Pages: 1 2 [3] 4 5 ... 22   Go Down

Author Topic: After Photokina Medium Format rumors (UPD. And The WINNAH is FUUUJIIII !!! )  (Read 56286 times)

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #40 on: August 28, 2018, 11:44:48 pm »

Does anybody know what the chips actually cost the manufacturer? And are the "FF" (53x40) chips massively more expensive than the 33x44 chips? Is A GFX twice the cost of a 24x36 mirrorless with nearly the same pixel count because the sensor makes up most of that, or is there a much higher margin (perhaps because of lower sales)? Is a Phase five times the cost of a GFX because that FF chip costs many thousands of dollars? I could see the chips getting massively more difficult to make as they got bigger, because one error ruins a whole chip, but I could also see the pricing being a function of a deliberate low-volume/high price strategy...

My feeling is that *at the moment* we're looking at under $1K for a crop sensor to $3K for a fullframe.

The differential depends on where one is on the yield curve for the tech. The general shape of the yield curve is well known, but the exact slope and the point where one is on it is the most carefully guarded production secret at a semiconductor company. AFAIK. I'm sure some specialist will share his opinion.

Also a camera with twice the number of pixels will need twice the storage and twice the bandwidth, so the sensor cost is compounded by a rise in the cost of the support electronics.

Last not least Phase has very high marketing costs and margins, in order to serve their institutional customers via a full-service dealer network. Hassy and especially Fuji have more realistic sales models.

As I explained in the section on Pentax in my post at the top of this thread, in my opinion fullframe mirrorless MF at a $10K pricepoint would be perfectly feasible; the existing players are however locked into high prices when it comes to Phase, and crop-format lens systems when it comes to Hassy XD and Fuji GFX. 

Edmund
« Last Edit: August 28, 2018, 11:57:10 pm by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #41 on: August 29, 2018, 04:10:17 am »

My feeling is that *at the moment* we're looking at under $1K for a crop sensor to $3K for a fullframe.

Edmund

For the benefit of this rumors thread let me give the guesses for P1 and Fuji economics:
------------
- P1 pays a sensor $3K.
- They add $1K-$1.5K at least for the for the infinity planetary computer or whatever they call it ie, a decent LCD and touchscreen, probably a microcomputer board based on an FPGA eg Xilinx Zink or something similar to interface with the sensor,  maybe some fast static RAM to read out the frame data, maybe a second microcomputer board to manage the user interface, a lot of RAM to store the images, some well designed power management circuitry, and military-grade chips everywhere, expensive circuit boards, very high grade internal connectors. The whole thing is avionics-grade  The electronic board assembly is subcontracted to a high-end small-scale manufacturing subcontractor who makes a batch, eg 500.

- Case costs eg. $100
- Protection glass cover etc. $50
- Final assembly costs when the sensors - delivered in very small batches - meet the backs "just in time". This involves an extensive calibration process, optical and electronic to make sure the focus plane is in the right place, and measuring the sensor's base characteristics. These will be burnt into the firmware, and retained for reference by the company. This final assembly process probably cannot be outsourced. Final assembly probably involves 1/2 day of highly skilled work for each back. Say $500.

So I'd guess that the marginal cost of each back that gets sold is around $5-$6K, and I'd guess they are pushed out the factory door with an invoice at $15-20K, which leaves some money on the table for the various company execs to pay themselves a salary, and for the techs to do the R&D for IQ5. The dealer invites the customer for a presentation cocktail and resells the thing at $40K or whatever they really make a customer pay these days for an IQ4-150
----------
For Fuji it's a different story.
My guess is they have crop-sensor costs of $800 or so, as they are doubtless now the largest buyer of these chips.
They will make their own simplified and miniaturised board, possibly using their own ASICS common to several cameras they manufacture.
They get really good deals for the LCD and RAM etc  ===> So $300-$500 for electronics
$300-500 for physical camera construction and assembly including shutter and lens mount
Final Assembly and testing is done on their own production line for a big batch, but in Japan. $300 final assembly costs per camera.

So Fuji end up with $2K marginal costs for their whole camera,  and they then ship it into their distribution system with an internal invoice around $5K and are perfectly happy. It will now go into a rabbit hole where creative accounting happens.
The box ends up at B&H who ship the thing out via Fedex and invoice the customer $7.5K
---------

All of the above are pure guesses. People with a more realistic view of things are welcome to comment.

