Ad
Ad
Ad
Pages: « 1 ... 4 5 [6] 7 »   Bottom of Page
Print
Author Topic: Nikon D800 - seems harder to justify Medium Format  (Read 30524 times)
gerald.d
Sr. Member
****
Offline Offline

Posts: 388


« Reply #100 on: May 07, 2012, 06:57:20 AM »
ReplyReply

I'd like to know in more detail too, it is a bit hard to find out since the Rodenstocks/Schneiders are rarely tested, one have to find shots here and there on the net and guess from that. The 24mm TS-E II has much better performance than the PC-E 24mm for sure, I have one myself. It is also not as handicapped in movements like the PC-E 24 either. What I do know is that the DSLR lenses get corner sharpness issues at high resolution (the TS-E 24 too), but I don't know for sure how large corner sharpness issues on the wides are for the Rodenstock/Schneiders, maybe problems are larger than I think there.

For me personally the comparison Nikon D800 + PC-E 24 vs Aptus 75 + Schneider 35 XL would be a great test. About the same amount of pixels, same FOV, both able to tilt and shift. I plan to get a 35 XL but don't have it yet unfortunately.

Comparing the 23mm Rodenstock with 70mm image circle with the TS-E 24 II is an interesting lens quality test (would not be surprised if the TS-E wins, it is a great lens), but system-wise it is more relevant to compare it with a 35mm or a 40mm with 90mm image circle. 135 digital only has 36x24mm sensors so they need shorter focal lengths and higher lpmm numbers than a 48x36 or 54x41mm sensor to achieve the same FOV and resolution.

I don't think it is that valuable to mix in price into the equation, speciality products like MF is very expensive and will always have poor price/performance compared to mass-market products. If MF has something to offer and is not too expensive for the professional then it is ok, but if it is just expensive without offering anything then it is worthless...

Thanks for taking the time to reply.

Regarding your third paragaph, I think we're both coming at this from slightly different perspectives.

My 23 vs 24 comparison is because a 23 Rodenstock on an Alpa with a IQ180 back will give (almost) the same FoV as a 24 TS-E on the HCam with the same IQ180 back. For me, this is one of the decisions I need to make in the coming months. I want to go wide angle with my IQ180, and need to decide which option to go with - I can't see myself affording (or indeed, wanting) both.

I recognise that if you're doing a MF vs 135 comparison, then it's not valid due to the different sensor sizes.

Regarding your final paragaph, I'm not so convinced that the professional and amateur approach to this subject is particularly different - both will certainly take price into consideration.

For the "true" professional (i.e. one whose primary source of income is derived from photography), there are of course fairly simple ROI calculations to determine whether or not it is worth investing in particular pieces of kit. At the end of the day, those calculations are very much impacted by the capital outlay necessary and depreciation over time, so price is important.

For an amateur (such as myself), MF of course has a barrier to entry based on price, but once you've decided that you're prepared to spend what it takes to join the game, there will always be choices to be made, and those choices will be driven primarily on bang for the buck. For me right now, it's Alpa + one lens, or HCam + several lenses, for the same investment. And if one of those lenses meets or beats the one lens on the Alpa, then (of course based on personal requirements) it's a no-brainer.

The existence of cameras such as the HCam really do break down the technical barriers between MF and 135, and give one the option of cherry-picking the best bits of kit from both worlds. Given that (apparently) there are a range of 135 lenses that are just as good as - if not better than - the best of the MF digital lenses, it wouldn't surprise me if there were further developments in this area.

« Last Edit: May 07, 2012, 07:09:47 AM by gerald.d » Logged
torger
Sr. Member
****
Online Online

Posts: 1588


« Reply #101 on: May 07, 2012, 07:57:00 AM »
ReplyReply

Ah, I understand. Stefan Steib may be able to provide files HCam + 24 TS-E, and at the Getdpi forum you would probably find someone that has IQ180 and a 23mm, and perhaps even some posted examples.
Logged
BJL
Sr. Member
****
Offline Offline

Posts: 5163


« Reply #102 on: May 07, 2012, 08:02:43 AM »
ReplyReply

I dont think that being lense-limited need to be such a bad thing.
Indeed, it seems to me that:

1. The best use of resources is a design that roughly balances the limitations from the various components such as lenses and sensors, rather than having any one vastly out-perform another, so that much of its good performance will be wasted.

