which allows me to, somewhat, answer my own question posed above: i.e. how it might compare as an alternative to the current class-leading landscape prime in the 24/25mm range - the Loxia 25mm.
Firstly, the Sigma is definately sharper than the Batis 25mm, and we know from Fred's tests that the Loxia 25mm is sharper than the Batis 25mm, so, I think its safe to assume that the Sigma is on par with the Loxia in this regard.
Micro-contrast advantage would still go to the Loxia, if, for no other reason, the Sigma has pretty bad distortion, which will require correction in most cases.
Flare performance is another win for the Loxia, as is its sunstar quality.
LOCA is better controlled in the Sigma, although the highlight-bokeh rendering of the Sigma in the corners is pretty colourful and rough.
That corner rendering on the Sigma also has distinct lemon shape, and increased DOF due to vignetting. From what I've seen, the Loxia is much better in this regard.
So, it seems, that no, overall the Sigma doesn't dethrone the Loxia, although it obviously has its other advantages.
Petegh wrote:
Micro-contrast advantage would still go to the Loxia, if, for no other reason, the Sigma has pretty bad distortion, which will require correction in most cases.
I wish that review sites would consider that factor. Tsdevine recently posted some interesting uncorrected/corrected image pairs to examine the loss in sharpness and microcontrast that result from correction for distortion:
scalanc2 wrote:
It is a matter of personal priorities.
To me , for instance, the Loxia is killing itself just being MF.
You're looking at it all wrong...the Loxia being MF only is a convenient way to justify TWO lenses in that focal range, one specialist MF lens (Loxia) and one AF lens (your choice!)
I'll be honest though, I mostly just use the Loxia in that range...it's generally landscape only for me at 24/25mm.
The 24GM is, I think, the best dual purpose lens in that FL in any mount. It's not near-perfect like the Loxia is, but it's really, really good...and has AF, is f/1.4 and has nice bokeh.
zugzwang2 wrote:
I wish that review sites would consider that factor. Tsdevine recently posted some interesting uncorrected/corrected image pairs to examine the loss in sharpness and microcontrast that result from correction for distortion:
It can be a difficult thing to quantify: if judged visually, its subjective, and very dependent on the experience of the viewer; if tested on Imatest, it can be more easily quantifed based on the dip in 50lp/mm curve, although it doubles the testers workload (before/after correction), which is probably why most reviewers don't bother.
IMHO, if a lens has a level of distortion that will require correction for the majority of subjects, then its peformance should be judged, post-correction; to me its performance pre-correction, is almost irrelevant, if I can't use it that way...
scalanc2 wrote:
It is a matter of personal priorities.
To me , for instance, the Loxia is killing itself just being MF.
My comments were made in the context of a dedicated landscape lens: here, manual focus, with hard stops and a DOF scale on the lens, can be advantageous, not disadvantageous, as compared to a AF lens. Of course, if Sony ever gets its act together and provides in-camera focus-stacking, the balance may shift back to AF lenses for landscape.
which allows me to, somewhat, answer my own question posed above: i.e. how it might compare as an alternative to the current class-leading landscape prime in the 24/25mm range - the Loxia 25mm.
Firstly, the Sigma is definately sharper than the Batis 25mm, and we know from Fred's tests that the Loxia 25mm is sharper than the Batis 25mm, so, I think its safe to assume that the Sigma is on par with the Loxia in this regard.
Micro-contrast advantage would still go to the Loxia, if, for no other reason, the Sigma has pretty bad distortion, which will require correction in most cases.
Flare performance is another win for the Loxia, as is its sunstar quality.
LOCA is better controlled in the Sigma, although the highlight-bokeh rendering of the Sigma in the corners is pretty colourful and rough.
That corner rendering on the Sigma also has distinct lemon shape, and increased DOF due to vignetting. From what I've seen, the Loxia is much better in this regard.
So, it seems, that no, overall the Sigma doesn't dethrone the Loxia, although it obviously has its other advantages.
