p.26 #1 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
The end result is what should be judged. You should inspect the SW-corrected output to see if the final quality is good or not. Sure SW correction will cause some degradation in IQ, but optical correction also makes the lens design more complex and can result in compromises of its own. It is quite possible that relaxing the need to optically correct distortions can result in an overall better optical design, which even after SW correction gives overall better quality.
p.26 #3 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
Fred Miranda wrote:
From the review. It's indeed high barrel distortion:
It looks crazy but, regrettably, it's not at all unheard-of. It reminds me very much of the uncorrected barrel distortion on the wide end of the Sony 18-135 zoom for APS-C. And the 24mm Tamron f/2.8 (and the even-worse 20mm). All of which are comparably hugely distorted... I imagine there are a few other lenses that show it, especially compact zooms(?). I suspect that just completely giving up on getting distortion right in-the-lens is a big help when trying to design a small or inexpensive wide-angle lens around 24 or 28mm.
"Just use the software distortion correction," some might say. But how do you frame your shot well at the time of taking it? You aren't taking what you think you're taking, you can't quite be sure what's in the frame and what isn't, and you can't be sure the lines and shapes you're seeing in the viewfinder are what you're really capturing, and inter-relate in the way you think they do... you can't be confident of what your composition really consists of. You can't be quite so intentional in creating your photo, there's a chaos and pot-luck factor.
To me that's the most offputting aspect of heavy distortion. The effective loss of sharpness is just an additional yuck-factor, and relatively minor.
It would be less of a problem if the live / viewfinder view also applies the distortion corrections in real-time, to the view you see as you're composing your image... not just to the saved JPEG or RAW afterwards. I don't think cameras currently do that, do they? But I'd be happy to be wrong about that!
p.26 #4 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
GHarris wrote:
It looks crazy but, regrettably, it's not at all unheard-of. It reminds me very much of the uncorrected barrel distortion on the wide end of the Sony 18-135 zoom for APS-C. And the 24mm Tamron f/2.8 (and the even-worse 20mm). All of which are comparably hugely distorted... I imagine there are a few other lenses that show it, especially compact zooms(?). I suspect that just completely giving up on getting distortion right in-the-lens is a big help when trying to design a small or inexpensive wide-angle lens around 24 or 28mm.
"Just use the software distortion correction," some might say. But how do you frame your shot well at the time of taking it? You aren't taking what you think you're taking, you can't quite be sure what's in the frame and what isn't, and you can't be sure the lines and shapes you're seeing in the viewfinder are what you're really capturing, and inter-relate in the way you think they do... you can't be confident of what your composition really consists of. You can't be quite so intentional in creating your photo, there's a chaos and pot-luck factor.
To me that's the most offputting aspect of heavy distortion. The effective loss of sharpness is just an additional yuck-factor, and relatively minor....Show more →
what you see in camera will already be digitally corrected for distortion so I don't understand your point regarding composition..
As already mentioned above, undercorrecting distortion optically and relying on software correction is just another compromise in optical design. Yes correcting in software is not lossless, but correcting it optically is also not lossless because you have to accept other compromises (e.g. increased size, weight, cost).
That being said the lenstip test results for the 24mm are not great.. which is a shame because it would be nice in a landscape kit together with the 40/50 + a longer lens.
p.26 #5 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
fjablo wrote: what you see in camera will already be digitally corrected for distortion so I don't understand your point regarding composition..
Is it? That's what I'd like to know for sure. I'm not talking about "is the distortion corrected when you review the image after taking it?" - when you review the JPEG or corrected-RAW on the LCD or viewfinder after taking the shot. I'm talking about the live view as you're setting the shot up. Is that distortion-corrected? I've thought it less likely to be so, since it's more computationally demanding to correct that live view nonstop than to correct a single image as it goes through its in-camera processing and save-to-card stages. I have up to now avoided these highly distorting wide-angle lenses so I can't test it for myself.
P.S. I notice that the Sony 28-60 zoom has somewhat heavy barrel distortion at 28mm, too (one source, for example). Not as heavy as this 24mm prime, but still substantial. I've always considered it a deal-breaker, not being sure that what I would see in the viewfinder is always what I'll really be getting. Would be happy to be mistaken about it.
p.26 #6 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
GHarris wrote:
Is it? That's what I'd like to know for sure. I'm not talking about "is the distortion corrected when you review the image after taking it?" - when you review the JPEG or corrected-RAW on the LCD or viewfinder after taking the shot. I'm talking about the live view as you're setting the shot up. Is that distortion-corrected? I've thought it less likely to be so, since it's more computationally demanding to correct that live view nonstop than to correct a single image as it goes through its in-camera processing and save-to-card stages. I have up to now avoided these highly distorting wide-angle lenses so I can't test it for myself.
P.S. I notice that the Sony 28-60 zoom has somewhat heavy barrel distortion at 28mm, too (one source, for example). Not as heavy as this 24mm prime, but still substantial. I've always considered it a deal-breaker, not being sure that what I would see in the viewfinder is always what I'll really be getting. Would be happy to be mistaken about it....Show more →
The reviewers said the Distortion Compensation was forced on with these which means it is applied to the live view. Sony seems to be making that menu option void. I would like to be able to disable it.