2. In this roughly balanced scenario, the limitations of several components will be significant: you should be able to see that image quality could be improved a bit by either a better lens or a better sensor. And for this, sensor resolution should be balanced against the best that the lens system can offer, meaning in the center of the image delivered by the best lenses at their optimum apertures, which will put the sensor resolution well beyond what you see near the edges with most lenses. If your sensor cannot clearly reveal edge and corner softness, it is not getting the most out of your lens system!

3. When the costs of further improvements are unequal, it probably makes sense to push more on the side where the costs are lower --- and this is almost certainly 35mm format sensors rather than lenses. (However, it might be the reverse in 645 format, if the cost of sensor progress there is as high as it seems to be.)

4. Oversampling the "signal" delivered by the lens is a good thing, for reasons much discussed of late. (Except perhaps if one almost never photographs any of the fine regular stuctures produced by man, or plant, or animal, nor any sharp lines of high contrast like the edges of harsh shadows.)
« Last Edit: May 07, 2012, 09:57:32 AM by BJL » Logged
TH_Alpa
Full Member
***
Offline Offline

Posts: 214



« Reply #103 on: May 07, 2012, 08:09:39 AM »
ReplyReply

Dear Gerald,

There are such tests for these lenses if you do some research, I've seen such on another forum.

But I would in your case really make my own testing, in my own shooting conditions. The Rodenstock 23mm HR is a fantastic lens, I've used it on many cameras. It is however a very short FL (so is the 24 mm TS) and needs extrem care and precision when focusing and setting movements like shifts, tilts or swings. I can only say that it needs a camera allowing maximum precision.

Concerning Alap, I would suggest you to get in touch with your local Alpa representative, or then to contact Alpa directly in Switzerland. Please also feel free to contact me, should you have specific questions.

Best regards
Thierry

Thanks for taking the time to reply.

Regarding your third paragaph, I think we're both coming at this from slightly different perspectives.

My 23 vs 24 comparison is because a 23 Rodenstock on an Alpa with a IQ180 back will give (almost) the same FoV as a 24 TS-E on the HCam with the same IQ180 back. For me, this is one of the decisions I need to make in the coming months. I want to go wide angle with my IQ180, and need to decide which option to go with - I can't see myself affording (or indeed, wanting) both.

I recognise that if you're doing a MF vs 135 comparison, then it's not valid due to the different sensor sizes.

Regarding your final paragaph, I'm not so convinced that the professional and amateur approach to this subject is particularly different - both will certainly take price into consideration.

For the "true" professional (i.e. one whose primary source of income is derived from photography), there are of course fairly simple ROI calculations to determine whether or not it is worth investing in particular pieces of kit. At the end of the day, those calculations are very much impacted by the capital outlay necessary and depreciation over time, so price is important.

For an amateur (such as myself), MF of course has a barrier to entry based on price, but once you've decided that you're prepared to spend what it takes to join the game, there will always be choices to be made, and those choices will be driven primarily on bang for the buck. For me right now, it's Alpa + one lens, or HCam + several lenses, for the same investment. And if one of those lenses meets or beats the one lens on the Alpa, then (of course based on personal requirements) it's a no-brainer.

The existence of cameras such as the HCam really do break down the technical barriers between MF and 135, and give one the option of cherry-picking the best bits of kit from both worlds. Given that (apparently) there are a range of 135 lenses that are just as good as - if not better than - the best of the MF digital lenses, it wouldn't surprise me if there were further developments in this area.


« Last Edit: May 07, 2012, 08:26:38 AM by TH_Alpa » Logged

torger
Sr. Member
****
Online Online

Posts: 1588


« Reply #104 on: May 07, 2012, 08:53:42 AM »
ReplyReply

I can only say that it needs a camera allowing maximum precision.

If you are going to use the lens at ~f/11 it will work out alright as it does on the DSLR, minor precision errors is hidden by DOF. If you want to use it at f/5.6 more precision is required to get focus distance accurate and avoid leftover tilt (it is easier to get it right when there should be tilt than when there should be no tilt), but then the TS-E 24 corner sharpness will be poor anyway so the lens is not a good choice for that. Possibly the Rodenstock 23 on an Alpa is up for f/5.6 use, I don't know.

