Canon APS-C Shooters Rejoice: Sigma to Launch the RF 17-40mm f/1.8 and RF 12mm f/1.4

[InternalRevenue]

Maybe more pertinently, it’s almost 300g(!!) less than the 18–35, which honestly makes me question the veracity of the rumor.

With the necessary adapter included, the EF version of the 18-35 actually hits 920 grams; granted the RF version of the 17-40 will likely be a bit heavier than the others as the 16-300 ended up being, but still that's a massive weight savings. I have both the Canon 18-35 and EFS 18-135 for my R7 and the 18-135 is light enough to be comfortable for extended wear, whereas the 18-35 is not.

As far as price goes, Sigma's recent full frame version of the 18-35 (the 28-45mm F1.8 DG DN ART) is currently $1,349 USD, before tariffs. I'll bet this new lens slides in at $200-300 less.

 

[Bob Howland]

Yes indeed. However these new Sigma lenses try to compete with aperture/speed, which is not a natural advantage of APSC. IMO, those looking for speed/selective focus are better served with a FF system. As such I bet Sigma would do better with an inexpensive and light zoom such as a 15-85. Right now it seems the APSC choices are zooms with limited range (e.g. 14-30, 18-45) or superzooms (e.g. 18-150, 16-300). Sigma, how about a standard standard zoom?

I own the Sigma 10-18 and 18-50 and use them on an R7. IMO, Sigma is very much competing on small size and light weight. Compare those lenses with the comparable OMD lenses for MFT. I'm not interested in the 17-40. I am somewhat interested in the 12 f/1.4 for video and have begged Sigma a couple times to introduce a small, light 50-135 f/2.8 DC.

 

[LCDVS00]

It's the time to waiting for 50-135 F1.8 DC and a UWA F1.8 zoom lens.
They will make a really worthy ultra trinity for R7M2.
Of course, a wishful thinking.

 

[Dragon]

I'm not sure I understand what you're saying. I do know that a speed booster doesn't change the number of photons going through the lens. It just concentrates them onto a smaller area. "Exposure" is the number of photons per unit area for a given time. Likewise, a Teleconverter doesn't change the number of photons. It just spreads them out onto a larger area.

The reason for faster lenses is often to get a better image in low light Alternatively it is to minimize DOF to highlight the subject. In either case, a smaller imager requires a faster lens due to equivalency. You correctly state that exposure is the number of photons per unit area, but if you have a larger area, you will capture more photons at the same illumination level. If you look at dynamic range comparisons here https://www.photonstophotos.net/Charts/PDR.htm and plug in any current FF and a current APS-c camera and you will see that there is a stop and change difference for equivalent dynamic range, which is the fundamental component of IQ. Note that a Canon aps-c imager has 1/2.56 the area of a FF imager, so the difference is a bit more than one stop, but for practical purposes, an image shot at f/2 on a FF camera will produce about the same IQ as an image shot at f/1.4 on an APS-c , all other settings being equal. Clearly, if there is enough light, or exposure time doesn't matter, then the ISO can be be adjusted to compensate and only at or near base ISO will the FF imager have a significant DR advantage.

The explanation of DOF is a bit more complicated, but the same f vs. imager area rule applies.

The speed booster is just a handy example that clearly demonstrates the above principles.

 

[Bob Howland]

It's the time to waiting for 50-135 F1.8 DC and a UWA F1.8 zoom lens.
They will make a really worthy ultra trinity for R7M2.
Of course, a wishful thinking.

135mm/1.8 means that the lens front element would have to be at least 75mm in diameter. No thanks. I want small and, especially, light. For 135/2.8, the front element can be 50mm.

 

[AJ]

the f/1.8 lens has a wider physical aperture at a given focal length so it passes more light (a lens with a smaller image circle isn’t the same as a speedbooster).

Yes, for a given focal length. But the focal lengths differ across formats. 17 mm APSC is roughly equivalent to 28 mm FF.
Now, aperture = FL/FSTOP. So for the APSC lens it is 17 mm / 1.8 = 10 mm. For the FF lens it is 28 mm / 2.8 = 10 mm.
The apertures are the same: 10 mm. The field-of-view is the same. So the amount of light that passes through is the same. Hence, equivalence.

