If the pixel size is identical ...

[deednets]

... how is it that so much is being read into the image quality?

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

What bugs me though are the relatively slow maximum aperture and the lack of IBIS. If what the rumour mill is correct that is? And then there s the pixel size:

• Fuji X-Pro3 3.76µm
• SONY A7RV 3.76µm
• Leica Q3 3.76µm
• Fuji GFX100 3.76µm

Isn't the APS-C, the FF just a crop of the MF camera? And if that is the case, how is it that MF would then have better quality, provided you look at your image at 100% rather than down-sampling an image?

Ignoring for a moment that the GFX has a different aspect ratio, if the pixels are the same, would that not mean there is only something to be gained if you down-sample an image, as any crop would be roughly the same?

Dunno but is there maybe also some other factor at play that I am not seeing?

Thanks

Deed

P.S.: Sorry if this has been asked gazillions of times before!

 

[Erik Kaffehr]

... how is it that so much is being read into the image quality?

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

What bugs me though are the relatively slow maximum aperture and the lack of IBIS. If what the rumour mill is correct that is? And then there s the pixel size:

• Fuji X-Pro3 3.76µm
• SONY A7RV 3.76µm
• Leica Q3 3.76µm
• Fuji GFX100 3.76µm

Isn't the APS-C, the FF just a crop of the MF camera? And if that is the case, how is it that MF would then have better quality, provided you look at your image at 100% rather than down-sampling an image?

Ignoring for a moment that the GFX has a different aspect ratio, if the pixels are the same, would that not mean there is only something to be gained if you down-sample an image, as any crop would be roughly the same?

Dunno but is there maybe also some other factor at play that I am not seeing?

Thanks

Deed

P.S.: Sorry if this has been asked gazillions of times before!

All those cameras use the same sensor technology, although the 'toppings' may differ.

Going to a larger format we get more pixels. With more pixels we can resolve more detail. If we downsample a higher resolution image to a lower resolution image, the resulting image will keep the modulation (sharpness) of the high resolution image and the also the noise advantage of having more pixels.

The lens in front of the camera matters. Canon, Nikon and Sony often prioritize fast lenses like f/1.4 and zooms while MF is often more restricted in aperture and the zooms have shorter range.

I don't think MF has a significant advantage when looking at actual pixel crops, but having more pixels is an underrated advantage, IMHO.

Best regards

Erik

 

[yaqbpl]

there is no more pixels factor if you look on the center of mf frame. crop it is identical to aps-c in this case. if you put same lens, image would be the same. the lenses make difference. bigger sensor allows bigger optics which could make advantage.

with same lense you could move closer with mf, getting less depth of field ans different perspective - but we could use equivalent lens so it is not really difference.

 

[Stan Disbrow]

Hi,

There isn't a difference within the constraints of the specified area. Meaning of you crop the MF image to the dimensions of the FF one, or the FF one to the APS-C one. (Or MF to APS-C to that matter)

The difference is what you have for a total from the full size image as the sensor becomes larger. And then that really only makes a difference if you print. The greater the total pixel count, the larger a print you can make. And the standard these days is 360 pixels to the inch.

In my case, there is no 60 MP sensor from Sony using the same chip fabrication as the 100 and 150 MP MF sensors use in my small format brand, which is Nikon.

But, for this discussion, let us say there is.

What I want to do is print 16x20. Sometimes larger, but that size is the sweet spot for having large prints on the walls without them being too large. So, what I need is a 5:4 aspect ratio image with around 45 MP left after the crop from whatever the native sensor gives me.

That's a greater crop from 3:2, in the case of Full Frame than from 4:3 in the case of Medium Format. However, there is enough left over from the FF crop to go with. So, that'd work.

Now, the 100 MP MF is overkill. And so you might be thinking I could use a 50 MP MF sensor just fine. And, you'd be correct. Except....

Except the sensor technology is of an older design, being Front Side Illuminated. So that's a minus. And there is a difference in regards to the microlenses used on top of the base silicon. Another minus. The bottom line is: a greater amount of aliasing. Now, that's one of those things one doesn't see until one does. And, then you can't unsee it! Much like the arrow in the middle of the FedEx logo.....

