itsastickup: Totally hopeless bokeh test.
For me it's a deal-breaker: bad bokeh means unusable portraits.
Closing down the aperture can get rid of hard edged rings but also double-line 'nissen' bokeh which make for poor/disturbing bokeh as in the two pics (I would call this bokeh 'poor' and unusable). Typically I shoot f1.4 lenses at f2 for bokeh reasons. But in addition, you are more likely to get rings where the focus is at a distance and the blur moderate, as with these pics, while at close portrait distances the rings may not be so hard; which is the type of photos I am most interested in.
On top of that: a closed down aperture can be useful (one doesn;t always want obliteration-bokeh) but the bokeh can deteriorate.
So effectively we need a range of apertures and distance to know what the bokeh situation is.
Bokeh is so neglected that I have to do a lot of research to work out whether a lens is any good. It's a pain.
Here is a comparison that actually means something:
Just because all wide primes have bokeh problems does not mean that they are equally bad.
ulfie: DxOMark test results/specs are great yet the actual image quality even for the "originals" are really no better than say the Panasonic 20 f1.7's which costs about 1/3 and is only a tad slower f-stopwise and a tad tighter FOVwise.
And who said anything about speed? Lenses do not have speed.
Until you understand that ISO 400 on 4/3 is equivalent to roughly speaking ISO 1600 on FF, you are going to remain confused.
A 20/1.7 lens has much smaller physical aperture than a 35/1.4 lens, as simple as that. The ratio is about 2.17 stops. This includes DOF, and total light captured. The total light determines the shot noise with the SAME speed.
You are confusing ISO, say 1600, with one format with ISO 1600 with another.
Photomonkey: For all the quacking about bokeh, I notice that very few of the samples can really muster much of an OOF background despite shooting wide open and being close to the subject.Those preoccupied by that issue should look to longer lenses.
Actually, most of the portraits show well blurred background, so you are wrong. The question is about its quality, not quantity. DPR removed the bad bokeh examples a few weeks ago. Any time I mention this, my comment gets deleted by them.
facedodge: If it weren't for the onion bokeh, this would beat the 35L.
Actually, double lines (due to onion bokeh) are very well visible without going to 100%. The 35L does this as well, but the Sigma is worse. There were direct comparisons in the Canon forum.
The Panasonic is MUCH slower. About 40/3.4 on FF.
The 35L has problematic bokeh as well but not as bad as the Sigma.
Erik Magnuson: I think this score will correlate pretty well with lens sharpness in general, i.e. a lens with a P-Mpix 15 (of 18) will be noticeably better than a lens with a score of 8. However, it may not correlate as well for small differences, e.g. a score of 14 and a score of 15 may mean lenses that merely have different strengths (i.e. center vs. edge, best vs. wide open performance) where the weighting used does not match the user preference.
If a user merely wants to know "what's an excellent vs. merely good lens", then this number will be useful. Buying the lens with the highest number is rarely going to be "wrong" but it may not be optimal. If you want to know "what the best lens for 35mm f/2 corner to corner" then this number will not be as useful. Digging down into the detail graphs may help, but if you have specific questions you should not expect average numbers to be the answer. Also remember YMMV so small differences in measurements may not be significant in field use.
They may not get a good lens. It is easy for the manufacturers to compromise the design to score higher at some aperture, in the center. Fast lenses have to deal with different challenges and that often hurts a bit sharpness at f/4-f/5.6 across the frame.
If you do this, you are giving an incentive to the manufacturers to compromise the design. You are the most powerful photo review site on the web, owned by the biggest photo equipment retailer. If the average Joe makes his purchase decisions based on that stupid number, the manufacturers would have to make him happy.
Entropius: There are a lot more things to a lens's performance than a one-dimensional measurement. What does this scheme make of a lens like the Panasonic 14/2.5, which is extremely sharp but has high chromatic aberration? What about the Canon 70-200/2.8 mk1 at 200/2.8, which has poor microcontrast but resolves a good amount of detail? The high sharpness and resolution, but bizarre bokeh, of the Zuiko 50-200?
What about a classic lens like my Olympus OM 50/1.8, which is surprisingly highly-resolving wide open in the middle of the frame (great for portraits) but without the high acutance of modern asphericals, and with a strong dropoff in sharpness toward the edge?
There's a whole lot more to a lens's performance than one number. That's why the old DPReview lens reviews were great, with the little checkerboards: you can see for yourself what the point spread function at different places in the frame is.
There is more they do not tell me. How they get that number. Where is the human vision involved in all this. It is an documented metric. It is part of a trend to dumb down the data to please the average reader.
"The Perceptual Megapixel measurement allows photographers to step away from manufacturers' announcements about resolution and to answer a key question when the time comes to change equipment: would it be better to buy a new camera or a new lens?"
They have been testing lenses+cameras and reporting combined resolution even before this. My guess is that they use the inverse square law to blur further the measurements according to what they think the human vision does for a fixed size and viewing distance. Many people can understand what MTF-50, for example, is; but almost nobody would understand the new metric - especially when they do not even bother to explain it.
Sharpness at what aperture? Where - center, borders?
Nonsense. What they need to do is to report what they measure, not to encrypt it.
Let us hope that this is just a marketing trick and they will still report the MTF data.
Donald Duck: "The larger the sensor the easier it is to get very shallow depth of field effects. The flip side of this is that it can be harder to get everything in focus when you do want to."
Incorrect, there is no flip side. You just need to use equivalent f-stops (and equivalent ISO, if needed). When you do want the same DOF, you should know a thing or two about that.
jj74e - You do not change anything. You do not start with a crop body, choose the settings, then switch to FF and "change" the settings. You directly shoot with FF. I thought that you have done this.
"The larger the sensor the easier it is to get very shallow depth of field effects. The flip side of this is that it can be harder to get everything in focus when you do want to."
Some of your f/1.4 images are a bit soft. Did you use a tripod?
Lu Heng: Is it just me, or your test pictures are getting better? I enjoyed some of them as photographs.And I think the lens are not bad at all.
The bokeh here: http://www.dpreview.com/galleries/reviewsamples/photos/2341157/pt3c9403?inalbum=sigma-35mm-f1-4-dg-hsm-preview-samples is terrible. The Canon would have problems there as well.
DPR - did you use a tripod here and in the other 1/30 sec shots?
DuxX: Mr. Cicala, sharpness and weight are not the only criteria for the evaluation of a lens. What about bokeh rendition? As far as I can see this Sigma isn't even close to Nikkor or Canon equivalents.
But without doubt. This is one great lens from Sigma and very good offer.
The number of blades does not affect the bokeh wide open, and is not a primary factor of its quality even when stopped down. It has more to do with the lens design and optimization. The lenses with the best bokeh are the ones which are not the sharpest. The sigma seems to go in exactly the opposite direction.
Kodachrome200: I dont get these crazy expensive m43 lenses. Small formats should not be a premium format. These lenses cost more than FF lenses. thats nuts
"Canon's new 35mm f2 was closer to $900..."
That is 17/1.0 on m43, with resolution on FF that the m43 can only dream about.