40 replies to this topic

[craig]

A good underwater camera is a lot more like a good studio camera than a good sports camera. Great high ISO is useful only occasionally and high frame rates even more rarely useful. If you double the pixel count of the D3, you could still expect good ISO 1600 and at least 4 frames per second. Both of those specs should be just fine. In exchange, width and height would be improved 40% yielding a very useful increase in resolving power. Perhaps that extra resolution isn't useful to you, but you don't get a second chance to reshoot some images and that's doubly true underwater.

Rarely does a underwater Canon pro body shooter choose a 1D over a 1Ds. The same reasoning would apply for a D3/D3x.

It's almost certainly true that the D3 will be the best ever Nikon for underwater. The D3x would likely be even better.

[vincentkneefel]

I agree with Liz. This is what Ken Rockwell noted about the number of megapixels on fullframe:

12MP vs 21 MP is usually meaningless. Above 12 MP on full frame, like above 6MP on DX, most of the pixels will be showing lens and technique flaws rather than more sharpness.

The D3 has 4,256 horizontal pixels while the 1Ds Mk III has 5,616. That's 32% more, which, presuming one has perfect optics, is starting to become visible in large prints of 20 x 30" (60 x 80 cm) and larger. It's not a night and day difference, or even visible at small sizes like 13 x 19" (30 x 50 cm).


Cheers,

Vincent

[james]

Hmm..

Hi Vincent - where does the Ken Rockwell quote end and your opinion begin? I want to make sure not to insult you when I laugh at what Ken Rockwell wrote. What lens and technique flaws do you think he's means exactly????

For one what does the pixel size have to do with shooting technique? If you use bad technique, you can be shooting a 3 megapixel camera and it won't make a bit of difference.

It's well accepted that all of the quality Canon lenses out-resolve even the highest megapixel FF sensors - so that's not an issue either. Or not one I understand in the context of this discussion.

I personally think the extra 32% pixels are tremendously useful. Look at it the other way - that's the difference between shooting a 100mm macro lens, or shooting a 70mm macro lens and being able to crop 30% That's a big difference IMO.

I agree that I'd take a GOOD 12 megapixel sensor over a crappy 22 megapixel sensor. But who says the 22 megapixel sensor is crap? I'm sure that per pixel, it will be at least as good as Canon's 16 megapixel sensor, which is excellent.

Cheers
James

[scorpio_fish]

Hmm..


For one what does the pixel size have to do with shooting technique? If you use bad technique, you can be shooting a 3 megapixel camera and it won't make a bit of difference.

Following this logic, if I own a 1ds and plan on shooting in sloppy manner, should I set the camera to lower resolution so it won't show up as much?

If your camera has so many pixels that it starts showing lens flaws, do we start adding blur while in PS?

Half of everything Ken Rockwell says is true.

[craig]

Yes, my goal in selecting a camera has never been to make sure it can't expose my lens and technique flaws. ;-) Otherwise, I could house an iPhone.

The good news is that Nikon has already announced all the camera Ken Rockwell will ever need. If only that would help him go away. Never again being treated to a lens review where Ken doesn't bother to look at the lens would be a small price to pay. Good thing we tried the 10-17 despite Ken's useless review.

"This is a special purpose, fun and special purpose lens. Did I remember to say special purpose? I'm not sure what to make of it until I get to play with one."

Now THAT'S a quality review!

While it's fair to say that the difference between 12 and 21MP is of no value to you or KR, it is indefensible to claim that no usable difference exists (as evidenced by objective testing that proves otherwise). Given the choice between extra pixels that I *may* not use and extra ISO/fps that I certainly *won't* use, I'll take the pixels.

[herbko]

If nothing else the 21MP will make the Dx crop or digital zoom feature that Nikon will put in more useful.

The advantages of the 12MP are cost and speed and most importantly better signal-to-noise and dynamic range. The cost in the required larger storage is negligible. If speed is not a factor to you then the only trade-off is the lost of about 1/2 stop in signal-to-noise and dynamic range.

You lose about a factor of 2 in signal-to-noise in making the photosites 1/2 as big. However, if you down sample back to the lower resolution you'll get 1/2 of that the back. So for a given print resolution doubling the number of photosites will only cost 1/2 stop in noise and dynamic range.

[loftus]

Someone should have warned Vincent never to mention K-- R------- on this site; I did it once and I'm still traumatised by the experience.
Can one of you guys who really understands this megapixel / noise / dynamic range relationship either explain it, or provide a good link that I can read about it?

