The Fuji GFX50s for Gearheads

[Frank Fremerey]

Shocking news for me. I was so sure Fuji would cut it but file quality below D810 standard is a killer argument. The wait for the D820 or whatever they might call it gets more urgent. I even contemplate to add a used D810 to my arsenal soon. Thank you, Michael. Big disappointment.

I guess the results would be better if Fuji would have sticked to what the know and can do. XTrans with superb JPEG and genious RAF.

[Michael Erlewine]

Shocking news for me. I was so sure Fuji would cut it but file quality below D810 standard is a killer argument. The wait for the D820 or whatever they might call it gets more urgent. I even contemplate to add a used D810 to my arsenal soon. Thank you, Michael. Big disappointment.

I guess the results would be better if Fuji would have sticked to what the know and can do. XTrans with superb JPEG and genious RAF.

That is just my experience. For all I know, perhaps I did something wrong in the post-processing, but I was carefully to set all of the raw file selections. Others may find a better experience, but the only reason I use raw is for the ability to recover from my various mistakes in lighting, etc. And I always expose to the left, not the right, so not being able to get the range in the blacks affected me a lot. I need that range for recovery, etc.

[simsurace]

And I always expose to the left, not the right, so not being able to get the range in the blacks affected me a lot. I need that range for recovery, etc.

Just a general remark: exposing to the left is not even a well-defined concept. What do you mean exactly? And why do you do it?
Anyway, if you have a choice, intentionally underexposing is a terrible idea. There is nothing at all to be gained from it. On the contrary: you are guaranteed to get the performance of a sensor that is several generations older (depending on how much you underexpose).
Please do yourself a favor and learn to expose to the right (it's not that hard). You are otherwise wasting a big chunk of the potential of whatever gear you are using.

As for being able to pull up shadows, a lot depends on how the controls are applied to the RAW data, which differs between RAW converters. You have to examine the RAW data itself e.g. in RawDigger or a similar program in order to see what's going on (e.g. whether there is a high noise floor due to read noise and thermal noise). It may well be that the next update for the RAW converter gives better control. But this has nothing to do with the sensor hardware.
Maybe you remember the day when Adobe released the 2012 process version for ACR: it suddenly became much easier to pull out detail even from older files.

[Michael Erlewine]

Just a general remark: exposing to the left is not even a well-defined concept. What do you mean exactly? And why do you do it?
Anyway, if you have a choice, intentionally underexposing is a terrible idea. There is nothing at all to be gained from it. On the contrary: you are guaranteed to get the performance of a sensor that is several generations older (depending on how much you underexpose).
Please do yourself a favor and learn to expose to the right (it's not that hard). You are otherwise wasting a big chunk of the potential of whatever gear you are using.

As for being able to pull up shadows, a lot depends on how the controls are applied to the RAW data, which differs between RAW converters. You have to examine the RAW data itself e.g. in RawDigger or a similar program in order to see what's going on (e.g. whether there is a high noise floor due to read noise and thermal noise). It may well be that the next update for the RAW converter gives better control. But this has nothing to do with the sensor hardware.
Maybe you remember the day when Adobe released the 2012 process version for ACR: it suddenly became much easier to pull out detail even from older files.

Of course I know how to expose to the right. I find that I like my results better when i don't expose too much to the right, but rather make sure my shadows (black) is ample. I should not have shared that with some of you folks, but in truth the fun in photography for me comes with the lowest legitimate ISO and the ability to have plenty of play in the blacks. You know what you like; so do I. 

[Ilkka Nissilä]

It is not so surprising; most cameras don't match the D810's base ISO dynamic range but I guess with a slightly larger sensor it should have been in the same ball park for the Fuji. However, from ISO 400 up it should be a really big surprise if the Fuji were not ahead. Perhaps soon there will be some formal tests of the raw data which can distinguish between the quality of the raw data and differences in processing. Since it's a Bayer sensor more sites can analyse it than with X-Trans.

[Michael Erlewine]

It is not so surprising; most cameras don't get close to the D810's base ISO dynamic range. However, from ISO 400 up it should be a really big surprise if the Fuji were not ahead. Perhaps soon there will be some formal tests of the raw data which can distinguish between the quality of the raw data and differences in processing. Since it's a Bayer sensor more sites can analyse it than with X-Trans.

In my case, I (seldom to) never use that high an ISO; I am a bottom feeder when it comes to dynamic range.

[simsurace]

Of course I know how to expose to the right. I find that I like my results better when i don't expose too much to the right, but rather make sure my shadows (black) is ample. I should not have shared that with some of you folks, but in truth the fun in photography for me comes with the lowest legitimate ISO and the ability to have plenty of play in the blacks. You know what you like; so do I. 