The third-party lenses don't force Distortion Compensation, and it is not applied to live-view even when enabled.
p.26 #9 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
philip_pj wrote:
Whoa indeed. We always want to quantify things, this is just extraordinary. The Lenstip review gives us the figure for full frame distortion - it's 8.62%! There is always a price to be paid when corners are cut - no pun intended.
'If it weren't for the Sony GM lenses, Sony would be behind Sigma'
Just the entree, literally. Just wait until the main course when Sigma brings the full force of their design expertise to bear on premium level optics for Sony, as they surely will. They have done the hard yards, the next releases will be something special.
I agree Philip_pj, I think Sigma has beat Sony at its own game here: the 'I' series lenses ooze class compared to these new compact offerings from Sony - with much smoother bokeh and rendering; sure, they're bigger and heaver - but represent a finely balanced set of compromises that those with trained eyes can clearly see. The Sony's look like a 'bridge too far' to me, with a slavish adherence to a minimum size and weight ethic, that has resulted in too much compromise in LOCA and rendering (and distortion in the case of the 24mm).
That said, context is everything - and I think these new Sony compacts will sell well as add-on sales to the less demanding demographic they are targeted for: need something wider for your A7C? here's the 24mm Sir; need something faster than the 28-60mm zoom - which would you prefer, 40 or 50mmf2.5?
p.26 #10 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
Petegh wrote:
That said, context is everything - and I think these new Sony compacts will sell well as add-on sales to the less demanding demographic they are targeted for: need something wider for your A7C? here's the 24mm Sir; need something faster than the 28-60mm zoom - which would you prefer, 40 or 50mmf2.5?
I agree but the pricing doesn't seem to fit into that market.
p.26 #11 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
I'd be interested to see a comparison of the 24mm to a good legacy 24mm f2.8, say a FD. I'm not sure the 'we had to compromise' fits in with the price despite the size. Distortion of that level seems to smack of setting a price and working out how little you can get away with providing for said price.
p.26 #13 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
Is there a comparison MTF review from the same reviewer of both the Sigma 24/3.5 DG DN and the Sony 24/2.8 G?
It is not given that correcting distortion digitally is worse than doing it optically, it’s the total package that counts. Think of teleconverter and how just a linear magnification can drop image quality. Why would a optical correction of barrel distortion fare any better with the image quality of a wide angle? Why would it not come at the expense of resolution, since it too will have to stretch the image, just like a digital correction would.
p.26 #17 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
NicholasRab wrote:
No question, but at least in the lens tip test the edge resolution really suffers, which could in part be due to this distortion correction.
My understanding is that lenstip results are based on uncorrected RAW - which is why they can see the distortion at the first place. So the edge weakness they see should not be due to digital correction.
Ironically it is quite possible that their edge sharpness results look worse because of using un-corrected output. Lenses with SW-correction based designs tend to have slightly wider field of view when uncorrected and some of that gets cropped during correction. So what lenstip is testing includes some edges of the image which will be cropped away in the final shot. So that could contribute to the weaker performance they report. My contention still is that they should test the lens as it has been designed, including all the SW corrections and then report the findings on the final result.
p.26 #18 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
Most people have probably seen this review, but it does include corrected comparisons between the Sony 24 2.8 and Sigma 24 3.5 wide open. At least at centre and corner they're similar, the Sony might be a bit better in the corner (corrected in camera for distortion and vignetting). As he points out, even after correction the Sony is slightly wider.
p.26 #19 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
curious80 wrote:
My understanding is that lenstip results are based on uncorrected RAW - which is why they can see the distortion at the first place. So the edge weakness they see should not be due to digital correction.
They use Tiff converted via dcraw, so could be uncorrected I guess. But they analyse distortion with and without profile applied, so don't know about that.
However, the specifically call out chromatic aberrations as driving factor behind the weak MTF results on the edge of the frame:
"The performance on the edge of the frame is a separate story, though. The lens fares atypically because the best results can be observed not on stopping down but close to the maximum relative aperture. It's a rare thing, appearing when e.g. the course of some of off-axis aberrations is very specific. In this case lateral chromatic aberration increases quite fast on stopping down, limiting efficiently the increase of the vertical component of MTF values – it will be described in more detail in the following chapter."
p.26 #20 · Pre-order Sony G lenses: FE 24mm f/2.8, 40mm f/2.5, 50mm f/2.5
curious80 wrote:
My understanding is that lenstip results are based on uncorrected RAW - which is why they can see the distortion at the first place. So the edge weakness they see should not be due to digital correction.
Ironically it is quite possible that their edge sharpness results look worse because of using un-corrected output. Lenses with SW-correction based designs tend to have slightly wider field of view when uncorrected and some of that gets cropped during correction. So what lenstip is testing includes some edges of the image which will be cropped away in the final shot. So that could contribute to the weaker performance they report. My contention still is that they should test the lens as it has been designed, including all the SW corrections and then report the findings on the final result. ...Show more →
Since it’s uncorrected raw, it’s not corrected for lateral CA either, which is easy to correct in digitally.
what you see in camera will already be digitally corrected for distortion so I don't understand your point regarding composition..