In MF I am myself an f/11 guy, and that has saved me a lot of pennies and headaches :-)
« Last Edit: May 07, 2012, 08:56:28 AM by torger » Logged
uaiomex
Sr. Member
****
Offline Offline

Posts: 1005


WWW
« Reply #105 on: May 07, 2012, 09:55:48 AM »
ReplyReply

 Cheesy

?
What about those of us shooting 44x33?
Logged
Gigi
Sr. Member
****
Offline Offline

Posts: 422


WWW
« Reply #106 on: May 07, 2012, 10:37:57 AM »
ReplyReply

The quote was referencing:
"It's only a very thin line separating MF shooters under 44X33 and MF shooters over 44X33. I think that whatever touches the former, eventually will touch the latter."

Some of us are shooting on the thinnest line of all, 44x33 exactly.
Logged

Geoff
Dick Roadnight
Sr. Member
****
Offline Offline

Posts: 1730


« Reply #107 on: May 07, 2012, 12:44:49 PM »
ReplyReply

...an image from a normal (not multishot) 22MP back has less resolution that given by a D800(E) with good enough lenses, and up-sampling will not overcome that.

¿Is this your opinion or whose?
¿Or the result of whose test?
¿Do they make "good enough" lenses for Nikon?... (To be good enough to get good res out of such a small sensor they would have to be better than the best MF lenses.)
Logged

Hasselblad H4, Sinar P3 monorail view camera, Schneider Apo-digitar lenses
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7678


WWW
« Reply #108 on: May 07, 2012, 02:43:22 PM »
ReplyReply

Hi Bernard,

My idea, which you quoted a bit out of context, is that we still have development going on. As we have 24 MP APS-C, from Sony and now Nikon, it is obvious that we can also build a 54 MP sensor. Such a sensor is simply an 24 MP APS-C sensor scaled up to full frame. There are certainly a lot of lenses that would match that resolution.

On the other hand we are approaching physical limits. I don't think we will go much beyond 3 micron pitch, that would be 96 MP. There are some advantages with decreasing pixel pitch. Aliasing will be reduced and real resolution is always better than interpolated resolution.

I don't know if there is need for more than 36 MP. If image quality is improved with 54 MP (or 96MP) than just fine.

Best regards
Erik





Not, it is not, but the question is whether it has to be or not?

- In terms of resolution, who needs more than 36 tack sharp pixels [please don't answer "crazy stitchers"...  Wink]?
- In terms of DR, who needs more?

All this is not done for the sake of improving camera technologies, but to solve practical issues met by photographers. Having spent one month with the d800, I see very few issues left to solve.

Cheers,
Bernard

Logged

BJL
Sr. Member
****
Offline Offline

Posts: 5163


« Reply #109 on: May 07, 2012, 02:51:36 PM »
ReplyReply

Dick,
    The resolution tests are out there, and for one thing they seem to show that the D800(E) has higher resolution that is possible with a 22MP back, due to the Nyquist limit. And yes, there are Nikon (and Zeiss) lenses with enough resolution to keep up. As has been pointed often in this forum, there are a good number of 35mm format lenses that keep up with the resolution on 16MP, 18MP and even 24MP in "APS-C" formats, and the D800 is no more demanding (in lp/mm) than a 16MP Nikon DX sensor, and less demanding than an 18MP Canon or a 24MP Sony in APS-C.

What would make you think that the 7,360 x 4,912 pixels of the D800(E) sensor would fail to deliver more "lines per picture height" than a 5344 x 4008 (22MP) sensor, given the difference of 22 - 38% in linear pixel count?
Logged
ErikKaffehr
Sr. Member
****
Online Online

Posts: 7678


WWW
« Reply #110 on: May 07, 2012, 11:04:38 PM »
ReplyReply

Hi,

The question is not only about resolution but also MTF, and than we would need add sharpening to the equation.

I have little doubt that the best primes would work just fine with the Nikon D800.

My guess is that the D800 would give better image quality than 20-25 MP MFDB. We also need to keep in mind that we may have different croppings. A square image on the D800 would be 24 MP.

