Great rendering! Another winner for Sigma.
@Richard Butler: Excellent article!
This topic has confused many minds and it is awesome to see it explained that well with such accuracy, conciseness, and supported with actual images.
Only two remarks:Page 3: You speculate that the more contrasty blur discs from the 1"-type sensor images come from lower LCA. Is it not much more likely, that the claimed "f/1.2" of the lens are not a true "f/1.2"? Have you tried to compare the "f/1.2" exposure to a "f/4" exposure and see whether you actually get the expected increase in intensity? The lens itself may even be "f/1.2" but extreme incident angles on the sensor could mean that the latter actually just "sees" an effective "f/1.8", or so. Camera manufacturers combat this "blindness" of sensors to very wide apertures with hidden ISO increases.
Page 4: You write "generally larger pixels...will tend to give cleaner images", but in one of your comments you demonstrate that you know that the pixel size is immaterial. I'd adjust the text.
ttran88: People all along were blaming Sigma and Tamron lenses for AF accuracy. But now we know it's really the body that they were mounted to.
@Rishi: How can you state that on-sensor PDAF is immune to problems with rays from the periphery of the lens?
On-sensor PDAF still uses the same principle of using different rays and if they are not from the periphery of the lens, the AF precision will be poor. PDAF needs a wide baseline for good precision and that implies involving the lens periphery.
Are you saying that on-sensor PDAF is imprecise and this is accounted for by final CDAF steps?
Class A: Could you not have chosen an AF point manually in your first test? The vast amount of shots were unusable with the nose being in focus but not the eyes.
It would be extremely disappointing if this were everything the camera could achieve with third-party glass.
BTW, if a lens requires AF microadjustments with a regular PDAF system then it will need the same adjustment with a sensor-based PDAF system. The only problem the latter can avoid is a constant offset which a regular PDAF system can compensate with one global adjustment as well.
If the A7R II is immune against FF/BF issues then because it employs CDAF in the final stages of focus acquisition.
@Rishi: Thank you for the clarification.
In my view the "focus on one eye" technique (exclusive to Sony lenses) demonstrated in the second video worked very well but your first video either demstrated a failing technology or a suboptimal use of it. The vast majority of photos were unsuccessful.
I don't think that pre-selection of a focus point hinders creativity in particular when there are 399 points to choose from. It is more convenient to have the camera follow a focus target automatically but the alternative isn't stifling in any way.
Guys, different path lengths ("misalignment") between a separate PDAF system and the sensor define one constant adjustment for all lenses. Such a mismatch is easy to figure out and fix. The difficult cases are caused by lens aberrations and will affect sensor-based PDAF as well.
BTW, I have never heard that temperature fluctuations can influence alignment dynamically.
Could you not have chosen an AF point manually in your first test? The vast amount of shots were unusable with the nose being in focus but not the eyes.
NomadCZM: Honestly, I think as technology improves, mirrorless will eventually replace dslr..
The EVF will eventually be as good or better than an OVF, tracking will eventually be on par with DSLR,.. Even the battery life issues can be solved.. So unless DSLR make some huge leap, I don't see them remaining the GO TO for pro's.
It most likely won't happen this year or next, but it will. And honestly, what's so bad about that?
Why do people get so upset when something can be eventually made better, you would think that's a good thing. So much brand/technology loyalty, people some how fail to realize it's not a bad thing.
I have two DSLR's because currently they do what I need them to do better.. If something can replace that with better features and performance AND be able to use my old lenses, that's great.. I would love it.
Some people are not upset about "good" being replaced with "bettern but rather about the characterisation of "mirrorles" as "better".
Even when EVF will get better, they will never allow test framing an image without turning on the camera and I don't see how they would ever be able to avoid white-balancing issues.
A mirrorless camera will always have to keep its sensor active for framing and focusing, thus consuming power and heating up, leading to more noise.
A mirrorless design will always need to deal with more oblique angles from lenses.
Lenses with good light gathering capabilities for nice bokeh and keeping shutter speeds low and noise down, need to be large. The small and lightweight mirrorless cameras of today will always be a mismatched partner for lenses that are used in a wide range of serious photography. A 70-200mm f/2.8 zoom is not that big and heavy because the camera behind it has a slightly longer registration distance (realised by a mirror box with zero weight). If you want the same optical performance for a mirrorless camera, the lens has to be just as big and heavy, there is no way around that. Now if you need such lenses, why wouldn't you match them with a camera that has the ergonomics to properly handle the camera/lens combination?
If a mirrorless camera is supposed to be smaller and lighter than a DSLR, it cannot feature as large a battery and in combination with the increased power consumption (main sensor + EVF), it will always provide less stamina.
A good compromise between DSLR and mirrorless cameras are cameras with decent grips and reasonable size to allow adequate handling of serious lenses, and have a hybrid VF, combining an OFV with EVF benefits. Now such a camera will not be mirrorless.
This promises to be another great Sigma lens. Hopefully Sigma will consider offering it in K-mount as well after a while. The lens would be a great companion to the announced Pentax FF camera (with built-in image stabilisation).
Good to see some balanced discussion in the article. DSLRs are by no means the "dinosaur technology" that they are often characterised as.