Notice that in the end it really doesn't matter a lot whether Phase pay their sensor $3K or$5K today, because in absolute terms this would make up 10% of variation in the invoiced factory unit price of a back they assemble "just in time" today and deliver tomorrow. They will just bill their dealers $2K more, but the customer won't notice it. The huge margins provided by the dealer system cushion the large costs in parts when a product is in its launch phase. Chips get much cheaper during the lifetime of a product.

Also, the Phase dealers don't really get to keep all of their huge margins. Nobody pays list. In the case of institutional sales there are a large number of "incidental" costs. The dealers may have to purchase dealership demo units, and loaners for customers, handing their own hard-earned money to Phase. They need to have a "Doug in the box(TM)" to help customers. They need to spend days and days teaching customers to use C1, help them install it etc.  It all mounts up.

In the case of Fuji, the company can still afford to lower its prices by a fair amount, because their distribution system does not *need* to be expensive, and a large company does not *need* to make a large profit on each unit shipped.. Of course, they will only risk lowering their pricing when they have amortised the R&D and are confident that they understand the product and will make economies of scale. But at that point they could come down to retail $5K fairly easily, in my opinion.

As I stated in the beginning, all of this is pure back-of the envelope guesswork. And my text shows that I don't understand that much about manufacturing and pricing, unfortunately. But maybe someone who does can help ...

Edmund

« Last Edit: August 29, 2018, 05:27:56 am by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

E.J. Peiker

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 891
    • http://www.ejphoto.com
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #42 on: August 29, 2018, 07:38:07 am »

Does anybody know what the chips actually cost the manufacturer? And are the "FF" (53x40) chips massively more expensive than the 33x44 chips? Is A GFX twice the cost of a 24x36 mirrorless with nearly the same pixel count because the sensor makes up most of that, or is there a much higher margin (perhaps because of lower sales)? Is a Phase five times the cost of a GFX because that FF chip costs many thousands of dollars? I could see the chips getting massively more difficult to make as they got bigger, because one error ruins a whole chip, but I could also see the pricing being a function of a deliberate low-volume/high price strategy...
Yes the large chip is drastically more expensive to manufacture than the cropped one.  I wrote an article many many years ago about this that explains it in terms of 35mm vs cropped - while the numbers would be different, the concept is identical:
https://www.naturescapes.net/articles/techniques/the-economics-of-digital-photo-sensors/
Logged

Gigi

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 548
    • some work
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #43 on: August 29, 2018, 09:30:59 am »

The cost summar is likely right on target, but there may be more to the picture. Admin/marketing/distribution has to be added, also R&D. Some may be allocated to software sales, but some also to the backs. Hard to know what that is.

How wacky can these other costs be? Consider for a moment the goofy selling of boutique water - the water probably costs very little, but its pricey on the shelves. Yes, probably includes excess profits, but even if you discounted the water say 50% for the retailer, this still suggests the cost of raw material may not be main driver.

In making and selling of books, some focus on printing costs, as its the easiest variable to identify. Yet the costs for  design, editorial, distribution and marketing can be equally significant. Its those darn invisibles that trip us up...

Logged
Geoff

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #44 on: August 29, 2018, 10:11:05 am »

Yes the large chip is drastically more expensive to manufacture than the cropped one.  I wrote an article many many years ago about this that explains it in terms of 35mm vs cropped - while the numbers would be different, the concept is identical:
https://www.naturescapes.net/articles/techniques/the-economics-of-digital-photo-sensors/

EJ, what do you think are the yield rates for fullframe 35mm, crop MF and full MF these days? And the manufacturing costs?

Edmund
« Last Edit: August 29, 2018, 10:27:18 am by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #45 on: August 29, 2018, 10:42:28 am »

At least my summary answers the issue of whether fullframe MF would be viable at the $10K market point eg for Fuji. The answer is yes, certainly.  And it also answers the issue of whether Phase should bother with crop sensors - the answer is "not worth the bother". Half of this has already played out, now let's wait for the other shoe to drop.Interestingly Hassy seems to be stuck in some demilitarized zone between Fuji and Hassy, they sell a prestige camera, have the same R&D costs, but don't incur the same sales costs.

Edmund

The cost summar is likely right on target, but there may be more to the picture. Admin/marketing/distribution has to be added, also R&D. Some may be allocated to software sales, but some also to the backs. Hard to know what that is.

How wacky can these other costs be? Consider for a moment the goofy selling of boutique water - the water probably costs very little, but its pricey on the shelves. Yes, probably includes excess profits, but even if you discounted the water say 50% for the retailer, this still suggests the cost of raw material may not be main driver.

