IQ5 rumors

[Paul Spinnler]

Yeah, there's definitely some buzz about an IQ5 200 possibly being in the works, but nothing concrete yet. Haven't spotted anything solid on Sony's roadmap either. It's mainly speculation at this point.

Instead of posting a random comment at the end of the thread maybe read through the thread where there’s a ton of discussion already about Sony’s big announcement of the 811 platform. By nomenclature and specs and place in the portfolio it is the successor of the 150 chip (411 platform).

Sony announced a 250 megapixel 54x36mm sensor this week; it’s either this one or a variant of it that we will see in the IQ5. Variant meaning 54x40 which would be 275 MP.

 

[hcubell]

If Sony will not produce the new sensor in any size on the short end bigger than 36mm (or Phase will not be willing to pay the extremely high costs for Sony to do it) and Phase wants to retain the 4:5 format, the sensor would be 45x36. The Hasselblad X and Fuji GFX lenses can probably cover that sensor size, or very close to it. That's the end of Phase's dominance on sensor size and MP. If the sensor is 54x36 and Phase tells you to just crop to 4:3, the answer is still the same in terms of effective MPs.
The dimensions of this new sensor may very well reflect the sea change that has occured over the past 8 years in the medium format digital market. Phase was historically the 800 lb gorilla that dominated and dictated the market for larger than FF sensors. It's an understatement to say that's no longer true today. Fuji likely sells more GFX bodies in a week than Phase sells IQ150 backs in a year. Hasselblad may sell more X2Ds and CFV 100Cs in a couple of months than Phase sells IQ5s in a year. Even if you factor in the industrial market, the numbers are unlikely to change that much.

 

[Paul Spinnler]

You are not considering the fact that Fuji and Hasselblad rather have expanded the market (35mm folks buying a Fuji with Fuji TSE for architecture for example, people moving from Sony to Hasselblad,etc.) while P1's user base has always been a niche core of enthusiasts and professionals with a few thousand units per year, as Steve rightly mentioned. Especially Fuji impacted them for sure, but I also think its not so that P1 is dead in any way should the 250/275 IQ5 arrive.

The crop discussion is already covered in the Poll thread. That's why I've said that it will be key for P1 to secure a custom order with Sony for a 54x40 variant to keep the large chip USP that they have been pushing over the last years since exiting crop MF. Its better for them to be at 275 vs. 180; worst case they are at 250 with crop MF at 220 in a 48x36mm chip scenario.

The question is whether the S4 will allow for a 48x36 sensor in its mount as the camera is being developed right now and they could still make last minute changes, most likely.

I think business will be great for P1 once they release an IQ5. Therer's thousands and thousands of IQ4s waiting to be upgraded.

I think its exciting that there will be a 250/275 MP back in all likelihood next year. A significant step up from 100 / 150.

 

[vjbelle]

FMPOV if this comes to fruition it will mean that new lenses are mandatory. Even the best f5.6 lenses will be taxed for sharpness and movements. Someone besides Rody will have to step up and produce lenses that are faster and somewhat competitive in price. Maybe this will be enough incentive for Schneider to get back into making lenses for MF - they certainly have the resources. It's one thing to shell out 40k, 50k for a DB but another to shell out 15k or 20k per lens. To me it always has been about the lenses.

And then there's the weight.....

Victor B.

 

[jng]

FMPOV if this comes to fruition it will mean that new lenses are mandatory. Even the best f5.6 lenses will be taxed for sharpness and movements. Someone besides Rody will have to step up and produce lenses that are faster and somewhat competitive in price. Maybe this will be enough incentive for Schneider to get back into making lenses for MF - they certainly have the resources. It's one thing to shell out 40k for a DB but another to shell out 15k or 20k per lens. To me it always has been about the lenses.

And then there's the weight.....

Victor B.

