Pages: 1 ... 11 12 [13] 14 15 ... 29   Go Down

Author Topic: Nikon D850: Announcement of an Announcement of Development  (Read 983309 times)

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #240 on: August 13, 2017, 02:27:41 am »

Not my experience with the Nikon 24 T/S.

.http://blog.kasson.com/the-last-word/stitched-panos-slide-or-spin/

Jim

Obviously - the Nikon PC-E 24 is well-known to have crap edges. It wouldn't take much to beat it.

Try the same with the PC-E 19 and it will be a different story.
Logged

Paul2660

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4066
    • Photos of Arkansas
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #241 on: August 13, 2017, 09:51:50 am »

As one who stitched almost a lot over the years with wides, I would disagree that the edges will not hold up to a similar image taken with a TS-E lens.  I stitch for increased FOV, with existing lens or creation of panoramas.  A normal TS-E PC-E lens or Rodenstock on Arca in 3 stitched of 12mm max will not IMO give you anything near a true panorama.  Even pushing the Arca to 20mm will not get there for me.  Just an opinion, but one from thousands of stitched exposures over the past 15 years. 

Also shifting either in 35mm or MF, is pushing a lens to the edges of the lenses Image circle, thus I can't ever see how a max shift of 12mm on 35mm will ever be as sharp as 3 rotational images from a good sharp 24mm lens.  Even the Rodenstock and Schneider tech lenses can start to show a bit of smearing as you hit the edges of the IC of these lenses, albeit they will hold up better than any 35mm PC-E or TS-E lens I have tried in a wide.  I have not tried the new Nikon 19mm, but have owned and used both of the wide Canon's 17mm and 24mm TS-EII lenses.  I also tried the Nikon 24mm PC-E and through many examples it was not a good lens when shifted past 5mm.  Nikon is long over due for a fix and I guess the 19mm was to be that. 

I used both Canon and Nikon tilt shift lenses for years, then Zork and MF and the Rodenstock and Arca with P1.  All can produce images that are very sharp edge to edge, but I can get the same degree of image quality using a Nikon 14-24 and D810 @ 24mm and vertical images stitched to create a landscape oriented image or @ 18mm even 14mm and horizontal pans.  The corners hold up fine for me.  I also get the same degree of sharpness with a P1 back and XF with the 35mm or 40-80 zoom. 

I agree that such shifting can and will cause problems if you are using any type of filter since your rotation will create different aspect of how the light is being transfer to the sensor, and a Cl-PL is by far the hardest.  Here the TS-E or Tech camera is an advantage.

Darker corners, and noise, were problems I agree again, but if you take a LCC frame and use C1 for the images the results are excellent.  LCC correct the light fall off and color cast issue even for a 35mm camera and TS-E glass and the noise is also much better.

Downsides are the same thing in that the LCC process is cumbersome, add much more work to the project and slows down the process totally.  But it's a necessary part of the workflow, which is why I prefer to stitch without such tools for landscape work. 

If I was working structures with known dimensions i.e. buildings, interiors, cars, etc, then I quickly fall back to the PC-E TS-E Arca tech solution but I don't shoot those very often if at all.

This is way off the D850 announcement I realize but the conversation has moved more to shifting and the like. 

Below is 3 part horizontal stitch, taken with a GFX and 32-64mm lens @ 32mm my point being the corners on the rotational shifts are as sharp as the centers.   
Note, as usual Flickr has added more sharpening to this image but you can still see my point I believe.

Paul Caldwell

cc 6 finger falls CC pano no1 DSCF1348 6 finger test 4 by paul caldwell, on Flickr]
« Last Edit: August 13, 2017, 10:22:37 am by Paul2660 »
Logged
Paul Caldwell
Little Rock, Arkansas U.S.
www.photosofarkansas.com

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #242 on: August 13, 2017, 07:31:12 pm »

It depends on the projection and angular dimensions of the final image.

With a rectilinear projection, a pixel that's 60 degrees off-axis will need to be stretched to double its dimensions, i.e. four times its area, for a geometrically-correct result. It only gets worse as you go wider.

There's obviously not much to gain (and a lot to lose) from shift-stitching an image with a narrower angle of view. But, for ultra-wide angles of view, that's a huge geometrical disadvantage to overcome. Whether it's doable or not also depends on the edge sharpness of the tilt-shift lens.

