Luminous Landscape Forum
Equipment & Techniques => Medium Format / Film / Digital Backs – and Large Sensor Photography => Topic started by: wing1 on February 05, 2016, 06:26:59 am
-
Hello, Do you know if with the new Capture One 9.03 the wide angles on technical cameras and iq3 100 are full employable? It seem that possible shift is less than the ccd back!
-
Hello, Do you know if with the new Capture One 9.03 the wide angles on technical cameras and iq3 100 are full employable? It seem that possible shift is less than the ccd back!
You won't get as robust performance as with the CCD-based IQ3 60MP. The IQ3 100MP performance with technical lenses is highly similar to the IQ250. If you use longer lenses you're fine, if you use wider lenses you will get significant crosstalk but if that matters or not is subject-dependent and also a matter of taste, but it is about fitting a square peg into a round hole. You rely on the huge capacity of this sensor in dynamic range and color separation to cover up the fact that the tech wides were not designed for this sensor.
Personally I don't think it's advisable to buy such an expensive digital back and not have lenses designed for it, but for sure some will do it anyway.
-
torger,
You won't get as robust performance as with the CCD-based IQ3 60MP. The IQ3 100MP performance with technical lenses is highly similar to the IQ250. If you use longer lenses you're fine, if you use wider lenses you will get significant crosstalk but if that matters or not is subject-dependent and also a matter of taste, but it is about fitting a square peg into a round hole.
Honestly, I think it has been about fitting a square peg into a round hole ever since the 40MP backs or your own play-nice 50MP back...it's simply that with time the peg has only increased in size.
-
Honestly, I think it has been about fitting a square peg into a round hole ever since the 40MP backs or your own play-nice 50MP back...it's simply that with time the peg has only increased in size.
I agree, the Rodenstock probably did not predict the microlens ripple or tiling issues with the Dalsa sensors (and Schneider wides weren't designed for the 6um Dalsa's at all), but I think if you use Rodenstocks and Capture One you get pretty robust performance which is good at cleaning up those issues. Unfortunately the larger crosstalk issue is not as easily fixed, but you can indeed hide it better than they do it currently by smearing the green channel. They never did for the IQ250 though, and as far as I know they don't do that in 9.0.3 either.
It's true that the peg has increased in size with time, moving a little bit up in size for each new generation. The 80MP Dalsa made the peg a little bit larger, and now the 100MP Sony a bit larger still. If you went from P45+ to P65+ to IQ180 and now to IQ3 100MP you have enlarged the peg just a little between each previous generation so perhaps you haven't noticed the huge gap between P45+ and IQ3 100MP...
The Kodaks were the last sensors that were actively designed for symmetric wide lenses and therefore had light shielded pixels and lacked microlenses. Then it became the other way around Rodenstock made the Digaron series for 6um sensor without light shields (obviously they did not calculate for microlens artifacts though), and then with smaller pixel sizes and still no light shields the situation has degenerated gradually.
But kept within limits it will work, and you can push a fair bit past the limit to most of the time. Just don't ask me exactly where the limits are, because it's subjective. For me personally it's past the messiness limit and I just don't think it would be enjoyable having to worry about these things.
If you want to shoot high end wide angle photography today there's no system without significant compromise. You can't have it all. You just need to choose which tradeoff that suits your shooting style the best.
-
On the Capture Integration site, there are some early testing examples:
https://captureintegration.com/phase-one-iq3-100mp-technical-camera-testing/ (https://captureintegration.com/phase-one-iq3-100mp-technical-camera-testing/)
They sampled with both the 32mm and 40mm. The files are a bit underexposed to my eyes, but that just helps with the testing both of the shadow recovery and recovery of the colors. One visual test for me is work up the files with movements without the LCC applied first and look to see just how much saturation and color is lost. To my eyes, and based on the IQ150 testing I did last year, the 100MP at 10mm or so of movement shows a bit less color cast than the IQ150 did. Again, something that everyone can test out for themselves and see if they find it to me an issue.
I tend to agree with Torger on the grounds that outwardly it appears the 100MP and 50MP both respond to movements the same way, i.e. visually you see quite a bit of red color cast on movements. But when you work the files in C1 9.0.3, you can easily see just how well the cast is removed. Then look at the shadow recovery and details in these areas. CI only tested the 100MP with the 32mm and 40mm but they did a comparison with the 60XL using the 280 3100 and a 60MP back. So you can download these files and take a look.
I have worked them all up now, and to me the corrections of the LCC help a lot, but on the shifts with the 32mm and 40mm, you can see the possible need to do a bit of color/sat work in the extreme corner to balance out the range.
I don't think Phase will be bringing anything else to the table, i.e. a separate tech camera designed back, but anything is possible. It seems that they would need a newer tech like BSI, or something similar. Anything is possible and Phase One tends to keep such info close to the vest. I will also say, based on the IQ250/150, what we see now, is pretty much what we will get later. Phase did not really make any huge changes to output (via firmware) to the 50MP backs, again at least to me. I am hoping that on the IQ100 files Phase will eventually create a LCC solution to remove the blothcy banding that still shows up on the 100MP backs, with movements. Visible in all the Alpa outdoor testing with the blue skies. You can't see it on the Capture Integration testing.
Best solution is keep looking for new downloadable files, hopefully some taken outdoors in natural light, and or locate a dealer and setup a demo of the back with your gear.
As someone who has waited for this solution for 4 years now, and made a huge wrong turn on the digital upgrade path by investing in an IQ260 (from a 160), I believe that this 100MP back could definitely get the job done.
Paul C
-
I waited patiently for a full-frame CMOS back, and was super excited when this was announced. However, after looking at files, I have decided to hold off. There additional tests DT is supposed to be posting so will wait and see the results of that as well.
Main issues are:
Performance of wides - 32HR in particular.
