I just shoot and scanned using the Top of the line Nikon 850 with prime art glass. I am getting amazing results, just beautiful. I rented this camera. I am looking into getting a loan to for gear to scan fossils and artifacts. Being greedy, in terms of quality, as post fixes in zbrush etc. work well, but the salary and time costs are more than any camera over a year’s time, I want to get the very best. So the question is has anyone had time to compare the Sony, Canon and Nikon 40+ SLRs for this work? I chose the Nikon because of the XQD solid state cards don’t heat up as fast. I am finding that 1200 images provide perfect results on giant skeletons and that is much easier to capture on big solid state recording media.
My understanding is that the RAW SLRs use some digital software techniques to produce such large images and this can be something that really holds no benefit to PG.
Has anyone been able to assess the difference between these and the true digital medium formats?
If you had a good budget for gear would you suggest medium format or the SLR Raw cameras?
i am trying to decide between Fujifilm GFX 50S and the Nikon 850, I cannot afford Hassleblad. any suggestions? A
The choice in full frame is between the d850 and sony a7riii. I could go in to specs but it sounds like you know your stuff.
What do you mean " RAW SLRs use some digital software techniques to produce such large images"?
Medium format is “better” but expensive. I think you will hit a plateau from capturing too much from one position. Could be wrong though never used medium format myself.
The RAW SLRs are all the new 40 mpx+ cameras, 5DR, the Sony one, and the NikonD850. My understanding is that the large mpx are more of a software trick like what happens with anamorphic lenses in film, but medium format is solid data. I was just wondering if anyone has done this very expensive experiment.
I have the Sony. It’s a true 42MP no aa filter. There is a mode called pixel shift that uses the stabilization to move the sensor 1 pixel in the up, down, left, and right plane to give each pixel value full RGB data. I think Fuji did it first and a lot of people thought this is like a 42MP camera taking 168MP photo. It’s more closer to having maximum color depth and shooting at ISO 25.
Other than that I know of no software tricks going on. I do want to try this on a subject, but haven’t found something worth the hassle. And it is a picky about movement in the frame and a sturdy tripod.
I am using a SONY A7R2 with sony/zeiss 16-35 f/4 and this is a great balance of resolution, weight, coverage/overlap, stabilization… the only caveat is a buffer that fills up a bit fast when you shoot in RAW+JPG quite fast.
I am considering a A7 III because I feel like 42 MP is maybe overkill in most cases. Still the 16-35 f/4 sony/zeiss zoom is my favorite for photogrammetry.
I also have a SIGMA Foveon DP0. Not a full frame, but the Foveon sensor is really fantastic for photogrammetry according to the few tests I did. But much slowler to use, and you are stuck at 100 iso, so only for outdoor bright day use.
And I am also using a CANON 5DSR 50MP but found out that at slow speeds (when shooting indoor around 1/40 to 1/100th), even with a stabilized lens, I had not so great results, this is a bit better with the sony. I suspect this is because of the mirror shock (the sensor being not stabilized) ? So now I am using it with a GIGAPAN head to take automatic panos of large spaces.
I think this is like asking which car is the best - it depends on so many things, like Jonathan hinted at. Everybody has different needs. I am not so certain that bigger is always better. A bigger sensor has not jut advantages. For example if you need a huge DOF, then a smaller sensor like APSC or even M43 might be a better option than a medium format.
I have heard and read repeatedly that the 850 is considered the camera with the best image quality for photogrammetry. But it comes at a price: money, weight and bulk. If those three are no issue, then why not?
I am aware that it is a large scoped and hard to answer question for a one size fits all. I am looking for people’s experience as stats mean nothing in the real world. The 850 seems like the answer for me, fast high quality cheaper high end glass (Sigma Art and Zeiss). The XQD cards are perfect, don’t overheat and cache well. The sensor is superb, batteries are great, manual controls are solid and not easily bumped mid scan. But before, I take out a loan, I wanted to take one swipe at seeing if people were getting better results from medium formats. Not a ridiculous notion.
My goal is to start with 850 for the field and then use medium format Fuji in the lab when we have the money.
