Clarifying a few misconceptions with medium format vs full-frame

Autoxave · Sep 11, 2023

I believe I have a few misconceptions about medium format vs full frame. Hoping to clarify this through the statements below. Are the statements below true or not true? Would be nice to have an explanation if the statements are not true.

Medium format images produces less noise in low light environments compared to fullframe presuming that available light is exactly the same and that the lenses are equivalent
Medium format images produces less depth of field, i.e. a better separation of the elements in an image, presuming that the field of view and distance to the element of interest is the same and lenses are equivalent
Medium format images produce less noise in post processing when increasing exposure in shadow areas of the image compared to fullframe presuming that available light was exactly the same when the image was captured, field of view is the same and that the lenses are equivalent
Medium format images produces less artifacts in post processing when sharpening an image presuming that the exposure and field of view is the same
Medium format cameras are superior (in terms of better image quality, less noise, more leeway in post processing) to fullframe when capturing long exposure seascape images presuming that available light is exactly the same and that the lenses are equivalent

areallygrumpyoldsod · Sep 11, 2023

I ill hope that others can answer your questions directly - but am happy to make the following statements. Which no doubt will be corrected.

The gap between the image quality one can obtain from the best full-frame sensors (be this the D850 or Z8/Z9 or A7RV/IV or Leica SL2) and equivalent resolution Medium Format Sensors (X1Dii or other bodies using the 50C back) has narrowed considerably since these new sensors use Back-Side Illumination. Stacked sensors also allow far far higher speed sensor reading.
Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least. Some FF bodies still only allow 12-bit so let us step past these.
Usually a 16-bit camera system will have higher dynamic range, lower noise, and better smooth gradients compared to a camera with lower bit depth, but that is not guaranteed to be the case.
A 12-bit RAW file can contain 4096 shades of red, green, and blue, contributing to a potential 68 billion colors. Raise the bar to 16-bit, and you're looking at 65,536 shades of red, green, and blue, and a potential 281 trillion colors.
The Sensor Size and larger throat size of Medium Format cameras (and their lenses) changes the geometric performance of lenses and DOF relative to similar focal length and aperture lenses in full frame.
1. Simply put the X2D-100C (with its 44x33mm Small Medium Format sensor) is a 0.787x Crop Factor -vs FF (36x24mm) - as a result the XCD 4/21 (which actually has a 21.8mm focal length) is equivalent to a 17.1mm Full Frame equivalent.
2. AND the 150mp Phase One IQ4 sensor is at 53.4 x 40mm provides a 0.634 crop factor
  The next point is more contentious -- an XCD 4/21 lens performs equivalent to a 3.2/17.1 lens on a full-frame sensor.
The depth of field of say an 85mm focal length lens on either a full-frame or Small Medium Format Sensor for the same aperture (f/1.9) and subject distance (9.84ft) calculated using DOFsimulator is as follows:
1. FF - 5.3 inches
2. SMF MF - 6.7 inches -- this large DOF is a result of an 85mm FL being equivalent to a 67mm focal length lens on a full-frame sensor
3. A 108mm focal length lens on a SMF sized sensor would be required to generate the same FOV BUT the DOF is reduced to 4.1 inches
4. The Phase One 110mm Blue Ring lens has a 35mm equ focal length of 68mm - a lens somewhere between the 120 macro LS f/4.0 (73mm eq) and 150mm LS f/3.5 (93mm eq) would be needed to replicate an 85mm on FF.
It would be hard to argue that any MF lens is built better or performs than say the best Leica APO lens, but as a rule the best MF lenses tend to be higher quality than most FF lenses.
I particularly refer to the best Schneider Kreuznach, Rodenstock and the original Hasselblad XCD lenses. I understand there is more variability in the newer V/P versions of the XCD lenses and in some of the Fuji GFX lenses.
So to are there quality variability in the Nikkor Z lenses -- with most S-line lenses being significantly superior to their F-mount predecessors; similarly Sony GM lenses.
Ignoring Leica lenses, most MF and SMF lenses are very expensive when compared to lenses for FF systems.
Larger sensors of similar resolution and design will provide better low light performance than a smaller sensor at the same ISO setting - while the answer probably is fair to say correct the reason is we wee bit more complex. Others like JimK and Thom Hogan normally weigh in here. There is HOWEVER, one vast difference that can make a huge difference. AND that is Auto-Focus. The X series Hasselblad do not currently have continuous AF nor do they have eye tracking. While i cannot speak about how well the GFX100/100s perform I can most certainly state that for rapidly moving subjects cameras like the Z8/Z9 and Sony A7RV will deliver vastly superior results when shooting with wide apertures.
Flash sync speeds -- Medium Format camera that use lenses with leaf shutters can work are vastly quicker flash sync speeds (1/2000th or even 1/4000) than those systems that use either electronic or focal plane shutters (1/160-1/200 and possibly 1/400th)
In camera tools -- here one has to look at what is possible with a hugely expensive system like the Phase One IQ4 with its Frame Averaging and 2 shot (+3EV) in body solutions. We are yet to see what of these tools will flow down to in camera solutions in the Fuji or Hasselblad bodies - there is some movement in a few FF bodies - most notably HiRes mode but this is also seen in the GFX bodies.