 

[SHK1]

I wonder where all the anti-Canon fanboys are waxing poetic about how this is too little too late.

 

[neuroanatomist]

Yes, for a given focal length. But the focal lengths differ across formats. 17 mm APSC is roughly equivalent to 28 mm FF.
Now, aperture = FL/FSTOP. So for the APSC lens it is 17 mm / 1.8 = 10 mm. For the FF lens it is 28 mm / 2.8 = 10 mm.
The apertures are the same: 10 mm. The field-of-view is the same. So the amount of light that passes through is the same. Hence, equivalence.

I see where you're going, which is basically full circle to where we started when I stated:

You might benefit from reading the articles on equivalence I linked above. You gain ~1.3 stops of light from a wider f/stop, you lose ~1.3 stops by throwing away 60% of the light the FF sensor can capture.
Or you can go ahead and buy a FF camera that you plan to use in crop mode…it’s your money to waste.

@Walrus suggested that by buying the 17-40/1.4 to use on an R5 in crop mode, he would get a similar MP count to his R6 but get the benefit of a 1.3-stop faster lens compared to the RF 28-70/2.8. Technically, I was wrong to say light was being 'thrown away' in that example, since the f/1.8 lens in crop mode is resulting in the same amount of light captured as the f/2.8 lens using the full sensor. Regardless, my point was that he would not be getting any benefit because the smaller area for light capture in crop mode cancels out the faster aperture of the f/1.8 lens...that's equivalence.

 

[AJ]

I see where you're going, which is basically full circle to where we started when I stated:

@Walrus suggested that by buying the 17-40/1.4 to use on an R5 in crop mode, he would get a similar MP count to his R6 but get the benefit of a 1.3-stop faster lens compared to the RF 28-70/2.8. Technically, I was wrong to say light was being 'thrown away' in that example, since the f/1.8 lens in crop mode is resulting in the same amount of light captured as the f/2.8 lens using the full sensor. Regardless, my point was that he would not be getting any benefit because the smaller area for light capture in crop mode cancels out the faster aperture of the f/1.8 lens...that's equivalence.

Yeah with the R5 you'd be throwing away unused pixels rater than light. And those pixels are expensive, so that would be a waste.
I see this 17-40/1.8 to be a lens you'd buy if you are a birder with a R7, and you are looking for a high-end standard lens for a crop body you already own.
I think most people understand the focal length equivalence of 17-40 to 28-70 FF. I hope they understand the fstop equivalence from 1.8 to 2.8 FF. I hope nobody is fooled into thinking they're getting an equivalent of the 28-70/2.0 L.

 

[DhlcadR6]

I wonder where all the anti-Canon fanboys are waxing poetic about how this is too little too late.

They'll just say don't care APS-C and demands FF 3rd party lenses.

 

[Dragon]

So far, there are no APS-c ART lenses, so I am questioning the DC DN suffix on this lens. An ART APS-c would be like Canon making an RFS L lens. Seems much more likely that it will be a FF lens and not for Canon. A 12mm contemporary is much more believable.

 

[InternalRevenue]

So far, there are no APS-c ART lenses, so I am questioning the DC DN suffix on this lens. An ART APS-c would be like Canon making an RFS L lens. Seems much more likely that it will be a FF lens and not for Canon. A 12mm contemporary is much more believable.

The DSLR era 18-35 f/1.8 that this lens would be replacing is an APS-C Art lens, and a popular one at that. No surprise that they would eventually create a mirrorless successor.