A week ago during the sales, I decided to take a punt at the Samyang 24mm f1.8 (should arrive in a few days). It reviewed well by Dustin Abbott, and Christopher Frost here:
I decided to try the Samyang in preference to the Sigma because of substantially lower vignetting and distortion, as well as being smaller and less expensive, and also surprisingly good corner sharpness. We’ll see how well their variation fairs 😏
Petegh wrote:
My comments were made in the context of a dedicated landscape lens: here, manual focus, with hard stops and a DOF scale on the lens, can be advantageous, not disadvantageous, as compared to a AF lens. Of course, if Sony ever gets its act together and provides in-camera focus-stacking, the balance may shift back to AF lenses for landscape.
Petegh wrote:
[Distortion] an be a difficult thing to quantify: if judged visually, its subjective, and very dependent on the experience of the viewer; if tested on Imatest, it can be more easily quantifed based on the dip in 50lp/mm curve, although it doubles the testers workload (before/after correction), which is probably why most reviewers don't bother.
IMHO, if a lens has a level of distortion that will require correction for the majority of subjects, then its performance should be judged, post-correction; to me its performance pre-correction, is almost irrelevant, if I can't use it that way...
A lens maven (I think it was DavidBM) noted here that physical or software corrections both exact a toll. That said, it does seem reasonable to compare the physical results from well-corrected lenses to the results that can be attained only via software correction with lenses designed with such corrections in mind. That's difficult to glean from lens reviews, because, as you said, "most reviewers don't bother."
By way of comparison of the new Sigma to the Loxia 25mm that you mentioned, the Sigma reaches 6.1% distortion per Lenstip, so perhaps four and a half to five times the maximum distortion of the Loxia lens. I've happily used the Loxia 25mm uncorrected for years, but I suppose that I would often correct images from the new Sigma. Others may differ.
In tsdevine's comparisons that I mentioned earlier, it's possible to squint hard enough to detect the loss of sharpness and contrast that results from correcting even the remarkably slight distortion (0.38% per Lenstip, so about 1/16 that of the Sigma) of the Voigtlander 50mm APO. That's just one of the possible trade-offs of lens design.
zugzwang2 wrote:
A lens maven (I think it was DavidBM) noted here that physical or software corrections both exact a toll. That said, it does seem reasonable to compare the physical results from well-corrected lenses to the results that can be attained only via software correction with lenses designed with such corrections in mind. That's difficult to glean from lens reviews, because, as you said, "most reviewers don't bother."
By way of comparison of the new Sigma to the Loxia 25mm that you mentioned, the Sigma reaches 6.1% distortion per Lenstip, so perhaps four and a half to five times the maximum distortion of the Loxia lens. I've happily used the Loxia 25mm uncorrected for years, but I suppose that I would often correct images from the new Sigma. Others may differ.
In tsdevine's comparisons that I mentioned earlier, it's possible to squint hard enough to detect the loss of sharpness and contrast that results from correcting even the remarkably slight distortion (0.38% per Lenstip, so about 1/16 that of the Sigma) of the Voigtlander 50mm APO. That's just one of the possible trade-offs of lens design. ...Show more →
Most probably won't care about having to rely on software distortion correction, as long as the corrected product still looks 'sharp enough' to them.
The problem for me, is that software correction creates a different 'kind' of sharpness, vs optical correction: software distortion corrected lenses have high edge accutance (what most people would call 'sharpness'), but only in the relatively low frequency ranges, as its the MTF of the higher frequency detail that takes a hit with software correction. This creates images that look 'sharp' to many (largely due to the preference the human visual system has for defining sharpness based on low frequency edge acutance), but the overall rendering takes a hit: it has a 'sharp, but without fine detail' kind of look to it; its not a look I like.
You need to remember that the Imatest lens testing software that LensTip, Dxo, Ephotozine, OpticalLimits etc use, is giving you the resolution at 50% MTF: so when LensTip says the lens is capable of resolving 86lp/mm, that's the resolution at the spatial frequency that has 50% contrast, so nowhere near the maximum resolving power of the lens - which would be around 10% MTF. So, the Imatest software may not show up the real extent of the damage being done to the higher frequencies, where micro-contrast resides.
Petegh wrote:
Most probably won't care about having to rely on software distortion correction, as long as the corrected product still looks 'sharp enough' to them.