Performance of some not so wides (symmetric) like 60XL
Blotchy patterns in skies (at least in the samples I have seen so far)
Some differences in specular highlights, but I need to play with this further
-
However, after looking at files, I have decided to hold off. There additional tests DT is supposed to be posting so will wait and see the results of that as well.
Was traveling to and attending a funeral this week, but should be able to post early next week.
Basic summary is "About the same as the 80mp regarding color cast and range of movement allowed, but with much better live view, more resolution, HDMI option, and better dynamic range."
-
Was traveling to and attending a funeral this week, but should be able to post early next week.
Basic summary is "About the same as the 80mp regarding color cast and range of movement allowed, but with much better live view, more resolution, HDMI option, and better dynamic range."
Thanks Doug, looking forward to your tests. Based on what I have seen so far, the 32HR is significantly better on the 80MP. In the CI tests, the right side was quite bad. Could have been a lens issue?
-
I agree, the Rodenstock probably did not predict the microlens ripple or tiling issues with the Dalsa sensors (and Schneider wides weren't designed for the 6um Dalsa's at all)
The Kodaks were the last sensors that were actively designed for symmetric wide lenses and therefore had light shielded pixels and lacked microlenses. Then it became the other way around Rodenstock made the Digaron series for 6um sensor without light shields (obviously they did not calculate for microlens artifacts though), and then with smaller pixel sizes and still no light shields the situation has degenerated gradually.
think it would be enjoyable having to worry about these things.
I'm wondering about this. Doesn't everyone involved - the sensor manufacturers, the back manufacturers, the lens manufacturers - okay now only Rodenstock - discuss the parameters within which a given product will work? Is it left to individual photographers to find out if something works or not? DT and CI do their tests - but that's after the deal is done.
-
I'm wondering about this. Doesn't everyone involved - the sensor manufacturers, the back manufacturers, the lens manufacturers - okay now only Rodenstock - discuss the parameters within which a given product will work? Is it left to individual photographers to find out if something works or not? DT and CI do their tests - but that's after the deal is done.
Just like Nikon is not obliged to support Sigma lenses, why should Phase One bother to push the sales of Rodenstock?
-
Another way to put it is "either you upgrade to the IQ3 100 and accept the limitations, or you're stuck in that CCD dead end" ;)
Waiting for a new sensor generation that reverses the trend is probably going to be a looooong wait. If we're lucky MFD sensor technology will start moving faster than it did in the CCD days, otherwise this CMOS will be reused in new products the coming 6-8 years.
If the trend is not reversed real soon I think people will start looking for other wide angle solutions, as live view and other CMOS features are just too sexy to ignore.
Will Rodenstock make a new lens line? I don't think so. Partly because return of investment is a problem, and partly because adding even more retrofocus makes it difficult to retain the desired "large format" qualities, lenses would just be unreasonably complex or otherwise not significantly outperform the standard SLR lenses.
It's hard to predict the future though.
If I was Phase One I'd invest in a crazy wide (20mm?) crazy sharp lens for the XF system, let it be the heaviest most complex and most expensive lens and then add digital correction on top. You get a one-size-fits-all for the architecture photographer that just crops as needed from the high resolution sensor.
-
Just like Nikon is not obliged to support Sigma lenses, why should Phase One bother to push the sales of Rodenstock?
Well, they have the A system in their lineup, Alpa plus Rodies.
I think it's not that much up to Phase One. They get what they get with little ability to customize.
Sony makes sensor technology that it can reuse in many markets and sell to many manufacturers. Making a tech cam compatible sensor with current technology means ditching microlenses and adding light-shields, which means losing some DR, losing quite significant high ISO, and increase aliasing. That is worsen performance for probably 98% of the sensor sales to satisfy 2%.
-
I'm wondering about this. Doesn't everyone involved - the sensor manufacturers, the back manufacturers, the lens manufacturers - okay now only Rodenstock - discuss the parameters within which a given product will work? Is it left to individual photographers to find out if something works or not? DT and CI do their tests - but that's after the deal is done.
The 32mm HR-W was designed especially for the 80MP Dalsa sensor, optimized at F8, this you can find on the Rodenstock main page. It takes some digging but when you find the literature on the 32mm, you will see specific mention to this. Optimization was at F8 as I understand it to handle diffraction issues past this due to the pixel size on the 80MP chips.
The bigger issue to me is that in 3 almost 4 years, the 100MP CMOS response to movements is basically the same as the 50MP, Net if you shift your shifted area goes red, very red, and this is not like the response the 80MP chips or 60MP chips give. So yes, either the glass or chip needed to change, and it's apparent that the chip did not change too much, it's yet to be seen if the glass will change, but I doubt it. Not sure how you would get around the issues already in place? i.e. retrofocus wides on tech cameras have to come in very close to the sensor to reach infinity which seems to induce a pretty harsh red color cast.
This screen shot shows the effect of 18mm of shift on the 60XL, untreated LCC's straight from the camera, and you can clearly see the red cast. To me this is outwardly an identical response that the 50Mp chips have. And unless a tweak to the chip has been made since announce it's fair to say this is the response you can expect from the 100MP chip To me it's not same as the 80Mp or 80Mp CCD chips and this really should be explained as each photographer needs to determine if they can work within these confines.
Paul C
-
The 32mm HR-W was designed especially for the 80MP Dalsa sensor, optimized at F8, this you can find on the Rodenstock main page. It takes some digging but when you find the literature on the 32mm, you will see specific mention to this. Optimization was at F8 as I understand it to handle diffraction issues past this due to the pixel size on the 80MP chips.
I haven't succeeded digging that out. When you dig out the press release for the 32HR it says: "A breathtaking super wide-angle lens with an image circle
for sensor up to 40x54 mm and even larger stitch formats and a resolution up to 60 megapixels (pixel pitch ≈ 6 μm)".
http://www.digitalback.com/wp-content/uploads/2015/03/HR_Dig-W_32_Neuh_2010_e_9930.pdf
But as it was released before the IQ180 they wouldn't mention that there I suppose.
