R7 vs. R6 II - Sensor Size

[DNBush]

Just in terms of the sensors, 32.5mp APS-C (R7) and 24.2mp FF (R6 II), which would generally be better in terms of IQ (low light, high ISO, DR, etc.)?

 

[EduPortas]

RF-S is Canon's third bite at the APS-C cherry and they are obviously now going for "small is beautiful".. Canon have the sales figures for their crop lenses over the last twenty years, so they will have a much better idea than any of us as to what actually sells

Yes, I completely agree.

But the M line never produced an R7. They were casual cameras. R7 is the real deal.

It's a shame Canon does not up the ante in the RF-S lens lineup bc a good chunk of sales are going directly to competitors that show some interest in the crop sensor glass.

I know I did, even though I still keep my tank-like 7DM2 with better than average lenses.

Canon have absolutely nailed FF in the MILC world (I own an R8 for video purposes). But APS-C is just the right spot for size/quality and competition is fierce. I preferred a tiny Z50 with small and fast glass made my themselves over Canon's own APS-C offerings.

 

[Ephemeris]

R6II will be better for what you seek. Like stated the R7 has an excellent sensor but like any tightly packed sensor it makes it light hungry.

How is it any more "light hungry" than any lower-MP Canon APS-C, or 1.6x crop mode from any Canon FF? You make it sound like the R7 has some special, unique noise problem, when the fact is that it has one of the CLEANEST 22.5mmx15mm sensor areas in the industry, with nothing better from Canon, all of whose FF cameras have similar (R6) or greater visible high-ISO noise in crop mode than the R7 has.

If I may interject. I understand your post John, but your paragraph below there is some misalignment

AF measured in pixels per target is more complex. On a car I may have 4MP but because I've gone high resolution I can't sufficiently target a deer at 300m. If I use the 1MP I can, but that then struggles with bicycles right by the vehicle.

Thinking AF is a simple as pixels per target is a long way from how we don't for cars, space, trams.

Maybe put hand held cameras donas you suggest - I know so very little about what that closed industry does.

It is true for AF, though, that the R7 will begin to fail sooner in low light with slow lenses, because the AF points on the sensor can't do much to integrate pixels for NR in real time, the way our brains do. Until a sensor of these dimensions comes in a stacked form, where NR can be applied to the smaller pixels, the AF will suffer with the same low light or slow lens.

However, if we think in terms of AF speed vs pixels-on-subject, rather than vs a body, then it is only when you are getting more pixels-on-subject that the R7 AF struggles more than other non-stacked sensors in low light (for a non-stacked sensor). Take the R6, R5, and R7. With the same slow lens, the R7 will start to struggle first around sunset, then the R5, then the R6, but if you had a 1.4x on the R5, and a 2x on the R6, you have similar "pixels-on-subject", and then the R7 is no longer the AF underdog in low light.

--
Beware of correct answers to wrong questions.
John
http://www.pbase.com/image/55384958.jpg

 

[Ephemeris]

R6II will be better for what you seek. Like stated the R7 has an excellent sensor but like any tightly packed sensor it makes it light hungry.

Light hungry meaning it won't do as well as a lower mp FF sensor in low light?

Correct. If the exposure is good it will be OK but you have a lot more leeway with the R6II

If you are getting 2.56x as much total light with the R6-II, then obviously, any under-exposure that is the same for both cameras will still have 2.56x as much light, so your statement is a bit of a tautology.

I think I expecting 2.56x more light available to the sensor has a few problems. The first perhaps the outer radius already has an attention to apply but how have you decided it's under exposed?

The question is, does the shooting situation allow for 2.56x the light?

Sometimes yes; sometimes no. "Yes" at base ISO with ample light. "Yes" when shutter speed needs prevent base ISO and lenses with larger pupils are available and you want or can work with shallower DOF. "No" when stopped-down DOF is needed. "No" when you crop away most of the FF image.

For both of the "No" scenarios, the R7 has Canon's best sensor for noise in very low light.

--
Beware of correct answers to wrong questions.
John
http://www.pbase.com/image/55384958.jpg

 

[Ephemeris]

Just in terms of the sensors, 32.5mp APS-C (R7) and 24.2mp FF (R6 II), which would generally be better in terms of IQ (low light, high ISO, DR, etc.)?