There's also the reality of physics: Unless you shoot at f/4, in terms of absolute resolving power you won't see much if any difference between the new sensor and the current IQ4 150 given that the 2.78 micron pixel pitch puts you right at Nyquist frequency (Airy disc @ f/4 = 5.3 microns). For comparison, the best resolving tech cam lens appears to be the Rodie 138 float, whose maximum aperture is f/6.3. Not to say that there aren't specific applications that wouldn't benefit from the increase in resolution, but otherwise I think we've reached the point of diminishing returns.

John

 

[vjbelle]

I shoot the Float almost always wide open to take advantage of the incredible resolution. But there sure is a price to pay when using that lens with respect to weight. The older Schneider lenses will struggle as they already do. I hope that newer lighter designs will come along to compliment a new sensor.

Victor B.

 

[Paul Spinnler]

This whole theoretical diffraction topic is completely overblown. It will work just fine on 2.8 pitch. My Sironar S lenses beyond 100mm from 30 years ago deliver tack sharp results at f11 on 150 megapixels. SK glass at f16. And a lot of HR and Apo Digitar glass is still extremely sharp where you can see they still have ample reserves which are worst case easily recoverable if there's a tiny bit loss of sharpness from diffraction.

In practice, you can easily sharpen in post and actually I find it very pleasing should there be an “analogue” slight loss of sharpness on 275 when stopping down. A 275 megapixel base resolution will provide a fantastic post processing base for any end resolution / print file and via sharpening techniques and grain sims you can mould a great tonality for a final print. The best analogue film simulations even add a tiny bit of blur to a digital file to give the image an organic film base look.

I am very curious for V3 of Tim Parkin’s 8x10 shootout with a 275 sensor.

 

[Jon Warwick]

This whole theoretical diffraction topic is completely overblown. It will work just fine on 2.8 pitch. My Sironar S lenses beyond 100mm from 30 years ago deliver tack sharp results at f11 on 150 megapixels. SK glass at f16. And a lot of HR and Apo Digitar glass is still extremely sharp where you can see they still have ample reserves which are worst case easily recoverable if there's a tiny bit loss of sharpness from diffraction.

In practice, you can easily sharpen in post and actually I find it very pleasing should there be an “analogue” slight loss of sharpness on 275 when stopping down. A 275 megapixel base resolution will provide a fantastic post processing base for any end resolution / print file and via sharpening techniques and grain sims you can mould a great tonality for a final print. The best analogue film simulations even add a tiny bit of blur to a digital file to give the image an organic film base look.

I am very curious for V3 of Tim Parkin’s 8x10 shootout with a 275 sensor.

I agree with your point re a less ostensibly “sharp” base for getting closer to an analogue feel. For example, when using my SL Summicrons (and GFX lenses), I’ve quite often needed to add some Blur in post to tone down the digital sharpness that I guess emanates from too high acutance. I like tons of fine detail, but in a “gentler edged” way - which is what 8x10 obviously provides if executed well. The most “analogue” digital camera I’ve used has been the Monochroms, something there about stripping out the CFA to give a less processed look to my eyes. And yes, I can imagine Tim doing a shootout versus 8x10 if a 275 sensor ever materialises, I always enjoyed his prior comparisons that appeared methodical, without hyperbole, and hence very useful, IMHO. When I’m not reading those comparisons, I’ve used him for his drum scanning service.

 

[Paul Spinnler]

Well what's interesting is that Tim did a 80 MPX, 150 MPX shootout and although the base detail is there, if you zoom in, film still has more information. Especially the border between 150 and 275 seems to be the final bit of difference I saw in his V2 test.

Fully agree with the Apo SL point. They are so sharp that I often mistake 200% on screen to 100%! A tiny bit of gaussian plus noise make them perfect.

There's a lot more room with modern optics and can't wait to work with better DR, higher res base files.

Leica S4 apparently has next gen APO F2 optics across the focal range. That's going to be crazy, literally at the limits of what's technically and physically possible.