In my case, I already have a sizable kit of MF lenses which happily adapt to the GFX. And I have a even larger kit of Nikon lenses. But, as I mentioned, no 60 MP Sony FF sensor equipped body in Nikon. And I don't fancy starting over in this regard.

I'm sure I could adapt my MF lenses to, say, a Sony 60 MP sensor equipped body. That might be a strain on the lens mount so I ruled that idea out.

This whole increased IQ with MF was quite obvious in the film days. Give me a larger negative and I need less enlargement for a given size print. So I was predisposed to just grab a MF camera when I wanted greater IQ. Which is what I did 5 years ago.

As you noted, this isn't as simple as it once was.

And we haven't even considered the differences in IQ from the differences in lens designs between systems. That's a much larger subject. One easy place to see how the lens designers do things differently is to look at the ranges of the zoom lenses. MF is typically restricted to a 2x zoom range. And those ranges tend to be greater as one reduces the sensor size. Because they don't have to cover the larger area.

Stan

 

[Ap0ll0n]

If you look at 100% then you are either looking at a smaller part of the image in medium format or you are looking at the same exact image but downsampled. The first would result in same iq for sure but for apples to apples comparison I don't see how you can get away from downsampling the MF image (ending up with better iq).

 

[DavidMillier]

The 100Mp vs 50MP MF sensor debate doesn't seem to be quite as cut and dried as it seems. Yes, more pixels means more resolution and larger prints. Yes, to less aliasing.

But, there appears to be some evidence that the 100MP sensor is more prone to hot pixels and then there is the PD pixel striping. It seems that these can be more of a problem in longer exposures. So, for me, who uses MF primarily for its long exposure capability, I'm put off the 100MP sensor.

I guess the same issue is likely true of the 60MP sensor as well. I have an A7Riv so I could actually test that theory - except I would have to go back to using a flappy cable release and an alarm clock, the absence of which is the very reasons I like to use my 50S....

 

[Stan Disbrow]

Hi,

I see the 50 MP aliasing to be more bothersome than the PDAF striping. That earns the YMMV or to each their own award, I suppose. In the end, I chose the 100 over the 50 to replace the 40 MP CCD.

Speaking of hot pixels, that CCD is prone to those. Little PIAs....

And, speaking of the CCD, either CMOS is a major upgrade.

Stan

 

[deednets]

If you look at 100% then you are either looking at a smaller part of the image in medium format or you are looking at the same exact image but downsampled. The first would result in same iq for sure but for apples to apples comparison I don't see how you can get away from downsampling the MF image (ending up with better iq).

 

[SrMi]

If you look at 100% then you are either looking at a smaller part of the image in medium format or you are looking at the same exact image but downsampled. The first would result in same iq for sure but for apples to apples comparison I don't see how you can get away from downsampling the MF image (ending up with better iq).

Yes, sure, but wasn't there some talk about the Mf look? Unless you crop a GFX100 photo to a 60Mpx, then the look is gone?

Bill Claff has a distinction on his website that suggests that an A7RV has less dynamic range when shot in APS-C mode?? Now, that's a wild one as if that high contrast part of the center, still running at 26Mpx, would lose some of the DR??

Because the magnification is higher with a crop, a crop loses DR. Comparison is not done at the pixel level but at the image level (same output size). Yes, the APS-C mode has less DR than the FF mode.

Likewise the centre 60Mpx crop from your GFX would gain DR? A 26Mpx crop from a GFX100, a Q3 and an X-Pro3 would differ in dynamic range, depending on Bill Claff's website?

Only throwing this in here as I felt from those other responses that it's a matter of downsampling or printing large?

But then, there was an only weakly opposed theory that a 26Mpx crop was better regarding IQ than any 26Mpx sensors out there????? If that was true, the GFX would then also be better because it's larger?

Not easy to grasp, but since there will be people pre-ordering a GFX100RF, why on Earth would they do that? The Greg uses a Q3 and the GFX, does he have an opinion??

Or Jim Kassons? He may have a theory of an aspect I am not seeing?

