15 hours ago
 |
Andy Westlake
Lives in
London, United Kingdom
Works as a
Technical Writer
Has a website at
www.dpreview.com
Joined on
Jan 28, 2008
|
   |
|
Connect with dpreview
|
JDThomas: I'm curious as to why this camera hasn't been officially announced in the US yet...
Nikon US are located in New Jersey and have been heavily affected by Sandy. It's quite possible they fully planned to announce in the US, but simply weren't able to finish the necessary preparatory work in time for this launch.
falconeyes: I think DPR does not understand an important point about lens tests: The influence of sharpening and a Bayer-AA filter. DxO therefore publishes full MTF results and provide valuable data. DPR however ONLY publishes MTF50 results where sharpening was off. This is not smart AT ALL!
Either publish MTF20 results (such as contained in DxO results), or use sharpening (such as done by photozone) or at least use test equipment without a Bayer-AA filter (D800E, K-5IIs etc.).
Otherwise results will clip at about 50% the sensor resolution and this will prevent the more expensive lenses to be compared effectively.
Sad DxO didn't consult DPR better ...
You seem to be labouring under the false assumption that I don't understand what you're saying. And I agree MTF50 is rather inefficient at comparing 'good' lenses on a fine level. But that's not the point - it *is* very good at distinguishing 'good' from 'not-so-good'. I also don't think it's remotely relevent how closely these MTF measurements correspond to those performed using an optical bench. The methodology we use is designed purely for internal comparison, and it works perfectly well in that respect.
falconeyes: I think DPR does not understand an important point about lens tests: The influence of sharpening and a Bayer-AA filter. DxO therefore publishes full MTF results and provide valuable data. DPR however ONLY publishes MTF50 results where sharpening was off. This is not smart AT ALL!
Either publish MTF20 results (such as contained in DxO results), or use sharpening (such as done by photozone) or at least use test equipment without a Bayer-AA filter (D800E, K-5IIs etc.).
Otherwise results will clip at about 50% the sensor resolution and this will prevent the more expensive lenses to be compared effectively.
Sad DxO didn't consult DPR better ...
MTF50 and MTF20 simply tell you slightly different things. MTF20 is close to 'resolution', MTF50 is more about perceived sharpness of the image as a whole. Is one automatically better than the other? That's a whole different question.
Your argument that MTF50 results should use sharpening seems odd though. All this does is inflate the MTF50 numbers in a systematic way. It doesn't make them any better comparable to each other.
There's certainly a strong case for using cameras without AA filters for testing lenses. But the problem is that you then have to do so for all systems, which isn't sensibly possible at the moment.
Ultimately no way of presenting data this complex is perfect. But if you really want to look at MTF20, you're free to follow the link to DxO's full data for each lens. That's what it's there for.
HubertChen: In the following question I am completely out of my field of expertise. I am wondering:
What is the point of this Nikon AF-S DX Nikkor 18-300mm F3.5-5.6G ED VR ?
It has VR, but it is not working on the long end in the shutter speeds you need. It works on a DSLR, but can not harvest the DSLR quality. It is an All in one lens, meaning people buy it for convenience, but it is so big and heavy that it is inconvenient. It is meant for the casual shooter, but it has so many issues which you need to be aware while shooting which usually only pros do. I hope I managed to write this without sounding cynical.
So here comes my question out of curiosity and ignorance: Would not be a dedicated small sensor super zoom camera serve the purpose better ? Or who would buy this thing doing what ? Nikon usually makes sense to me. What am I not getting ? Why reboot the lens review with this thing ? I am really curious :-)
"Why reboot the lens reviews with this thing?" We had to start somewhere, superzooms are popular, and we didn't know how it would perform before we started. That, after all, is what reviews are all about.
Simon Joinson: could someone please complain about us doing reviews?
@Nightwings - if you're not colour blind, how do you know that our gradient display is useless to those who are - and what would it matter to you anyway? And why do you assume colour-blind readers couldn't just read the graph? (I'm not criticising SLRGear's representation, BTW, I just don't understand your logic.)
dbscsp: All in all, that's a realistic review. In reality, it's a lot of fun to work with the 18-300. My hands are large enough to handle a D7000 with this kind of lens a day long. In-camera correction of distortion works OK, and in most cases, a small amount of remaining barrel distortion is benefitial on wideangle. Use DxO or whatever to completely correct this. Use 18-200 if your hands are not large. Use dedicated primes for low light or maximum quality. The stabilisation issue described here... don't know, I will do some tests. At 300 mm in many cases you will have mild to heavy blurring because of the atmosphere you are looking through, remember that.
