I very much appreciate the review - and as usual, I learned a lot. I'm not sure I understand why you kept the in camera or in PS distortion correction off for the Canon lens (and presumably for the Nikon ?), though. The lens is designed to be used with distortion correction on - and it's going to be used that way by almost every person who encounters it, simply because it's a native lens to the camera system. I do realize that other lenses in this comparison don't have that limitation, but those are presumably not going to be used with camera correction. I would prefer to see the lens used as intended - and if that in camera correction results in reduced or smeared detail near the corners, that's certainly a limitation that needs to be kept in mind. I certainly do NOT approve of Canon's decision to create lenses that don't function properly without distortion correction on, either. But, I'm not sure that dislike of the decision should be used this way in a review.
Good and fair question and definitely something I struggled with on the Canon especially because Canon does seem to be designing lenses to work with in-camera software. And yes, I did leave lens corrections off on the Nikon, too. Here's a glimpse at what I thought about on that very topic for this video: So I decided to leave in-camera corrections off because I wanted to make as even a field as possible for all the lenses. I felt that if one lens had a software-based advantage that the test would no longer be about the lens but about the lens as well as the software used on one camera. And that would not have been an even comparison then. Also, the next big question becomes how much does the camera affect performance? Would my RP do a worse job with in-camera corrections than, say, an R5? I suspect yes, but have no way to prove that. So to that end, does my selection of camera give a lens an unfair edge or maybe even cause a bit of detriment to it? Would distortion correction introduce softness outside of the APS-C area because of pixel stretching (definitely possible) that would cause another element of lens performance to be judged unfairly? Those were some of the questions I asked myself as I say down to do this, as well the overarching issue that if I were to enable lens corrections for each camera I'd have no way to evaluate how much of the lens' performance was camera-based and how much was lens-based, and with three systems being tested, that was a problem for me to reconcile. I suppose that, in theory, I could use each camera with a lens I could mount on all three and then do some A-B testing on in-camera lens corrections (assuming they don't require direct input from the lenses themselves) to see how that would work. For at least the RP and Z6, IIRC, they will not offer lens corrections with unchipped lenses (such as the three Leica M lenses in this video) and that would mean that I couldn't establish a performance comparison baseline to use for evaluating how much the cameras actually altered lens performance in the first place. And then the question is that the Nikon performed stunningly well without in-camera corrections, which means an affordable lens of this spec can be made. If I'm going to compare a Canon lens to a similar modern lens and the Canon engineers are taking optical shortcuts to save money (my assumption) on the lens build, I don't know that I can let myself give them a pass on that with in-camera corrections when I'm not doing the same for the other lenses (especially since none of the other lenses needed in-camera corrections to perform well.) And since the other lenses, especially the Nikon, don't need that to perform wonderfully, well, I'm not going to give Canon a pass on taking shortcuts and making a sub-par product (compared to other lenses of the same spec) because when they do that it is, to my way of thinking, not in the best interests of their customer. I would have been mad, very mad, about the Canon lens' performance had I bought it. I would not have felt comfortable with myself of this video had I put my finger on the scale for the Canon by enabling in-camera corrections as that wouldn't actually make the lens itself better and it would be facilitating Canon's approach to selling consumer-tier lenses of making a product which is not as good as what their competitors make and selling it for the same or nearly the same price.
It’s a tricky situation. With a lens like the Canon, or more noticeably with the Canon RF 16/2.8, they were made with huge flaws and designed to be corrected in-camera. For myself, I’d rather see reviews with corrections, while noting that the lens is bad without them. All optics are tradeoffs, and digital tradeoffs are part of that now. The RF16 is smaller, lighter, and cheaper by far than anything else similarly wide, but it’s garbage without in camera correction. If the flaws are manageable in real-world use with correction, that might be a reasonable tradeoff. Just by the spec sheet, the RF24/1.8 is more than a stop faster, focuses closer, has better stabilization than the EF 24/2.8, while being no bigger and the same price, and the EF was not distortion-free either. Big performance gains without increase in bulk or cost. But those gains rely on in camera correction. For most people in most situations, there’s an upside to that tradeoff larger than the downside in real-world use. It’s import to note, since even if physically possible, you probably could never adapt an RF lens like one of these and get acceptable results, while you could with an EF, or Leica lens. But at the same time, it does feel like setting the lens up to fail when looking at performance how it was intended to be used. Not quite the same as testing one lens with a diffusion filter and complaining about sharpness. I’d rather see the results from the lens how they were meant to look, how the folks designed it to work, and stick a pin on it to acknowledge the big red flag of extensive digital correction. That feels truest to real-world use, IMHO.