Deed

 

[Doppler9000]

If you look at 100% then you are either looking at a smaller part of the image in medium format or you are looking at the same exact image but downsampled. The first would result in same iq for sure but for apples to apples comparison I don't see how you can get away from downsampling the MF image (ending up with better iq).

Yes, sure, but wasn't there some talk about the Mf look? Unless you crop a GFX100 photo to a 60Mpx, then the look is gone?

The MF look is, essentially, a manifestation of magical thinking.

Bill Claff has a distinction on his website that suggests that an A7RV has less dynamic range when shot in APS-C mode?? Now, that's a wild one as if that high contrast part of the center, still running at 26Mpx, would lose some of the DR??

Likewise the centre 60Mpx crop from your GFX would gain DR? A 26Mpx crop from a GFX100, a Q3 and an X-Pro3 would differ in dynamic range, depending on Bill Claff's website?

Only throwing this in here as I felt from those other responses that it's a matter of downsampling or printing large?

But then, there was an only weakly opposed theory that a 26Mpx crop was better regarding IQ than any 26Mpx sensors out there????? If that was true, the GFX would then also be better because it's larger?

Not easy to grasp, but since there will be people pre-ordering a GFX100RF, why on Earth would they do that? The Greg uses a Q3 and the GFX, does he have an opinion??

Or Jim Kassons? He may have a theory of an aspect I am not seeing?

Deed

I will suggest an analytical algorithm.

 

[Ap0ll0n]

If you look at 100% then you are either looking at a smaller part of the image in medium format or you are looking at the same exact image but downsampled. The first would result in same iq for sure but for apples to apples comparison I don't see how you can get away from downsampling the MF image (ending up with better iq).

Yes, sure, but wasn't there some talk about the Mf look? Unless you crop a GFX100 photo to a 60Mpx, then the look is gone?

I think you meant to say unless you downsample, not crop (if you crop, the look is gone for sure).

Bill Claff has a distinction on his website that suggests that an A7RV has less dynamic range when shot in APS-C mode??

Because cropping introduces (amplifies) noise as there is no free lunch. You can crop an apsc 800 ISO image by say 4x and observe it being very noisy.

Now, that's a wild one as if that high contrast part of the center, still running at 26Mpx, would lose some of the DR??

Likewise the centre 60Mpx crop from your GFX would gain DR?

In my original reply I said the center crop, as in the single pixel, would have same noise (or set of pixels, same DR) as any other sized pixel from any other sensor, assuming same sensor tech of course. So again, cropping a 100MP to 60MP reduces DR, downsampling the entire 100MP image on the other hand (to view it on your monitor) would have more DR than the same image (content, magnification) taken with apsc.

A 26Mpx crop from a GFX100, a Q3 and an X-Pro3 would differ in dynamic range, depending on Bill Claff's website?

Only throwing this in here as I felt from those other responses that it's a matter of downsampling or printing large?

But then, there was an only weakly opposed theory that a 26Mpx crop was better regarding IQ than any 26Mpx sensors out there????? If that was true, the GFX would then also be better because it's larger?

Not easy to grasp, but since there will be people pre-ordering a GFX100RF, why on Earth would they do that? The Greg uses a Q3 and the GFX, does he have an opinion??

Or Jim Kassons? He may have a theory of an aspect I am not seeing?

Deed

--
Apollon
http://www.flickr.com/photos/apollonas/

 

[Doppler9000]

... how is it that so much is being read into the image quality?

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

What bugs me though are the relatively slow maximum aperture and the lack of IBIS. If what the rumour mill is correct that is? And then there s the pixel size:

• Fuji X-Pro3 3.76µm
• SONY A7RV 3.76µm
• Leica Q3 3.76µm
• Fuji GFX100 3.76µm

Isn't the APS-C, the FF just a crop of the MF camera? And if that is the case, how is it that MF would then have better quality, provided you look at your image at 100% rather than down-sampling an image?

Ignoring for a moment that the GFX has a different aspect ratio, if the pixels are the same, would that not mean there is only something to be gained if you down-sample an image, as any crop would be roughly the same?

Dunno but is there maybe also some other factor at play that I am not seeing?