The FF sensor is more than 2 and a half times the size of the crop sensor.

Ah, is that right ? Looking here:

https://www.canon-europe.com/get-inspired/tips-and-techniques/aps-c-vs-full-frame/

"However, a full-frame image sensor is physically about 63% or 1.6x larger than an APS-C format image sensor".

Not 2.5 times.

Just another case of linguistic shortcuts causing confusion.

1.6x the linear dimensions = 2.56x the area. No reason for any confusion here, if you make it clear which you are talking about.

I think it's confusing. It doesn't happen in any other sensor world of mine than photography, but perhaps that's a good thing.

--
Beware of correct answers to wrong questions.
John
http://www.pbase.com/image/55384958.jpg

 

[Ephemeris]

R6II will be better for what you seek. Like stated the R7 has an excellent sensor but like any tightly packed sensor it makes it light hungry.

How is it any more "light hungry" than any lower-MP Canon APS-C, or 1.6x crop mode from any Canon FF? You make it sound like the R7 has some special, unique noise problem, when the fact is that it has one of the CLEANEST 22.5mmx15mm sensor areas in the industry, with nothing better from Canon, all of whose FF cameras have similar (R6) or greater visible high-ISO noise in crop mode than the R7 has.

I knew you'd likely respond and I don't possess the detailed technical knowledge you do. Yes the R7 is a very clean sensor and there is nothing wrong with it.

It is true for AF, though, that the R7 will begin to fail sooner in low light with slow lenses, because the AF points on the sensor can't do much to integrate pixels for NR in real time, the way our brains do. Until a sensor of these dimensions comes in a stacked form, where NR can be applied to the smaller pixels, the AF will suffer with the same low light or slow lens.

I do have slow lenses, the 100-500 and when you place a TC on it really slows down. I have no technical explanation except for an overall assessment. Comparing all my files and all shooting situations I prefer my R6II. There is nothing technical to that. Just observation. I do realize this is a technical form so likely not as helpful.

We use a 1.4x Tele on the 100-500 lens regularly. It's clear it takes more time to focus, refocus and onwards. It does this even in the Arabian dessert we did some work recently.

The 100-300 F2.8 seems to make little user experience difference adding. Tele.

When we build systems there will be embedded cal files, which control elements and arms of the model.

Perhaps that 100-500 needs slowing down, perhaps a glass mass is high, a sensor a bit slow etc. either way it's done by design.

I have the RF 24-105. If my wife is singing in her indoor Christmas concerts I don't even consider my R7. If I had no choice I use the R7 and make it work.

However, if we think in terms of AF speed vs pixels-on-subject, rather than vs a body, then it is only when you are getting more pixels-on-subject that the R7 AF struggles more than other non-stacked sensors in low light (for a non-stacked sensor). Take the R6, R5, and R7. With the same slow lens, the R7 will start to struggle first around sunset, then the R5, then the R6, but if you had a 1.4x on the R5, and a 2x on the R6, you have similar "pixels-on-subject", and then the R7 is no longer the AF underdog in low light.

Agreed.

--
I roll with pleasing colour

 

[John Sheehy]

Would this be accurate?…

Generally speaking, the size of a sensor and number of megapixels determines the size of photosites which are primarily responsible for gathering light on the sensor. A full frame camera has a larger sensor. it's surface area is larger. The R6II has a lower pixel density giving it the ability to have larger photosites which in turn gives it the ability to gather light more effectively. The R7 has a smaller APS-C sensor, a higher pixel density and smaller photosites making it less sensitive to light and in turn a little more noisy.

A higher pixel density does not reduce sensitivity to light, except that smaller pixels can be a little more inefficient at very low f-numbers up to around f/2, with front-side-illuminated sensors.

 

[Ephemeris]

Would this be accurate?…

Generally speaking, the size of a sensor and number of megapixels determines the size of photosites which are primarily responsible for gathering light on the sensor. A full frame camera has a larger sensor. it's surface area is larger. The R6II has a lower pixel density giving it the ability to have larger photosites which in turn gives it the ability to gather light more effectively. The R7 has a smaller APS-C sensor, a higher pixel density and smaller photosites making it less sensitive to light and in turn a little more noisy.

