interact with color aberration in interesting ways, for example turning longitudinal color aberration into what looks like transverse color aberration (TCA) as seen here. We can see that the two stacks have been registered differently, so that's one candidate.

On the other hand, it's conceivable that the two sensors interact differently with the lens, so that even the original frames show different amounts of TCA. I wouldn't expect that to be nearly this dramatic, so it would be very interesting to know if that's what is happening.

Looking at some of the corresponding single frames should help resolve this.

--Rik

[acerola]

Anybody guess why the top one shows so much more chromatic disturbance than the bottom one?

Do NikonUser used the same aperture settings for the two set of images? Different apertures can result different CA.

[P_T]

Very interesting discussion.

I have a question. Are these two cameras from the same line of DSLR? ie. is it like comparing two top end DSLR (eg. Canon EOS 1D MkIII vs MkII) or is it like comparing a top end DSLR to a midrange DSLR, (eg. Canon EOS 40D)?

[DaveW]

No, the D2xs was Nikon's top of the range highest megapixeled APS-C sensored Professional camera when it was introduced. It has not really been replaced yet except the more modern APS-C sensored D300 has the same megapixels and most of it's features, but in a Prosumer body.

The D2Xs is expected to be superceeded by a full frame FX sensored "D3X", or whatever they choose to call it, using Sony's new 24 megapixel sensor. The new full frame D3 is Nikon's high speed but "lower megapixeled" sports camera. There were a lot on show at the Olympics, but it is really a replacement for the high speed APS-C sensored D2Hs.

The APS-C sensored D70 was mid range, just above entry level with the 10 megapixel Prosumer cameras like the D200 between it and the D2Xs.

DaveW

[NikonUser]

.......... In fact the new D300 prosumer camera, at a far smaller price, now beats the pants off the D2Xs although having a similar sensor size from all accounts, which is what another 14 months of sensor development and it's surrounding circuitry does.

DaveW

I spent several years assembling a photo reference collection of local flies (OK, so I need to gat a life) using flat bed scanners (probably prosumer level); very satisfied. Then I got into macrophotography with the Nikon D70 and stacking. Realised how much better these stacked images were and thus had to start over again. I then upgraded from the 6MP D70 to the 12MP D2Xs and noticed a significant difference in large prints. Start again. Now DaveW tells me that the new Nikon D300 "beats the pants off the D2Xs". This bothers me.
I find it hard to see how I could get significantly better images with the D300 and can't afford the time to again start again. It's more than a 1 lifetime project and I have less than 1 lifetime left.

Original Helicon Focus stacked image is 70 MB; cropping out the blank sides reduced it to 36MB; the head is the true size of the image.

So my question: how would a D300 image "beat the pants off" this D2Xs image?

It's a Picture-winged fly (Family: Ulidiidae); 8 mm long and typical of the size of flies I photograph - makes a very nice print on 13x19 paper.

[DaveW]

The D300 has a similar 12 megapixel sensor as the D2Xs, but obviously other features on the camera have been improved in the meantime and it is far cheaper being a Prosumer rather than a Pro camera. Whether you would see a lot of difference in the image quality is another matter, but you would be now paying far less to get it. (The prices now being quoted for the D2Xs are for a discontinued clearance model and do not take inflation into account as to what it's equivalent price would now be if just introduced)

http://www.kenrockwell.com/nikon/d300.htm

Make allowances for Ken though as he likes to crank his cameras up to get cartoon or Disneyesque colours. We in the UK prefer them more muted and natural!

The high speed D3 is a similar 12 megapixeled camera, but with these spread over the larger full frame sensor, so even better image quality due to larger pixel sites.

http://www.dpreview.com/reviews/compare ... 3&show=all

The full frame 24 megapixel Pro replacement for the D2Xs is supposed to be released this year. Some say Nikon is now intending to call it the D4 not the D3X, but nobody really knows. However Sony has just brought out the Sony Alpha A-900 with the same 24 megapixel sensor that will probably be in the new Nikon. At $3000 (£1700) this is way below other lesser megapixeled Pro cameras, so this may force the pricing of Pro cameras down in future.

http://www.dpreview.com/previews/sonydslra900/

The rate digital cameras are evolving, on average a replacement model every 18 months, few of us can afford to keep up with technology, we have to put up with what we have.

