What’s Coming Next from Canon?

[bloblom]

In terms of signal/noise a 800mm f/7.1 is equivalent to a 500mm f/5.6, ie 0.7 stops faster than than 500mm f/7.1.

Aren't you mixing up things? Where does 800mm f/7.1 come from here?

So, in fact the signal/noise advantage of the RF 200-800mm on FF is greater than 1.3 stops, not less, and is actually up to 2 stops better than 500/7.1 on crop.

Where does that come from? Aperture dictates light intensity (photons per surface area), even if the entrance pupil is bigger, the focal length is bigger too, we're still talking about f9 vs f7.1 so that's 0.67 stops against FF.
APS-C doesn't work like a teleconverter. You're not taking half the light and projecting it onto the same surface like a TC does, you're taking half the light and projecting onto half the surface, light intensity at the sensor plane remains exactly the same. So it's not 1.3 + 0.67 but rather 1.3 - 0.67 = ~0.67 in favor of FF.

Initially I was talking about the measured DR between the R5II and the R7. At 6400 the R7 has 5.09 stops, at 12800 the R5II has 5.
I see that 1.3 stop figure repeated in this thread. Not saying it's wrong, in a perfect world, with perfect cameras, there would be a flat 1.3 stop delta across the entire ISO range due to sensor size difference between FF and APS-C. In reality, it's almost never 1.3 stops, and it's not constant across the entire ISO range. Physics are real, but physics also say cameras aren't isolated systems.

 

[Bob Howland]

Aren't you mixing up things? Where does 800mm f/7.1 come from here?


Where does that come from? Aperture dictates light intensity (photons per surface area), even if the entrance pupil is bigger, the focal length is bigger too, we're still talking about f9 vs f7.1 so that's 0.67 stops against FF.
APS-C doesn't work like a teleconverter. You're not taking half the light and projecting it onto the same surface like a TC does, you're taking half the light and projecting onto half the surface, light intensity at the sensor plane remains exactly the same. So it's not 1.3 + 0.67 but rather 1.3 - 0.67 = ~0.67 in favor of FF.

Initially I was talking about the measured DR between the R5II and the R7. At 6400 the R7 has 5.09 stops, at 12800 the R5II has 5.
I see that 1.3 stop figure repeated in this thread. Not saying it's wrong, in a perfect world, with perfect cameras, there would be a flat 1.3 stop delta across the entire ISO range due to sensor size difference between FF and APS-C. In reality, it's almost never 1.3 stops, and it's not constant across the entire ISO range. Physics are real, but physics also say cameras aren't isolated systems.

When using a FF lens on an APS-C sensor, the light intensity on the APS-C sensor is correct for the selected F-Stop. But that extra light around the edges is simply wasted, just like if the FF image was cropped in post processing. That's where Neuro is coming up with 1.3 stops loss, although I think it's 1.0. The speed booster takes the output of a FF lens and reduces the size of the sensor onto which it projects and reduces the apparent focal length of the lens.. (A teleconverter works in exactly the opposite way.) The light intensity on the APS-C sensor is now 1 stop higher than it would be with a FF lens. The speed booster is why my 100-400 f/5-6.3 becomes a 71-284 f/3.5-4.5 and why my 150-600 f/5-6.3 becomes a 106-424 f/3.5-4.5. The EXIF data sent to the camera reflect that fact. With the 150-600, the result is so good that it's just plain eerie.

 

[neuroanatomist]

Aren't you mixing up things? Where does 800mm f/7.1 come from here?

The long end of the 100-500, when used on APS-C, gives an FoV equivalent to 800mm on FF, but the f/number remains f/7.1. Thus 800mm f/7.1.

I’ll leave the rest of the explanations to @AlanF if he chooses to reply, but his numbers and reasoning are completely correct.

 

[bloblom]

When using a FF lens on an APS-C sensor, the light intensity on the APS-C sensor is correct for the selected F-Stop. But that extra light around the edges is simply wasted, just like if the FF image was cropped in post processing. That's where Neuro is coming up with 1.3 stops loss, although I think it's 1.0. The speed booster takes the output of a FF lens and reduces the size of the sensor onto which it projects and reduces the apparent focal length of the lens.. The light intensity on the APS-C sensor is now 1 stop higher than it would be with a FF lens. The speed booster is why my 100-400 f/5-6.3 becomes a 71-284 f/3.5-4.5 and why my 150-600 f/5-6.3 becomes a 106-424 f/3.5-4.5. The EXIF data sent to the camera reflect that fact. With the 150-600, the result is so good that it's just plain eerie.

Yes, for total light gathered (light intensity * exposure * sensor size). If aperture and exposure are the same, that's log2(1.6²) = 1.3561 stop of difference.

 

[neuroanatomist]

That's where Neuro is coming up with 1.3 stops loss, although I think it's 1.0.

1.6 squared is 2.56. The base 2 log (i.e., stops) of 2.56 is 1.356144.

I round that to 1-1/3 so it translated to camera-based 1/3-stop exposure settings.

 

[bloblom]

The long end of the 100-500, when used on APS-C, gives an FoV equivalent to 800mm on FF, but the f/number remains f/7.1. Thus 800mm f/7.1.

I’ll leave the rest of the explanations to @AlanF if he chooses to reply, but his numbers and reasoning are completely correct.

For FoV but it's not the core of the message here.

Regarding the "2 stops more" I sincerly doubt his numbers are correct.
You let more light with a bigger pupil, true, but you spread it over a greater surface (focal length/magnification), that's the entire point of having f-stop (focal length/entrance pupil) as a universal value for exposure, it's already taken into account.

If we take the total light gathered formula (light intensity * exposure * sensor size), how can reducing light intensity by 2/3rd of a stop increase the dynamic range by another 2/3rd of a stop?

Let's simplify to total light to ignore exposure, assuming it's the same on both systems.

Light intensity at the focal plane is defined by 1/N² where N is the aperture of the lens (this is directly derived from the formula that allows you to calculate the F-number of said lens)

For the R7:
Intensity = 1/7.1² = 1/50.41 = 0.01984
Sensor size = 22.3 * 14.8 mm = 330.04 mm²

0.01984 * 330.04 = 6.547

For the R5:
Intensity = 1/9² = 1/81 = 0.01235
Sensor size = 36 * 24 mm = 864 mm²

0.01235 * 864 mm² = 10.67

10.667 / 6.547 = 1.6298

That's log2(1.6298) = 0.7047 stop more light