Fujifilm GFX100 Pixelshift Firmware is out

[drevil]

FUJIFILM GFX100

Download the latest firmware for FUJIFILM GFX100 from here.

fujifilm-x.com

Here are my first impressions


AWESOME but...

absolute steadiness is required, you easily get artifacts.

we have 2020, and the software to combine the images is using the CPU to process? COME ON FUJI! i didnt get a amd xt 5700 rx just for gaming

it would be nice to put the 16 images, each into a new "ps" folder, to keep the images apart.

also even if its time consuming, give us the option to have the camera combine the images. but yeah, the resulting 1,6gb files are crazy

and phase one needs to optimize C1, its taking crazily long to load the images at 100%, photoshop is nearly instantly loaded

 

[Felipe]

Does the Pixel Shift work with flash?

 

[drevil]

yes, there is a delay mode, which gives a flash enough time to recycle between shots.
thats what the manual says

 

[gerald.d]

My understanding is pixel shift uses the electronic shutter. You will be severely limited regarding sync speed.

Kind regards,


Gerald.

 

[Jared]

I know people were looking forward to this. Why? This is not just baiting—I want to know what one hopes to get. Reduced aliasing? More resolution? For what intent? Future proofing for 12K monitors? Cropping? Wall sized landscape prints?

It’s certainly cool, but what would one actually use it for? I know, one can never have too much horsepower or too many megapixels, but beyond bragging rights (with either) is there a practical use? I assume it is nearly impossible to get benefits outdoors since the scene must be absolutely static. I have the same feature on my Leica SL2 (with half the megapixel count, of course) and have only employed it in test shots. But maybe it’s a “Me” issue and the feature has more benefits than I’m aware of.

 

[rdeloe]

I know people were looking forward to this. Why? This is not just baiting—I want to know what one hopes to get. Reduced aliasing? More resolution? For what intent? Future proofing for 12K monitors? Cropping? Wall sized landscape prints?

It’s certainly cool, but what would one actually use it for? I know, one can never have too much horsepower or too many megapixels, but beyond bragging rights (with either) is there a practical use? I assume it is nearly impossible to get benefits outdoors since the scene must be absolutely static. I have the same feature on my Leica SL2 (with half the megapixel count, of course) and have only employed it in test shots. But maybe it’s a “Me” issue and the feature has more benefits than I’m aware of.

The biggest benefit will likely be for people who need to reduce aliasing. I could see it being useful for people doing repro work in controlled conditions.

...have only employed it in test shots.

That gave me a chuckle. I spent some time figuring out the workflow for flat stitching to make files in 4:3 aspect ratio with much greater resolution than I can get in one shot (so not for panoramas). I've used the technique exactly once in an actual non-testing setting, and then I didn't end up using the resulting file. Some new developments are a solution looking for a problem.

 

[gerald.d]

I know people were looking forward to this. Why? This is not just baiting—I want to know what one hopes to get. Reduced aliasing? More resolution? For what intent? Future proofing for 12K monitors? Cropping? Wall sized landscape prints?

It’s certainly cool, but what would one actually use it for? I know, one can never have too much horsepower or too many megapixels, but beyond bragging rights (with either) is there a practical use? I assume it is nearly impossible to get benefits outdoors since the scene must be absolutely static. I have the same feature on my Leica SL2 (with half the megapixel count, of course) and have only employed it in test shots. But maybe it’s a “Me” issue and the feature has more benefits than I’m aware of.

I always find this kind of response a little odd.

You ask, "why?", and then without even breaking into a sweat list a number of perfectly valid reasons "why" (and there are of course plenty more).

Out of interest, at what point in the last couple of decades did the resolution of cameras reach your personal nadir?

Was it when digital sensors with 5MP resolution became available? 10? 24? 50? 100?

Perhaps if you could take the time to explain why you only require "x" resolution, then people who are more than happy to take advantage of more will be willing to go into great depth as to why it benefits them.

(I've created images as large as 45 gigapixels in the past, so have quite a bit of insight on the subject.)


Kind regards,


Gerald.

 

[Jared]

Wasn’t suggesting that it was valueless. Was wondering which of the possible values people were looking for? And what the use case was? I know my use cases do not match everyone else’s.