A higher pixel density does not reduce sensitivity to light, except that smaller pixels can be a little more inefficient at very low f-numbers up to around f/2, with front-side-illuminated sensors.

 

[Zeee]

I used an EF 24-105 F4 on crop bodies for 6 years.

So?

So I did. Back to you. So?

 

[Ali]

In this past year Canon went ahead and allowed third parties to start delivering RF-S lenses. Some pretty good ones are already on the market.

No thanks.

Well they seem pretty good. In addition I was surprised to see that you can even firmware update the Sigma right on the camera - that is a tight level of integration.

 

[R2D2]

Would this be accurate?…

Generally speaking, the size of a sensor and number of megapixels determines the size of photosites which are primarily responsible for gathering light on the sensor. A full frame camera has a larger sensor. it's surface area is larger. The R6II has a lower pixel density giving it the ability to have larger photosites which in turn gives it the ability to gather light more effectively. The R7 has a smaller APS-C sensor, a higher pixel density and smaller photosites making it less sensitive to light and in turn a little more noisy.

A higher pixel density does not reduce sensitivity to light, except that smaller pixels can be a little more inefficient at very low f-numbers up to around f/2, with front-side-illuminated sensors.

+1 Sensor AREA and Entrance Pupil are much bigger factors.

R2

 

[kodakrome]

Just in terms of the sensors, 32.5mp APS-C (R7) and 24.2mp FF (R6 II), which would generally be better in terms of IQ (low light, high ISO, DR, etc.)?

The FF sensor is more than 2 and a half times the size of the crop sensor.

Ah, is that right ? Looking here:

https://www.canon-europe.com/get-inspired/tips-and-techniques/aps-c-vs-full-frame/

"However, a full-frame image sensor is physically about 63% or 1.6x larger than an APS-C format image sensor".

Not 2.5 times.

Just another case of linguistic shortcuts causing confusion.

1.6x the linear dimensions = 2.56x the area. No reason for any confusion here, if you make it clear which you are talking about.

I think it's confusing. It doesn't happen in any other sensor world of mine than photography, but perhaps that's a good thing.

In this case, people could get confused because of an inaccurate statement in an article from Canon-Europe. The writer wrote this: However, a full-frame image sensor is physically about 63% or 1.6x larger than an APS-C format image sensor. That is not accurate because he is specifically referring to the size of a sensor, and not just a single linear dimension. I would not dismiss that as just a linguistic shortcut.

 

[John Sheehy]

In this past year Canon went ahead and allowed third parties to start delivering RF-S lenses. Some pretty good ones are already on the market.

No thanks.

Well they seem pretty good. In addition I was surprised to see that you can even firmware update the Sigma right on the camera - that is a tight level of integration.

If the third parties are operating with Canon's blessing and with technical information that they don't have to reverse-engineer, then I would think that it is safer than ever to go 3rd-party with lenses, for the life of the RF mount and protocol, or any future mount and protocol that is supposed to support RF lenses.

I would not expect any problems like my Tamron G1 150-600 dropping into manual focus with recent Canon bodies, or my old Sigmas giving a dancing frame and extra blur if IBIS is enabled, due to fudged lens identities.

 

[EduPortas]

So I did. Back to you. So?

This is a good thing? Starting at 38mm FOV on APS-C?

 

[Zeee]

So I did. Back to you. So?

This is a good thing? Starting at 38mm FOV on APS-C?

Didn’t bother me. I liked that range for street photography. Some shoot with 35mm and 50mm exclusively. I was just pointing a FF lens works fine on a crop body. That lens wasn’t that great at 24mm on FF anyway. I dumped it for the EF 24-70 2.8 II when I got a FF body.

 

[EduPortas]

So I did. Back to you. So?

This is a good thing? Starting at 38mm FOV on APS-C?

Didn’t bother me. I liked that range for street photography. Some shoot with 35mm and 50mm exclusively. I was just pointing a FF lens works fine on a crop body. That lens wasn’t that great at 24mm on FF anyway. I dumped it for the EF 24-70 2.8 II when I got a FF body.

Nice.