 

[HMC11]

The reason for faster lenses is often to get a better image in low light Alternatively it is to minimize DOF to highlight the subject. In either case, a smaller imager requires a faster lens due to equivalency. You correctly state that exposure is the number of photons per unit area, but if you have a larger area, you will capture more photons at the same illumination level. If you look at dynamic range comparisons here https://www.photonstophotos.net/Charts/PDR.htm and plug in any current FF and a current APS-c camera and you will see that there is a stop and change difference for equivalent dynamic range, which is the fundamental component of IQ. Note that a Canon aps-c imager has 1/2.56 the area of a FF imager, so the difference is a bit more than one stop, but for practical purposes, an image shot at f/2 on a FF camera will produce about the same IQ as an image shot at f/1.4 on an APS-c , all other settings being equal. Clearly, if there is enough light, or exposure time doesn't matter, then the ISO can be be adjusted to compensate and only at or near base ISO will the FF imager have a significant DR advantage.

The explanation of DOF is a bit more complicated, but the same f vs. imager area rule applies.

The speed booster is just a handy example that clearly demonstrates the above principles.

I am uncertain if my reasoning is valid. Consider an aperture of fixed diameter. If there is no lens in front of it, i.e. light is simply shined on it (such that it covers a larger area than the aperture before reaching the aperture), then the amount of light passing through the aperture would be the same for the same exposure time. However, when a lens is placed in front of the aperture, and let’s assume that the back of the lens covers exactly the diameter of the aperture, then the larger the (front) lens elements, the more light it would collect and pass through the aperture. In an actual camera system, I would imagine that the total light collected would be spread over an area larger than the largest aperture in the lens, meaning that light would be increasingly lost as the aperture is reduced. Nevertheless, larger lens element collecting and sending on more light might still hold. If so, then a FF lens with larger lens element could have better ISO performance (for the same aperture, not f-stop) on an APSC camera compared to the equivalent APSC lens.

 

[Dragon]

I am uncertain if my reasoning is valid. Consider an aperture of fixed diameter. If there is no lens in front of it, i.e. light is simply shined on it (such that it covers a larger area than the aperture before reaching the aperture), then the amount of light passing through the aperture would be the same for the same exposure time. However, when a lens is placed in front of the aperture, and let’s assume that the back of the lens covers exactly the diameter of the aperture, then the larger the (front) lens elements, the more light it would collect and pass through the aperture. In an actual camera system, I would imagine that the total light collected would be spread over an area larger than the largest aperture in the lens, meaning that light would be increasingly lost as the aperture is reduced. Nevertheless, larger lens element collecting and sending on more light might still hold. If so, then a FF lens with larger lens element could have better ISO performance (for the same aperture, not f-stop) on an APSC camera compared to the equivalent APSC lens.

ISO performance is a function of the sensor and subsequent circuitry, not the lens. A FF lens will illuminate a larger area than an APS-c lens and thus the extra light that the FF lens (with the same f stop as an APS-c lens) collects will be lost on an aps-c sensor, so the the light that strikes an Aps-c sensor will always be somewhat less than half the light that strikes a FF sensor with a lens of the same f number. If you add elements to the FF lens to concentrate the light so it only covers that aps-c sensor, you will effectively have shortened the lens and lowered the f number so the aps-c sensor gets all the light that previously reached the FF sensor. This is exactly what a speed booster does, but the speed booster is typically a 1 stop boost, which will still slightly over-illuminate an aps-c sensor and thus still not quite equal the performance of a FF sensor with the original lens. Hope that helps.

 

[HMC11]

ISO performance is a function of the sensor and subsequent circuitry, not the lens. A FF lens will illuminate a larger area than an APS-c lens and thus the extra light that the FF lens (with the same f stop as an APS-c lens) collects will be lost on an aps-c sensor, so the the light that strikes an Aps-c sensor will always be somewhat less than half the light that strikes a FF sensor with a lens of the same f number. If you add elements to the FF lens to concentrate the light so it only covers that aps-c sensor, you will effectively have shortened the lens and lowered the f number so the aps-c sensor gets all the light that previously reached the FF sensor. This is exactly what a speed booster does, but the speed booster is typically a 1 stop boost, which will still slightly over-illuminate an aps-c sensor and thus still not quite equal the performance of a FF sensor with the original lens. Hope that helps.