In summary -- 16-bit provides a vast step up in potential image quality over what is possible with FF bodies. SMF/MF systems are far slower to use, typically have relatively low tech approaches to AF (so struggle with AF tracking) and since these bodies have high resolutions the sensor read time is LONG, which can lead to staggering rolling shutter issues - but less so when a mechanical shutter is used.

NAwlins Contrarian · Sep 11, 2023

Autoxave said:
I believe I have a few misconceptions about medium format vs full frame. Hoping to clarify this through the statements below. Are the statements below true or not true? Would be nice to have an explanation if the statements are not true.

Medium format images produces less noise in low light environments compared to fullframe presuming that available light is exactly the same and that the lenses are equivalent

False, if by low light we mean that the exposure conditions are such that shutter speed is a constraint (moving subject and/or shaking camera). Medium format only has a noise advantage when there's plenty of light.

To be clear, "the lenses are equivalent" is a key point. E.g., if you shoot a GFX 63mm at f/2.8, then the FF equivalent is a 50mm at f/2.2 (or thereabouts, depending on the final output aspect ratio). Failing to observe / maintain that equivalence results in an apples-to-oranges comparison.

Autoxave said:
Medium format images produces less depth of field, i.e. a better separation of the elements in an image, presuming that the field of view and distance to the element of interest is the same and lenses are equivalent

False, as long as equivalent apertures are used. In fact, because FF generally has substantially faster lenses available, generally the lowest achievable depth of field is with FF, not medium format.

Autoxave said:
Medium format images produce less noise in post processing when increasing exposure in shadow areas of the image compared to fullframe presuming that available light was exactly the same when the image was captured, field of view is the same and that the lenses are equivalent

False, as stated above.

Autoxave said:
Medium format images produces less artifacts in post processing when sharpening an image presuming that the exposure and field of view is the same

Maybe / probably, given that medium format typically has a resolution advantage, and often has lenses that are on the whole better. Also, if (but only if) there's plenty of light, then medium format typically has a noise and dynamic range advantage.

Autoxave said:
Medium format cameras are superior (in terms of better image quality, less noise, more leeway in post processing) to fullframe when capturing long exposure seascape images presuming that available light is exactly the same and that the lenses are equivalent

I think this is a can of worms, and long exposures may be something of a special case. I really don't know, but I suspect this is an 'it depends' situation.

Medium format shines where there's plenty of light, such as where (1) the outdoor subject is static and you can put the camera on a sturdy tripod, set the desired aperture, and leave the shutter open as long as necessary to achieve ETTR exposure at base gain (ISO 100 or whatever) and (2) the studio exposure is produced by strobes with plenty of power for ETTR exposure at base gain and whatever aperture is desired.

JimKasson · Sep 11, 2023

areallygrumpyoldsod said:
Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

GFX 100S shadow performance with 14 and 16 bit precision - the last word

This is the second in a series of posts about the Fujifilm GFX 100S. If you look on the right side of this page, scroll down to the Categories section, and select GFX 100S from the dropdown list, you can see all the reviews. When I tested the GFX 100, I found the 16-bit mode

blog.kasson.com

Rich42 · Sep 11, 2023

The issue regarding “vast improvement in image quality” due to 16 bit files from MF equipment vs 14 bit files from FF is another can of worms.

