A pretty slick visual guide to format size implications by quezra

[sportyaccordy]

I have been trying to articulate what this infographic lays out pretty cleanly for months. Thanks quezra for pulling this together from DxOMark data.

Lot of implications to be made but I will let you guys have fun with it first.

 

[Martin.au]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

 

[Giklab]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

Not to mention that APSC has been capable of 14 EV before that tech came to FF.

 

[sportyaccordy]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

ISO100 on MFT is not the same as ISO100 on FF, with regards to SNR etc. No MFT camera has the dynamic range or color depth of a current gen full frame.

1" DoF is wrong.

That's just off the top of my head.

1" DoF is right- fastest 1" lens is F/1.8 which translates to F/4.8 full frame.

Sometimes it's better to be right than to be first.

 

[sportyaccordy]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

Not to mention that APSC has been capable of 14 EV before that tech came to FF.

What tech?

 

[Mr Olympian]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

I thought there were 3 F/.95 micro four thirds lenses.

It uses F-stops instead of T-stops which means its unreliable.

Someone should add a chart showing where DoF is too small to be usable in certain situations. Or Focus hit rate as DoF gets shallower.

 

[Giklab]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

Not to mention that APSC has been capable of 14 EV before that tech came to FF.

What tech?

That sensor design. The K-5 passed 14 EV in 2010, with the D7000 coming just slightly behind. The first FF camera that broke 14 was the D800 (and, of course, the E) 2 years later.

 

[pixelpushing]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

Not to mention that APSC has been capable of 14 EV before that tech came to FF.

What tech?

That sensor design. The K-5 passed 14 EV in 2010, with the D7000 coming just slightly behind. The first FF camera that broke 14 was the D800 (and, of course, the E) 2 years later.

The K-5 also applied NR to RAW, which skewed DxO. Same with Nikon.

I've had 16MP Sony sensor cameras (same as K-5) and they aren't a patch on FF of this or previous gen.

 

[mosswings]

I have been trying to articulate what this infographic lays out pretty cleanly for months. Thanks quezra for pulling this together from DxOMark data.

Lot of implications to be made but I will let you guys have fun with it first.

If Edward Tufte (The Visual Display of Quantitative Information) would have had a look at this, he would have spat out his coffee. The graphical elements of this infographic convey no information by their areas, and it's dang hard to figure out how to operate cameras from each of the formats to obtain IQ and CoC equivalence to FF cameras, which is the whole point of this.

The majority of this chart can be represented by 6 numbers: 1.3, 2, 3. These are the sensor area ratios, expressed in stops (yes, T-stops are more correct, but they vary from product to product. The should appear as a range of variation about the theoretical nominal).

What I'd consider a more useful definition of "shooting envelope" is the combination of f-stop, ISO, and aperture required for an equivalent image brightness and SNR with a particular level of motion-induced blur, expressed angularly for easy normalization. This would reveal the operational limitations for not only noise levels and DOF but also resolution.

 

[quezra]

Well I didn't want to put this chart up because I knew I'd have to end up spending the rest of the week fielding questions and/or flames

But my basic point with this chart is anyone can take great photos keeping their cameras above those dotted lines (roughly). The chart's purpose was to illustrate that you get a wider shooting envelope as sensor size goes up (which is a partial answer to the perennial "why FF" question that comes up in the Sony forums), and that is basically irrefutable.

Regarding criticisms:

• I looked up basically AF lenses (somehow people failed to notice I start at f/1.2 when surely I'm aware of the legendary Noctilux and Canon 50/0.95s?), though for some there's a kind of judgement call - e.g. Canon have two f/1.2 AF lenses (actually FF), so I kind of included f/1.7 in the APS-C envelope, though I'm not really sure many people would ever have that experience. The problem with including MF lenses is you get these exotic f/0.85 lenses that can theoretically be put on an APS-C or MFT mirrorless but (1) hardly anyone spends that kind of money on an APS-C lens, (2) their corner performance is awful and nowhere close to what you think of when you think of an FF f/1.2 lens, and (3) they are made by very obscure third party makers so hardly accessible or well-known as far as the "common" experience goes. Should I have included Nikon 1's 32/1.2? Maybe, but they literally only have one lens with that aperture speed, so is it really part of the "shooting envelope"? There were lots of questions like that could not be resolved so think of it as a rough average rather than a hard edge. I lean towards "representative" rather than "physically possible" as far as the DoF control goes.
• Extended ISO 100 doesn't count for SNR and DR purposes I'm afraid. See DxO scores. For the same reason I don't include ISO 50 (extended) on FF cameras. I tried to base it off Sony sensors (to avoid sensor "generation" controversy), so I'm pretty sure I was looking at EM5 or 10 and Nikons when looking at the DxO scores (though you can see from the ISO score I did peep at A7S though I don't seriously consider it as part of the general FF "envelope" since it's such a specialist camera).
• F/stops give DoF measure. T-stops do not. You misunderstand the purpose (please read the column headings), it has nothing to do with light transmission, only depth of field.
• Unfortunately this chart was thrown together in the midst of a conversation in the Sony forums and was not really meant to be some sort of definitive textbook on shooting envelopes, so inaccuracies are all my fault.

I would like to add in Medium Format once DxO rates the Pentax 645Z or Hasselblad H5D (if they ever do), as that would be quite interesting (oddly, they don't have better DoF-control lenses than FF typically does)

 

[March Of The Tripods]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

Not to mention that APSC has been capable of 14 EV before that tech came to FF.

What tech?

That sensor design. The K-5 passed 14 EV in 2010, with the D7000 coming just slightly behind. The first FF camera that broke 14 was the D800 (and, of course, the E) 2 years later.

The K-5 also applied NR to RAW, which skewed DxO. Same with Nikon.

I've had 16MP Sony sensor cameras (same as K-5) and they aren't a patch on FF of this or previous gen.

The K5 only used NR past ISO 1600, which is obviously not the end we are talking about. They did not use NR at lower ISO, you know, the ones where DR reached 14Ev. There is also a big difference in sensor output after it's tweaked by a MFG, which is why the exact same sensor in the D7k performed slightly different than in the K5, despite being of the same generation and nearly the same release period.

So you are wrong about the K5 DR and your info on your Sony 16mp camera is irrelevant.

 

[mosswings]

Well I didn't want to put this chart up because I knew I'd have to end up spending the rest of the week fielding questions and/or flames

But my basic point with this chart is anyone can take great photos keeping their cameras above those dotted lines (roughly). The chart's purpose was to illustrate that you get a wider shooting envelope as sensor size goes up (which is a partial answer to the perennial "why FF" question that comes up in the Sony forums), and that is basically irrefutable.

The DR aspect is the big one. You can maintain equivalence in IQ up to the point where you run out of well capacity, and you have 1.3 / 2 / 3 stops less of it as you shrink sensor size (given the same sensor architecture and process technology). But a LOT of pictures only need the DR of the 1" sensor, which gives you a heck of a lot of room to play with in the larger sensor sizes.

Regarding criticisms:

• I looked up basically AF lenses (somehow people failed to notice I start at f/1.2 when surely I'm aware of the legendary Noctilux and Canon 50/0.95s?), though for some there's a kind of judgement call - e.g. Canon have two f/1.2 AF lenses (actually FF), so I kind of included f/1.7 in the APS-C envelope, though I'm not really sure many people would ever have that experience. The problem with including MF lenses is you get these exotic f/0.85 lenses that can theoretically be put on an APS-C or MFT mirrorless but (1) hardly anyone spends that kind of money on an APS-C lens, (2) their corner performance is awful and nowhere close to what you think of when you think of an FF f/1.2 lens, and (3) they are made by very obscure third party makers so hardly accessible or well-known as far as the "common" experience goes. Should I have included Nikon 1's 32/1.2? Maybe, but they literally only have one lens with that aperture speed, so is it really part of the "shooting envelope"? There were lots of questions like that could not be resolved so think of it as a rough average rather than a hard edge. I lean towards "representative" rather than "physically possible" as far as the DoF control goes.
• Extended ISO 100 doesn't count for SNR and DR purposes I'm afraid. See DxO scores. For the same reason I don't include ISO 50 (extended) on FF cameras. I tried to base it off Sony sensors (to avoid sensor "generation" controversy), so I'm pretty sure I was looking at EM5 or 10 and Nikons when looking at the DxO scores (though you can see from the ISO score I did peep at A7S though I don't seriously consider it as part of the general FF "envelope" since it's such a specialist camera).