I can absolutely agree that it provides flexibility for cropping, that it will reduce aliasing, and that the additional megapixels will have more info. I wanted to know how those possible benefits will be used by those who own the camera. I hadn’t thought of reproduction work—that’s certainly a situation where it could be helpful.

As to my own requirements... Aside from cropping, I stopped seeing benefits beyond 24 megapixels. And I hadn’t yet found a use in my own photography for pixel shift, but I mostly shoot outdoors where wind is a limitation. That’s just me, though.

 

[MGrayson]

Product photography. Possibly architecture with a lot of final image cleanup. But Topaz Gigapixel does a great job. I tried the multi-shot just now. Doubt I will use it again, as I don't shoot perfectly static scenes. As for how many MP is enough? I have yet to find an instance where the S(007)'s 37.5MP wasn't enough. I have never shot a picture that was "better" because of the 100MP of the GFX. IBIS? Yes. AF? Yes. But MP? No.

 

[Jared]

(I've created images as large as 45 gigapixels in the past, so have quite a bit of insight on the subject.)


Kind regards,


Gerald.

Very cool. What does one do with a 45 GP image (or any very high pixel count image)? And how does one make one? Panoramas stitched together or using pixel shift or a combination?

I’m not asking anyone to justify their needs or desires—just explain them because I’m curious.

 

[Paul2660]

FUJIFILM GFX100

Download the latest firmware for FUJIFILM GFX100 from here.

fujifilm-x.com

Here are my first impressions


AWESOME but...

absolute steadiness is required, you easily get artifacts.

we have 2020, and the software to combine the images is using the CPU to process? COME ON FUJI! i didnt get a amd xt 5700 rx just for gaming

it would be nice to put the 16 images, each into a new "ps" folder, to keep the images apart.

also even if its time consuming, give us the option to have the camera combine the images. but yeah, the resulting 1,6gb files are crazy

and phase one needs to optimize C1, its taking crazily long to load the images at 100%, photoshop is nearly instantly loaded

Thanks for the info. Can you share a bit more info on a few questions?

From reading the info, it appears the 16 raws are just stored on the card? not in a separate folder on the card, so can it become hard to figure out which images apply to a particular shift or do they have a specific numbering?

Once shot, the images have to be loaded into the Fuji software "Fujifilm combiner software", where I guess the 16 images are combined into 1 large DNG.

Does C1 see the dng (C1 20) or 21 see the dng with a fuji color profile? Or just a raw dng profile like how a DJI drone file is loaded into C1 (which required picking a color profile)

How to Adobe LR or CC/ACR handle the same dng?

Lastly, I assume that subject motion (wind) would be very problematic? Just curious as with the old Pentax K1 subject motion caused issues but was very dependent on the software used to open the file. Adobe was terrible, Silky Pix much better, C1 never worked on them.

Thanks

Paul C

 

[Steve Hendrix]

Very cool. What does one do with a 45 GP image (or any very high pixel count image)? And how does one make one? Panoramas stitched together or using pixel shift or a combination?

I’m not asking anyone to justify their needs or desires—just explain them because I’m curious.

There are many other types of photography than the types most commonly discussed on this forum. But certainly, since the very early days of digital, back when we had 4mp and 6mp multi shot capture solutions, multi shot has been utiized to provide superior color, reduce aliasing, moire and other artifacts, and maximize resolution. They have traditionally been used for studio photography - especially in the early days. Those sub 10mp single shot captures seemed impressive at the time, but had glaring shortcomings - you didn't dare photograph textiles, unless they already had some sort of rainbow pattern that the moire would maybe just happen to perfectly blend in with.

With advanced resolution and improved color science across the board, most (but not all) product photography is produced via single shot capture today. There are exceptions, however. And cultural/historical reproductions are one such important and noteworthy category, but there are others. You don't realy want aliasing or color artifacting when trying to reproduce tiny text from a historical document that is centuries old.

Medium Format Multi Shot solutions from Eyelike, Sinar, Imacon/Hasselblad, Leaf (early Leaf) have supported advanced photography for decades. For our part, we are happy to see that Hassleblad and Fuji are continuing with this technology.


Steve Hendrix/CI

 

[gerald.d]

As to my own requirements... Aside from cropping, I stopped seeing benefits beyond 24 megapixels. And I hadn’t yet found a use in my own photography for pixel shift, but I mostly shoot outdoors where wind is a limitation. That’s just me, though.