If they did design it for the Dalsa 80MP they surely ignored the microlens ripple and indeed also the crosstalk which becomes pretty bad with the 32HR with large shift, I've studied this closely a couple of years back. I remember one user that actually downgraded to a 60MP back because of this.
So indeed I have about the same question as Kumar, what do the lens manufacturers actually know about coming sensor technology properties when they make the lenses? You get the feel that they design for some coarse spec they got from a phone call and then hope for the best.
-
I haven't succeeded digging that out. When you dig out the press release for the 32HR it says: "A breathtaking super wide-angle lens with an image circle
for sensor up to 40x54 mm and even larger stitch formats and a resolution up to 60 megapixels (pixel pitch ≈ 6 μm)".
http://www.digitalback.com/wp-content/uploads/2015/03/HR_Dig-W_32_Neuh_2010_e_9930.pdf
But as it was released before the IQ180 they wouldn't mention that there I suppose.
If they did design it for the Dalsa 80MP they surely ignored the microlens ripple and indeed also the crosstalk which becomes pretty bad with the 32HR with large shift, I've studied this closely a couple of years back. I remember one user that actually downgraded to a 60MP back because of this.
So indeed I have about the same question as Kumar, what do the lens manufacturers actually know about coming sensor technology properties when they make the lenses? You get the feel that they design for some coarse spec they got from a phone call and then hope for the best.
They have totally redone their site and the wording is a bit different, but they still refer to the pixel pitch of the 80Mp back, but I might have that wrong. If so and the 32mm HR-W was designed for 60MP, then the situation is worse than expected.
http://www.rodenstock-photo.com/en/products/professional-lenses-digital/hr-digaron-w (http://www.rodenstock-photo.com/en/products/professional-lenses-digital/hr-digaron-w)
Personally, I would be surprised to see a total redesign, however if it did happen, the cost and weight would be extreme. Look the mass and weight of Phase/Schneider's 40-80 and 35LS, lenses, albeit designed around a mirror box, but the 32mm is approaching the weight limit for a copal shutter anyway. So I guess you can add some form of e shutter to the laundry list.
Paul C
-
To be honest, I can't understand how anyone lives with "accepting" what the manufacturers really should be taking care of BEFORE they put a product out on the market...
I mean, why not shoot film and scan it? Film's cheap, especially when you consider the depreciation factor of the technology you're willing to spend huge amounts of capital on...
You won't have to worry about doing LCC's, color shifts due to moving a lens 5mm too far to the right or left, or all this hullabaloo mumbo jumbo hoops y'all seem to enjoy jumping through...
Perhaps I'm simply not seeing it, or choosing to not "see it", but I can output a 150+ "megapixel" file from a 4x5 piece of well-exposed film, without much effort whatsoever. No dynamic range bracketing needed with today's color negative films either. Portra and Ektar are amazing, especially when used with "technical" lenses designed for systems such as the ALPA.
-Dan
-
To be honest, I can't understand how anyone lives with "accepting" what the manufacturers really should be taking care of BEFORE they put a product out on the market...
I mean, why not shoot film and scan it? Film's cheap, especially when you consider the depreciation factor of the technology you're willing to spend huge amounts of capital on...
You won't have to worry about doing LCC's, color shifts due to moving a lens 5mm too far to the right or left, or all this hullabaloo mumbo jumbo hoops y'all seem to enjoy jumping through...
Perhaps I'm simply not seeing it, or choosing to not "see it", but I can output a 150+ "megapixel" file from a 4x5 piece of well-exposed film, without much effort whatsoever. No dynamic range bracketing needed with today's color negative films either. Portra and Ektar are amazing, especially when used with "technical" lenses designed for systems such as the ALPA.
-Dan
Scanning has a HUGE range of issues in itself. But yeah, for very low volume and artistic reasons large format film is still a VERY viable choice. Glad it is still being made.
Also, some people that post in this forum will ALWAYS find something to complain and worry about, its what they do.
Rather than see what the gear can do they look for what it can't.
-
Scanning has a HUGE range of issues in itself. But yeah, for very low volume and artistic reasons large format film is still a VERY viable choice. Glad it is still being made.
Also, some people that post in this forum will ALWAYS find something to complain and worry about, its what they do.
Rather than see what the gear can do they look for what it can't.
I guess that can mean me, if so sad statement. I don't see anything wrong with allowing a person wanting to use a tech camera have a full understanding of the optical issues that they will have to contend with. They were pointed out during the 50MP chip rollout, and I made the decision to purchase, but due to the crop factor returned the back. I am in line for this back, along with lots of other folks, however I am going into the purchase with both feet wet. I would hope that you would want the same level of understanding before making such a purchase. The are many positives to full frame CMOS, but the user of a tech camera does need to really demo this back or do a lot of research before making this back a purchase.
Each technology here will have issues with a tech camera, CCD has it's own set of problems, CMOS fixes some of those but then creates others.
And I will go on the record with saying that I am seeing the pretty much the exact same response as with the 50Mp chip, and after 3 years of general knowledge of this, I had hoped that a better solution for movements with wide lenses would be developed. But it does appear that for now that is not going to happen at least from Phase One.
I sure don't see that as complaining.
Paul C
-
Hi Paul,
I would guess that you are not the one addressed, could be me…
I would guess that we are living in a changing world with users having different requirements depending on things like shooting conditions, travel and wallet dimensions.
Anyway, I am most thankful for folks sharing experience!
Best regards
Erik
I guess that can mean me, if so sad statement. I don't see anything wrong with allowing a person wanting to use a tech camera have a full understanding of the optical issues that they will have to contend with. They were pointed out during the 50MP chip rollout, and I made the decision to purchase, but due to the crop factor returned the back. I am in line for this back, along with lots of other folks, however I am going into the purchase with both feet wet. I would hope that you would want the same level of understanding before making such a purchase. The are many positives to full frame CMOS, but the user of a tech camera does need to really demo this back or do a lot of research before making this back a purchase.