Rodie HR was APO F4, Leica will be APO F2 ... with 2025 technology in AF and precision in construction based off of learnings from 2017 gen APO SL manufacturing techniques.

So SL is mid 2010s tech, S4 will be mid 2020s tech. Another leap all built-in house at Leica.

 

[Jon Warwick]

I don’t want to misquote Tim’s findings (especially after all his hard work), but my takeaway from his various articles was 4x5’s ability to capture fine detail is similar to 80-100mp. That feels right based on my observation - before selling my 4x5, I compared identical scenes with 4x5 Ektachrome (funnily enough, drum scanned by Tim) and using my 110mm Super Symmar XL …..vs my GFX100S. Crops off the final images that I sized to 60” were remarkably similar in terms of final detail. I think his conclusion for 8x10 vs 150mp was they were similar’ish up to a certain print size, but then 8x10 had more info in the tank still to yield more fine detail at very very large prints. I suspect flawless capture is rather easier and user friendly off the digital back than 8x10 though!

 

[Paul Spinnler]

And this is why the 275 vs. 8x10 test will be so interesting because that fine detail bit - I think this is where digital finally catches up! Also super curious to do one shot scanning of 6x7 with it. Tonality will be amazing.

 

[dougpeterson]

Love the strong feelings about theoretical aspect ratio in a theoretical product – great holiday reading.

For what it’s worth I’ve been cropping most of my marque images to 16:9 over the last few years. Maybe out of appreciation for cinema, maybe because it looks nicer on a wall to me, maybe because it looks nice on phones at social gatherings, maybe as a reaction to instagrams square-everything. Couldn’t tell you. There is a certain gut factor when it comes to aspect ratio and no one can, or should, try to talk you out of your preference.

On the other hand on a tech camera and fixed subject matter (eg landscape or architecture) there is no such thing as a fixed aspect ratio. You want ultra wide? A Left-Right in horizontal-orientation can get you an X Pan level of wide. Want a square? A Left-Right in vertical-orientation can get you a square. And while some may obsess about sensor size down to a difference of a mm or two, you can also crop. One of the best things about shooting with an IQ4 150mp today is that even aggressive crops (eg cropping a vertical from an originally horizontal frame) leave you tons of pixels to print. Sensor size is a big deal but it’s just one of several reasons that shooting an IQ on an XT or XF is a fundamentally different tool and experience than a GFX or iPhone or whatever.

 

[dougpeterson]

I don’t want to misquote Tim’s findings (especially after all his hard work), but my takeaway from his various articles was 4x5’s ability to capture fine detail is similar to 80-100mp. That feels right based on my observation - before selling my 4x5, I compared identical scenes with 4x5 Ektachrome (funnily enough, drum scanned by Tim) and using my 110mm Super Symmar XL …..vs my GFX100S. Crops off the final images that I sized to 60” were remarkably similar in terms of final detail. I think his conclusion for 8x10 vs 150mp was they were similar’ish up to a certain print size, but then 8x10 had more info in the tank still to yield more fine detail at very very large prints. I suspect flawless capture is rather easier and user friendly off the digital back than 8x10 though!

Speaking from my experience designing the systems that are used for the vast majority of still film scanning at US libraries, museums, and archives – precious little large format film was shot in a way that maximized its possible technical quality. The reasons varied - Film holders weren’t flat, lenses weren’t always modern-sharp, diffraction from very high apertures, wind was blowing bellows, focus planes were estimated by old eyeballs through imperfect loupes etc etc. Didn’t stop them from making gorgeous images and documenting the state of humanity and our environment and in many cases changing the world. But very very few photographers back then were making enormous prints - the cost and time, skill, and facilities to make a 40” print (let alone a 60” or 120”) were hugely different than today. In an era where a 16x20” was considered a very large print, the level of scrutiny of micro detail was very different compared to an era where you almost always jump to a 100% pixel view to review focus/detail. Heck, a good amount of 8x10” was shot in no small part so that they could be contact printed at 8x10” a print size many here would now consider “consumer grade”.