Yep. Base ISO bottom limit. I tend to look at the higher ISOs - 2 - 3 stops above base - as the most practical range for comparison, since with slower lenses that's where people will be shooting a lot of the time. It puts Canon back in the running (I don't shoot Canon, but I'm a fair-minded Nikonian)

• F/stops give DoF measure. T-stops do not. You misunderstand the purpose (please read the column headings), it has nothing to do with light transmission, only depth of field.

True enough. Recordable diffraction effects, though, are a nuanced thing. I believe that you're assuming equal resolution sensors?

• Unfortunately this chart was thrown together in the midst of a conversation in the Sony forums and was not really meant to be some sort of definitive textbook on shooting envelopes, so inaccuracies are all my fault.

Understand, and we will have SportyAccordy summarily drawn and quartered for taking this graphic out of context

I would like to add in Medium Format once DxO rates the Pentax 645Z or Hasselblad H5D (if they ever do), as that would be quite interesting (oddly, they don't have better DoF-control lenses than FF typically does)

 

[J A C S]

In a nutshell - bigger is better. ;-)

 

[quezra]

Well I didn't want to put this chart up because I knew I'd have to end up spending the rest of the week fielding questions and/or flames

But my basic point with this chart is anyone can take great photos keeping their cameras above those dotted lines (roughly). The chart's purpose was to illustrate that you get a wider shooting envelope as sensor size goes up (which is a partial answer to the perennial "why FF" question that comes up in the Sony forums), and that is basically irrefutable.

The DR aspect is the big one. You can maintain equivalence in IQ up to the point where you run out of well capacity, and you have 1.3 / 2 / 3 stops less of it as you shrink sensor size (given the same sensor architecture and process technology). But a LOT of pictures only need the DR of the 1" sensor, which gives you a heck of a lot of room to play with in the larger sensor sizes.

I actually mentioned exactly that in the original thread this was from: most everyday shooting only involves 7-8 EV of DR and so JPEGs and crop sensors handle all that just fine.

• F/stops give DoF measure. T-stops do not. You misunderstand the purpose (please read the column headings), it has nothing to do with light transmission, only depth of field.

True enough. Recordable diffraction effects, though, are a nuanced thing. I believe that you're assuming equal resolution sensors?

If I'm not wrong, diffraction limits are universal, resolution only changes when your camera can detect the difference (so higher resolution detects it before lower resolution, but they all start at the same absolute point).

 

[The Incredible Hoke]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

1" DoF is wrong.

That's just off the top of my head.

Not to mention that APSC has been capable of 14 EV before that tech came to FF.

What tech?

That sensor design. The K-5 passed 14 EV in 2010, with the D7000 coming just slightly behind. The first FF camera that broke 14 was the D800 (and, of course, the E) 2 years later.

The K-5 also applied NR to RAW, which skewed DxO. Same with Nikon.

I've had 16MP Sony sensor cameras (same as K-5) and they aren't a patch on FF of this or previous gen.

The K5 only used NR past ISO 1600, which is obviously not the end we are talking about. They did not use NR at lower ISO, you know, the ones where DR reached 14Ev. There is also a big difference in sensor output after it's tweaked by a MFG, which is why the exact same sensor in the D7k performed slightly different than in the K5, despite being of the same generation and nearly the same release period.

So you are wrong about the K5 DR and your info on your Sony 16mp camera is irrelevant.

This last line is the most perfect DPreview quote ever. It perfectly captures why no one should ever want to post anything definitve on here.

There is always some mouth breather in his Mom's basement set to refute anything you post. Well done, Tripod guy.

--

Brooklyn

 

[mosswings]

Well I didn't want to put this chart up because I knew I'd have to end up spending the rest of the week fielding questions and/or flames

But my basic point with this chart is anyone can take great photos keeping their cameras above those dotted lines (roughly). The chart's purpose was to illustrate that you get a wider shooting envelope as sensor size goes up (which is a partial answer to the perennial "why FF" question that comes up in the Sony forums), and that is basically irrefutable.