In making and selling of books, some focus on printing costs, as its the easiest variable to identify. Yet the costs for  design, editorial, distribution and marketing can be equally significant. Its those darn invisibles that trip us up...
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

E.J. Peiker

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 891
    • http://www.ejphoto.com
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #46 on: August 29, 2018, 01:29:22 pm »

EJ, what do you think are the yield rates for fullframe 35mm, crop MF and full MF these days? And the manufacturing costs?

Edmund
I've been out of the industry for too long now to have any accuracy as I did back when I wrote that but it would very much be a function of wafer size in use.  If they are still using 300mm wafers which they were in 2017 so that's probably a decent assumption but it is possible that they have moved to 450mm, and assuming a $1500 wafer cost then they are probably getting 3 good sensors or so per wafer so the processing cost alone would be $500, then you add assembly, bayer array, wafer sort (separating good die from bad), assembly cost, final test, QC, and all of the other overhead and I could easily see the raw sensor cost, when purchased in volume of $1000 or more.  But again, I've been out of the game too long.  In 2009 I could have predicted it down to within a dollar or so accuracy due to the job I had and the data I could get ;)

I am reaching out to some old contacts to get a better estimate of wafer cost in 2018 and will update this if it changes materially.
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #47 on: August 29, 2018, 01:45:23 pm »

I've been out of the industry for too long now to have any accuracy as I did back when I wrote that but it would very much be a function of wafer size in use.  If they are still using 300mm wafers which they were in 2017 so that's probably a decent assumption but it is possible that they have moved to 450mm, and assuming a $1500 wafer cost then they are probably getting 3 good sensors or so per wafer so the processing cost alone would be $500, then you add assembly, bayer array, wafer sort (separating good die from bad), assembly cost, final test, QC, and all of the other overhead and I could easily see the raw sensor cost, when purchased in volume of $1000 or more.  But again, I've been out of the game too long.  In 2009 I could have predicted it down to within a dollar or so accuracy due to the job I had and the data I could get ;)

I am reaching out to some old contacts to get a better estimate of wafer cost in 2018 and will update this if it changes materially.

That would be interesting. What I would also be interested in knowing is what the yield numbers for one of the known sensors sizes would be.

Edmund
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

heinrichvoelkel

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 394
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #48 on: August 29, 2018, 03:13:05 pm »

Some students from a different branch of my university, are trying to build a medium format camera prototype. Buying a sensor from Sony (1 piece) is about 3k in Euro.
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #49 on: August 29, 2018, 03:24:44 pm »

Some students from a different branch of my university, are trying to build a medium format camera prototype. Buying a sensor from Sony (1 piece) is about 3k in Euro.

Yes, these are sample prices.
What they should do is ask Sony for an "engineering grade sample" of the sensor which is a chip that has some imaging defects but is functional.
AFAIK these are handed out for almost nothing.

Edmund
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

E.J. Peiker

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 891
    • http://www.ejphoto.com
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #50 on: August 29, 2018, 08:46:31 pm »

That would be interesting. What I would also be interested in knowing is what the yield numbers for one of the known sensors sizes would be.

Edmund

I did say that in my response, my estimate based on a 300mm wafer size and common defect densities for this type of process would be around 3 D/W.
Logged

BJL

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6600
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #51 on: August 29, 2018, 09:05:58 pm »

Some students from a different branch of my university, are trying to build a medium format camera prototype. Buying a sensor from Sony (1 piece) is about 3k in Euro.
Which model? In particular, 44x33mm of 54x40mm?
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #52 on: August 30, 2018, 12:18:38 am »

I wonder how many of the new 100MP and 150MP chips are really available ... Maybe just enough to do some announcing and for Phase to supply a handful of customers. We may be heading for "sales season" where everybody is remaindering 50MP MF products like crazy, or possibly offering upgrade deals where you buy the 50 and get the 100 later for a low "upgrade" price.

Edmund
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

capital

  • Full Member
  • ***
  • Offline Offline
  • Posts: 222
    • Website
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #53 on: August 30, 2018, 03:03:19 am »

Yes, these are sample prices.
What they should do is ask Sony for an "engineering grade sample" of the sensor which is a chip that has some imaging defects but is functional.
AFAIK these are handed out for almost nothing.