Thanks

Deed

P.S.: Sorry if this has been asked gazillions of times before!

Think about this question in the film realm. If quality was invariant to “sensor” size, a Minox image would contain the same information as an 8x10 camera image. This might be true at 100% but is untrue at any output size larger than the Minox negative.

Think about the problem in terms of magnification or enlargement.

 

[JimKasson]

If you look at 100% then you are either looking at a smaller part of the image in medium format or you are looking at the same exact image but downsampled. The first would result in same iq for sure but for apples to apples comparison I don't see how you can get away from downsampling the MF image (ending up with better iq).

Yes, sure, but wasn't there some talk about the Mf look? Unless you crop a GFX100 photo to a 60Mpx, then the look is gone?

Bill Claff has a distinction on his website that suggests that an A7RV has less dynamic range when shot in APS-C mode?? Now, that's a wild one as if that high contrast part of the center, still running at 26Mpx, would lose some of the DR??

Bill's PDR metric is normalized to picture height in pixels.

Sensor Analysis Primer – Engineering and Photographic Dynamic Range

ISOlessness comparisons across resolutions - the last word

I left off yesterday's post with two questions about comparing shadow noise performance of cameras with different resolutions: What should we pick for a common resolution, and what value should we pick for a barely-acceptable SNR? Thinking that I couldn't be the first person to ask these...

blog.kasson.com

Likewise the centre 60Mpx crop from your GFX would gain DR? A 26Mpx crop from a GFX100, a Q3 and an X-Pro3 would differ in dynamic range, depending on Bill Claff's website?

Only throwing this in here as I felt from those other responses that it's a matter of downsampling or printing large?

But then, there was an only weakly opposed theory that a 26Mpx crop was better regarding IQ than any 26Mpx sensors out there????? If that was true, the GFX would then also be better because it's larger?

Not easy to grasp, but since there will be people pre-ordering a GFX100RF, why on Earth would they do that? The Greg uses a Q3 and the GFX, does he have an opinion??

Or Jim Kassons? He may have a theory of an aspect I am not seeing?

Deed

 

[JimKasson]

The MF look is, essentially, a manifestation of magical thinking.

Not completely:

Format size and image quality - the last word

The format rhetoric seems to be heating up again. Maybe it never cooled off. In the last 7 or 8 years, I’ve owned and made many images with APS-C (with crop factors between 1.3 and 1.5), full frame (FF), and 33x44 (MF, or crop MF, depending on your level of precision – and maybe your

blog.kasson.com

 

[SeeingSomeThingsMissingMany]

deednets wrote

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

From reading Jim's blog :

Format size and image quality - the last word

The format rhetoric seems to be heating up again. Maybe it never cooled off. In the last 7 or 8 years, I’ve owned and made many images with APS-C (with crop factors between 1.3 and 1.5), full frame (FF), and 33x44 (MF, or crop MF, depending on your level of precision – and maybe your

blog.kasson.com

It's easier to design build a terrific f/4 lens than a faster lens;

It's easier to design build a terrific lens for a larger sensor than for a smaller sensor.

Combining these two then could in theory make Gfx100Rf f/4 a superior image generator than Rx1 versions, Q, Q2, X100F.

 

[Ap0ll0n]

deednets wrote

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

From reading Jim's blog :

https://blog.kasson.com/the-last-word/format-size-and-image-quality/

It's easier to design build a terrific f/4 lens than a faster lens;

It's easier to design build a terrific lens for a larger sensor than for a smaller sensor.

Combining these two then could in theory make Gfx100Rf f/4 a superior image generator than Rx1 versions, Q, Q2, X100F.

 

[deednets]

... how is it that so much is being read into the image quality?

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

What bugs me though are the relatively slow maximum aperture and the lack of IBIS. If what the rumour mill is correct that is? And then there s the pixel size:

• Fuji X-Pro3 3.76µm
• SONY A7RV 3.76µm
• Leica Q3 3.76µm
• Fuji GFX100 3.76µm

Isn't the APS-C, the FF just a crop of the MF camera? And if that is the case, how is it that MF would then have better quality, provided you look at your image at 100% rather than down-sampling an image?