Each technology here will have issues with a tech camera, CCD has it's own set of problems, CMOS fixes some of those but then creates others.
And I will go on the record with saying that I am seeing the pretty much the exact same response as with the 50Mp chip, and after 3 years of general knowledge of this, I had hoped that a better solution for movements with wide lenses would be developed. But it does appear that for now that is not going to happen at least from Phase One.
I sure don't see that as complaining.
Paul C
-
I wonder if it would be possible to put a couple of different scenarios together for the tech camera users of how to deal with the current set of choices out there. For example -
- say you want workable live view, then its CMOS, which works well really with these lens groups, and moderately so with this other group (say super wides, or symmetrical lenses) ...
- or if those other lenses are of import, then go this way....
- and if shifting is important within a modest range, then...
- unless full range shifting (say 15-20mm on wide angle lenses) is key, then these are your choices....
There's some pretty good knowledge posting here, and getting some summary opinions would be real useful to those of us who struggle with all the different parts and pieces.
Thanks.
-
I guess that can mean me, if so sad statement. I don't see anything wrong with allowing a person wanting to use a tech camera have a full understanding of the optical issues that they will have to contend with. They were pointed out during the 50MP chip rollout, and I made the decision to purchase, but due to the crop factor returned the back. I am in line for this back, along with lots of other folks, however I am going into the purchase with both feet wet. I would hope that you would want the same level of understanding before making such a purchase. The are many positives to full frame CMOS, but the user of a tech camera does need to really demo this back or do a lot of research before making this back a purchase.
Each technology here will have issues with a tech camera, CCD has it's own set of problems, CMOS fixes some of those but then creates others.
And I will go on the record with saying that I am seeing the pretty much the exact same response as with the 50Mp chip, and after 3 years of general knowledge of this, I had hoped that a better solution for movements with wide lenses would be developed. But it does appear that for now that is not going to happen at least from Phase One.
I sure don't see that as complaining.
Paul C
Not directed to you at all. You mostly talk about photo gear in the context of photography unlike others which get into the technical realm and stay there mostly. Would rather them work and talk with the manufacturers directly.
-
To be honest, I can't understand how anyone lives with "accepting" what the manufacturers really should be taking care of BEFORE they put a product out on the market...
I mean, why not shoot film and scan it? Film's cheap, especially when you consider the depreciation factor of the technology you're willing to spend huge amounts of capital on...
You won't have to worry about doing LCC's, color shifts due to moving a lens 5mm too far to the right or left, or all this hullabaloo mumbo jumbo hoops y'all seem to enjoy jumping through...
Perhaps I'm simply not seeing it, or choosing to not "see it", but I can output a 150+ "megapixel" file from a 4x5 piece of well-exposed film, without much effort whatsoever. No dynamic range bracketing needed with today's color negative films either. Portra and Ektar are amazing, especially when used with "technical" lenses designed for systems such as the ALPA.
My digital system, a Linhof Techno, Schneider Digitar lenses and a Hasselblad/Kodak 50MP back does need an LCC (there's no digital solution out there that wouldn't need it) but has zero of those other artifacts. So I'm pretty pleased. However there's no secret that my system is a dead end. If I want to improve anything in my system I must sacrifice something else.
So yes I've been thinking about film for the future, especially if digital degenerates into an electronics festival taking away the zen feel. But really, I prefer the "middle way" I have today.
While I guess you could say you can get 150MP from 4x5" if you look through the grain, if we count grain-free resolution it's not even 40MP, as demonstrated nicely in Tim Parkin's test: http://static.timparkin.co.uk/static/tmp/cameratest-2/800px.html
I don't really mind the graininess of film, but I wouldn't really see it as an upgrade from my current system in terms of image quality, and certainly a downgrade in handling. People tend to forget how much easier even the dumbest digital back is compared to film. So rather than shooting film my speciality could become to shoot with really old digital backs as long as there's life in them :)
-
I wonder if it would be possible to put a couple of different scenarios together for the tech camera users of how to deal with the current set of choices out there. For example -
- say you want workable live view, then its CMOS, which works well really with these lens groups, and moderately so with this other group (say super wides, or symmetrical lenses) ...
- or if those other lenses are of import, then go this way....
- and if shifting is important within a modest range, then...
- unless full range shifting (say 15-20mm on wide angle lenses) is key, then these are your choices....
There's some pretty good knowledge posting here, and getting some summary opinions would be real useful to those of us who struggle with all the different parts and pieces.
It's difficult to do as it's so much subjectivity involved. "Test for yourself" or let a dealer help you is the standard response. Unfortunately it's a bit of a science project to test these systems. I ended up with the legacy systems as it fits my needs and tastes best.
You can use the full Digaron lens range with the IQ3 100MP or a CFV-50c or what you choose. Residual issues aren't visible most of the time.
I'm more negative to this type of pushing-the-limits use than most others, as I 1) don't want to have to rely on super-advanced LCC algorithms for archival reasons 2) don't like to worry about occasional hard-to-predict quality issues jumping out at me and 3) why spend this amount of money on a system that is not designed to work together, it just feels wrong, I don't like it. 4) When there are fine symmetrical lenses I like to be able to use them, they're just so elegant. The 32HR is indeed very sharp, but is a monster.
So I'm stuck with the legacy systems. But if you have a whole set of different views on the issue you may be perfectly satisfied with the IQ3 100MP with movements, on the Digaron series.
As I see it something went wrong already with the introduction with Dalsa 6um, then the dismantling of "large format digital" genre started. It took a couple of years before folks started to see all the new artifacts, and eventually people got used to it and accepted it as normal. Just use less movements, just clean up in photoshop if required. With the Sony CMOSes we're now pushing the limit even further, but eventually people will see that as normal too.