Obviously, if you go with a non-rectilinear projection, this doesn't apply.
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #243 on: August 14, 2017, 11:04:26 am »

It depends on the projection and angular dimensions of the final image.

With a rectilinear projection, a pixel that's 60 degrees off-axis will need to be stretched to double its dimensions, i.e. four times its area, for a geometrically-correct result. It only gets worse as you go wider.

While true, it's going to look shrunk/correct again when viewed from the proper perspective point. And oversampling is also not that hard to do, it just takes more tiles and a longer focal length (usually already better optically corrected) to achieve, and good Panostitcher resampling quality also solves a lot in case of sub-optimal viewing conditions. One can even stitch high quality upsampled tiles (e.g. produced with Photozoom which increases edge resolution). There are many ways that lead to Rome.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #244 on: August 14, 2017, 11:12:23 am »

Obviously - the Nikon PC-E 24 is well-known to have crap edges. It wouldn't take much to beat it.

Try the same with the PC-E 19 and it will be a different story.

You said:

A shift-stitched panorama from a 24mm tilt-shift is going to have sharper corners than a rotational panorama of the same angle of view taken using a 24mm lens, or even a 35mm lens.

This is a thread about the Nikon D810, so I assumed you meant the Nikon 24 T/S, or possible both the Nikon and Canon ones. Now I think you're saying that you didn't mean the Nikon. Did you perhaps mean only the Canon one? That was certainly not clear.

I'm not sure why you're mentioning the 19 in this context. I have no opinion on whether it's better to spin or slide it.

JIm

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #245 on: August 14, 2017, 03:32:21 pm »

You said:

This is a thread about the Nikon D810, so I assumed you meant the Nikon 24 T/S, or possible both the Nikon and Canon ones. Now I think you're saying that you didn't mean the Nikon. Did you perhaps mean only the Canon one? That was certainly not clear.

I'm not sure why you're mentioning the 19 in this context. I have no opinion on whether it's better to spin or slide it.

JIm

I was discussing tilt-shifts in general, their availability and quality on Nikon-mount (19mm is good, 24mm is terrible when shifted, 45 and 85mm are equal to their much-older Canon equivalents, which are about to be replaced), and the utility of the shift function in landscape photography (i.e. does it matter that the PC-E 24mm has terrible edges).

This was in the context of discussing a Nikon-Sony hybrid setup, with a 46MP/9fps D850 for action/wildlife and a higher-resolution 60-80MP A9r or A7r3 for nonmoving subjects, using a (mostly) shared set of lenses.
Logged

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #246 on: August 14, 2017, 05:05:40 pm »

While true, it's going to look shrunk/correct again when viewed from the proper perspective point. And oversampling is also not that hard to do, it just takes more tiles and a longer focal length (usually already better optically corrected) to achieve, and good Panostitcher resampling quality also solves a lot in case of sub-optimal viewing conditions. One can even stitch high quality upsampled tiles (e.g. produced with Photozoom which increases edge resolution). There are many ways that lead to Rome.

Cheers,
Bart

You can't generate real detail that wasn't there in the first place - that is, it must have been oversampled in the first place.

In order to have the same sensor area contributing to a point 60 degrees off-axis, a rotational panorama would need to be shot with a lens with twice the focal length of a shifted tilt-shift. Any less and there will be more noise in the corner than the shift-stitched image. The focal length needed for equivalent detail is less, since the midfield of a sharp, standard lens is going to be sharper than the shifted edge of a tilt-shift, but still more than that of the tilt-shift lens.

Oversampling is easy enough, but, the more frames you need, the more risk there is of image-destroying flaws. All it takes is a bit of vibration, or a gust of wind rustling leaves and grass to make a single frame blurry, and the entire image becomes useless. This is even more the case when multiple exposures are needed for dynamic range. Happens far too often. If I can capture sufficient resolution using two or three frames (increasing sensor resolution and sharper lenses only make this easier) I'd rather that than needing 6-7 exposures to come out movement-free.
Logged

Eric Brody

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 487
    • http://www.ericbrodyphoto.com
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #247 on: August 14, 2017, 05:41:31 pm »

Shadowblade, no intelligence insult intended, but please remember that so far as I'm aware, there's currently no way to adjust the aperture of any  Nikon E (electronic aperture) lens with any available adapter. So the Nikon Sony hybrid idea will be a non-starter with any Nikon E lens, including the PC-E tilt shift lenses. If you know of a solution, I'd be really happy to hear it. One needs a Nikon body to adjust the aperture, an expensive and bulky solution. (I'd LOVE to be wrong on this, especially for Nikon-Fuji combinations).
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8913
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #248 on: August 14, 2017, 06:15:48 pm »

You can't generate real detail that wasn't there in the first place - that is, it must have been oversampled in the first place.