I don’t have much experience with medium format (I just took a few pictures with the FUJI at my local camera store), but from what I read, there are slower to use than the high end 24x36 FF. And going medium format will mean less Depth of Field also, which is not what we want in photogrammetry usually. And of course, bodies and lenses are heavier than the regular 24x36 FF (especially the sonys hybrids), so this can count at the end of the day after thousands of pictures taken handheld… And lenses and bodies are much more pricier… So for me, the high end 24x36 FF are still the best choices for photogrammetry.
Pixel Shift looks great, but I am afraid this is not that practical on the field when you have to shoot lots of pictures in a limited time ? And is the quality raise really worth the extra time / memory / storage … ? I am not that convinced.
don’t get me wrong, I am all for asking questions!
What keeps me from wanting an 850 is the bulk and the mirror, which has it’s own issues.
Jonathan, do you think pixel shift takes any additional time? If, then it should be within the scope of the exposure and if that isn’t a second or more, then it should be below a relevant threshold. But it does need post processing as it seems…
But even without this technique the A7III (or it’s R cousin) should be on par with the 850, at least in terms of noise and image quality.
Oh, and for the case Fuji uses an X-Trans in the medium format as well, then you should be aware that this sensor needs a different de-mosaicing than the normal bayer sensors which many well established Raw processors cannot handle as well, or in case of DxO not process at all…
Jonathan, do you think pixel shift takes any additional time? If, then it should be within the scope of the exposure and if that isn’t a second or more, then it should be below a relevant threshold. But it does need post processing as it seems…
Yes, and if you shoot handheld I am not sure it works so well, because the camera will have moved a little between the shots, event if this is 1/15th total… On tripod this should be easier.
Steven, did you test the pixel shift on the A7R3 ? what is your feedback about it ?
Götz You are making me take a serious look at the A7R3, your issue with mirror is the physical mechanics in general, not a specific Nikon issue? I ldo ike the XQD cards and their speed and capacity for fast field shoots, and the Sigma art glass which is very reasonable for the price and quality. Availability of and choice of glass is consideration. Also, I want to mount to stabilizers and drones soon.
I have never had an issue with uhs-II overheating or speed issues with the a7riii. The pixel shift is 4 captures at minimum 0.25 apart. You can make it longer, but this is the lowest time you can set.
The a7riii and the d850 are both quite close in image quality. At base ISO the d850 wins out because the base ISO is 64. The A7RIII is better at higher ISO’s though.
The price is also a consideration. Sony’s bodies are cheaper and more bang for the buck, but the glass is more expensive generally. What I have heard time and time again, and after years know to be true, is buy the glass first and then the body to match. Your results will depend on the lens more than any other factor.
You should also know Sony makes the Nikon sensor. They make almost all sensors from the Iphone to the Phantom. Canon is the only real competition in this department.
Pixel shift takes a long time. The files are huge. I can see the improved results with my eyes though. It gives better color and divides the ISO of whatever you shot at by 4 (so ISO 100 is now ISO 25, 4 times less noise) Its not some marketing gimmick, and I would love to find a subject worth committing the time to. I can say that if your lens cannot resolve the megapixels of the camera sensor though it will not make a difference. You have to have the glass to match.
Overall, I’ll say even though I haven’t used it in photogrametry yet, I’m glad it’s in my bag of tricks for those times it matters. Just like anything in photography and computing (life in general) you can trade between quality, time, and money.
Some people would say that Fuji is also a competitor with their X-Trans !
There are some benefits and also some disadvantages, like with everything.
Also, it doesn’t just depend on the sensor but also the software behind it. I’ve read people say that Nikon can get better results with the same hardware than Sony. But that’s a whole other issue and I have no personal experience. Canon can be a bit tricky for photogrammetry since some sensors apparently alter the geometry on a pixel level.
The mirror issue is a very general one - it needs to be moved before each shot and that needs to be handled properly (CAN cause vibrations). Just another variable to think about and imho not neccessary any more at all. Plus it’s another moving part which can break or wear down. No idea why “they” still make DSLRs these days.