-
I'm perfectly happy with my Sinarback 54H which I use in one-shot mode, Betterlight Super 6K-2 and Acros 4x5 and rollfilm. I was just wondering about you poor souls who need to go through all this after having spent a boatload of cold cash!
-
It looks like Torger's reports may help me save a very large chunck of cash! ;)
Cheers,
Bernard
-
It looks like Torger's reports may help me save a very large chunck of cash! ;)
Cheers,
Bernard
+1
-
It looks like Torger's reports may help me save a very large chunck of cash! ;)
Heh, my perspective is one of many though. It differs a lot between people in how big a problem those artifacts are considered to be.
It is a dry fact though that Sony sensors and and Rodenstock wide angle lenses are not designed to work together. There is no question about that. You can surely push it and you may find the results to be adequate.
To me the largest issue is that dry fact that they're not designed to work together, and the trend so far has been that it becomes worse, or at least not better, for each new generation. Why get into a digital back upgrade cycle when support for my camera system is not considered a factor? I rather then use a legacy system that while primitive it's actually designed to function together.
I don't really blame the manufacturers that it has become this way. While Phase One and Hasselblad et al surely can customize small things like pixel black masks, and minor filter adjustments they can't on their own make Sony change their pixel design. And Sony, they need to consider where the sales are in all manufacturers they sell to. We're not doing BSI or organic sensors yet at this size, so to make today's technology work well with tech wides you need to employ the old-school design, removing microlenses and adding light shields. That does not come for free. You reduce full well capacity (less DR), get worse high ISO and increase aliasing, and maybe you need to increase pixel size a little too and reduce resolution down to about 60 megapixels. While I would love such a tradeoff while waiting for the BSI and organic sensors, it's obvious that it wouldn't be a big seller in the big markets.
What you could complain about is that Phase One seems to pretend that this issue doesn't exist and push the problem out to dealers and users. If I had technical responsibility on Phase One I wouldn't sell the A series system with IQ3 100, even for center frames I don't think the back is up for the 23HR. But they do sell it, which I think hurts their credibility when it comes to taking image quality seriously, at least in my eyes. If you do take it seriously you shouldn't push it to the user to evaluate, you should use your in-house expert eyes and technical expertise and make the decision what's good and what's not. This time it seems like their desire to sell is stronger than their will to maintain the highest image quality standard, and for that I think they deserve some criticism. I think they will get away with it though as the problem is subtle, even undetectable with typical subjects (except possibly that blotchiness issue some talk about, I have not studied that myself) and tech cam users have always been more interested in corner pixel peep sharpness than in color response and tonality.
When Kodak was around there was the concept of making two types of sensors, one that was designed for working with those symmetrical lenses, and one to be used with mirror box cameras. This has been lost and I doubt that it's coming back. I do think that there's a high chance that angular response is coming back either with BSI and organic sensors (as you then get it "for free"), but it could be 6-10 years until it's available in MFD.
My intention is not to trash talk the IQ3 100MP with use with technical cameras, I just want my fellow users to know that there are issues and that you should approach it with care, just as I think users should know that the tech cam future is an uncertain one, meaning that if you're the type that like to upgrade and have the latest and greatest you may end up having to switch camera system sooner than you'd like. I don't think users were well-informed about what would happen with lens compatibility issues when upgrading from P45+ to P65+, and not from P65+ to IQ180 either. I think/hope this time around we're more well-prepared to make a long critical look before jumping on the upgrade.
-
I wonder if it would be possible to put a couple of different scenarios together for the tech camera users of how to deal with the current set of choices out there. For example -
- say you want workable live view, then its CMOS, which works well really with these lens groups, and moderately so with this other group (say super wides, or symmetrical lenses) ...
- or if those other lenses are of import, then go this way....
- and if shifting is important within a modest range, then...
- unless full range shifting (say 15-20mm on wide angle lenses) is key, then these are your choices....
There's some pretty good knowledge posting here, and getting some summary opinions would be real useful to those of us who struggle with all the different parts and pieces.
Thanks.
I will offer a slightly different position on all this than Torger, although I agree pretty entirely with everything he contributes. I see a distinction between 40mm and up vs 35mm and down. I will attempt to propose some criteria:
- 40mm is the widest you need to shift/rise/fall, and you are satisfied with the 40hr vs the sk43 or 35.
- Shift/rise/fall up to 15-18mm, 10-12mm on the 40hr.
- The 32hr can be used, but for the most part on center
If you read the list above and it sounds terribly limiting, then the CMOS backs will be a challenge. The best current back being sold in that case may be the IQ360. If the above sounds like criteria within which you normally shoot, then perhaps you should start saving. :)
These limitations are not to be taken lightly; they could be quite limiting to some and of no issue to others. And of course some will say these are too conservative while others will say they are full of issues. That's where the "demo and try it yourself" response comes in.
I don't have the IQ160/260/360, but I do have the IQ180. I tested the 3100 and felt it had about the same limitations as the 180*. Although the actual color cast was very different between the two, the LCC process had about the same success correcting issues. There is a good percentage of technical camera users (maybe 50%?) who prefer the 60 mp series to the 80 mp because of the increased latitude for movements. I see that group as the one that will be cautious with CMOS. I have been thrilled with the results using the 180, so I represent the other half (?) who don't mind the limitations. Note that doesn't mean I will get in line for the 3100. I'm still on the fence.
Anders is in the conservative side of the 60mp group because of his sound preference for symmetrical lenses and his sensitivity to any and all aliasing issues. That is a perfectly legitimate and respectable concern.
Dave
*Tested with the 40hr, 60xl, 90hrsw, sk150
-
The thing is that live view was most awaited for use with tech cameras and that tech cameras are mostly useful since they allow movements. ;)
Without all this perfectly operationnal, the differentiator of the phaseone system vs a 4 times cheaper Pentax system gets less and less visible.