Fortunately, we can (generate more detail/resolution)! It would take this thread too far off topic to discuss that here, but PhotoZoom Pro does that on edge and line detail (discussed in another thread here on LuLa, I'll see if I can find it). We do not even have to use supersampling techniques.

Quote
In order to have the same sensor area contributing to a point 60 degrees off-axis, a rotational panorama would need to be shot with a lens with twice the focal length of a shifted tilt-shift.

Yes, although a 120-degree FoV in rectilinear projection is 'stretching the limits' of an acceptable perspective under normal viewing conditions. It would take a 10.3 mm lens on a 36 mm wide sensor, image quality would probably leave a lot to be desired at the edges/corners.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #249 on: August 14, 2017, 10:56:12 pm »

Fortunately, we can (generate more detail/resolution)! It would take this thread too far off topic to discuss that here, but PhotoZoom Pro does that on edge and line detail (discussed in another thread here on LuLa, I'll see if I can find it). We do not even have to use supersampling techniques.

Yes, although a 120-degree FoV in rectilinear projection is 'stretching the limits' of an acceptable perspective under normal viewing conditions. It would take a 10.3 mm lens on a 36 mm wide sensor, image quality would probably leave a lot to be desired at the edges/corners.

Cheers,
Bart

Interesting - real detail using deconvolution techniques, to counter diffraction and motion blur, or fake detail using fractal methods?

You'll reach 120 degrees in the corners of a 12 5mm single shot on full-frame, or in the corners of a shift-stitched image from a 17mm or 19mm tilt-shift. A 24mm tilt-shift with 12mm of shift gets you to 107 degrees - this works out to a 1.68x stretching in each dimension at the corners and would require a 40mm rotational stitch to duplicate with the same sensor area contributing to the corners.
Logged

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #250 on: August 14, 2017, 10:58:28 pm »

Shadowblade, no intelligence insult intended, but please remember that so far as I'm aware, there's currently no way to adjust the aperture of any  Nikon E (electronic aperture) lens with any available adapter. So the Nikon Sony hybrid idea will be a non-starter with any Nikon E lens, including the PC-E tilt shift lenses. If you know of a solution, I'd be really happy to hear it. One needs a Nikon body to adjust the aperture, an expensive and bulky solution. (I'd LOVE to be wrong on this, especially for Nikon-Fuji combinations).

At present, anyway. Currently there is no need to buy a Sony if you have Nikon lenses, since the D810 exists.

If Sony releasea a 70MP camera and Nikon is stuck at 46MP, I'd imagine a working adapter will come out within months.
Logged

BernardLanguillier

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 13983
    • http://www.flickr.com/photos/bernardlanguillier/sets/
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #251 on: August 15, 2017, 01:14:07 am »

At present, anyway. Currently there is no need to buy a Sony if you have Nikon lenses, since the D810 exists.

If Sony releasea a 70MP camera and Nikon is stuck at 46MP, I'd imagine a working adapter will come out within months.

I know you have been writing for months about your concerns regarding the ability of Nikon to source good sensors (not sure what exactly worries you since you apparently don't own any Nikon glass), but facts at this point seem to indicate they are doing pretty well.

Owning today both a 100mp medium format beast and a 36mp D810, I am really not sure that a 70mp Sony would get me excited to the point of buying one... ;)

- For "casual shooting", the 45mp of the D850, being close to 4x5 quality, will enable exhibition grade A1 prints,
- For "serious shooting", I am frankly not sure that the different between 45mp and 70mp will be that super visible in prints. Let's not forget that we are now much closer to the peak resolution of many lenses than we were when Nikon went up from 24mp to 36mp. A move from 45mp to 70mp with the same sensor size is going to be incremental for a majority of applications.