Best regards
Erik


Dick,
    The resolution tests are out there, and for one thing they seem to show that the D800(E) has higher resolution that is possible with a 22MP back, due to the Nyquist limit. And yes, there are Nikon (and Zeiss) lenses with enough resolution to keep up. As has been pointed often in this forum, there are a good number of 35mm format lenses that keep up with the resolution on 16MP, 18MP and even 24MP in "APS-C" formats, and the D800 is no more demanding (in lp/mm) than a 16MP Nikon DX sensor, and less demanding than an 18MP Canon or a 24MP Sony in APS-C.

What would make you think that the 7,360 x 4,912 pixels of the D800(E) sensor would fail to deliver more "lines per picture height" than a 5344 x 4008 (22MP) sensor, given the difference of 22 - 38% in linear pixel count?
Logged

hjulenissen
Sr. Member
****
Offline Offline

Posts: 1683


« Reply #111 on: May 08, 2012, 12:45:30 AM »
ReplyReply

As we have 24 MP APS-C, from Sony and now Nikon, it is obvious that we can also build a 54 MP sensor. Such a sensor is simply an 24 MP APS-C sensor scaled up to full frame. There are certainly a lot of lenses that would match that resolution.
I think that scaling a given sensel infinitely may be practically and economically difficult.

If the probability of a section of the sensor being catastrophically damaged is some "p", then increasing the area of the chip leads to a drastic increase in the number of chips that have to be trashed. If your wafer is circular, then increasing the size of (a rectangular) chip leads to more wasted wafer area. Those are the classic microprocessor arguments against large chips.

It is my understanding that the photolithography used to make camera sensors and microprocessors is quite expensive and specialized gear, and that the camera sensor manufacturers have to borrow equipment/processes originally developed for microprocessors. There is supposedly some upper limit on the area that can usually be formed in "one go", somewhat smaller than 24x36mm, meaning that large image sensor manufacturers have to rely on ad-hoc solutions ("stitching"?) that no doubt further increase cost and time to market.

All based on internet hearsay...

-h
Logged
Fine_Art
Sr. Member
****
Offline Offline

Posts: 1099


« Reply #112 on: May 08, 2012, 02:16:59 AM »
ReplyReply

Unless they are small wafers where you use one per chip, making the most of the round lens output.
Logged
Phil Indeblanc
Sr. Member
****
Offline Offline

Posts: 1107


« Reply #113 on: May 08, 2012, 03:41:26 AM »
ReplyReply

I think he may have been "thrown off" by him saying how heavy the D800 is....At least thats what it sounded like?

While it maybe heavy...if you only use your MF on a stand and not even hold it. Otherwise the SLR is gonna be a lot less weight.
Logged

If you buy a camera, you're a photographer...
ondebanks
Sr. Member
****
Offline Offline

Posts: 833



« Reply #114 on: May 08, 2012, 05:22:21 AM »
ReplyReply

...so now all high-end MF backs have microlenses...

Actually, almost no MF backs have microlenses these days. The last microlensed MF backs were the P30+, H4D31 and H4D40, and a couple of Sinars (eSprit65, eVolution86H).

All the Phase One backs > 31MP, the Hasselblad backs > 40MP, and all the Leaf backs (of any size) lack microlenses. For Phase and Leaf, that's every one of their "high-end" backs, including the IQ and Credo series, as well as all of the higher-megapixel P+ and Aptus ones.

Since about 2009, there's been an interesting DALSA 48MP, 6 micron, 48x36mm sensor (FTF-6080C) which does have microlenses, and impressively low (for MFD) readout noise of 12 e-. I know I tend to kick DALSA MF sensors a lot - they deserve it for being so far behind Kodak in dark noise, q.e., IR sensitivity, and (to a much lesser extent) readout noise - but this one actually looks pretty good for visible light! Apart from it still suffering the DALSA Achilles' heel of high dark noise limiting long exposures, it is closely analagous to the Kodak 50MP, 6 micron, 49x36mm sensor in the H4D-50/CFV-50, but with the addition of microlenses and slightly more DR. The microlenses take its q.e. higher than the Kodak 50MP sensor...but still way short of the q.e. of the equivalent microlensed 6 micron Kodak sensor (the 40MP in the Pentax 645D and H4D40).