The problem for me, is that software correction creates a different 'kind' of sharpness, vs optical correction: software distortion corrected lenses have high edge accutance (what most people would call 'sharpness'), but only in the relatively low frequency ranges, as its the MTF of the higher frequency detail that takes a hit with software correction. This creates images that look 'sharp' to many (largely due to the preference the human visual system has for defining sharpness based on low frequency edge acutance), but the overall rendering takes a hit: it has a 'sharp, but without fine detail' kind of look to it; its not a look I like.
You need to remember that the Imatest lens testing software that LensTip, Dxo, Ephotozine, OpticalLimits etc use, is giving you the resolution at 50% MTF: so when LensTip says the lens is capable of resolving 86lp/mm, that's the resolution at the spatial frequency that has 50% contrast, so nowhere near the maximum resolving power of the lens - which would be around 10% MTF. So, the Imatest software may not show up the real extent of the damage being done to the higher frequencies, where micro-contrast resides....Show more →
I”m just curious. But what do you use your pictures for?
scalanc2 wrote:
I”m just curious. But what do you use your pictures for?
Image quality expectations have been increasing exponentially in recent years, largely driven by 4K and now 8K TV, HDR etc,etc. Why be 'blinkered', and limit your IQ to current 'very good' rather than 'exceptional' standards - this requires the best lenses.
I want my landscapes to look good in 10, 20 or more years, when we all have wall sized OLED monitors in our homes, like in the sci-fi pic's!
Just traded in a bunch of camera and video stuff for a Lox 85.
Now I have in the 85-90 range:
85/1.8G
85/4 ZM
85/2.4 Lox
90/3.5 CV SL-II
90/2.8i Sigma
Going to have a big 85/90mm comparison and then thin the herd. Likely the 85/4ZM and 90/3.5 CV will be the ones to go.
The 85/1.8 will be for vacations
The 90/2.8 Sigma for backpacking
the 85/2.4 Lox for dedicated landscape/night shots
All fine lenses and hate to see the CV and ZM go, but that's just too many 85/90mm lenses to have.
I look forward to your comparison. I am curious as to why you would get both the Sony 85/1.8 and the Sigma 90/2.8. I would like to have both too but not sure I can justify it.
jhinkey wrote:
Just traded in a bunch of camera and video stuff for a Lox 85.
Now I have in the 85-90 range:
85/1.8G
85/4 ZM
85/2.4 Lox
90/3.5 CV SL-II
90/2.8i Sigma
Going to have a big 85/90mm comparison and then thin the herd. Likely the 85/4ZM and 90/3.5 CV will be the ones to go.
The 85/1.8 will be for vacations
The 90/2.8 Sigma for backpacking
the 85/2.4 Lox for dedicated landscape/night shots
All fine lenses and hate to see the CV and ZM go, but that's just too many 85/90mm lenses to have.
smpetty wrote:
I look forward to your comparison. I am curious as to why you would get both the Sony 85/1.8 and the Sigma 90/2.8. I would like to have both too but not sure I can justify it.
Because when I go on vacations typically we are inside dark museums, churches, etc. and f/1.8 comes in handy and sometimes a f/1.8 candid shot is way cool. It is a bit bulky though. So the sigma 90/2.8 might be used for such things if I don't have the room. I always carry on my camera gear, so compactness is usually pretty paramount.
Can't see ever taking the 85/2.4 Lox on a vacation due to the weight, unless it's someplace cool like Iceland where it could really be put to great use.
Also, the 85/1.8 has been with me for a long time and the Sigma was a recent (this summer) purchase specifically for backpacking. I've never taken the 85/1.8 backpacking as it's just too huge (relatively speaking).
I think that with a 3 lens kit like 24 2.5, 45 2.8 and 90 2.8, one is almost done for situations where light conditions are not critical.
This set weights around 700 gr..
Very high IQ with sharpness, rendering and bokeh seconds to few other sets.
Family pic from this morning with the 90/2.8 Sigma on my A7RIII.
This lens is more than sharp enough across the frame wide open for most uses - far and near.
Wide open.
...