Cheers,
Bernard
-
The thing is that live view was most awaited for use with tech cameras and that tech cameras are mostly useful since they allow movements. ;)
Without all this perfectly operationnal, the differentiator of the phaseone system vs a 4 times cheaper Pentax system gets less and less visible.
Cheers,
Bernard
That is so true. Or perhaps an interesting offering from Hasselblad lurking in the wings that could be used with a nice manual 501...
-
Thanks for broadening the views Dave. I am indeed the "conservative" type, and there's other ways to see it which makes it so difficult to provide one recommendation that suits all.
-
torger,
While Phase One and Hasselblad et al surely can customize small things like pixel black masks, and minor filter adjustments they can't on their own make Sony change their pixel design.
Why not? If Leica can have a sensor fabricated to its specification by CMOSIS then why can't Phase One/Hasselblad et al. by Sony?
-
The thing is that live view was most awaited for use with tech cameras and that tech cameras are mostly useful since they allow movements. ;)
Without all this perfectly operationnal, the differentiator of the phaseone system vs a 4 times cheaper Pentax system gets less and less visible.
Cheers,
Bernard
Not entirely because even when used only on center the tech camera lenses are MUCH better than any Pentax lens. The difference is quite significant. Also tilt is available with a tech cam so even if you do not use shift (or very little) it is still more than you can do with a Pentax 645z specially with wide angle lenses. For working on a tripod the tech camera solutions are still the best when you want better optical quality.
It would be interesting to see the 100mp A-Camera test results before making any conclusions.
-
torger,
Why not? If Leica can have a sensor fabricated to its specification by CMOSIS then why can't Phase One/Hasselblad et al. by Sony?
Volume maybe? Hassy and Phase work with much lower sales volumes than Leica.
-
Why not? If Leica can have a sensor fabricated to its specification by CMOSIS then why can't Phase One/Hasselblad et al. by Sony?
It's indeed a good question, and I have just as little details as anyone else so it's about semi-educated guesses. I don't think Sony work as CMOSIS, but anyway even CMOSIS also have their set of pixel technologies and you can choose from that and package it at a custom size (within certain manufacturing limits), but if it comes to making a something entirely new then it's associated with very large development costs. It's best if you can make one pixel technology and reuse in many different packages.
It would be interesting to know why Kodak put so much effort into making such wide angular response, was it only for the large format lenses, or was there some other application too? I don't know. In any case whatever the reason was it doesn't seem like it's as important any longer when new sensors are made, at least not important enough to make trade-offs with other performance aspects.
And then it comes down to sales, how large is the wide angle tech cam niche, and within this niche how large is the group of "conservatives" that actually holds off buying due to compatibility issues? If the niche is small and the backs sell anyway to that niche why care? It would probably require an idealist that cares more about the traditional format as such than maximizing profit, and I'm not sure MFD has the luxury to think that way...
-
Hi,
To be correct, the Leica sensor was designed by CMOSIS, it is fabricated by STMicroelectronics.
Now, it is a good question if Phase or Hassy could make requirements for a Sony made sensor, that could be feasible, but both Hasselblad and Phase One are selling backs foremost for their own systems which are DSLRs.
With Leica, it was different as their M-series cameras operate with extreme beam angles. So they needed to have a sensor for supporting those lenses.
The CMOSIS/STM sensor is said to have some issues that need software fixes. Edmund has worked with some sample chips and says so.
Best regards
Erik
torger,
Why not? If Leica can have a sensor fabricated to its specification by CMOSIS then why can't Phase One/Hasselblad et al. by Sony?
-
What also has to be considered, is the law of diminishing returns. Phase One, to their credit saw this chip coming over a year ago. Did they push for a technical camera favorable design, or a chip with great DR and low noise. No one will know but a few.
Phase One has been slowly working up their lens lineup, with the 35LS, they have a winner as with most of the blue label glass. These lenses can work with the new sensor, and work with the older CCD chips. Phase has a good line up of glass and again to their credit, with the XF body and the vibration reduction setting they have done a very good job. It's not VR in the lens, or on the chip, but something different, but it works and works well. Schneider may have seen the writing on the wall, and tossed in their cards on tech lenses. But they are still right in the mix as all the new glass from Phase One is designed in partnership with Schneider, albeit, made in Japan probably in the old Mamiya plant? Still Schneider is showing their abilities for sure with the lenses they are producing for Phase One.
If I was on the board at Phase One, I would be pretty happy right now. The market I am going after is pretty limited anyway, and I can offer it a full range of products.
There was a period of 3 years where Sony and Phase I am sure had a lot of conversations on the other side, the Tech Camera.
But Phase and Sony had a lot less control on that piece. You have a small spattering of companies making the equipment, Arca, Cambo, Alpa, Linholf and Rodenstock and Schneider. Phase One can't control much if any of this, and when you consider the total number of digital backs sold (new or as upgrades) world wide, I can't see it being more than 20K if that. More Nikon and Canon Pro DSLR's and now Sony cameras are sold in a month than this or maybe in a week. Many photographers have moved to the Sony A7rII and adapters that allow the Canon shift lenses to work or solutions like the Universalis Arca or Cambo Aptus. Both give the use of the tech glass in a very price affordable package, (but not much wide support).
I have to believe based on the results of the 100MP chip that Sony and Phase did the best they could within the technology available today. And Phase One has done a lot of work on the LCC to go along with the camera. Just look at any of the test shots on the market now, where movement was involved and you can judge for yourself just how much work is going on under the covers to get that amount of red cast out.
Will Rodenstock come out with a new lineup? Anyone's guess. They probably make a lot more in the eye glasses market then camera glass. It's all about the bottom line. Copal is gone, the supply will dry up soon. The 32mm HR-w has already reached a point where the mass of glass in front of the copal is too much (very delicate and easy to mis align the shutter even Rodenstock has pointed this out). So will new glass be any lighter? I can't see that happening, but of course maybe e shutter will finally be out there.