While a file may be in 16 bit format, the information contained is usually limited to 14 bits or slightly more depending on the sampling used. But I don’t believe any MF camera provides true 16 bit depth information.

While I (think) I enjoy increased image quality (sometimes) comparing my MF images to my FF images (on a very good day, when the light is just right and I’ve taken better care to achieve the shot with my 100S than I have with my D850), it has nothing to do with “16 bit files” vs “14 bit files.”

JimKasson · Sep 11, 2023

Autoxave said:
I believe I have a few misconceptions about medium format vs full frame. Hoping to clarify this through the statements below. Are the statements below true or not true? Would be nice to have an explanation if the statements are not true.

I'm going to assume the same pixel pitch and pixel design and same sized prints in my answers.

Autoxave said:
Medium format images produces less noise in low light environments compared to fullframe presuming that available light is exactly the same and that the lenses are equivalent

If you're shutter speed limited, the visible noise in equivalent FF and MF images is the same.

Autoxave said:
Medium format images produces less depth of field, i.e. a better separation of the elements in an image, presuming that the field of view and distance to the element of interest is the same and lenses are equivalent

At equivalent settings and lenses, the DOF observed in a same size print is the same.

Autoxave said:
Medium format images produce less noise in post processing when increasing exposure in shadow areas of the image compared to fullframe presuming that available light was exactly the same when the image was captured, field of view is the same and that the lenses are equivalent

Nonlinear noise reduction is more effective with more pixels, which is what you get with MF vs FF at same pixel pitch.

Autoxave said:
Medium format images produces less artifacts in post processing when sharpening an image presuming that the exposure and field of view is the same

Sharpening is more effective with more pixels, which is what you get with MF vs FF at same pixel pitch.

Autoxave said:
Medium format cameras are superior (in terms of better image quality, less noise, more leeway in post processing) to fullframe when capturing long exposure seascape images presuming that available light is exactly the same and that the lenses are equivalent

Same shutter speed, same ND filter? No. But if you're using a less dense ND filter to get the same shutter speed, yes.

JimKasson · Sep 11, 2023

Rich42 said:
The issue regarding “vast improvement in image quality” due to 16 bit files from MF equipment vs 14 bit files from FF is another can of worms.

While a file may be in 16 bit format, the information contained is usually limited to 14 bits or slightly more depending on the sampling used. But I don’t believe any MF camera provides true 16 bit depth information.

Both the GFX 100x and H2D provide, in 16 bit mode, a true 16-bit ADC. But the read noise is about 2 LSBs.

Rich42 said:
While I (think) I enjoy increased image quality (sometimes) comparing my MF images to my FF images (on a very good day, when the light is just right and I’ve taken better care to achieve the shot with my 100S than I have with my D850), it has nothing to do with “16 bit files” vs “14 bit files.”

Yep.

Erik Kaffehr · Sep 11, 2023

Autoxave said:
I believe I have a few misconceptions about medium format vs full frame. Hoping to clarify this through the statements below. Are the statements below true or not true? Would be nice to have an explanation if the statements are not true.

Medium format images produces less noise in low light environments compared to fullframe presuming that available light is exactly the same and that the lenses are equivalent

That would presume same aperture and shutter time. Equivalent aperture would need a longer shutter time.

Autoxave said:
Medium format images produces less depth of field, i.e. a better separation of the elements in an image, presuming that the field of view and distance to the element of interest is the same and lenses are equivalent.

Not when using equivalent aperture.

Autoxave said:
Medium format images produce less noise in post processing when increasing exposure in shadow areas of the image compared to fullframe presuming that available light was exactly the same when the image was captured, field of view is the same and that the lenses are equivalent

No, it doesn't apply for equivalent aperture, but would apply if exposure would be kept the same.

Autoxave said:
Medium format images produces less artifacts in post processing when sharpening an image presuming that the exposure and field of view is the same

That depends mostly on the pixel pitch, pixel aperture and presence of OLP filter. As an example, my P45+ back that has 6.8 micron pitch, around 0.75 in linear fill factor, is an aliasing machine.