But strangely, only one MFD manufacturer is using this chip, and that's in a tethered-only, multishot back (Sinar eVolution 86H) - no LCD, no card, no battery! It's very surprising that those who are currently 100% wedded to DALSA (Phase & Leaf) don't use it in a normal portable back. Anyway, that's yet another tangent to this thread...

Ray 
Logged
torger
Sr. Member
****
Online Online

Posts: 1588


« Reply #115 on: May 08, 2012, 05:30:49 AM »
ReplyReply

Actually, almost no MF backs have microlenses these days. The last microlensed MF backs were the P30+, H4D31 and H4D40, and a couple of Sinars (eSprit65, eVolution86H).

All the Phase One backs > 31MP, the Hasselblad backs > 40MP, and all the Leaf backs (of any size) lack microlenses. For Phase and Leaf, that's every one of their "high-end" backs, including the IQ and Credo series, as well as all of the higher-megapixel P+ and Aptus ones.

Uh, no? The IQ180 etc have the same problems but less so, my understanding was that this was due to better microlens design, not removal of them. I have seen myself those vertical streaks in IQ180 files for shifted non-retro wides, if that is not from the microlenses, what is it?

From http://www.phaseone.com/en/Search/Article.aspx?articleid=1221&languageid=1

"Furthermore the P 65+/IQ160/IQ180 is not recommended with extreme wide non-retro focus lenses like the Schneider Kreuznach 35mm XL and even the Schneider Kreuznack 47 mm XL. Even though these lenses has more than 70 mm image circle, they are producing a very steep angle incoming light along the edges, occasionally causing thin lines to show up in the image."
Logged
ondebanks
Sr. Member
****
Offline Offline

Posts: 833



« Reply #116 on: May 08, 2012, 05:42:30 AM »
ReplyReply

- large and bright viewfinder***
- touch screen interface (some bodies); hard to find a system you can check 100% focus on faster on a specific part of the image than an IQ or Credo
- tools like auto-horizon and auto-keystone which correct the level and pitch of the image in software based on the electronic levels in the back, making every horizon straight and every vertical parallel without manual tweaking
- Flash sync speed with standard strobes rather than dinky flashes (up to 1/1600th)
- More tactile lens response when manually focusing (large focus barrel, actual lens gearing*)
- aspect ratio (some prefer 4:3 or 1:1, especially for verticals)
- waist level viewfinder (some bodies)
- ability to shoot vertical without rotating camera (some backs)
- low ISO without ND filters (useful for dragging shutter in some styles)
- ability to shoot film with same system as digital (some bodies)
- ability to turn sensor on/off independent of the shutter/flash firing (allows to build up exposure with strobes without excessive ambient light, even in bright conditions e.g. interiors)
- ability to crop a vertical and horizontal from the same frame (even 36mp in 3:2 is not enough for many applications when cropped to a vertical)
- ability to use on specific legacy cameras (some folks just plain love Contax, Hassy 500)
- ability to use on speciality equipment like Aerial, industrial, art-repro systems (obviously a niche)
- ability to use on tech cameras
---- rise/fall/shift/swing/tilt on every lens (if IC allows)
---- fully mechanical/traditional shooting
---- absolute best glass, period
---- ground glass (some prefer it regardless of other options)
- compatibility with view cameras
---- close focus possible with many lenses, not just select macros
---- rise/fall/shift/swing/tilt on every lens, not just select TS lenses
---- ground glass (some prefer it regardless of other options)
- less frequent updates required (we still have many happy studio shooters using H25 backs users, don't know many happy Canon 1D shooters)**
- longer software support (original Phase One Lightphase from 1998 is still fully supported tethered in OSX 10.7 and Capture One 6, while the Canon 5D from 2006 isn't even officially supported tethered in LR4 or EOS Utility in OSX 10.7, nor 1Ds II in Windows 7 64 bit)
- consistent shooting speed; an IQ or Credo can maintain it's frame-rate indefinitely with a fast CF card, any Canon/Nikon can shoot much faster but unless you restrain yourself you can easily hit a buffer and the camera won't fire when you think it should. The IQ or Credo will be slower (around 1.2fps for the 40mp model) but it is reliably consistent - you know when you can shoot next and can develop a rhythm.
- larger bodies (for some this will be a big negative, but for others their hands are simply too large to comfortably use a camera like the D800, even with the optional vertical grip)

And honestly it's 6am and I haven't had my coffee yet so I imagine I'm missing quite a few.