Anyway NikonUser your pictures are great, so what are you moaning about? Even in another ten years will digital cameras have evolved that much in image quality they will look any better on a computer screen, maybe if you are printing poster sized prints from them but otherwise I doubt it.

DaveW

[Joseph S. Wisniewski]

4T close-up lens + MF 105mm Micro Nikkor @ full extension +16 cm bellows; 1/250s @ f8 on lens, ISO 200, single SB800 flash; stack of 18 images @0.1mm, Helicon Focus 4.1.
(The D2Xs will go down to ISO 100 which may give a better resolution but the D70 goes only to ISO 200).

OK, by the numbers...

Assuming that a manual focus 105mm micro Nikkor (unlike the AF version) has no pupillary magnification factor.

The aperture is 105mm / f8 = 13.1mm

The focusing extension of a MF 105mm is 105mm/2 = 52.5mm

The effective length is 105mm + 52.5mm + 160mm = 317.5mm

The effective aperture is 317.5mm / 13.1mm = f24.2

Let's call it f22, that's within 10%.

Now, taking a quick trip to the Cambridge In Color diffraction calculator...

http://www.cambridgeincolour.com/tutori ... graphy.htm

We check the "Set Circle of Confusion = Twice Pixel Size?" and set the camera type to "digital SLR with CF of 1.5X"

Check both 6mp and 12mp

The Airy disc diameter for f22 is 29.5um

At 12mp the pixel size is 5.7um, the max circle of confusion is 11.3um

At 6mp the pixel size is 8.0um, the max circle of confusion is 16.0um

It simply doesn't matter which camera you use, both of them have higher resolution than your lens at that extension.

Anything you like about one image more than the other has to do with other characteristics of the sensor, like the better IR immunity of the D2X. Since you're using flash, the issue of D2X having mirror lockup and D70 not having it is pretty trivial.

If you really want to see a clear difference between a D70 and a D2X, as far as detail, you need to get down to an effective f11. That's a 14.8um Airy disc, the D2X can take advantage of that, the D70 can't.

Unfortunately, that means opening up wide open, which probably isn't the best idea with that 4T on the 105mm.

Now, for the magnification:

The 4T is (from the handbook) 2.9 diopters.

The MF 105mm Micro Nikkor is 1000mm/105mm = 9.5 diopters.

The combination is 1000mm/(2.9d + 9.5d) = 80.6mm

The magnification is (317.5mm / 80.6mm) - 1 = 2.9x

That's "easy" magnification. Got an old 55mm f2.8 micro Nikkor floating around? I've seen them for around $100 on the bay of 'e'.

The lens (at least the "rear node") needs to be 55mm * (2.9 + 1) = 214.5mm from the sensor for your same 2.9x magnification. With no 4T needed

Mounted reversed on the bellows, you can easily get the lens 214.5mm from the sensor.

The aperture is 55mm / f2.8 = 19.6mm.

Wide open, at 214.5mm from the sensor, you're effective aperture is 214.5mm / 19.6mm = f10.9.

That's going to be a whole lot cleaner than a 105mm f4 with a 4T.

Less CA, so no color fringes in the image.

[Joseph S. Wisniewski]

Anybody guess why the top one shows so much more chromatic disturbance than the bottom one?

Were both images shot raw?

Processed by the same software?

Identical exposure, and no blown channels on the background on either?

For example, Nikon Capture has a wonderful automatic CA correction feature (if you remember to enable it) that disengages when the entire periphery of the image is blown. So an image with a bright background will show much more CA if you blow the background.

Now, on a happier note, when using a 4T on a macro lens at high extension, you will get reduced CA if you mount the 4T to the 105mm micro Nikkor "backwards", using a coupling ring, instead of "forwards" by just screwing it into the filter ring of the 105mm.