Autoxave said:
Medium format cameras are superior (in terms of better image quality, less noise, more leeway in post processing) to fullframe when capturing long exposure seascape images presuming that available light is exactly the same and that the lenses are equivalent

Not really necessarily. With same exposure, like f/8 at f/125, the larger sensor may have an advantage, assuming that quantum efficiency is similar.

Assuming that the basic sensor technology is similar, and that the capabilities of the sensor can be fully utilized, a larger sensor will have a benefit.

If we have good light or shoot on tripod, the larger format will have an advantage. But, doubling sensor size will just improve SNR by 41%.

I would think that the most obvious advantage of larger sensors may be that they may have more pixels.

When shooting wide open, in order of blurring the background, smaller formats mostly have faster lenses, like many 24x36 mm lenses having f/1.4 aperture, while fast MFD lenses are more rare.

If we consider f/1.4 on 24x36 mm, that would correspond to f/1.8 on 33x44 mm f/2.2 on 54x41 mm. My hunch may be that older f/1.4 lenses for 24x36 mm may not have been so great and say f/2.8 on 54x41 mm was quite OK.

Doing proper tests is not particularly easy, far to easy to miss a factor or two.

Best regards

Erik

themediocreenthusiast · Sep 11, 2023

Autoxave said:
I believe I have a few misconceptions about medium format vs full frame. Hoping to clarify this through the statements below. Are the statements below true or not true? Would be nice to have an explanation if the statements are not true.

Medium format images produces less noise in low light environments compared to fullframe presuming that available light is exactly the same and that the lenses are equivalent

Medium format images produces less depth of field, i.e. a better separation of the elements in an image, presuming that the field of view and distance to the element of interest is the same and lenses are equivalent

Medium format images produce less noise in post processing when increasing exposure in shadow areas of the image compared to fullframe presuming that available light was exactly the same when the image was captured, field of view is the same and that the lenses are equivalent

Medium format images produces less artifacts in post processing when sharpening an image presuming that the exposure and field of view is the same

Medium format cameras are superior (in terms of better image quality, less noise, more leeway in post processing) to fullframe when capturing long exposure seascape images presuming that available light is exactly the same and that the lenses are equivalent

Your questions are more perceptive (thankfully not very numbery-technical) and I guess there's no right or wrong then.

Some people here have been medium format for over 70yrs (Jim / Rich?) ; some people here have "been there, done that" - slowly climbed up the digital sensor ladder from 1/1.7" CCD to now medium format sensors *mostly here to stay now

Personally, before my foray into MF ... For FF - I had a Sony a7s 12mp , a7rII 42mp , Sigma fp 24mp, Nikon Df 17mp, and Canon 5D.

Foveon : Sigma SD14, DP1 & 3 Merrill, DP1 Quattro

I wake up at 530am everyday and walk to office with 1 or 2 everyday camera, so I'm quite used to seeing > shooting Golden hour / blue hour ideally with a good 24mm or 16-35mm lens (or iPhone lol). So from my personal experience after flirting with various cameras - my eyes and heart do agree with what your point (2) & (5)

- Medium format images produces less depth of field, i.e. a better separation of the elements in an image, presuming that the field of view and distance to the element of interest is the same and lenses are equivalent

- Medium format cameras are superior (in terms of better image quality, BUT MORE noise, more leeway in post processing) to fullframe when capturing long exposure seascape images presuming that available light is exactly the same and that the lenses are equivalent

I still "feel" Medium format has better tonality, better IQ & sharpness, better color layers or more color separation - whichever you like to call it.

Lastly , I feel the best part owning being part of a MF system is "what you see is what you get" . Transparency, facts and less B-S. ........

For example - I don't have to worry about spending $1-2k on a native lens but only to find out the lens is OK for low-res FF sensors but not good enough to resolve the high resolution stacked FF sensor - if you know what i mean

SrMi · Sep 11, 2023

JimKasson said:
areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Unfortunately, both Fuji and Hasselblad claim that 14 vs. 16 bits is about color, not noise.