Edmund

I don't believe this to be accurate for sensors, perhaps if its a sample request for a lower end part from say Texas Instruments, Sony etc.
 I inquired 11 years ago with Kodak, here was their quote:

KAF-39000CE Image Sensor 
Color with Sealed AR Coated Glass
  Standard Grade     Catalog # 4H0762 Product Name: KAF-39000-CAA-DD-AA
  Engineering Grade  Catalog # 4H0763 Product Name: KAF-39000-CAA-DD-AE

Evaluation Board
  20MHz 12-Bit Digital Catalog # 4H0845 Product Name: KEK-4H0845-KAF-39000-12-24

KAF-39000CE (M35) (Color)
Total Quantity Purchased
Jan. 1, 2005 to present                  Color 
                                    Photographic Grade
                                    Price Each ($US)
Standard Grade                         $5900
Engineering Grade                      $1200
Evaluation Board (Includes EG Sensor)  $3700
====================================================================================

Engineering Grade Policy:

Engineering Grade devices are fully operational parts which may have more cosmetic defects (points, clusters, or columns) than standard production grade parts.    Quantities are strictly limited and the parts are sold on an "as available" basis and ONLY in conjunction with the purchase of a production grade imager of the same type or an evaluation board set.  Engineering Grade devices are intended for uses such as testing new board designs to avoid risking a high value production grade part during the development phase of a new camera project.  There is no warranty associated with Engineering Grade parts, and they are never authorized for use in camera systems sold to end user customers.

The Evaluation Board Kit includes the following:
Imager Board
Timing Board
40-pin Connector
Board Interface Cable
Co-axial Cable
8-Position Power Supply Plug
Documentation CD

The kit order  also includes an Engineering Grade device in your choice of configuration at no charge.  Please specify the configuration required when ordering.

The evaluation board kit connects to a PC via a frame grabber.  The recommended frame grabber is the National Instruments PCI-1424 (LVDS).  You should also be able to get the connecting cable from them as well  (NI 185012A-02 Type D100100). Check www.ni.com to locate a distributor you can work with if you need to purchase the frame grabber and cable. 


All prices are in US dollars. Prices include air shipment to your location.  Prices quoted are valid for 60 days from the date of quotation. Normal lead time for ISS is 30 days from receipt of order.  Due to high demand at this time, some products can not be delivered in that time frame.  The best available delivery date will be provided with the confirmation at the time the order is placed. For customers that do not have a formal credit relationship with Eastman Kodak, lead time is 30 days from receipt of payment.  Please see the attached Terms of Sale for warranty and other information.


« Last Edit: August 30, 2018, 03:30:54 am by capital »
Logged

eronald

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6642
    • My gallery on Instagram
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #54 on: August 30, 2018, 07:43:04 am »

I'm sorry to have used the term "almost nothing"; as you indicate 20% of the price of a photo-grade sensor was still a lot of money.
I never figured out how things work with samples, what you really pay for them varies with the realtionship you establish with the supplier. The last design I did (a measurement device) I paid the going rate for an eval board for the very novel sensor, but then when I later requested a batch of samples for a set of prototype products, a company VP suddenly became available for a meeting in my location and came bearing gifts :)

Edmund

I don't believe this to be accurate for sensors, perhaps if its a sample request for a lower end part from say Texas Instruments, Sony etc.
 I inquired 11 years ago with Kodak, here was their quote:

KAF-39000CE Image Sensor 
Color with Sealed AR Coated Glass
  Standard Grade     Catalog # 4H0762 Product Name: KAF-39000-CAA-DD-AA
  Engineering Grade  Catalog # 4H0763 Product Name: KAF-39000-CAA-DD-AE

Evaluation Board
  20MHz 12-Bit Digital Catalog # 4H0845 Product Name: KEK-4H0845-KAF-39000-12-24

KAF-39000CE (M35) (Color)
Total Quantity Purchased
Jan. 1, 2005 to present                  Color 
                                    Photographic Grade
                                    Price Each ($US)
Standard Grade                         $5900
Engineering Grade                      $1200
Evaluation Board (Includes EG Sensor)  $3700
====================================================================================

Engineering Grade Policy:

Engineering Grade devices are fully operational parts which may have more cosmetic defects (points, clusters, or columns) than standard production grade parts.    Quantities are strictly limited and the parts are sold on an "as available" basis and ONLY in conjunction with the purchase of a production grade imager of the same type or an evaluation board set.  Engineering Grade devices are intended for uses such as testing new board designs to avoid risking a high value production grade part during the development phase of a new camera project.  There is no warranty associated with Engineering Grade parts, and they are never authorized for use in camera systems sold to end user customers.