@@thebitterfig9903 that is my feeling also. It's not just Canon designing lenses like this, either. I don't know if anything is gained in terms of size/weight/image-quality, or if it just locks the lens to its designed platform by making it too difficult to adapt, or perhaps both.
@@selkiemaine I feel like it's just most extreme with Canon, tho. Everyone uses In-Camera Correction to paper over a few flaws, make slight tweaks, get the last 5%. Canon often leaves the last 15% to digital, and having the camera be one more element in lens design. The goal might partly be to keep the platform closed--my understanding is that Sony shooters extensively used EF lenses in the early days--but a lot of the RF lenses are reasonably small and cheap for the specs. The gains are less for the 28/2.8, which isn't much smaller or more affordable, but still. For comparisons like this... having most tests run with corrections, but adding another category for uncorrected performance might make sense. Canon loses that, but if the performance with corrections is good enough that real-world use doesn't really notice, I think that matters.
@@thebitterfig9903 I saw a recent Nikon lens that was just as bad. Sorry, but I don't remember which one, as I don't shoot Nikon. I also feel that which lens makes a difference - if it was a complex, fast, zoom, I would find some distortion more forgivable. But, to have this much distortion in a simple prime just feels ridiculous, especially when they're already using aspherical elements - one would think it correctible without additional manufacturing cost .
I figured the Leica and Zeiss would be the top performers. The Nikon surprised me, and good for them. I've been so disappointed in Canon since the R mount was released, and for the sake of brevity, I won't list all those reasons here. But I'm slowly moving away from Canon all together. I might have to look into Nikon as an excellent replacement, alongside my Leica gear. Thank you for this comparison test! I know it's a lot of work.
Thank you! The Nikon was solidly the star. I ALMOST bought a Z6II and that lens today. Very close. Anyway, Given that it's about $400, and was right there in a dead heat with a $1,100 and $2,800 lens, yeah, that's huge. Honestly, the amazing close focus it has is, for me, the biggest deal.
I recently got the Brightin Star 28mm, I adapt it to my sigma LPL. I love the super compact design. I use a close focus adapter giving me a lot more flexibility on how close I can get to my subject. I find this eliminates the strange focus in the corners allowing for much smoother corner to corner bokeh when shooting close objects.
I do love how small it is and it looks great, definitely. And yes, using it with a macro adapter is the way to go because that makes it a lot more versatile.
Shouldn't have. I try not to use AF for test photos because I know where I want the focus to be. So I typically use manual focus with magnification for focusing.
It's like my fictional Grandpa Stockwell always told me: "Always put your money on a German optical company which has been in business since the mid-1800s." 📷
Blind guess: Nikon, Leica, Canon, Zeiss, BS End result - wow! The Zeiss held up far better than I'd expected, given the older optical formula. The Canon is apparently just as bad as the 16mm RF that I returned. Great stuff David!
I very much appreciate the review - and as usual, I learned a lot.
I'm not sure I understand why you kept the in camera or in PS distortion correction off for the Canon lens (and presumably for the Nikon ?), though. The lens is designed to be used with distortion correction on - and it's going to be used that way by almost every person who encounters it, simply because it's a native lens to the camera system. I do realize that other lenses in this comparison don't have that limitation, but those are presumably not going to be used with camera correction. I would prefer to see the lens used as intended - and if that in camera correction results in reduced or smeared detail near the corners, that's certainly a limitation that needs to be kept in mind.
I certainly do NOT approve of Canon's decision to create lenses that don't function properly without distortion correction on, either. But, I'm not sure that dislike of the decision should be used this way in a review.