JimKasson · Sep 11, 2023

SrMi said:
JimKasson said:

areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Click to expand...

Unfortunately, both Fuji and Hasselblad claim that 14 vs. 16 bits is about color, not noise.

Except for black point, I see no color differences.

SrMi · Sep 11, 2023

Erik Kaffehr said:
Autoxave said:

I believe I have a few misconceptions about medium format vs full frame. Hoping to clarify this through the statements below. Are the statements below true or not true? Would be nice to have an explanation if the statements are not true.

Medium format images produces less noise in low light environments compared to fullframe presuming that available light is exactly the same and that the lenses are equivalent

Click to expand...

That would presume same aperture and shutter time. Equivalent aperture would need a longer shutter time.

Erik Kaffehr said:

Medium format images produces less depth of field, i.e. a better separation of the elements in an image, presuming that the field of view and distance to the element of interest is the same and lenses are equivalent.

Click to expand...

Not when using equivalent aperture.

Erik Kaffehr said:

Medium format images produce less noise in post processing when increasing exposure in shadow areas of the image compared to fullframe presuming that available light was exactly the same when the image was captured, field of view is the same and that the lenses are equivalent

Click to expand...

No, it doesn't apply for equivalent aperture, but would apply if exposure would be kept the same.

Erik Kaffehr said:

Medium format images produces less artifacts in post processing when sharpening an image presuming that the exposure and field of view is the same

Click to expand...

That depends mostly on the pixel pitch, pixel aperture and presence of OLP filter. As an example, my P45+ back that has 6.8 micron pitch, around 0.75 in linear fill factor, is an aliasing machine.

Erik Kaffehr said:

Medium format cameras are superior (in terms of better image quality, less noise, more leeway in post processing) to fullframe when capturing long exposure seascape images presuming that available light is exactly the same and that the lenses are equivalent

Click to expand...

Not really necessarily. With same exposure, like f/8 at f/125, the larger sensor may have an advantage, assuming that quantum efficiency is similar.

Assuming that the basic sensor technology is similar, and that the capabilities of the sensor can be fully utilized, a larger sensor will have a benefit.

If we have good light or shoot on tripod, the larger format will have an advantage. But, doubling sensor size will just improve SNR by 41%.

I would think that the most obvious advantage of larger sensors may be that they may have more pixels.

+1

On the other hand, why are people buying 50MP MF cameras? Higher quality lenses?

Erik Kaffehr said:
When shooting wide open, in order of blurring the background, smaller formats mostly have faster lenses, like many 24x36 mm lenses having f/1.4 aperture, while fast MFD lenses are more rare.

If we consider f/1.4 on 24x36 mm, that would correspond to f/1.8 on 33x44 mm f/2.2 on 54x41 mm. My hunch may be that older f/1.4 lenses for 24x36 mm may not have been so great and say f/2.8 on 54x41 mm was quite OK.

Doing proper tests is not particularly easy, far to easy to miss a factor or two.

Best regards

Erik

areallygrumpyoldsod · Sep 11, 2023

JimKasson said:
areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Jim

I do not know how the GFX works and what it does when in 14-bit -vs- 16-bit (which I assume is native) - are the 14-bit downsampled files? Either way the blog is not really to my point.

It is not all about Shadow recovery and noise - Look at skin tones, colour correct product shots and repro/heritage shoots and the differences between the results when using a say a FF 14-bit RAW and a SMF 16-bit RAW to obtain ultra high quality results.

Be assured there are really compelling reasons why the IQ4 has really captured this market for those clients with budget that value and are willing to pay for this level of output. Not just hubris.

The GFX100s has become a top camera for those studio shooters with a wee bit smaller budget and a desire to shoot quicker and H6D or XF -- tell me the 16-bit depth is not a key part of this (sure the 102MP is also a plus)

Have a look at this from Fuji

As I note yes FF sensors have improved a lot - I have some of the best Nikon use, and own/use Hasselblad X2D for work that needs the little bit extra -- even given the limitations of slower shooting and only AF-S or Manual focussing.