Interesting.

So what benefits did the SL2 bring you that made it worthwhile buying over the SL?

Clearly it can't be down to the increase in resolution.

Presumably there is a considerable price differential between the two cameras. How do you rationalize that capital outlay?

(The subject of this thread is a free firmware upgrade - absolutely no need to do a cost/benefit analysis to justify it.)

Kind regards,


Gerald.

 

[Paul2660]

After giving this some thoughts, here is a bit of feedback.

First and foremost, I do have to bow my head to Phase One, and be a bit less negative towards them on the IQ4, as when you compare what the IQ4 can process internally, with dual exposure and frame averaging all being done on the back. Frame averaging can easily involve a lot more then 16 shots and as I understand it all the shots are full frame exposures taken with the ES and processed in the back into one raw file.

Back to Fuji.

After taking this long to bring the 400MP output to out, and I do congratulate them on this, it surprises me that more options were not brought forward. If it takes 16 frames to get to 400MP, then is it far to to consider:

4 frames would give you "true" color at 100MP, i.e. no color arising, and thus in theory a better overall image as no interpolation for color would be done.
8 frames would give you "true" color and 200MP
12 frames would give you "true" color and 300MP.

The first two options make a lot of sense to me and should not have been an problem at all to add to the firmware unless the math won't work. I don't believe any movement can be tolerated (wind in tree limbs, or water moving) so really using this outdoors will be very limited. I would love to have seen the 8 frame option and 4 frame option if those were possible. These lesser number of total frame sequences might have a bit more luck with slight motion.

Also, from what I can tell, the images in a pixel shift are just added to the card? Not stored in a separate folder on the card. That surprised me. Do the files have the same .RAF extension and if they are just numbered in sequence with normal shots, then what a mess trying to figure out which 16 image apply to a pixel shift sequence. You almost would have to take a lens cap dark frame to create a blank frame to let you know where to start and stop.

Hopefully Fuji will add some features in the future, if possible giving a 4 and 8 shot combination.

Paul C

 

[MGrayson]

Frame averaging is about the easiest thing to do in the back - it's two multiplications and one addition per pixel per exposure. No storing multiple exposures - you just need a one frame buffer. I'd MUCH rather have had that added than the multi shot, but it wasn't my call.

 

[JK12]

After giving this some thoughts, here is a bit of feedback.

First and foremost, I do have to bow my head to Phase One, and be a bit less negative towards them on the IQ4, as when you compare what the IQ4 can process internally, with dual exposure and frame averaging all being done on the back. Frame averaging can easily involve a lot more then 16 shots and as I understand it all the shots are full frame exposures taken with the ES and processed in the back into one raw file.

Back to Fuji.

After taking this long to bring the 400MP output to out, and I do congratulate them on this, it surprises me that more options were not brought forward. If it takes 16 frames to get to 400MP, then is it far to to consider:

4 frames would give you "true" color at 100MP, i.e. no color arising, and thus in theory a better overall image as no interpolation for color would be done.
8 frames would give you "true" color and 200MP
12 frames would give you "true" color and 300MP.

The first two options make a lot of sense to me and should not have been an problem at all to add to the firmware unless the math won't work. I don't believe any movement can be tolerated (wind in tree limbs, or water moving) so really using this outdoors will be very limited. I would love to have seen the 8 frame option and 4 frame option if those were possible. These lesser number of total frame sequences might have a bit more luck with slight motion.

Also, from what I can tell, the images in a pixel shift are just added to the card? Not stored in a separate folder on the card. That surprised me. Do the files have the same .RAF extension and if they are just numbered in sequence with normal shots, then what a mess trying to figure out which 16 image apply to a pixel shift sequence. You almost would have to take a lens cap dark frame to create a blank frame to let you know where to start and stop.

Hopefully Fuji will add some features in the future, if possible giving a 4 and 8 shot combination.

Paul C

I'm with Paul. Although this is a welcomed update, it would be just as good if not better to have offered a 4 shot and an 8 shot mode. Hopefully Fuji will do this.

 

[gerald.d]

After taking this long to bring the 400MP output to out, and I do congratulate them on this, it surprises me that more options were not brought forward. If it takes 16 frames to get to 400MP, then is it far to to consider:

4 frames would give you "true" color at 100MP, i.e. no color arising, and thus in theory a better overall image as no interpolation for color would be done.
8 frames would give you "true" color and 200MP
12 frames would give you "true" color and 300MP.