None of this is of any particular value to mention. Just something I found fascinating as I’ve become deeply involved with the archives at places like Library of Congress and National Geographic Society. I remember exactly the first time I realized Migrant Mother is not “properly” focused; that didn’t stop it from being the iconic representation of the Dust Bowl experience.

Technical quality still matters to me. But it’s only ever icing on the cake. The image itself, and the experience of crafting and sharing it - those are the core of photography.

So anyway that’s what’s running through my head when I read modern comparisons to 8x10 - it’s moderately challenging to maximize the potential image quality of an IQ4 while staying focused on the image and the experience, but it’s nearly impossible for 8x10. So the comparison is interesting to me, but only in an academic sense.

 

[Paul Spinnler]

It is about one-shot. Stitching yes. But 3:2 is an inefficient use of a mount vs. 4:3, but of course efficient for Sony to cut it down to 35mm, crop MF formats. And stitching gets annoying with 36mm as it means you need to make three shots with an STC say for a pano in pano orientation. 36mm means more stitching is needed to get to a same FoV than with the IQ4. Its nice that you can shift the IQ4 left right 18mm on an STC in portrait mode and have a 2 shot stitch in PS. The more stitches you need, the more risk for artifcats. That's why for critical work (e.g. gallery print without errors) its easier to have a nice one shot file where 4:3 is a lot more flex.

This means for non static - people (environmental), aerial, scenes with people, wind, water, etc. you lose format flexibility in 3:2. I already can see the dealer's spin on it, with crop being still higher res than previous gen full frame, but besides one-shot flexibility you also lose a tiny bit of perspective differentiation vs. crop MF.

4:3 is also reminiscent of large format, widely perpetuated in fine art. I like the amount of foreground and sky one gets in environmental full body portraiture for example vs. 3:2 which looks like *cheaper 35mm". It is as you say subjective.

I hope they can get a custom order going 54x40.

There's also still catch up potential with portra film re DR. 1 to 1.5 more stops in the highlights would be awesome.

One side aspect of a 3:2 ratio would be that shorter focal lengths get longer via crop, meaning if your favorite focal length was 40 it is now 32 and 23 becomes the new 32, etc. 90 becomes even more longer in practice etc.

So it could be a bit of a revival for the 32, 40, 50 core of HR rodies plus 43 and 60 XL and 90, 120 being less used.

The poll is super clear in that people want 54x40 - I hope P1 takes note. They have the clout to commit to an exclusive order of a few thousand units over the next years across aerial, repro and photo., I hope.

 

[f8orbust]

I've always wondered just how many sensors actually resolve their full potential, so IMHO any DB with > 200MP only makes sense if it comes with some form of stabilisation. A 2.81 pitch means each pixel is 0.00281mm wide - I defy anyone preventing a camera without image stabilisation moving that much during a 'typical' exposure. In really tightly controlled conditions (e.g. a lab, high speed flash etc.) - maybe - anywhere else, no chance. I think it's why (historically) the images that 'fat pixel' DBs produce are so compelling. There's simply a greater chance that each pixel (sensel) is able to uniquely resolve the light hitting it, and so micro-smearing is mitigated significantly for 'typical' exposures.

 

[Paul Spinnler]

A lot of testing ahead ...

Another point with 3:2 is that the XT is limited to 12mm in any direction; that's fine for 54x40mm, but if precisely you still have 54 on one end and 36 on the other you cannot fully exploit stitching with a limited shift body in order to get to a 4:3 ratio via stich. In the STC its solved by doing three shots. Wth the XT it stays at two and your stuck with 3:2 native stitched.

Last element is that Leica is in midst of designing the S4; so they could really swoop in now with a sensor that's bigger than 44x33 to take the wind out of the sails of P1 and to be differentiated vs. the other crop MF players.