Ignoring for a moment that the GFX has a different aspect ratio, if the pixels are the same, would that not mean there is only something to be gained if you down-sample an image, as any crop would be roughly the same?

Dunno but is there maybe also some other factor at play that I am not seeing?

Thanks

Deed

P.S.: Sorry if this has been asked gazillions of times before!

Think about this question in the film realm. If quality was invariant to “sensor” size, a Minox image would contain the same information as an 8x10 camera image. This might be true at 100% but is untrue at any output size larger than the Minox negative.

Think about the problem in terms of magnification or enlargement.

I am sure I don't follow. If you take an environmental portrait:





... then extend the canvas to 100Mpx. The image quality and the dynamic range goes up?

Really?

Bear with me for a moment, the pixel size being equal, how would the SAME shot taken with a GFX100 be different?

The minox example would only be a valid one if you cropped the 24x36 to the same millimeter size of the Minox. Provided the film grain was identical, like the pixel size.

Deed

 

[JimKasson]

deednets wrote

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

From reading Jim's blog :

https://blog.kasson.com/the-last-word/format-size-and-image-quality/

It's easier to design build a terrific f/4 lens than a faster lens;

It's easier to design build a terrific lens for a larger sensor than for a smaller sensor.

Combining these two then could in theory make Gfx100Rf f/4 a superior image generator than Rx1 versions, Q, Q2, X100F.

Keeping in mind that a FF f/1.8 lens stopped down at f/4 should in general be better at f/4 than a MF lens at f/4 or thereabouts, right?

Why do you make that assertion?

Maybe worst after that, TBD (that's just the lens, the MF sensor will certainly add to the superiority).

 

[deednets]

Bill's PDR metric is normalized to picture height in pixels.

https://www.photonstophotos.net/Gen...ngineering_and_Photographic_Dynamic_Range.htm

https://blog.kasson.com/the-last-word/isolessness-comparisons-across-resolutions/

Here is an example regarding my thought process, (your link to a D300 "Sensor Analysis Primer" wasn't particularly helpful as a response as you didn't address the pixel size):



The original shot was taken with an A7RV, to stay within my comparison, I will then add more canvas to get to 100Mpx. Since the pixel size is identical, this should then not look too different from the GFX100 shot? I would then use a GFX100 and a 110mm, equalize the aperture to F2.8.

So what are you actually saying? That the mere fact that you use a larger sensor would get you higher dynamic range? But possibly not by just changing the canvas size?

Since in both cases the full size, non-cropped, A7RV and the cropped GFX100 would use IDENTICAL pixel size, how can the dynamic range possibly be different? I mean: at all?

Photographic Dynamic Range versus ISO Setting

Note: DO get the resizing! But what I am talking about is the DR in any discernible area, after all the image could be partially black:





How would then the dynamic range change by cropping out the black bit? Or white, or whatever?

So in short: I am not buying this! If you and me, will never happen, but just as an example, we both shot that monkey me with an X-Pro3 and you with the GFX100, if the monkey took up the same pixel size in the centre, how could then be the DR of your GFX100 miraculously dropping only because you crop the image? How would cropping affect the brightest and darkest pixel in the monkey shot? Or altering the canvas in some other ways?

You follow? No idea how to make this any clearer but know that one can get a bit snow blind when looking to often at charts etc. Not you possibly, but a little link to a D300 website didn't really shed any light onto the pixel size issue.

Deed

 

[deednets]

deednets wrote

With some interest about the GFX100RF I tried to figure out whether this could be a camera for me, as I had been using all 3 versions of the RX1, the Q, Q2 and the X100F ... so in short: I am kind of used to the concept of a fixed lens camera.

From reading Jim's blog :

https://blog.kasson.com/the-last-word/format-size-and-image-quality/

It's easier to design build a terrific f/4 lens than a faster lens;

It's easier to design build a terrific lens for a larger sensor than for a smaller sensor.

Combining these two then could in theory make Gfx100Rf f/4 a superior image generator than Rx1 versions, Q, Q2, X100F.

But if the pixel size is identical?