You can always dial back the exposure in the RAW converter. But if you start out with the brightest part of the image 2 stops below the clipping point, you are giving up 2 stops of dynamic range, and they will never come back.
Of course exposing to the right does not mean bumping the ISO, it means giving more exposure.

[simsurace]

It is not so surprising; most cameras don't match the D810's base ISO dynamic range but I guess with a slightly larger sensor it should have been in the same ball park for the Fuji. However, from ISO 400 up it should be a really big surprise if the Fuji were not ahead. Perhaps soon there will be some formal tests of the raw data which can distinguish between the quality of the raw data and differences in processing. Since it's a Bayer sensor more sites can analyse it than with X-Trans.

Shouldn't the Fuji have roughly 2/3 of a stop advantage based on sensor size alone? Of course performance can be worse than what is predicted from the increase in sensor size if for some reason there are other noise sources. But to me it would be quite a big surprise and a serious failure in design if the Fuji came out worse than the D810 at the same ISO. I'm also looking forward to some more 'scientific' tests.
On the other hand, this same argument would predict that we are unlikely to see a one stop advantage or more in dynamic range. Thus any dynamic range advantage over the D810 can only be exploited by using optimal shooting parameters.

[Ilkka Nissilä]

The D810 has ISO 64 as base ISO, the Fuji has 100.

[Jack Dahlgren]

Sad to hear that the Fuji didn't scratch your itch. I guess it shows that your current equipment is really quite good.

[Erik Lund]

The problem for the GFX sensor/in-camera RAW file processing, seems to be that it is sensitive or unforgiving with regards to highlights, then ETTR is flawed/hindered since your only option is to turn down the exposure to avoid blooming etc,,, Sort of how it was in the beginning with Nikon D1

The sensor in the D810 seems to be extremely forgiving with regards to blown highlights,,, recovery in PP is mind boggling - To me it seems like Nikon has stumbled upon a sensor and in-camera RAW firmware processing Fine line that is extremely capable even though its a 36MP FF sensor.

We can hope Fuji is able to tweak the firmware on the GFX, but so far it doesn't look promising,,, all test I have seen has had these blocked shadow/dark areas more or less,,,

Sony has similar Fine line performance for the A7R II,,, really an impressive camera!

[Michael Erlewine]

The problem for the GFX sensor/in-camera RAW file processing, seems to be that it is sensitive or unforgiving with regards to highlights, then ETTR is flawed/hindered since your only option is to turn down the exposure to avoid blooming etc,,, Sort of how it was in the beginning with Nikon D1

The sensor in the D810 seems to be extremely forgiving with regards to blown highlights,,, recovery in PP is mind boggling - To me it seems like Nikon has stumbled upon a sensor and in-camera RAW firmware processing Fine line that is extremely capable even though its a 36MP FF sensor.

We can hope Fuji is able to tweak the firmware on the GFX, but so far it doesn't look promising,,, all test I have seen has had these blocked shadow/dark areas more or less,,,

Sony has similar Fine line performance for the A7R II,,, really an impressive camera!

My understanding is that the Nikon D810 base ISO was a specially-designed hardware implementation in order to feature that particular ISO 64, and it works really well. I found it hard to be without that kind of quality, which I did not (to the best of my ability) find in the GFX. Nikon does a lot of things right. They just are behind the curve (for me) in updating their high-end landscape DSLR.

[simsurace]

The D810 has ISO 64 as base ISO, the Fuji has 100.

Thanks!

[Erik Lund]

Not to be pessimistic but sometimes new is not always better in all regards/ on all specs,,,

This is clearly demonstrated with the D5 on Dynamic range, where the D4s is clearly and by far superior at the lower ISO values - Yes I know its PJ action camera but still,,,

So a new D820 might not even be better dynamic range or what ever, than the D810 at base ISO 64,,,

BTW I also by far prefer to capture my images at ISO 64 disregarding  ETTR or ETTL,,,

[Michael Erlewine]

Not to be pessimistic but sometimes new is not always better in all regards/ on all specs,,,

This is clearly demonstrated with the D5 on Dynamic range, where the D4s is clearly and by far superior at the lower ISO values - Yes I know its PJ action camera but still,,,

So a new D820 might not even be better dynamic range or what ever, than the D810 at base ISO 64,,,

BTW I also by far prefer to capture my images at ISO 64 disregarding  ETTR or ETTL,,,

Well, if the D810 update gives up the ISO 64, I will just buy a second D810 or look for whatever next mirrorless medium-format  beast comes down the pike.