[craig]

Right, herb, and I'd add, for a given shot, it's not clear what the threshold for useful dynamic range/noise is. It's only a disadvantage to give up that stop or half stop if the image actually benefits from the difference. To put it in simpler terms, if you are always shooting at base ISO then noise is probably not an issue unless every shot is a sunburst.

[vincentkneefel]

No more reffering to K.. R...... for me

James, thank you for your constructive reply. I am also very keen on learning more about the megapixel / noise /dynamic range relationship.

Someone should have warned Vincent never to mention K-- R------- on this site; I did it once and I'm still traumatised by the experience.
Can one of you guys who really understands this megapixel / noise / dynamic range relationship either explain it, or provide a good link that I can read about it?

[james]

Thanks Vincent,

I had a look at your website and enjoyed it very much. I especially liked the shots from the Azores - great marine mammals. Also the mobula rays from the Sea of Cortez. How did you get that jumping Mobula shot???? WOW.

Cheers
James

[vincentkneefel]

Thank's a lot James. The Azores was my first U/W photography trip. Still had to get used to everything, but was a lot of fun I hope to return next year with some more experience and a Fish Eye lens. The Azores is a really amazing place to go to, one of the best places I have visited so far. If you want any more information on the Azores, please let me know.

The Mobulas are amazing, you need to go out in a panga and look for the big shools. It's not hard to find them when they are there, because they are jumping most of the time. Sometimes there are more than (an estimated) 1000 of them. I found it very hard to take (good) pictures of them when I was in the water, because it was just pitch black all over with Mobulas. The one (sharp) Mobula jumping I got was when we followed a school from a distance (manual focussed and tried to predict when they jumped). Unfortunatey I only have a 55-200, I am sure with a 400mm you can get stunning shots.

Gr.

Vincent

Thanks Vincent,

I had a look at your website and enjoyed it very much. I especially liked the shots from the Azores - great marine mammals. Also the mobula rays from the Sea of Cortez. How did you get that jumping Mobula shot???? WOW.

Cheers
James

[craig]

The thing with Ken Rockwell is that he plays the devil's advocate in order to convert controversy into page hits and advertising dollars. He's definitely a "no such thing as bad publicity" type. KR makes no attempt to be an authoritative source, or even a correct one, and his opinions should read with that in mind ;-)

I made the mistake of referring to him once and got burned as well. I didn't notice that his lens review contained SIMULATED test shots! Ooops.

[Liz Hanks]

The advantages of the 12MP is cost and speed and most importantly better signal-to-noise and dynamic range. The cost in the required larger storage is negligible. If speed is not a factor to you then the only trade-off is the lost of about 1/2 stop in signal-to-noise and dynamic range.

You lose about a factor of 2 in signal-to-noise in making the photosites 1/2 as big. However, if you down sample back to the lower resolution you'll get 1/2 of that the back. So for a given print resolution doubling the number of photosites will only cost 1/2 stop in noise and dynamic range.

OK, this is useful. So if we had a 21MP D3X, we could favor high resolution in shots that had low noise and dynamic range (losing about a stop in each one of those), and down sample images that had noise/DR issues and only lose 1/2 stop compared to the D3.

I can see that if 1/2 stop isn't really noticeable on either front, this combination gives us more flexibility than a lower resolution camera, although not necessarily higher image quality across the board. Also, I don't think comparisons between either the Canon 1D/1Ds or the Nikon D2H/D2X tell the whole story. On the Canon side, I believe there has always been a sensor size/crop factor difference between the high and low res versions of cameras in the same generation. With Nikon, the PJ cameras have all been very low res by today's standards.

With the D3, even though PJ/Sports photographers are ecstatic about its release, wedding photographers are also salivating, and they care about several of the same issues we do. Nikon seems to have broken new ground here and created a camera that isn't at all just for the PJ/Sports crowd. Hopefully it won't be dismissed as such just because it happens to allow us to shoot at 9 fps and ridiculously high ISO values...

Cheers,

Liz

[Paul Kay]

Alex

I've summarised a few points I'd make below:

1. I'm not sure that many Pros will house a D3 unless they actually plan on using one topsides - in my experience most Pros use a 1DS series camera underwater because this is what they already own/use above water! I have said it before and no doubt will again, that I PERSONALLY know no Pros using Nikons at the moment out of choice over the 1DS series cameras, and do know many who have shifted from Nikon to Canon. The reasons are varied and I'm not going to go into them here, but Canon do seem to have the high ground at the moment and Nikon are playing catch up - and please accept that I'm not trying to be controversial merely factual!