One point that rarely gets made is that image quality is largely determined by the lens and that lenses that provide high-quality images even in bad conditions tend to be large (that's just physics that no engineering can get around). A tiny mirrorless body more often than not then just provides an inadequate grip for such lenses, yielding a combo that is imbalanced and not well-matched, with the overall size of a quality kit negating the small body size and weight advantage a mirrorless camera offers.
Surely mirrorless cameras have their perfectly matched application areas, but so do DSLRs. This sustained hailing of mirrorless cameras as the future technology that is still misunderstood by customers got old quite a while ago already.
audiobomber: Hopefully the K-3 II review will be written by Dan Bracaglia and Rishi Sanyal. They did a great job on the A77 II review, especially the focus testing. The K-3 was sabotaged by all the wrong settings for the AF-C test.
@Rishi Sanyal: Just wanted to give you a big thumbs up for showing interest in what may have gone wrong in the K-3 AF test.
Unfortunately, even advice from Ricoh directly may not have been optimal, depending on who your contact was.
I personally don't know the test well enough to comment any further, but I nevertheless wanted to applaud your attitude and would encourage you to listen to falconeyes. Please contact him, if his suggestions were not 100% clear to you. He really knows what he is talking about and you have a chance to become an industry leader if you can come up with an adequate CAF testing approach.
Class A: "...we're disappointed by ... the disappearance of the built-in flash..."
Really? P-TTL is so bad at controlling more than one off-camera flash, why would you put up with such a bad optical triggering system (that does not even support HSS through the built-in flash and requires line-of-sight)?
Using a dedicated radio trigger on camera with direct control over manual flash levels right from the camera position is so much better and adequate for a camera in this price bracket.
@falconeyes: I share your concerns. I surmise GPS was considered a good fit to the profile of an "outdoors" camera.
Still, I would have left GPS to an external unit and would have thought Wi-Fi integration to support composition in the field, photo sharing, and image transfer, made more sense.
I also agree that Pentax should have developed a radio-based flash system by now. However, at least third-party solutions are available already. While they don't offer automatic flash exposure support, they nevertheless provide competent control of multiple off-camera flashes with predictable results.
Even flash systems better than P-TTL are limited when they use optical triggering.
WIth a pop-up flash you won't have much reach and will struggle with flashes within light modifiers (e.g., softboxes).
Using a dedicated master flash on camera increases reach but still requires line of sight. Joe McNally only puts up with this nonsense when he promotes CLS.
So no, any way you dice it, a measly pop-up flash is justified on a point and shoot, but not on a top class DSLR.
Sony and Canon have had models without a pop-up flash, so Ricoh is not exactly breaking new ground here.
"...we're disappointed by ... the disappearance of the built-in flash..."
chillgreg: Does the pixel shift technology mean that this camera can emulate a Foveon sensor's output?
Yes, with improved sensitivity and better colour separation.
However, this will only work optimally with the camera on a very stable tripod and a completely still scene.
A Foveon sensor does not require such constraints to capture full colour information for each pixel.
Can't help but think adding Wi-Fi would have been more useful than GPS (that hopefully can be turned off in order to save battery life).
Otherwise, it is a very nice improvement on an already great camera.
P.S. I'd never miss a built-in flash. I used the one on my K100D as a noob a couple of times and am not even sure my K-5 II has one. :)
I trust this will be a great performer and AF won't be missed on a macro lens. The version for Pentax supports automatic metering so it pretty much ticks all the boxes.
I wish, however, it would have been a 200mm f/4 design. That would have provided more stand-off distance. Hopefully, Samyang will offer such a design in the future.
In my experience, the automatic generation of a camera profile from a shot of a Colorchecker does not work optimally.
Perhaps Adobe have optimised this process for the CFA characteristics of Canon cameras, but when using this approach with my Pentax, I got different colours but not all of them were correct.
For anyone needing 100% accurate colours, it would be worth their effort to take a picture of a Colorchecker in every condition and then tweak the image later (through a camera profile or in post) so that the colours are reproduced faithfully.
For everyone else, I'd say it is much more fruitful to tweak the colours in post to your liking and where you start from isn't terribly important.
Thanks a lot for this very interesting inside view on a factory producing some of the most exciting DSLR lenses.
I have number of Sigma lenses that are not just excellent performers but are also extremely affordable.
I hope Pentax's FF camera will create some new momentum for their camera sales so that Sigma will consider producing more of their lens models in K-mount again.
A Pentax DSLR combined with one of the excellent Sigma lens models provides incredible "value for money" that is not rivalled by any other combination*.
* depends on application areas, of course.
It will be interesting to see how well "internet memory" will work.
Can Triggertrap really expect to continue as a company as if nothing had happened? Will they be able to continue selling products or will they face huge adversity from many disappointed backers recommending potential customers to stay away from Triggertrap products?
Community power can cut both ways, these days...
While kickstarter projects in general cannot be expected to always result in the desired product, I'm surprised to see this particular endeavour fail. After all, this wasn't really exploring uncharted territory in terms of technology. Similar products already exist and it seems that a responsible use of the money should have resulted in a product.
Big kudos to Richard Butler who mentioned the upcoming Pentax FF model at the "Gear Day" feature at CreativeLive!
Thanks, Richard, for supporting diversity, competition, and speaking of Pentax as a manufacturer of "photographer's cameras" with "well-positioned controls".
You made this Pentaxian's day. :)