Thanks. I am actually comparing two lenses and assume that the one that collects more light and direct it to the image circle would then require a lower ISO setting, hence 'better' in ISO performance. This all assumes the same aperture, and f-stop can certainly change depending on focal length (ie 50mm at f5 for apsc would have the same aperture diameter as 80mm at f8 for FF).

 

[SDFilmFan]

Even though the zoom ain't the replacement for my EF-S 15-85mm that I'm (also) eagerly waiting for, it looks like I should start saving for two new lenses

I'm glad that I am not the only voice in the wilderness asking for an RF version of the EFS 15-85. It is the best non-L zoom I have had, and is the reason I still keep my SL1 after buying the R6 and R10. The SL1 with 15-85 kicks butt over the R10 with 18-150.

 

[Denevans]

The lenses look very interesting, but not sufficiently so to tempt me to buy an APS-C R body to use them.

To those who understand equivalence, the advantages of APS-C remain lower cost and size/weight. The FFeq of 17-40/1.8 is 27-64/2.9, so my RF 24-105/2.8 is ‘better’. Likewise, 12/1.4 is equivalent to 19/2.2 and the extra 1-1/3 stops of my 20/1.4 is worth more than the 0.5 mm difference (to me, based on DxO correction of barrel distortion).

 

[Denevans]

The f-stop doesn’t change with APSC. The ratio of lens diameter to focal length stays the same. The diameter shrinks but so does the focal length leaving the f-stop the same. Thats why u can get smart phone cameras with an f-stop of 1.4 or so.

 

[neuroanatomist]

The f-stop doesn’t change with APSC. The ratio of lens diameter to focal length stays the same. The diameter shrinks but so does the focal length leaving the f-stop the same. Thats why u can get smart phone cameras with an f-stop of 1.4 or so.

Note that I stated, “To those who understand equivalence…,” it seems you need to work on that. Try the links in this earlier post.

Also note that I did not state that the aperture or focal length change with sensor size, neither of them do. But if you want to compare sizes, you need to understand that focal length is not the only factor for which you need to compensate. That's what equivalence is all about.

The focal length of a lens is the distance between the rear nodal point and the sensor, when the lens is focused to infinity. If you find a lens where that value varies with the size of the sensor behind it please let us know, until then my belief in the laws of physics will remain strong. No, the focal length does not 'shrink'. The f/number is that value divided by the maximum diameter of the aperture. Like focal length, it is an intrinsic property of the lens and does not change based on the size of the sensor behind that lens.

Maybe 'u' can get smartphone cameras with an f/1.4 lens because those cameras are magical items that shrink the focal length and defy the laws of physics. Out here in the real world, the rest of us can get smartphone cameras with fast lenses because the focal lengths of those lenses are very short. For example, the main camera in the iPhone 16 has a focal length of 6.9mm, meaning its f/1.78 lens requires a maximum physical aperture diameter of 3.9 mm. The reason the focal length can be short and the aperture can be small is because the sensor is small.

Unlike focal length and aperture that are intrinsic to the lens, field of view, depth of field, and image noise are affected by the size of the sensor. When you say 'the focal length shrinks' with a smaller sensor, what you really mean is that the field of view for a given focal length is directly proportional to the size of the sensor, i.e. a smaller sensor has a narrower FoV for the same focal length. That's why the 6.9mm lens of the iPhone 16 (~3.5x crop sensor), a 15mm lens on a Canon APS-C camera (1.6x crop sensor) and a 24mm lens on a FF camera all give the same field of view. That's equivalence in terms of field of view, or if you prefer equivalence in terms of focal length.

The thing you apparently don't understand is the just like FoV changes with sensor size, depth of field and image noise also change with sensor size. There is no free lunch. A smaller sensor doesn't magically decrease the FoV (increasing the equivalent focal length) but not affect other parameters. Equivalence.