--
areallygrumpyoldsod
Nikon and Hasselblad shooter -- wildlife and and --
https://www.andymillerphoto.co.uk/
https://www.flickr.com/photos/ajm057/
I do not respond to PMs or messages via my website

JimKasson · Sep 11, 2023

areallygrumpyoldsod said:
JimKasson said:

areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Click to expand...

Jim

I do not know how the GFX works and what it does when in 14-bit -vs- 16-bit (which I assume is native) - are the 14-bit downsampled files?

No. The sensor in the GFX 100x and X2D is the same sensor, and it has 14-bit and 16-bit precision (as well as others) as options.

areallygrumpyoldsod said:
Either way the blog is not really to my point.

It is not all about Shadow recovery and noise - Look at skin tones, colour correct product shots and repro/heritage shoots and the differences between the results when using a say a FF 14-bit RAW and a SMF 16-bit RAW to obtain ultra high quality results.

Now you are moving the goalposts. I'm comparing MF 14 and 16 bit precision.

areallygrumpyoldsod said:
Be assured there are really compelling reasons why the IQ4 has really captured this market for those clients with budget that value and are willing to pay for this level of output. Not just hubris.

Maybe so, but 16-bit precision is minimally significant.

areallygrumpyoldsod said:
The GFX100s has become a top camera for those studio shooters with a wee bit smaller budget and a desire to shoot quicker and H6D or XF -- tell me the 16-bit depth is not a key part of this

OK, the 16-bit precision is not the reason for spring for the GFX. There are plenty of other reasons to do so.

areallygrumpyoldsod said:
(sure the 102MP is also a plus)

JimKasson · Sep 11, 2023

SrMi said:
JimKasson said:

areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Click to expand...

Unfortunately, both Fuji and Hasselblad claim that 14 vs. 16 bits is about color, not noise.

References?

What is the supposed mechanism whereby the difference between 14 and 16 bit precision can affect color in regions of the photon transfer curve where the SNR is dominated by photon noise?

Note: this discussion is about ADC precision, not the Hasselblad calibration technique that turns 14-bit values into 16-bit ones.

--
https://blog.kasson.com

Rich42 · Sep 11, 2023

JimKasson said:
Note: this discussion is about ADC precision, not the Hasselblad calibration technique that turns 14-bit values into 16-bit ones.

SrMi · Sep 11, 2023

areallygrumpyoldsod said:
JimKasson said:

areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Click to expand...

Jim

I do not know how the GFX works and what it does when in 14-bit -vs- 16-bit (which I assume is native) - are the 14-bit downsampled files? Either way the blog is not really to my point.

It is not all about Shadow recovery and noise - Look at skin tones, colour correct product shots and repro/heritage shoots and the differences between the results when using a say a FF 14-bit RAW and a SMF 16-bit RAW to obtain ultra high quality results.

Be assured there are really compelling reasons why the IQ4 has really captured this market for those clients with budget that value and are willing to pay for this level of output. Not just hubris.

The GFX100s has become a top camera for those studio shooters with a wee bit smaller budget and a desire to shoot quicker and H6D or XF -- tell me the 16-bit depth is not a key part of this (sure the 102MP is also a plus)

Have a look at this from Fuji

As I note yes FF sensors have improved a lot - I have some of the best Nikon use, and own/use Hasselblad X2D for work that needs the little bit extra -- even given the limitations of slower shooting and only AF-S or Manual focussing.

In addition to what Jim wrote:

GFX100 and X2D can set ADC to produce either 16 or 14-bit output.

Many Nikon cameras allow setting ADC output to either 14 or 12-bit. The only difference is that 14 bits make less noise at the lowest ISOs. No difference in color can be observed.

IQ4 has a larger sensor, hence better DR or less noise. Less noise improves everything.

Both Hasselblad and Fuji use 16-bits to claim better colors over 14-bits FF. Many feel that 16-bits gives them a better IQ. However, I do not know of any serious comparison showing any difference between 14 and 16 bits on the same camera/lens combination.

SrMi · Sep 11, 2023

JimKasson said:
SrMi said:

JimKasson said:

areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Click to expand...

Unfortunately, both Fuji and Hasselblad claim that 14 vs. 16 bits is about color, not noise.

Click to expand...

References?