Their new line of APO F2 optics should be able to exploit the new sensor, potentially even with IBIS.

 

[jng]

This whole theoretical diffraction topic is completely overblown. It will work just fine on 2.8 pitch. My Sironar S lenses beyond 100mm from 30 years ago deliver tack sharp results at f11 on 150 megapixels. SK glass at f16. And a lot of HR and Apo Digitar glass is still extremely sharp where you can see they still have ample reserves which are worst case easily recoverable if there's a tiny bit loss of sharpness from diffraction.

In practice, you can easily sharpen in post and actually I find it very pleasing should there be an “analogue” slight loss of sharpness on 275 when stopping down. A 275 megapixel base resolution will provide a fantastic post processing base for any end resolution / print file and via sharpening techniques and grain sims you can mould a great tonality for a final print. The best analogue film simulations even add a tiny bit of blur to a digital file to give the image an organic film base look.

I am very curious for V3 of Tim Parkin’s 8x10 shootout with a 275 sensor.

Fair enough. And yes, of course the images will look just fine, as they do when I use my legacy Hasselblad-Zeiss lenses on the IQ4 150 (the corollary to: increased sensor resolution will not make the image from a given lens look worse). Don't get me wrong, I never met an extra megapixel I didn't like. That said, as the last stitched image I processed worked out to 19,200 on the long side, I'm not sure how additional pixels would have helped (certainly not my composition), especially given the other tools we have at our disposal as you point out, so again my contention that we've reached the point of diminishing returns. But that's just me. YMMV.

John

 

[Paul Spinnler]

Fair enough. And yes, of course the images will look just fine, as they do when I use my legacy Hasselblad-Zeiss lenses on the IQ4 150 (the corollary to: increased sensor resolution will not make the image from a given lens look worse). Don't get me wrong, I never met an extra megapixel I didn't like. That said, as the last stitched image I processed worked out to 19,200 on the long side, I'm not sure how additional pixels would have helped (certainly not my composition), especially given the other tools we have at our disposal as you point out, so again my contention that we've reached the point of diminishing returns. But that's just me. YMMV.

John

Diminishing returns is true, but then again the pixel pitch jump if a lot more significant than expected. Going from 150 to 275 or 250 is a significant jump and as it is always mentioned in this context it gives you creative freedom via cropping.

If anything, this should put into focus the lower end of the Rodie HR linup even more so. The 23 HR all of a sudden becomes the ultimate architectural lens as you can do one shot and crop, same with 32 HR. If indeed they do 54x36, which I hope they don't, it will move the whole focal lens range to the right to get similar FoV.

My favorite FoV is 40 HR rn say - tha't the 32 HR in the new cropped IQ-250 world.

 

[dchew]

Fair enough. And yes, of course the images will look just fine, as they do when I use my legacy Hasselblad-Zeiss lenses on the IQ4 150 (the corollary to: increased sensor resolution will not make the image from a given lens look worse). Don't get me wrong, I never met an extra megapixel I didn't like. That said, as the last stitched image I processed worked out to 19,200 on the long side, I'm not sure how additional pixels would have helped (certainly not my composition), especially given the other tools we have at our disposal as you point out, so again my contention that we've reached the point of diminishing returns. But that's just me. YMMV.

John

Within the contex of "none of this really matters in a print," I also wonder if the smaller sensor is good or bad for some of our other lenses that are not 138/120asph/90hr shiftable. On the one hand, pixels are smaller so airy disks will creep into the equation earlier. On the other hand, we could be shifting with fewer mm, so the outskirts of the image may improve. It will depend on the lenses we have and how we use them.

Dave

 

[Jacob Buchowski]

Although most of what we do in photography is subjective, I much prefer the 4:3 ratio for most of my work. Heck, even when using my Nikon gear, I crop to 4:3. I hope IQ5 keeps the 4:3 format!