The first two options make a lot of sense to me and should not have been an problem at all to add to the firmware unless the math won't work. I don't believe any movement can be tolerated (wind in tree limbs, or water moving) so really using this outdoors will be very limited. I would love to have seen the 8 frame option and 4 frame option if those were possible. These lesser number of total frame sequences might have a bit more luck with slight motion.

Also, from what I can tell, the images in a pixel shift are just added to the card? Not stored in a separate folder on the card. That surprised me. Do the files have the same .RAF extension and if they are just numbered in sequence with normal shots, then what a mess trying to figure out which 16 image apply to a pixel shift sequence. You almost would have to take a lens cap dark frame to create a blank frame to let you know where to start and stop.

Hopefully Fuji will add some features in the future, if possible giving a 4 and 8 shot combination.

Paul C

All of the examples you list as desirable are of course possible to generate from the 16-shot files.

Realistically, I would imagine that almost everyone who is going to be taking advantage of this new feature will be using it whilst tethered. When shooting tethered using the supplied Pixel Shift Combiner software, the process of creating the multi-shot file is completely automatic. You can see the process in the following video -

Kind regards,


Gerald.

 

[Jared]

There are many other types of photography than the types most commonly discussed on this forum. But certainly, since the very early days of digital, back when we had 4mp and 6mp multi shot capture solutions, multi shot has been utiized to provide superior color, reduce aliasing, moire and other artifacts, and maximize resolution. They have traditionally been used for studio photography - especially in the early days. Those sub 10mp single shot captures seemed impressive at the time, but had glaring shortcomings - you didn't dare photograph textiles, unless they already had some sort of rainbow pattern that the moire would maybe just happen to perfectly blend in with.

With advanced resolution and improved color science across the board, most (but not all) product photography is produced via single shot capture today. There are exceptions, however. And cultural/historical reproductions are one such important and noteworthy category, but there are others. You don't realy want aliasing or color artifacting when trying to reproduce tiny text from a historical document that is centuries old.

Medium Format Multi Shot solutions from Eyelike, Sinar, Imacon/Hasselblad, Leaf (early Leaf) have supported advanced photography for decades. For our part, we are happy to see that Hassleblad and Fuji are continuing with this technology.


Steve Hendrix/CI

Yes, the repro requirement makes sense to me. Rob mentioned that as well. The improvement in aliasing, in particular, would be of value. Product photography surprises me a bit, but I don't have clients in that area so I'm certainly not an expert.

 

[Jared]

Interesting.

So what benefits did the SL2 bring you that made it worthwhile buying over the SL?

Clearly it can't be down to the increase in resolution.

Presumably there is a considerable price differential between the two cameras. How do you rationalize that capital outlay?

(The subject of this thread is a free firmware upgrade - absolutely no need to do a cost/benefit analysis to justify it.)

Kind regards,


Gerald.

The major benefits to the SL2 for me were (in no particular order):
- IBIS
- Ability to turn off long exposure noise reduction (I actually use it for wide field astrophotography)
- Improved viewfinder
- Better UI

Others may have been after the higher megapixel count. I wasn't. As to justifying the cost, I can't. I don't expect it recoup its cost. I do a little professional work, but not enough to cover my photography expenses as a whole.

Gerald, I really wasn't trying to attack anyone with my question. I realize people get pretty defensive on photography forums because there is so much vitriol and so many ad hominem attacks. You are right that it's a free firmware upgrade, and I applaud Fuji (and quite a few companies these days) for continuing to improve their cameras well after initial release. That's great. I was just wondering how people feel this new feature will improve their photos. No need for a "justification," just an explanation of how it will be helpful. I just couldn't think of a practical application. But my photography requirements are not everyone else's. For example, I have been wanting Leica to remove the long exposure noise reduction requirement for years. For most people this would be a totally trivial matter that affects none of their photos. For me? This is the difference between a camera that is suited to my photography and one that is not.

 

[mristuccia]

By the way, the 100% crop of the example shown on DPReview doesn't look that great...

https://www.dpreview.com/news/65224...-its-gfx-100-camera-with-3-00-firmware-update

it seems out of focus, maybe back-focused.