How GFX100 Camera Gives You Perfect Colors and Tones

Why you should switch to FUJIFILM GFX100’s 16-bit RAW mode for exceptional image quality

fujifilm-x.com

X2D 100C - Hasselblad

An unrivalled merge of form and function – The X2D 100C establishes itself as Hasselblad’s flagship mirrorless medium format camera. A 100-megapixel CMOS sensor and advancements in image quality, focusing, stabilisation, responsiveness, and storage, are built into iconic Scandinavian design. The...

www.hasselblad.com

JimKasson said:
What is the supposed mechanism whereby the difference between 14 and 16 bit precision can affect color in regions of the photon transfer curve where the SNR is dominated by photon noise?

I believe that the claim was formulated by PR. PR often does not use a scientific approach.

JimKasson said:
Note: this discussion is about ADC precision, not the Hasselblad calibration technique that turns 14-bit values into 16-bit ones.

Yes, discussing X2D and GFX100.

P.S.: I do not believe that 16-bits matter for color.

Rich42 · Sep 11, 2023

SrMi said:
JimKasson said:

SrMi said:

JimKasson said:

areallygrumpyoldsod said:

Only Medium Format sensors offer 16-bit rather than 14-bit RAW image recording and this is a HUGE differentiator for me at least.

Click to expand...

In the GFX 100x, the DR differences between 14 bit precision and 16 bit precision are miniscule.

https://blog.kasson.com/gfx-100s/gfx-100s-shadow-performance-with-14-and-16-bit-precision/

Click to expand...

Unfortunately, both Fuji and Hasselblad claim that 14 vs. 16 bits is about color, not noise.

Click to expand...

References?

Click to expand...

https://fujifilm-x.com/en-us/exposure-center/how-gfx-100-camera-gives-you-perfect-colors-and-tones/

https://www.hasselblad.com/x-system/x2d-100c/

SrMi said:

What is the supposed mechanism whereby the difference between 14 and 16 bit precision can affect color in regions of the photon transfer curve where the SNR is dominated by photon noise?

Click to expand...

I believe that the claim was formulated by PR. PR often does not use a scientific approach.

. . . See, there are facts, and then there are alternative facts . . .

SrMi said:
SrMi said:

Note: this discussion is about ADC precision, not the Hasselblad calibration technique that turns 14-bit values into 16-bit ones.

Click to expand...

Yes, discussing X2D and GFX100.

P.S.: I do not believe that 16-bits matter for color.

paulraphael · Sep 11, 2023

areallygrumpyoldsod said:
The GFX100s has become a top camera for those studio shooters with a wee bit smaller budget and a desire to shoot quicker and H6D or XF -- tell me the 16-bit depth is not a key part of this (sure the 102MP is also a plus)

It is not, as Jim said.

Digital sampling theory can be counter-intuitive; here's an analogy that might clear it up (or make you want to punch me, who knows).

Suppose you're trying to measure water in a measuring cup, but your accuracy is limited by vibrations that slosh the water around. You can't improve your accuracy by switching to a measuring cup with finer gradations.

That's essentially what you're doing by switching to a longer digital word length (from 14 to 16 bits) when the noise floor of the analog signal is already higher than the quantizing noise of 14 bit sampling. You're trying to improve the part of the signal chain that isn't the problem.

Now, in the case of the GFX 100 megapixel sensor, testing shows a very slight improvement at 16 bits. Like, if you boost the bejeezus out of the shadows and squint at the noise. This suggests that the analog signal maybe has the equivalent of 14.2 bits of SNR. I'm making up that number, but you get the idea.

If the sensor could really capture 16 bits worth of SNR in the analog domain, then switching to 16 bit mode would make a much bigger difference. But it doesn't seem like sensor technology is at that point.

Clarifying a few misconceptions with medium format vs full-frame

Forum Enthusiast

Senior Member

Forum Pro

Community Leader

Senior Member

Community Leader

Community Leader

Veteran Member

Senior Member

Veteran Member

Community Leader

Veteran Member

Senior Member

Community Leader

Community Leader

Senior Member

Veteran Member

Veteran Member

Senior Member

Senior Member

Similar threads

Keyboard shortcuts