Choosing between Medium Format and 35mm digital platforms

Digital Camera Scanning
1

Hi everyone,
I apologize if this is a stupid question, feel free to remove if not allowed.

I was wondering if there were significant technical benefits to 16bit vs 14bit capture platforms, (ie GFX100s vs z8) and medium format sensor size vs 35mm sensor size when doing camera scanning and reproduction work. I have been camera scanning with a z9 and am very happy with the results. I am wanting to future proof my setup for the eventual death of my nice lab scanner (Fuji Lanovia Quattro) and was curious if anyone has used both the z8 and the GFX100s and could speak to the technical differences. Is the 16bit vs 14bit really worth the hassel of upgrading a camera scanning setup to medium format. I am also curious if people see an increase in color fidelity when using the pixel shift modes on the z8 so that each pixel has rgb data and how that impacts color inversion fidelity.

TLDR
Is the upgrade from to medium format from a 35mm sensor with the same pixel density worth it
Is 16bit vs 14bit a big difference with camera scanning (gfx100s vs z8)
Does RGB pixel shift increase color fidelity when doing color negative inversions

For reference I have been camera scanning thus far with a z9, 75mm 1:1 Rodenstock D, and full valoi setup, looking to upgrade to the gfx100s and the contax 120mm macro to gain 16bit and pixel shift capabilites, but curious if i should just get a z8 and use that instead given i already have everything for a 35mm setup.

2

I have not used the specific equipment you mention, but I have seen the results of Phase One cameras versus Sony a7r4 and quite frankly Iā€™ve come to the conclusion that all these imaging systems are so good that the limits of human visual perception and the limits of colour negative film are likely to be the main constraints on actually seeing systematic differences of quality between various cameras once you get into the quality range at issue here. The two things that really matter are the sizes at which you intend to print the eventual output and the quality and properties of lens you will be using.

1 Like
3

I agree with what @Mark_Segal wrote above, investing in medium format cameras just for scanning might probably be out of proportions.

One thing though: Depending on what you scan, the wasted number of pixels can differ quite a lot. Most of my (old) analog material is in 645 format with an aspect ratio of 4:3. Here, a GFX would increase the number of usable pixels considerably. Scanning 135 format is best done with a 3:2 proportioned sensor. Iā€™ve commented this in several threads, one of them is this in post: Do I need to change my lens/camera to scan medium format? - #8 by Digitizer

Paper prints have a range of about 7 bits, whether you start with 14 or 16 bits seems negligible - unless you do really crazy things or have to handle heavily underexposed material.

All in all (and apart from the aspect ratio considerations) weā€™re in that part of 80:20 where cost/benefit is not budget friendly.

4

Itā€™s an interesting series of questions but Iā€™ve not seen anyone give definitive answers to any of them, perhaps they will here.

The 4-shot ā€˜Bayer-cancellingā€™ pixel shift mode should in theory provide an improvement but again Iā€™ve not seen it demonstrated, and thatā€™s not confined to the GFX of course, Mark D Segal has that on his Sony A7R IV so Iā€™m sure he would be making use of it if he felt there was an improvement, though having to process them in separate software counts against it in terms of an efficient workflow.

Certainly true that a 100MP Fuji should have an advantage for medium format 645, 6x6 & 6x7 though.

You havenā€™t mentioned whether you want to scan medium format, for 35mm the Z8/Z9 would be so good already that any improvements would be almost impossible to quantify. However, for 35mm copying are you confident that the Contax 120mm Macro on the GFX would resolve as well as the 75mm Apo-Rodagon D 1x does on full frame?

5

RE: Pixel Shift - I have done some pixel shift testing on B&W film (and some of us have discussed at length on the forum) to look at sharpness, ā€œresolutionā€ and color noise. I used a Lumix S1r. I found and posted a lot of interesting information online both theoretical (with math) and practical (with examples) that went beyond my own testing in the thread. I have scanned 35mm slides extensively as well, but the aim was simply print size for now on these.