It seems likely that the D850 at 45mp with their brilliant 70-200 f2.8 is probably going to be nearly as good as a 70mp Sony with their average 70-200 f2.8.

I may be getting older, but I am finding it more and more difficult to get excited about sensor resolution in 35mm photography. Similarly, a move from 100mp to 150mp in medium format won't prevent me from sleeping at night. We are deep in nitpicking territory and more than ever before, spherical stitching is the only way to get really significant increases of image quality. And for stitching, 45mp and 70mp are very close to identical.
- I would dare to say that if you use T/S lenses to stitch, the 45mp and 70mp will be 99% identical because both will outdo the actual resolution of T/S lenses when shifted a lot,
- For spherical stitching, you are talking about adding a few frames in a pano, with next to zero practical negative impacts.

This is essential about bragging rights in terms of who owns the highest resolution body... and I don't think that Sony will be the only brand with such sensors. At the price point we can expect the a9r to sell for, they will be small series items and it would be in the best interest of Sony to have these sensors sold to Nikon lens owners also.

Cheers,
Bernard
« Last Edit: August 15, 2017, 02:05:33 am by BernardLanguillier »
Logged

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #252 on: August 15, 2017, 03:40:28 am »

I know you have been writing for months about your concerns regarding the ability of Nikon to source good sensors (not sure what exactly worries you since you apparently don't own any Nikon glass), but facts at this point seem to indicate they are doing pretty well.

Owning today both a 100mp medium format beast and a 36mp D810, I am really not sure that a 70mp Sony would get me excited to the point of buying one... ;)

- For "casual shooting", the 45mp of the D850, being close to 4x5 quality, will enable exhibition grade A1 prints,
- For "serious shooting", I am frankly not sure that the different between 45mp and 70mp will be that super visible in prints. Let's not forget that we are now much closer to the peak resolution of many lenses than we were when Nikon went up from 24mp to 36mp. A move from 45mp to 70mp with the same sensor size is going to be incremental for a majority of applications.

It seems likely that the D850 at 45mp with their brilliant 70-200 f2.8 is probably going to be nearly as good as a 70mp Sony with their average 70-200 f2.8.

I may be getting older, but I am finding it more and more difficult to get excited about sensor resolution in 35mm photography. Similarly, a move from 100mp to 150mp in medium format won't prevent me from sleeping at night. We are deep in nitpicking territory and more than ever before, spherical stitching is the only way to get really significant increases of image quality. And for stitching, 45mp and 70mp are very close to identical.
- I would dare to say that if you use T/S lenses to stitch, the 45mp and 70mp will be 99% identical because both will outdo the actual resolution of T/S lenses when shifted a lot,
- For spherical stitching, you are talking about adding a few frames in a pano, with next to zero practical negative impacts.

This is essential about bragging rights in terms of who owns the highest resolution body... and I don't think that Sony will be the only brand with such sensors. At the price point we can expect the a9r to sell for, they will be small series items and it would be in the best interest of Sony to have these sensors sold to Nikon lens owners also.

Cheers,
Bernard

I own a 14-24, which has been used on several different 5D2s, the A7r and the A7r2. Bought it because the Canon 16-35 II was crap. It works with an adapter, but, obviously, is not an electronic aperture lens.

Obviously, a 70MP sensor will be much more demanding on lenses than a 46MP sensor. You may not be able to get the most out of the sensor with all lenses, due to the sensor outresolving the lens. But that's actually the ideal situation,  particularly in the absence of an AA filter - it reduces the risk of aliasing artifacts. Furthermore, even though you may not gain much spatial resolution due to lens limitations, you will gain colour resolution, while both chroma and luminance noise will be finer-grained (even if the total noise is the same) and less obtrusive in large prints. Finally, many lenses can take advantage of the resolution, at least in the centre - anything that can use a 28MP APS-C sensor can use a 70MP full-frame sensor.

Yes, the Sony 70-200 f/2.8 is crap. On the other hand, a Sony 70MP camera with a 100-400 GM will likely be far better than a Nikon 46MP camera with an 80-400 on it. And both cameras can take a manual-focus Otus.

With regards to tilt-shift lenses, I'd be very interested to see the new TS-E 50mm. It'd basically be the same as a 50mm lens for medium format in terms of image circle, and those are typically sharp right to the edge.