Good and fair question and definitely something I struggled with on the Canon especially because Canon does seem to be designing lenses to work with in-camera software. And yes, I did leave lens corrections off on the Nikon, too. Here's a glimpse at what I thought about on that very topic for this video:
So I decided to leave in-camera corrections off because I wanted to make as even a field as possible for all the lenses. I felt that if one lens had a software-based advantage that the test would no longer be about the lens but about the lens as well as the software used on one camera. And that would not have been an even comparison then. Also, the next big question becomes how much does the camera affect performance? Would my RP do a worse job with in-camera corrections than, say, an R5? I suspect yes, but have no way to prove that. So to that end, does my selection of camera give a lens an unfair edge or maybe even cause a bit of detriment to it? Would distortion correction introduce softness outside of the APS-C area because of pixel stretching (definitely possible) that would cause another element of lens performance to be judged unfairly? Those were some of the questions I asked myself as I say down to do this, as well the overarching issue that if I were to enable lens corrections for each camera I'd have no way to evaluate how much of the lens' performance was camera-based and how much was lens-based, and with three systems being tested, that was a problem for me to reconcile.
I suppose that, in theory, I could use each camera with a lens I could mount on all three and then do some A-B testing on in-camera lens corrections (assuming they don't require direct input from the lenses themselves) to see how that would work. For at least the RP and Z6, IIRC, they will not offer lens corrections with unchipped lenses (such as the three Leica M lenses in this video) and that would mean that I couldn't establish a performance comparison baseline to use for evaluating how much the cameras actually altered lens performance in the first place.
And then the question is that the Nikon performed stunningly well without in-camera corrections, which means an affordable lens of this spec can be made. If I'm going to compare a Canon lens to a similar modern lens and the Canon engineers are taking optical shortcuts to save money (my assumption) on the lens build, I don't know that I can let myself give them a pass on that with in-camera corrections when I'm not doing the same for the other lenses (especially since none of the other lenses needed in-camera corrections to perform well.) And since the other lenses, especially the Nikon, don't need that to perform wonderfully, well, I'm not going to give Canon a pass on taking shortcuts and making a sub-par product (compared to other lenses of the same spec) because when they do that it is, to my way of thinking, not in the best interests of their customer.
I would have been mad, very mad, about the Canon lens' performance had I bought it. I would not have felt comfortable with myself of this video had I put my finger on the scale for the Canon by enabling in-camera corrections as that wouldn't actually make the lens itself better and it would be facilitating Canon's approach to selling consumer-tier lenses of making a product which is not as good as what their competitors make and selling it for the same or nearly the same price.
It’s a tricky situation. With a lens like the Canon, or more noticeably with the Canon RF 16/2.8, they were made with huge flaws and designed to be corrected in-camera. For myself, I’d rather see reviews with corrections, while noting that the lens is bad without them.
All optics are tradeoffs, and digital tradeoffs are part of that now. The RF16 is smaller, lighter, and cheaper by far than anything else similarly wide, but it’s garbage without in camera correction. If the flaws are manageable in real-world use with correction, that might be a reasonable tradeoff. Just by the spec sheet, the RF24/1.8 is more than a stop faster, focuses closer, has better stabilization than the EF 24/2.8, while being no bigger and the same price, and the EF was not distortion-free either. Big performance gains without increase in bulk or cost. But those gains rely on in camera correction. For most people in most situations, there’s an upside to that tradeoff larger than the downside in real-world use.
It’s import to note, since even if physically possible, you probably could never adapt an RF lens like one of these and get acceptable results, while you could with an EF, or Leica lens. But at the same time, it does feel like setting the lens up to fail when looking at performance how it was intended to be used. Not quite the same as testing one lens with a diffusion filter and complaining about sharpness.
I’d rather see the results from the lens how they were meant to look, how the folks designed it to work, and stick a pin on it to acknowledge the big red flag of extensive digital correction. That feels truest to real-world use, IMHO.
@@thebitterfig9903 that is my feeling also. It's not just Canon designing lenses like this, either. I don't know if anything is gained in terms of size/weight/image-quality, or if it just locks the lens to its designed platform by making it too difficult to adapt, or perhaps both.