Others on the forum talk about creations of sensors like they are just something that can be made in a few weeks, from start to finish, without the factor of figuring out the recovery of that R&D in sales. This market is very small and sadly getting smaller and the technology is surrounding it from all sides putting a lot of pressure on the few companies that continue to product the products.
Phase One, now really has 2 products, the 50MP and 100MP CMOS, sure there will be sales of the 80MP, 60MP 40MP CCD backs, but you have to wonder how many of them will be new and or upgrades vs used/certified used sales. Will Phase come out with a 75MP or 60MP CMOS designed for tech cameras? Hard to say when Canon and Sony are at 42MP and 50MP right now, and a lot of people already don't see any difference in the output from the cameras.
The only possible mistake I have seen is where Phase One priced their 50MP back. I believe that a more reasonable price point could have really lead to a market share growth for them. But they have their own ideas on where they want this back to go. It would be interesting to see sales figures for IQ150 vs Hasselblad 50c, personally I don't believe the IQ150 had many sales, even Adobe was un aware of it until myself and a few other folks pushed for months to have it added to LR camera raw support.
It will be interesting to see what happens as the year rolls on.
Paul C
-
Paul,
Just look at any of the test shots on the market now, where movement was involved and you can judge for yourself just how much work is going on under the covers to get that amount of red cast out.
To the detriment of file quality?
This market is very small and sadly getting smaller...
If memory serves, according to Doug, Phase One sales have increased year on year.
Will Phase come out with a 75MP or 60MP CMOS designed for tech cameras?
I suspect that a 50MP full-frame CMOS sensor would be extremely popular with technical camera users, combining as it would good angular response from wide-angle lenses.
It will be interesting to see what happens as the year rolls on.
It will indeed.
-
Paul,
To the detriment of file quality?
If memory serves, according to Doug, Phase One sales have increased year on year.
I suspect that a 50MP full-frame CMOS sensor would be extremely popular with technical camera users, combining as it would good angular response from wide-angle lenses.
It will indeed.
I take the sales numbers with a grain of salt, until you see something in writing, which unfortunately I don't think will ever happen, (wonder why?), just asking. Also, as far as I know there are 4 main dealers for Phase in the US, CI, DT and Bear imaging, maybe Samy's I haven't checked their site in a while. There are 319 Million folks just in the US, and I can bet you over half of them have a camera (camera phones included as the vast majority of people don't understand the difference anymore). The market is changing pretty fast and the perception of that a photograph is and what quality is has changed with it.
The issue of loss of quality, that's up to the user.
There is still a ton, yes a ton, of stuff being done under the covers with a IQ260 file and a tech camera. Microlens ripple, tiling, color cast are just a few of the issues. Probably the last Phase One back that did not have this many issues and a tech camera was the P45+
When I used the IQ150, I did not see the quality fall off that some see, I saw a lot of good work done with a LCC that gave a very good file. I just did not like the crop factor and still don't. But you make a valid point, as something new is always around the corner for sure and a more favorable chip may still hit the market next year.
Take a look at the CI tech camera shots, they are on their website to download. I don't see that big an issue with the 32HR-W shots, besides a focus/detail problem in the right corner and more than likely that is a lens issue (alignment). Corrected out, the shots past muster for me.
Paul C
-
Paul,
The issue of loss of quality, that's up to the user.
The issue of acceptability of loss of quality is up to the user - loss of quality resulting from required application of an LCC is not.
I don't see that big an issue with the 32HR-W shots... Corrected out, the shots past muster for me.
My criteria is more onerous: results from the 23HR must pass muster.
-
I assume all interested parties have already seen this thread at getdpi.com:
http://www.getdpi.com/forum/medium-format-systems-and-digital-backs/57712-dt-tech-cam-samples.html
Thanks a lot to Doug and his team for the work done here.
Having checked a few samples, my personnal take is:
- The gap of image quality between the Rodenstock tech camera lenses non shifted and the Schneider SLR lenses is vastly exagerated. Yes, the Roddys may be a bit cleaner and a tiny bit sharper, but it is very far from being the night and day difference some have been boasting about,
- Even the supposedely perfect 90mm shows a very significant degradation of sharpness shifted to the extend that using a wider lens cropped may end up delivery similar image quality (and both will be far behind a stitch when manageable),
- I don't see too much of the color issues/loss of saturation mentioned by Anders even with some of the wider lenses,
- the wider lenses (23mm and 32mm) are pretty much unusable without a center filter.
Comments welcome. ;)
Cheers,
Bernard
-
I haven't looked at the raws myself, as I've pretty much lost the interest in the system.
They've learnt a lesson from the IQ250 tests to not push the sensor to obvious breakdowns but tried to show it in the useful range. And they're discussing the method to combining small shifts with software perspective corrections, or using XF + 35ls instead and just do software perspective corrections. We're starting to see a "paradigm shift" in how MF manufacturers think you should make high end architecture photography, where tech cameras are eventually replaced by standard cameras without optical shift. Traditionalists like myself that are attracted by the "large format shooting style" think it's a bit boring, but I can't stop technical development.
Crosstalk-related color issues are typically very subtle unless you push it way too far or have some exotic color combination or light. I've described in another thread a while ago why it's pretty messy to evaluate that aspect, and I'm probably not doing more of that now. I already know as much that it has the same problems as the IQ250 in about the same extent. I don't want to spend a large amount of money on a digital back and having to worry about pushing the sensor past its designed limits even if it usually doesn't show much. I don't have any problem with others doing that, but it's not my thing.
Leica kept the long-time technically obsolete manual range-finder design just because there's a niche of users that like the shooting experience. In a similar way I'm hoping that there will be a market for traditionalist tech cam with optical movements but I'm afraid the concept won't survive another 6-8 years with digital backs that doesn't play well with it.
-
I assume all interested parties have already seen this thread at getdpi.com:
http://www.getdpi.com/forum/medium-format-systems-and-digital-backs/57712-dt-tech-cam-samples.html
Thanks a lot to Doug and his team for the work done here.