The DR aspect is the big one. You can maintain equivalence in IQ up to the point where you run out of well capacity, and you have 1.3 / 2 / 3 stops less of it as you shrink sensor size (given the same sensor architecture and process technology). But a LOT of pictures only need the DR of the 1" sensor, which gives you a heck of a lot of room to play with in the larger sensor sizes.

I actually mentioned exactly that in the original thread this was from: most everyday shooting only involves 7-8 EV of DR and so JPEGs and crop sensors handle all that just fine.

• F/stops give DoF measure. T-stops do not. You misunderstand the purpose (please read the column headings), it has nothing to do with light transmission, only depth of field.

True enough. Recordable diffraction effects, though, are a nuanced thing. I believe that you're assuming equal resolution sensors?

If I'm not wrong, diffraction limits are universal, resolution only changes when your camera can detect the difference (so higher resolution detects it before lower resolution, but they all start at the same absolute point).

Diffraction limits are universal, though I would not call them limits as the transition is quite gradual. However, the ability to record diffraction effects is related to sensor resolution, yes.

 

[Martin.au]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

ISO100 on MFT is not the same as ISO100 on FF, with regards to SNR etc. No MFT camera has the dynamic range or color depth of a current gen full frame.

No, but ISO 100 on MFT is the same as ISO 400 on FF, not ISO 800 (as per equivalence). The GH4 has ISO 100 (and probably numerous other cameras).

1" DoF is wrong.

That's just off the top of my head.

1" DoF is right- fastest 1" lens is F/1.8 which translates to F/4.8 full frame.

Nikon Australia - 1 NIKKOR 32mm f/1.2 -

What's this then?

Sometimes it's better to be right than to be first.

Irony.

 

[March Of The Tripods]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

ISO100 on MFT is not the same as ISO100 on FF, with regards to SNR etc. No MFT camera has the dynamic range or color depth of a current gen full frame.

No, but ISO 100 on MFT is the same as ISO 400 on FF, not ISO 800 (as per equivalence). The GH4 has ISO 100 (and probably numerous other cameras).

Goodness knows I wouldn't agree with the Honda man if I didn't have to, but I believe ISO 800 would be the more accurate comparison as MFT ISO 100 is an extended ISO. At least that's what my GX7 had, ISO 200 was the lowest native ISO, so two stops more would be 800.

 

[sportyaccordy]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

ISO100 on MFT is not the same as ISO100 on FF, with regards to SNR etc. No MFT camera has the dynamic range or color depth of a current gen full frame.

No, but ISO 100 on MFT is the same as ISO 400 on FF, not ISO 800 (as per equivalence). The GH4 has ISO 100 (and probably numerous other cameras).

1" DoF is wrong.

That's just off the top of my head.

1" DoF is right- fastest 1" lens is F/1.8 which translates to F/4.8 full frame.

Nikon Australia - 1 NIKKOR 32mm f/1.2 -

What's this then?

Sometimes it's better to be right than to be first.

Irony.

I was thinking of the RX100s. You know, cameras/lenses people actually buy. You could have an RX100 III and change for the price of that 1 lens, which is equivalent to 85 1.8. Heck, you could buy an APS-C kit + 50 1.4 for that money. We might as well start throwing MF Samyangs in the mix. No, pedantry does not win here.

 

[Martin.au]

Shame it's wrong. I'd usually try and get the numbers right before making them pretty.

M4/3s SnR/ISO numbers are wrong. (Lots of M4/3s cameras do ISO 100, eg: GH4, E-M1)

ISO100 on MFT is not the same as ISO100 on FF, with regards to SNR etc. No MFT camera has the dynamic range or color depth of a current gen full frame.

No, but ISO 100 on MFT is the same as ISO 400 on FF, not ISO 800 (as per equivalence). The GH4 has ISO 100 (and probably numerous other cameras).

Goodness knows I wouldn't agree with the Honda man if I didn't have to, but I believe ISO 800 would be the more accurate comparison as MFT ISO 100 is an extended ISO. At least that's what my GX7 had, ISO 200 was the lowest native ISO, so two stops more would be 800.

--
"People who think they know everything are a great annoyance to those of us who do." - Isaac Asimov

Why don't you just go and look.