@@selkiemaine I feel like it's just most extreme with Canon, tho. Everyone uses In-Camera Correction to paper over a few flaws, make slight tweaks, get the last 5%. Canon often leaves the last 15% to digital, and having the camera be one more element in lens design.
The goal might partly be to keep the platform closed--my understanding is that Sony shooters extensively used EF lenses in the early days--but a lot of the RF lenses are reasonably small and cheap for the specs. The gains are less for the 28/2.8, which isn't much smaller or more affordable, but still.
For comparisons like this... having most tests run with corrections, but adding another category for uncorrected performance might make sense. Canon loses that, but if the performance with corrections is good enough that real-world use doesn't really notice, I think that matters.
@@thebitterfig9903 I saw a recent Nikon lens that was just as bad. Sorry, but I don't remember which one, as I don't shoot Nikon.
I also feel that which lens makes a difference - if it was a complex, fast, zoom, I would find some distortion more forgivable. But, to have this much distortion in a simple prime just feels ridiculous, especially when they're already using aspherical elements - one would think it correctible without additional manufacturing cost
.
I figured the Leica and Zeiss would be the top performers. The Nikon surprised me, and good for them. I've been so disappointed in Canon since the R mount was released, and for the sake of brevity, I won't list all those reasons here. But I'm slowly moving away from Canon all together. I might have to look into Nikon as an excellent replacement, alongside my Leica gear. Thank you for this comparison test! I know it's a lot of work.
Thank you! The Nikon was solidly the star. I ALMOST bought a Z6II and that lens today. Very close. Anyway, Given that it's about $400, and was right there in a dead heat with a $1,100 and $2,800 lens, yeah, that's huge. Honestly, the amazing close focus it has is, for me, the biggest deal.
I have the Elmarit and while I have no other 28’s to compare it to, it simply feels and shows like a good performer.
It is. It's a really good lens.
I recently got the Brightin Star 28mm, I adapt it to my sigma LPL. I love the super compact design. I use a close focus adapter giving me a lot more flexibility on how close I can get to my subject. I find this eliminates the strange focus in the corners allowing for much smoother corner to corner bokeh when shooting close objects.
I do love how small it is and it looks great, definitely. And yes, using it with a macro adapter is the way to go because that makes it a lot more versatile.
The Leica Elmarit-m is impressive. I am just curious how much better it would be compared to modern zoom lenses like the Tamron 28-75mm f/2.8 G2.
Good question and it's hard to say. Zoom and prime lenses are hard to compare in some ways.
Your videos are very entertaining AND informative. I have both the Nikon and the Leica 28 mm lenses. Great discussion.
RS. Canada
Thank you, Richard! Did my experiences with those two lenses jibe with yours?
I must admit, I wish one of the Minolta 28s' made an appearance. Either the CLE Minolta 28, AF 28, or MD
I was trying to keep this video to lenses currently in production. But, the Minolta AF 28mm f/2.8 will have its own video on June 10.
Did the af use cause any reduction in quality?
Shouldn't have. I try not to use AF for test photos because I know where I want the focus to be. So I typically use manual focus with magnification for focusing.
Canon EF 28mm F2.8 IS USM is missing.
@@QuietOC these were all mirrorless camera lending. The EF is a DSLR lens.
It's like my fictional Grandpa Stockwell always told me: "Always put your money on a German optical company which has been in business since the mid-1800s." 📷
You know, that's solid advice.
the elmarit eats them all
Yeah, it performed way better than I expected it to.
Blind guess: Nikon, Leica, Canon, Zeiss, BS
End result - wow! The Zeiss held up far better than I'd expected, given the older optical formula. The Canon is apparently just as bad as the 16mm RF that I returned. Great stuff David!
Solid guess and the Nikon, especially given the price, was something else.
I think you’re a Nikon lover 😂
I'm pretty brand agnostic and shoot a lot of different gear. My big thing is how much I like an individual piece of equipment.
Not one bad lens in the lineup besides the BS.
Oh man, I might say two. I hated the Canon. Honestly, I'd take the BS over the Canon.