• If you match the FoV at a given focal length you will be further away from the subject with the smaller sensor, so the DoF will be deeper.
• If you maintain subject distance at a given focal length, DoF will actually be (slightly) shallower with the smaller sensor (because of the smaller circle of confusion...but if equivalence already confuses you, best to leave CoC for another time).
• For a given scene and f/number, the smaller sensor will give the same exposure setting because that is determined by light per unit area hitting the sensor, but the noise in the resulting image will be inversely proportional to the sensor size for a given ISO. Smaller sensors are noisier.

That last point is what confuses a lot of people. What we generically call ISO is in reference to the International Organization of Standards (ISO) 12232:2019 standard (updated twice since its first release in 1998). That standard is why different sensor sizes give the same exposure for a given scene and f/number, they are calibrated based on that standard to do so. Recall the bucket in the rain analogy – for a given scene and aperture, a smaller sensor will collect less light. If you want that scene to appear to be the same brightness in your output image, regardless of the sensor size (that is the point of the ISO standard), that means the lower amount of light captured by the smaller sensor must be amplified relatively more. More amplification means more noise.

 

[Denevans]

Note that I stated, “To those who understand equivalence…,” it seems you need to work on that. Try the links in this earlier post.

Also note that I did not state that the aperture or focal length change with sensor size, neither of them do. But if you want to compare sizes, you need to understand that focal length is not the only factor for which you need to compensate. That's what equivalence is all about.

The focal length of a lens is the distance between the rear nodal point and the sensor, when the lens is focused to infinity. If you find a lens where that value varies with the size of the sensor behind it please let us know, until then my belief in the laws of physics will remain strong. No, the focal length does not 'shrink'. The f/number is that value divided by the maximum diameter of the aperture. Like focal length, it is an intrinsic property of the lens and does not change based on the size of the sensor behind that lens.

Maybe 'u' can get smartphone cameras with an f/1.4 lens because those cameras are magical items that shrink the focal length and defy the laws of physics. Out here in the real world, the rest of us can get smartphone cameras with fast lenses because the focal lengths of those lenses are very short. For example, the main camera in the iPhone 16 has a focal length of 6.9mm, meaning its f/1.78 lens requires a maximum physical aperture diameter of 3.9 mm. The reason the focal length can be short and the aperture can be small is because the sensor is small.

Unlike focal length and aperture that are intrinsic to the lens, field of view, depth of field, and image noise are affected by the size of the sensor. When you say 'the focal length shrinks' with a smaller sensor, what you really mean is that the field of view for a given focal length is directly proportional to the size of the sensor, i.e. a smaller sensor has a narrower FoV for the same focal length. That's why the 6.9mm lens of the iPhone 16 (~3.5x crop sensor), a 15mm lens on a Canon APS-C camera (1.6x crop sensor) and a 24mm lens on a FF camera all give the same field of view. That's equivalence in terms of field of view, or if you prefer equivalence in terms of focal length.

The thing you apparently don't understand is the just like FoV changes with sensor size, depth of field and image noise also change with sensor size. There is no free lunch. A smaller sensor doesn't magically decrease the FoV (increasing the equivalent focal length) but not affect other parameters. Equivalence.

• If you match the FoV at a given focal length you will be further away from the subject with the smaller sensor, so the DoF will be deeper.
• If you maintain subject distance at a given focal length, DoF will actually be (slightly) shallower with the smaller sensor (because of the smaller circle of confusion...but if equivalence already confuses you, best to leave CoC for another time).
• For a given scene and f/number, the smaller sensor will give the same exposure setting because that is determined by light per unit area hitting the sensor, but the noise in the resulting image will be inversely proportional to the sensor size for a given ISO. Smaller sensors are noisier.

That last point is what confuses a lot of people. What we generically call ISO is in reference to the International Organization of Standards (ISO) 12232:2019 standard (updated twice since its first release in 1998). That standard is why different sensor sizes give the same exposure for a given scene and f/number, they are calibrated based on that standard to do so. Recall the bucket in the rain analogy – for a given scene and aperture, a smaller sensor will collect less light. If you want that scene to appear to be the same brightness in your output image, regardless of the sensor size (that is the point of the ISO standard), that means the lower amount of light captured by the smaller sensor must be amplified relatively more. More amplification means more noise.