O yeah and the look, and the image quality which, yes, is still better :-).

*As opposed to e.g. the Canon 85/1.2 with fly-by-wire focusing and a dinky focus barrel
**This is not just a question of cost since of course the 1D owner could have updated to a 1DsII and a 1DsIII and spent about the same; some photographers just dislike the hassle of switching cameras - new batteries, new chargers, new cables, new settings, new button locations, new software, new look (forcing them in some cases to expend time/energy getting the new camera to produce the look of the old camera). Some photographers love getting new gear, some despise it.
***I never understood why this isn't mentioned/discussed more often; you have to look through the viewfinder for nearly every frame you take - it's your portal to the world you are capturing.


Had Doug had his coffee, he might also have thought of this one:

- You have a special optic, like a telescope or microscope or ancient Petzval lens, with high image quality (or poor but interestingly funky image quality!) over a generous image circle. What camera should you strap on the back of it? Obviously, the one with the bigger sensor, taking the most advantage of the overall image delivered by the optic.

It is for this very reason that companies like Takahashi, Borg, Astro-Physics, William Optics, etc. etc. make and have always made adapters for both small and medium format cameras to their high-end APO telescopes.

You see, debates like D800 vs. MFD have always taken place in the assumption that whatever optics MFD has, there's an equivalent scaled down lens from the 35mm mfgr which can deliver the same lp/ph or look on their smaller, denser sensor.

But sometimes the optic is fixed, unique and unscaleable...and the nod must go to Medium Format.

Ray
Logged
ondebanks
Sr. Member
****
Offline Offline

Posts: 833



« Reply #117 on: May 08, 2012, 06:00:04 AM »
ReplyReply

Uh, no? The IQ180 etc have the same problems but less so, my understanding was that this was due to better microlens design, not removal of them. I have seen myself those vertical streaks in IQ180 files for shifted non-retro wides, if that is not from the microlenses, what is it?

From http://www.phaseone.com/en/Search/Article.aspx?articleid=1221&languageid=1

"Furthermore the P 65+/IQ160/IQ180 is not recommended with extreme wide non-retro focus lenses like the Schneider Kreuznach 35mm XL and even the Schneider Kreuznack 47 mm XL. Even though these lenses has more than 70 mm image circle, they are producing a very steep angle incoming light along the edges, occasionally causing thin lines to show up in the image."

I gather that such problems can occur at very steep incident angles, even without microlenses. Over on getDPI, Yair once pointed out to me that the old non-microlensed 9 micron Kodak sensors had pixel crosstalk of this nature, when lenses were shifted.

I have never seen a reference to these DALSA 40/60/80 MP sensors using microlenses, and the low base ISO specs (35/50) of the backs certainly indicate that they lack them.

In the lists of "don't use the following backs with tech cameras - because of their microlenses!" that dealers helpfully provide, I don't recall seeing any backs using the DALSA 40/60/80 MP sensors.

But I could be wrong. But you know what would really eliminate all doubt and put this to bed, once and for all? If we had the DALSA 40/60/80 MP CCD datasheets! A month or two ago, Steve Hendrix promised to source them for us...maybe this will act as a reminder!

Ray
Logged
torger
Sr. Member
****
Online Online

Posts: 1588


« Reply #118 on: May 08, 2012, 06:42:11 AM »
ReplyReply

But I could be wrong.

And me too! What a dreadful thought... :-)
Logged
Peter Devos
Sr. Member
****
Offline Offline

Posts: 465


« Reply #119 on: May 08, 2012, 10:30:40 AM »
ReplyReply

I think the best thing to do is to shoot a few weeks with a D800 and get to know it.
« Last Edit: May 08, 2012, 10:35:49 AM by Peter Devos » Logged
Pages: « 1 ... 4 5 [6] 7 »   Top of Page
Print
Jump to:  

Ad
Ad
Ad