The Evaluation Board Kit includes the following:
Imager Board
Timing Board
40-pin Connector
Board Interface Cable
Co-axial Cable
8-Position Power Supply Plug
Documentation CD

The kit order  also includes an Engineering Grade device in your choice of configuration at no charge.  Please specify the configuration required when ordering.

The evaluation board kit connects to a PC via a frame grabber.  The recommended frame grabber is the National Instruments PCI-1424 (LVDS).  You should also be able to get the connecting cable from them as well  (NI 185012A-02 Type D100100). Check www.ni.com to locate a distributor you can work with if you need to purchase the frame grabber and cable. 


All prices are in US dollars. Prices include air shipment to your location.  Prices quoted are valid for 60 days from the date of quotation. Normal lead time for ISS is 30 days from receipt of order.  Due to high demand at this time, some products can not be delivered in that time frame.  The best available delivery date will be provided with the confirmation at the time the order is placed. For customers that do not have a formal credit relationship with Eastman Kodak, lead time is 30 days from receipt of payment.  Please see the attached Terms of Sale for warranty and other information.
« Last Edit: August 30, 2018, 07:47:15 am by eronald »
Logged
If you appreciate my blog posts help me by following on https://instagram.com/edmundronald

Christopher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1499
    • http://www.hauser-photoart.com
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #55 on: August 30, 2018, 12:21:39 pm »

Erik just as a note. All the Phase One lenses I currently own are exceptional some even better than my GFX lenses.

Don’t get me wrong the Fuji lenses are amazing especially at that price point, however, lenses like the 35/45/120/150 from Phase are exceptional as well and have no problem resolving the 100Mp and pretty certain the 150 as well. They are as good or better than all these praises Rosenstock lenses.


Sent from my iPhone using Tapatalk
Logged
Christopher Hauser
[email=chris@hauser-p

Gerd_Peters

  • Full Member
  • ***
  • Offline Offline
  • Posts: 106
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #56 on: August 30, 2018, 01:06:59 pm »

That's exactly how I see it...

Greeting Gerd

PS: And I also have the GFX / 645Z and (almost) all lenses.
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11311
    • Echophoto
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #57 on: August 30, 2018, 03:41:51 pm »

Hi Christopher,

It is difficult to compare lenses on systems that don't have the same resolution. It would be easy, would we have MTF data, that characterizes lens without the sensor. But we have not so much that data and presentation may differ a lot.

I have seen a lot of very good analysis on the GFX lens line, due to the valiant efforts of Jim Kasson.

Once we have 100 MP on the GFX and 150 MP on 54x41 we may have accurate comparisons.

Jim Kasson had a sample of HC lenses for the "Blad" and also V-series lenses and he did test them and they were no where close the GFX lenses.

Little is known about the Phase One lenses. Capture Integration did publish some MTF data for some of the LS-lenses. They were not so easy to compare say Hasselblad MTF data, as they measured at 15,30 and 60 lp/mm, while HB measures at 10,20,40 lp/mm. The data were nice, but it did not blow your socks off.

An intressant observation may be that a well corrected lens achieves optimum performance around f/4 or f/5.6. Why? Because optimum performance is sort of a race between aberration control and diffraction, that sets the limit of the performance of the lens. So, a lens that needs to be stopped down to f/8 for maximum sharpness will be a lot behind a lens that achieves optimum  at f/4.

it seems that GFX lenses measured with the GFX sensor are pretty darned good. How they perform at 100 MP, that is something we will see when 100 MP is around.

I am no betting man. But, would I bet, I am pretty sure I would bet on the GFX at 100 MP versus 54x41 mm at 150 MP. That said, I may be wrong... We will know better once adequate samples are available from both.

Best regards
Erik


Erik just as a note. All the Phase One lenses I currently own are exceptional some even better than my GFX lenses.

Don’t get me wrong the Fuji lenses are amazing especially at that price point, however, lenses like the 35/45/120/150 from Phase are exceptional as well and have no problem resolving the 100Mp and pretty certain the 150 as well. They are as good or better than all these praises Rosenstock lenses.