 

[Thomas Hegna]

I think there are some confusion regarding FF sensors and the statement that they capture 2,56 times more light. (1,6 x 1,6)

-Yes FF sensors are 2,56 times bigger (area)

-exposure is always the same when settings are the same. (And the sensors are equally sensitive to light)

-Its all about pixel size. If you have a crop sensor and a FF sensor with the same individual pixel size. They would be equally as good. (If made the same way) The problem is that crop sensors are always packed tighter compared to FF sensors when regarding individual pixel size. This is why FF sensors generally perform better.

 

[EduPortas]

This is why FF sensors generally perform better.

That and a ton of advances in other parts of the camera like A/D converters, processors, integrated software, etc.

No one in their right mind would compare a 5D classic to a R5 in terms of ISO latitude and sensitivity. Most current APS-C models have easily outperformed the 5D classic in that same regard. Even M43 cameras, if we're being honest.

 

[Sittatunga]

I think there are some confusion regarding FF sensors and the statement that they capture 2,56 times more light. (1,6 x 1,6)

-Yes FF sensors are 2,56 times bigger (area)

-exposure is always the same when settings are the same. (And the sensors are equally sensitive to light)

-Its all about pixel size. If you have a crop sensor and a FF sensor with the same individual pixel size. They would be equally as good. (If made the same way) The problem is that crop sensors are always packed tighter compared to FF sensors when regarding individual pixel size. This is why FF sensors generally perform better.

Yes, the exposure is the same whatever the size of the sensor, but that's a measure of light intensity, not of the number of photons forming the image. Noise at the pixel level would be the same if the pixels were the same size for two sensors of different formats, but we don't look at pictures at the pixel level if we want to see the image the photographer intended. For the same exposure a full-frame sensor will use 2.6× (as Canon's crop factor is more like 1.62 than the often quoted 1.6) or 2¼× (everybody else's sensors) the number of photons than an APS-C sensor will to record the same image. Individual photons, being quantum particles, are completely random though we can predict the likely behaviour of large populations of photons, and the noise from their random nature is the basis for and the largest part of the image noise that the camera records. Their random nature means that we can use a Poisson distribution to describe the effects of a set of them, and that's what tells us that their random noise is inversely proportional to the square root of the number used to form the image. In other words, when you look at the same image captured with the same framing and exposure, the image noise will be just about in proportion to the crop factor between the two formats. That's the reason larger formats perform better; they just use more photons to form the image.

 

[Alastair Norcross]

I think there are some confusion regarding FF sensors and the statement that they capture 2,56 times more light. (1,6 x 1,6)

-Yes FF sensors are 2,56 times bigger (area)

-exposure is always the same when settings are the same. (And the sensors are equally sensitive to light)

-Its all about pixel size.

No, it isn't.

If you have a crop sensor and a FF sensor with the same individual pixel size. They would be equally as good. (If made the same way)

Equally as good at what? Gathering light? No, they wouldn't. The FF sensor would gather 2.6 times as much light (the crop factor is actually 1.6216, just rounded down to 1.6). They would be "equally as good" at the pixel level, not the image level. The only sense in which they would be "equally as good" is if you were always cropping the FF sensor by a factor of 1.62. But if you were always doing that, why would you have a FF camera in the first place?

The problem is that crop sensors are always packed tighter compared to FF sensors when regarding individual pixel size. This is why FF sensors generally perform better.

No, it isn't. It's about total light. It's amazing that people on this forum still get this wrong.

--
“When I die, I want to go peacefully in my sleep like my grandfather. Not screaming in terror, like the passengers in his car.” Jack Handey
Alastair
http://anorcross.smugmug.com
Equipment in profile

 

[Alastair Norcross]

In this past year Canon went ahead and allowed third parties to start delivering RF-S lenses. Some pretty good ones are already on the market.

No thanks.

Your loss. I have recently bought the Sigma 10-18 and 18-50 for my R7. They are both really nice lenses. Better than Canon's RF-S 10-18 (which is surprisingly good for the price) and 18-45. I have kept my RF-S 18-150 for the extra reach, but the Sigma 18-50 is better for everything up to 50mm. Also, the Sigma 56 F1.4 is a simply superb lens. For a while it was my most used lens on my M6II. I might get the RF version for my R7. The 16 F1.4 is a great lens too, but the size meant I didn't use it much on my M6II.