I just hate having my panos ruined by some random vibration or movement in a single frame, which isn't evident until it's loaded in the RAW converter on a large screen...

There was an interview with Sony a few months back (or was it last year?) where they said they would not be making their best sensors available to other manufacturers. Which would certainly be consistent with no-one else using the 42MP sensor (the Pentax was released after it became available, but still used the old 36MP sensor; also, the 42MP sensor doesn't show up on Sony's own list of commercially-available sensors, whereas the 36MP does). Not sure if this only applies to sensors designed and commissioned by Sony itself, though - it certainly suggests that, if Sony came up with a top sensor for its own cameras, they wouldn't share it, but leaves it open as to whether they'd design one for someone else on commission. Certainly, if Nikon or anyone else came up with their own design and asked Sony to make it, I'd imagine they would - if they didn't,  someone else would do it anyway.

At this stage, we don't know if the decision to go with 46MP/9fps was made explicitly to make it an action camera with good cropping potential, or because they couldn't get a higher-resolution sensor (i.e. 'we can't match Canon/Sony's next generation for landscape/studio, but we have the AF system and bandwidth to make it shoot action too, so let's make it the best body out there for that'). It looks like a great camera for wildlife and distant action, but does leave a (relative) hole in Nikon's camera lineup - if anything, it's more in the mould of a D750 successor or super-D750 than a D810 successor. Certainly, many of Nikon's lenses are capable of making use of many more than 46MP. I guess this will be answered by whether they bring out a slow-shooting, ultra-high-resolution body at some stage, although the maximum price of such a body is constrained by the decreasing cost of MF bodies (currently using the Sony 50MP sensor, but there would almost certainly be a higher-resolution sensor in the works for the next generation).
Logged

davidgp

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 758
    • davidgp fotografia
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #253 on: August 15, 2017, 04:03:08 am »


There was an interview with Sony a few months back (or was it last year?) where they said they would not be making their best sensors available to other manufacturers. Which would certainly be consistent with no-one else using the 42MP sensor (the Pentax was released after it became available, but still used the old 36MP sensor; also, the 42MP sensor doesn't show up on Sony's own list of commercially-available sensors, whereas the 36MP does). Not sure if this only applies to sensors designed and commissioned by Sony itself, though - it certainly suggests that, if Sony came up with a top sensor for its own cameras, they wouldn't share it, but leaves it open as to whether they'd design one for someone else on commission. Certainly, if Nikon or anyone else came up with their own design and asked Sony to make it, I'd imagine they would - if they didn't,  someone else would do it anyway.

If I remember correctly, that interview had an error and was deleted from the webpage because the author didn't double check with Sony if it was ok to published... it was later republished with an apology from the website owner and the corrections from Sony engineers about their words... if my memory does not fail me.


http://dgpfotografia.com

davidgp

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 758
    • davidgp fotografia
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #254 on: August 15, 2017, 04:08:25 am »

If I remember correctly, that interview had an error and was deleted from the webpage because the author didn't double check with Sony if it was ok to published... it was later republished with an apology from the website owner and the corrections from Sony engineers about their words... if my memory does not fail me.


http://dgpfotografia.com

Here it is the infamous interview... after the first question there is a long explanation about what it really meant Sony
http://www.imaging-resource.com/news/2017/03/17/sony-thailand-factory-tour-qa-mapping-out-the-future-of-the-interchangeable


http://dgpfotografia.com

Jim Kasson

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2370
    • The Last Word
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #255 on: August 15, 2017, 10:14:13 am »


Obviously, a 70MP sensor will be much more demanding on lenses than a 46MP sensor.

Much more? That's 23% greater resolution. My results with top FF lenses indicate that, on-axis,  the lens is not the most important factor until well over 200 MP.

Jim
« Last Edit: August 15, 2017, 02:03:42 pm by Jim Kasson »
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #256 on: August 15, 2017, 01:47:33 pm »

Here it is the infamous interview... after the first question there is a long explanation about what it really meant Sony
http://www.imaging-resource.com/news/2017/03/17/sony-thailand-factory-tour-qa-mapping-out-the-future-of-the-interchangeable
Thanks. The extensive added explanation makes it clear that the statement is simply about the fact that in some cases, Sony camera designers work with Sony sensor designers to customize sensor designs, and these "bespoke" sensor designs are for use only in Sony cameras. In this context, I note that over the years, Nikon's camera designers have similarly worked with Sony sensor designers, providing some Nikon-developed sensor design ideas for sensors used exclusively in Nikon cameras.