On the slides I have scanned recently, sometimes the histogram shows more color data but the preview images are indistinguishable. Perhaps they are more flexible to edit or have finer color gradation? Pixel shift theoretically brings the color depth closer to the full bit depth of the sensor as I understand the mathā€¦ but can our eyes even distinguish 11.8bits from 13.2bits on a screen using a 14bit sensor? How about from a 16bit sensor? One video I found showed the photographer got cleaner shadow details but again I have not tried it on color negative film as he did, only B&W and slides.

Pixel Shift is currently a hassle for most brands because they do not produce a RAW file in-camera and so it becomes even more time consuming if you intend to do it at high volumes unless you can create an action/automate it in some way. That said, there seems to be usefulness to the process for some needs. When I get a break from big work projects, I will do another similar test with color neg and slides larger than 35mm ā€¦ hopefully 2-3 more lenses, too.

Is the pixel shift on a 16bit camera even more capable? Mathematically yes, but realistically I donā€™t know yet! Would love the chance to compare with people though.

@Digitizerā€™s 80/20 comment applies strongly here. Is it worth it? I would say no if it is exclusively for scanning at home and not making truly large prints. I would say yes if it is for an institution doing cultural heritage work, especially if they have the budget for it and the proper lens. I have yet to find an image that stumped the 14bit capture of my R5 and S1r. I use the pixel shift sparingly for top portfolio level images.

And @Harry is right, there are no definitive answers yet, I think. But I do see that most large institutions and cultural heritage capture is done with fancy 16bit, 100+MP specialty rigs nowā€¦ at the same rate the very companies who advise for the standards for cultural heritage capture also sell the gear to do soā€¦ it isnā€™t exactly an independent standard in that regard.

For now the benefits I have personally seen are: less color noise, larger total pixel count without stitching, and I think it did a nicer job on the B&W grain resolution. But again, this is really chasing the dragon, most people for most uses wonā€™t notice any of this.

EDIT: @Mark_Segalā€™s point is a more succinct version of what I said. Sorry Mark! You said it and I missed it :slight_smile:

1 Like
6

Your work on Pixel-shift with the Panasonic S1R and the thread that it generated is a very valuable addition to the discussion, the only reason that I didnā€™t mention it here was because the Panasonic ā€˜onlyā€™ does 8 shot so increases the resolution at the same time, it would have been very interesting to see the result had it been able to do 4-shot, and in-camera processing is such an advantage on the workflow (and file size) front.

Reading it all again I see that Iā€™ve done Mark D Segal a disservice, he did share his test of 4-shot mode and found that it didnā€™t give him an advantage, at least in terms of the resolution of the detail in the dye clouds. I suppose the conclusion is that the processing of Bayer images is so good that any that theoretical difference that 4-shot pixel-shift might give is extremely difficult to demonstrate.

https://forums.negativelabpro.com/t/pixel-shift-vs-regular-capture-quick-comparison-w-s1r/8358/424

That also might mean that 4-shot is not seen to be a useful sales feature for manufacturers, and possibly why Panasonic chose not to offer it at all. Thatā€™s also probably why the Canon R5 MkII has moved from pixel-shift to AI interpolation. AI is such a marketing buzz word now and thereā€™s none of that difficult science to have to explain, never mind that the actual results are shown to be inferior.

7

This is probably true for every question about everything that is not completely framed :grimacing:

The one advice I could possibly give without hesitating:

  • List your starting points: what do I already have, know and suppose
  • Create a list of criteria that define your goals or expectations.
    Donā€™t mix/confound requirements with something that might be part of a solution!
  • Check out possibilities/implementations/make/buy
  • Create a shortlist of implementations that more or less deliver what your criteria ask for
  • Calculate the cost and effort of getting and using the implementations
  • Which is the implementation that you can afford
  • Which is a lower cost implementation that checks almost the same number of boxes
  • Buy it/make it, use it
  • Never look backā€¦unless you want to go through the pain again!

Requirement or solution or else?

  • I need a Phase One CH solution!
  • I can spend USD 50ā€™000
  • I need 10 ft high prints that show no pixels when viewed from > 20in distance
  • I only do B&W
8

I think in this case the OP has done that, even laid out the actual choice of equipment that they are considering, a Fuji GFX100 medium format camera body with a Contax 120mm Macro against buying a 45MP Z8 Nikon body and using the lenses they have already been using on a (borrowed/hired?) Z9.