Having checked a few samples, my personnal take is:
- The gap of image quality between the Rodenstock tech camera lenses non shifted and the Schneider SLR lenses is vastly exagerated. Yes, the Roddys may be a bit cleaner and a tiny bit sharper, but it is very far from being the night and day difference some have been boasting about,
- Even the supposedely perfect 90mm shows a very significant degradation of sharpness shifted to the extend that using a wider lens cropped may end up delivery similar image quality (and both will be far behind a stitch when manageable),
- I don't see too much of the color issues/loss of saturation mentioned by Anders even with some of the wider lenses,
- the wider lenses (23mm and 32mm) are pretty much unusable without a center filter.
Comments welcome. ;)
Cheers,
Bernard
You are definetly right. As an owner of 7 Schneider Digitars from 35XL to 180mm I can attest that the Schneider's are easy to shoot shifted. Sure they need an LCC but so do the Rody's. The 35XL is very sharp and with the latest version of C1 I'm able to get a clean 10mm of shift or rise on my IQ180. I have no doubts that when I receive my 100MP that it will perform as well or better.
Victor
-
You are definetly right.
In fact my comment wasn't about the Schneider tech lenses, but about the Phaseone Schneider Mamiya mount ones, obviously not shifted.
Cheers,
Bernard
-
Yes, it's safe to say that the new Schneider lenses, 35LS, 40-80LS, and 120mm Macro are very good indeed. They are priced in the same range as the Rodenstocks, so they need to be. The 55mm and 80mm are also very good optics.
Paul C
-
In fact my comment wasn't about the Schneider tech lenses, but about the Phaseone Schneider Mamiya mount ones, obviously not shifted.
Cheers,
Bernard
It does pay to read carefully.. ;D
Victor
-
Should be said that the Schneider XF system lenses get precise software correction in Capture One, which the tech lenses don't get, so they're cheating a bit ;), but indeed the results are very good with cheats applied and it's the result that counts.
-
The Schneider's get corrections, but most of the Rodenstock's do now also, but as far as I can tell the corrections are standard distortion type adjustments.
What would be nice, for 6700.00 or 8700.00 (price of 35LS and 40-80 respectively) would be that Phase One develop a way to handle diffraction softness as Fuji and now Canon are doing with their lens optimization routines.
Sadly the Fuji Corrections can only be used on in camera jpg, not raw (sure would be nice if C1 or LR figured out a way to use it) and Canon has announced this on the 1DS MKII, not sure how it will in implemented i.e. only on in camera jpgs or raw or both (will C1 and LR be able to use the data?)
With 100MP backs or larger, this type of optimization is what I am hoping to see in the future. I don't see this working with tech lenses are they are pretty dumb, in regards to giving any feedback. And I only hope that it can be added to the modern Schneider/Phase optics via firmware.
These are the small changes that will possible make using such high MP backs a bit easier.
Paul C
-
What would be nice, for 6700.00 or 8700.00 (price of 35LS and 40-80 respectively) would be that Phase One develop a way to handle diffraction softness as Fuji and now Canon are doing with their lens optimization routines.
Sadly the Fuji Corrections can only be used on in camera jpg, not raw (sure would be nice if C1 or LR figured out a way to use it) and Canon has announced this on the 1DS MKII, not sure how it will in implemented i.e. only on in camera jpgs or raw or both (will C1 and LR be able to use the data?)
So far as I know this is glorified in-camera sharpening. The same deconvolution sharpening could be applied to raws from any camera; they are just doing it in camera for those who shoot JPG.
-
The Schneider's get corrections, but most of the Rodenstock's do now also, but as far as I can tell the corrections are standard distortion type adjustments.
Distortion, sharpness falloff (though, obviously, very little is needed), and chromatic aberration mapping (again, very little to be done) are all present for the characterized Rodenstock lenses.
-
Distortion, sharpness falloff (though, obviously, very little is needed), and chromatic aberration mapping (again, very little to be done) are all present for the characterized Rodenstock lenses.
I suppose you need to remember your tilt, shift and aperture settings though so you can put that into the correction algorithm in C1?
-
So far as I know this is glorified in-camera sharpening. The same deconvolution sharpening could be applied to raws from any camera; they are just doing it in camera for those who shoot JPG.
Maybe, but looks like Canon seems to feel it's more than "glorified in camera sharpening" and if you read about a bit on the Fuji side, there quite a bit going on that Phase One might be able to consider. At 100MP even F11 may start showing diffraction issues and wouldn't it be better to correct this at the time of the raw conversion vs later on with a third party software like Focus Magic or Piccure+. Both do a great job but it just seems it would be better during the initial processing of the raw itself.
As far as sharpness fall off, in C1 that is available in an adjustment slider for the lenses mentioned. Are you saying that there is also some sharpness fall off being done in the raw conversion (not sure how C1 could do this as for a tech lens as it's unaware of the lens from the meta data) and you have to manually select the lens in the lens tab. I could see it being done for the A series maybe as more communication seems to be there at least for the LCC process. And for the new Schneider LS optics as there is communication between the back/and XF and lens and I guess firmware in the lenses also.
Just curious.
Paul C
-
I suppose you need to remember your tilt, shift and aperture settings though so you can put that into the correction algorithm in C1?
This is the one reason why I am not looking forward to upgrading and having to deal with the Roddie wides. For me, it is not too bad, since I almost always shoot tethered.
However, there are plenty of times when shooting exteriors I just shoot to the card, and having to write down the movements is kind of annoying and time consuming.
Considering that Acra has a sensor in the camera to assist with focusing on the eModule, I guess they could install sensors to record shift and tilt. However, this would be on Arca's side and I am not sure how that would get translated to the metadata of the image.
This being said, dealing with a CF is a pain, and I would rather have the Roddie 40mm in an IQ #60 back then my SK 35 on a P45+.