In fact, it also seems that some Sony sensors are exclusively for Olympus M43 cameras. I mention this just to emphasize that Sony Semiconductor Solutions Corporation offers both bespoke and prét-a-porter sensors.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #257 on: August 15, 2017, 02:00:01 pm »

Much more? That's 23% greater resolution. By results with top FF lenses indicate that, on-axis,  the lens is not the most important factor until well over 200 MP.

Jim
Yes, it is worth noting that—once the market-speak of pixel counts is put aside—over the last fifteen years, there has been only a doubling of resolution (l/mm) in digital ILC's: 4/3" has gone from 5MP in 2003 to 20MP; "APS-C" has gone from 6MP in 2002 to 24MP, and 36x24mm format has gone from 11MP (Canon) and 13.5MP (Kodak) in 2002 to 42MP (Sony) and 51MP (Canon).

And at least some lenses from back then are still performing fine at the doubled resolution.
Logged

davidgp

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 758
    • davidgp fotografia
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #258 on: August 15, 2017, 02:40:00 pm »

Thanks. The extensive added explanation makes it clear that the statement is simply about the fact that in some cases, Sony camera designers work with Sony sensor designers to customize sensor designs, and these "bespoke" sensor designs are for use only in Sony cameras. In this context, I note that over the years, Nikon's camera designers have similarly worked with Sony sensor designers, providing some Nikon-developed sensor design ideas for sensors used exclusively in Nikon cameras.

In fact, it also seems that some Sony sensors are exclusively for Olympus M43 cameras. I mention this just to emphasize that Sony Semiconductor Solutions Corporation offers both bespoke and prét-a-porter sensors.

Yes, not sure if it is the same in other sensor sizes but you can see this behaviour in MF market.  Phase ONE usually puts a lot of effort developing some CMOS sensors with Sony, Phase ONE released the sensor, other manufacturers have to wait around 6 months to be able to release a product with it. This was the case for the first CMOS 50mpx sensor (the one now used by everybody in the MF market) and the 100 MPx sensor (now also used by Hasselblad). I think that some interview of Kevin with the Phase ONE CEO, the Phase ONE CEO mentioned that fact for the 100 MPx sensor... I suppose you have the exclusivity depending on how much you are willing to pay.

P.D.: A roadmap of Sony sensors for MF was released by Sony... maybe from now on, with some many players, there is not much exclusivity for new sensors for all MF manufacturers.

shadowblade

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2839
Re: Nikon D850: Announcement of an Announcement of Development
« Reply #259 on: August 15, 2017, 03:23:05 pm »

Much more? That's 23% greater resolution. My results with top FF lenses indicate that, on-axis,  the lens is not the most important factor until well over 200 MP.

Jim

23% is quite significant. It means that, in the space where the lens would have had to resolve 100 lines, it now has to resolve 123.

Not a problem in the centre of most sharp lenses, but definitely a consideration in the corners, particularly on wide lenses.

EDIT: Just noticed that you said 'on-axis'. That's the issue here. What you need depends on what you shoot (it's also why one-number sharpness scores are worse than useless). I primarily shoot landscapes/cityscapes, so corner sharpness is the most relevant number (since I need corner-to-corner sharpness and it represents the weakest part of the field and, thus, the minimum sharpness I can expect across the frame - this is why I rate the Nikon 24-70 VR above the 24-70 non-VR, despite the latter's better central sharpness). Particularly corner sharpness in the f/5.6-f/8 range, at infinity focus distance.  A wildlife photographer, whose corners will be out of focus anyway, cares primarily about central sharpness for fur and feather detail. Someone buying a 500mm lens to be used most of the time with a 1.4x TC isn't going to care about the outer third of the image circle. For these people, sensor performance is unlikely to be limited by lens performance for a long time. But a limitation which doesn't apply to the centre, and doesn't impact wildlife/people photographers, is much more challenging at the corners, where it is applicable to landscape and architectural photographers, and also things like art reproduction.
« Last Edit: August 15, 2017, 03:42:09 pm by shadowblade »
Logged
Pages: 1 ... 11 12 [13] 14 15 ... 29   Go Up