2. The underwater photo world is conservative and traditional (yes I'm now being controversial!) in that Nikon have been dominant for a long time because of ttl Nikonos fitting flash units which originated with the Nikonos 5 and its horrid flash connector and then became the norm with film SLR cameras resulting in Nikon having number one position underwater. This is now changing slowly in as much as Canon dSLRs are starting to show up more in housings and the compact market is shifting the traditional allegances with regard to those who upgrade to dSLRs.

3. The D3 has a lot going for it in terms of lenses for potential wide-angle use underwater. The 20mm and 18mm lenses offer the potential for careful optimisation behind dome ports which could seriously challenge Canon's existing lenses for high end use, however.....

4. I can see lens flaws using a 1DS camera both above and below water and adding in a dome port simply makes things worse. Increasing MPixels gives better central image quality but does not deal with edge quality issues so I'd actually say that the D3 currently sounds like a well and viably specified camera to house, but....

5. Nikon do need to deal with their macro lens line up - I hear many complaints about the new 105 VR not being as good as the older lens (although I haven't tried it myself) and the used price of the none VR has increased (if you can get one). The 60 is a superb lens but is looking old and needs to be replaced with a faster internal focus version. And...

6. Just how many shooters will see advantages of the D3 over the DX series will depend on how and what they shoot and if they do see FF advantages for their style, they are probably already shooting Canon!

Lastly, I personally would be interested in a housed D3 but am far from sure that it would provide sufficient useful advantages to really make one worth my while buying. And a rumoured D3X has severe competition to face in the 1DSMkIII whatever its high ISO advantages as Nikon will have to provide a full series of high quality 'L' type lenses to compliment it and compete with those available from Canon.

Just a few thoughts!!!

[Alex_Mustard]

Here is a fun point of view on the debate:

http://www.bythom.com/topten.htm

Alex

[Paul Kay]

I have always thought that a small clock fitted into a body cap might give many cameras a new lease of life!

[Dan Schwartz]

I'll ask Mautner when I see him at PhotoPlus in 4 days...

Mainly, are the photos he's showing shot the lazy way with JPEG; or are they CCD-RAW? If CCD-RAW, is he using Nikon's software? ACR? CaptureOne? Each has it's inherent weaknesses & strengths...

But also, Watch Out for so-called high ISO settings, as sometimes "noise reduction" (a/k/a blurring) is used, either in the JPEG engine or in the RAW conversion software...

Been there, done that, have the tee shirt...

I'm not too sure that Canon will be the clear leader in low-noise images after the release of the D3. Have y'all seen the test drive w/ images by wedding photographer Cliff Mautner:

http://cliffmautner.typepad.com/

Check out the JPEG quality (these are all from the Nikon JPEG engine I believe) images at ISO2000 and ISO3200? Wow!

[craig]

Any RAW files will be converted with NX at this point but I'm sure Adobe has seen samples by now. Most images are JPG because that's all that Nikon has been allowing. Rumor has it that firmware is finalized and restrictions on images have been lifted. It doesn't help only viewing postage stamp-sized examples though.

What I've gleaned from dpreview is that images are good to 3200 and passable at 6400. That seems far more than adequate for underwater. Double the pixel count and you should still be good at 1600. I bet the 1Ds3 will have ISO performance totally adequate for underwater use.

What must it be like to be a housing manufacturer right now and feel a serious need to house 4 different yet compelling new camera bodies? Then there's the new wide angle lenses that may be problematic.

[Alex_Mustard]

Here are some high ISO 100% crops from a reliable contributor on DPReview. No noise reduction, apparently.

http://forums.dprevi...essage=25199140

For housing manufacturers I still think that the D300 should be worked on first (once the Mk3 housings are done) and then start on the D3 housings. This makes the most sense financially to me.

I guess we will see what mock-ups there are at DEMA/Antibes in the next few weeks.

Alex

p.s. Craig, I am yet to see any decent tests of Nikon's FX format with wide angle. Have you seen any?

[loftus]

Alex, not sure what you mean by decent, but there are some with the new 14-24 here;
http://www.craigsact...3FirstTest.html

Edited by loftus, 16 October 2007 - 04:10 AM.