So they are looking for advice on whether there might be significant advantages with the change to 14 bit to 16 bit capture and the pixel shift capabilities of the GFX. I believe that the Z8 has pixel shift but the Z9 doesnā€™t, not sure why that would be or whether that has subsequently been added to the Z9 as a firmware update.

Itā€™s the word ā€˜significantā€™ which is of course open to subjective assessments.

9

Supplementary question - why do I have striped bobble hat (?) next to my name now?

10

Itā€™s a piece of cake with a cherry on top.
You can try to edit that in your forum account settings.

I get the impression that there was a change in the forum. The name field that I use to outline my gear (Canon and Apple) is simply not showing.

:man_shrugging:

11

Thanks, a piece of cake, of course, Iā€™m not sure which I prefer. Maybe I clicked on something somewhere, Iā€™ll have to look.

12

My experience is that the limiting factor is usually the film itself. There is a marginal but noticeable difference between a 12 MP camera scan of a 35mm negative and a 24 MP camera scan of the same negative. Scan again with a 50 MP camera, and the difference is negligible. With larger negatives, you do benefit from more MP (24 to 100).

I would feel perfectly comfortable using a Lumix GH7 Micro 4/3 camera in Sensor Shift mode for 100 MP files, if I needed to go really big (i.e.; a 60 inch by 40 inch print), with something that would be viewed critically from the diagonal dimension of the print. But if Iā€™m going to examine the same 60x40 print from 8 to 12 inches, that would make me want 120/220, 4x5, or even 8x10 film originals!

Would 16-bit files from great big sensels be better? In a side by side comparison, maybe. Is it worth a whole new setup? Would paying customers care? In my estimation, not very often. For images printed up to 11x17 inches or displayed on 4K monitors, Iā€™d stick with what you have. If youā€™re curating film collections for a museumā€™s digital archives, MAYBE you can justify it.

Note that the use cases for viewing large prints from such ridiculously close distances are few and far between! Aerial spy photography, photographing HUGE groups of people, and copying large format films to make big inkjet prints are a few of them.

13

I own a GFX 100 II, a Nikon Z8, and have tested both for scanning. I shoot mostly 6x7 120 negatives and havenā€™t come across a scenario where one looks ā€œbetterā€ than the other (when viewed at fit-to-screen zoom resolutions). When I crop to a 4:5 ratio there is a decent loss of resolution with best case is approx 33MP files after cropping. For situations where I might have wanted to do more extensive cropping then the GFX does win for the resolution. Also, I print very regularly at 16x20 on a Canon Pro 1000 and for that size the Z8 is fine for camera scans.

When shooting digital I do see a pretty major difference in those two specific cameras particularly with ETTR style metering and highlight recovery - the GFX is significantly better than the Z8.

1 Like
14

Excellent ā€˜on the moneyā€™ information, I get the impression that you havenā€™t felt the need to make use of pixel-shift on either camera for ā€˜camera scanningā€™?

15

Well it seems I had a forum anniversary on the 16th, the following day, maybe it had something to do with that. Anyway, itā€™s gone now.

16

Possibly a little thank you indeed. And less conspicuous than :birthday:

17

I shoot portraiture, documentary, and on-set photos for TV so pixel-shift isnā€™t necessary for my application. Perhaps if I were to do a massive Andreas Gursky size print, then I would consider it. As it is, I feel the GFX 100 II is probably overkill, but I already owned the camera so I figured ā€˜why not?ā€™. Over the top can be fun.

18

Quite soā€¦if you have a decent lens for macro too. Whatā€™s the one you (plan to) use?

If Iā€™d be in your position, Iā€™d try to test this lens.

19

That looks cool. Iā€™m using an adapted Pentax 645 120mm macro.

20

Sure, I like the sound of that kind of ā€˜over the topā€™. I was thinking in terms of the OP doing camera scanning with a Z8 should they choose to go for that over the GFX, itā€™s possible that when they needed more resolution, for cropped MF perhaps, they could use pixel-shift. Thatā€™s not something youā€™ve needed to do of course.