Sent from my iPhone using Tapatalk
« Last Edit: August 30, 2018, 03:46:12 pm by ErikKaffehr »
Logged
Erik Kaffehr
 

siddhaarta

  • Full Member
  • ***
  • Offline Offline
  • Posts: 130
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #58 on: August 30, 2018, 04:49:00 pm »

It is a pity that among the manufactures of premium lenses (Zeiss, Leica, Hasselblad, Fuji, Rodenstock, etc.), only Schneider Kreuznach chooses not (anymore) to publish MTF data. Obviously, MTF does not say all about lens performance, but it helps understanding the basics  ...

Among the few who publishes Imatest data on resolution of MF lenses is (strangely) PCmag, not very detailed and thorough, but better than nothing. Two comparisons seem interesting to me:

1) Two standard lenses: SK BR 80mm (rated by DT as B - good) vs Fuji GF 63mm: the Fuji starts at 4755/3744 lines per picture height in the centre/corner at f2.8 and improves to 5064/3744 at f4. The SK starts at 4539/3970 lines at f/2.8 and improves to 4809/4650 at f4.

Seems that these results are quite good for both lenses, the Fuji being better in the centre and the SK being better in the corners. Now the point is that these results are measured with the 50MP sensor for the Fuji and the 100MP sensor for the SK, meaning that the SK is not able to extract more detail on the bigger sensor than the Fuji on the smaller sensor. To put this into perspective, the Otus 85mm resolves in the centre at about 3800-3900 lines on a 36MP sensor between f2.8 and f/4.

Note: the non BR SK 80mm has also been tested with the 50MP back: at f/2.8 it reaches only 2910/2212 lines, and at f/4 it improves to moderate 3411/2714 lines (centre/corner).

2) Two wide angle lenses: SK BR 35mm (rated by DT as A++ - extraordinary) vs Hasselblad XCD 21mm: the Hasselblad starts at 4979/4647 lines (centre/corner) at f/4 and improves to 5196/4999 at f/5.6. The SK starts at 5311/3283 at f/4 and improves to 6278/4562 at f/5.6. Whereas the Hasselblad is very strong in the centre and corners, the SK is even better in the centre but quite weak in the corners and needs substantial stopping down to f/8 to improve substantially. Again these results compare the Hasselblad on a sensor with 50MP resolution to the SK on a sensor with 100MP.

Note: There is also data for the SK 28mm on the 50MP sensor available: 3880/2084 at f/4.5 and 4045/2618 at f/5.6. This lens would match better the FoV of the XCD 21mm, but is a quite old design (rated by DT as C - okay).

In both cases it can also be seen that the Fuij/Hasselblad lenses have their peak performance at f/5.6, whereas the SK at f/11, I think Erik explained why.

Question is what does that mean when we compare in the near future system performance between 100MP "crop sensor" (with newly designed Hasselblad or Fuji lenses) and 150MP "full sensor" (with partially aged SK lenses).

But then: the IQ4 series is supposed to be "much more than resolution".

 
« Last Edit: August 30, 2018, 07:38:31 pm by siddhaarta »
Logged

Christopher

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1499
    • http://www.hauser-photoart.com
Re: Photokina rumor and announcement Medium Format thread (UPDATING)
« Reply #59 on: August 30, 2018, 05:57:11 pm »

I think lenses are quite difficult to test and it’s really hard to judge them. Everyone has a different opinion and I for myself am not a big fan of MTF curves. In general they should be really great to judge, however, I have had to many lenses and copies of these lenses that did not match my experience the MTF curve I saw and performance I was hoping and looking for.

I can compare a few lenses at 50mp (GFX vs iq350) and here most of my Phase lenses, I now own, are better than anything I have for the GFX.

There are some lenses like the Phase BR 150 2.8, which are huge and heavy, but some of the finest glas I have ever used and that includes quite a lot from Nikon/Zeiss/Fuji and Leica.

I have been shooting the GFX next to the IQ3100 quite a while now and while I absolutely love my Fuji camera and would never give it up, it‘s not in the same ballpark compared to the IQ3100.

This might change when comparing 100vs150 but I still think the IQ4 150 with the right lenses will have no problem keeping a distance to the GFX2.

I won’t argue, that when doing difficult hikes and tours I will be more than happy with a 100Mp Fuji and will leave my Phase equipment at home.

I love gear, but in the end the stuff has to get the job done and make some money. That’s currently for me a Fuji/Phase combination. I looked at the new D850 and Sony’s, but they just don’t provide the Detail im looking for.


Sent from my iPhone using Tapatalk
Logged
Christopher Hauser
[email=chris@hauser-p
Pages: 1 2 [3] 4 5 ... 22   Go Up