I'm not saying you can't get good and detailed macro shots with this lens - after all one of the Pros in the review is 'Decent close-up capability'. I'm just saying that you can't assume the VR system will work anywhere near as well at these high magnifications. But this doesn't matter if you have lots of light, or use a tripod.
dbscsp: All in all, that's a realistic review. In reality, it's a lot of fun to work with the 18-300. My hands are large enough to handle a D7000 with this kind of lens a day long. In-camera correction of distortion works OK, and in most cases, a small amount of remaining barrel distortion is benefitial on wideangle. Use DxO or whatever to completely correct this. Use 18-200 if your hands are not large. Use dedicated primes for low light or maximum quality. The stabilisation issue described here... don't know, I will do some tests. At 300 mm in many cases you will have mild to heavy blurring because of the atmosphere you are looking through, remember that.
Atmospheric blurring will be visible in long shots, but in mid-range shots - maybe wildlife or sports - it's unlikely to be a problem. The IS problem around 1/80sec, however, can affect shots taken at any distance.
Remember also that very few IS systems work well for macro work at all. For this you need to be able to detect and correct translational movement, i.e. the lens's entrance pupil changing position relative to the subject. Only Canon's 'Hybrid IS' and Olympus's '5-axis' IS on the E-M5 are designed to counter this. So it would be a mistake to expect any superzoom to be much good for hand-held macro at slower shutter speeds.
Leonard Shepherd: Was the review good?
It says VR was not as effective at 300mm as shorter focal lengths, but used a constant focus distance, not allowing for the obvious magnification increase at 300mm.
Safe hand held speeds increase with image magnification - with any focal length.
Nikon indicate in the 105 VR instructions a subject 8 feet wide or narrower needs 1 shutter speed faster for sharp results. This implies a need for 2 speeds faster by 4 feet wide and so on.
Increased magnification at 300mm same focus distance relative to 28mm leaves less VR benefit available for camera shake reduction.
The review gives a fair summary of the VR results a relative novice might get. An advanced worker might know about the image magnification issue using VR.
Was there a slight tendency to misfocus at 300mm? The example shown used a fine detail subject with which all Nikon DSLR instructions indicate AF may not be particularly accurate. Was the reviewer paying enough attention to the quality of the AF target?
You've clearly not read the text about the stabilisation testing, which states that the 300mm tests used a longer distance. Nor have you noticed that the criticism is highly specific, and based on an inability to stabilize properly at 1/80sec. This is a clear difference to other lenses I've tested using exactly the same methodology.
As for misfocusing, that was consistent across a large number of real-world shots, using 2 different lenses. The sample shown simply illustrates it particularly well. But while the 100% crops show fine low-contrast detail, the image itself clearly has plenty of high-contrast patterns for the AF system to work with, and that's what I aimed at when focusing. (I've done this stuff before, you know.)
Phil Askey: Great to see lens reviews make a return to dpreview, to see photographs of Tower Bridge and the surrounding area again, the first review reads well and has lots of good technical data. Kudos Andy and the rest of the team.
Thanks Phil, and full credit to Josh for writing the widget - it's all-too-easy for the developers to go unnoticed.
rfsIII: The only criticism I have is that you shouldn't test lens distortion by taking pictures in a city that's more than 2,000 years old.
So many of London's buildings are beyond ancient and the rest are built atop the unsteady debris of thousands of years of human occupation; I cannot imagine that there is a straight wall or level floor anywhere within that venerable metropolis.
What you see as evidence of "severe" barrel distortion is more likely just a medieval wall bowed outward by the relentless forces of gravity and time.
Thanks for the laugh. Did you look at the test data and example images before posting?
NetMage: The Help for the new Lens Reviews doesn't explain the X axis for the chromatic aberration chart. What does the X axis represent?
Exactly the same thing as for the MTF graph - the two charts share the same x-axis.
Mattoid: Announcing the DEVELOPMENT? Anyone can do that! They don't even have to release them for the statement to be true. I can personally announce today that I am developing a medium format compact camera with a 16-800mm f1.0 stabilized zoom lens that weighs only 150g and fits in a shirt pocket.
What you're missing here, perhaps, is that Nikon won't announce the development of a lens they don't fully intend to deliver. Conceptually this announcement isn't so different to the lens 'roadmaps' offered by other manufacturers.
Sam Bennett: I would actually argue that while lens can't produce the same amount of "background blur" as an equivalent FF lens/body combo, it actually can provide as much or similar "subject isolation".
Subject isolation for me isn't just about DoF - it's also about the focal length you're using, how it influences your distance to your subject and hence the perspective of the shot and how much background "clutter" is present. With the same amount of DoF present in two photos shot with different focal lengths (and different distances to maintain a consistent subject size) the wider focal length will result in less subject isolation and more background clutter.
In my experience, focal length is actually the more important factor in terms of isolation - that's why in practice the 45-200mm actually provides a good amount of isolation at long focal lengths, despite it's relatively slow aperture.
The fundamental problem with this argument is that all the stuff you're talking about will look almost exactly the same shooting at 200mm F5.6 on full frame as at 100mm F2.8 on Four Thirds (or 135mm F3.5 on APS-C). So if you're carrying a full frame camera with a 70-200mm F2.8 (or even F4) and a Micro Four Thirds camera with the Panasonic 35-100mm F2.8, the only situation you can possibly construct where the MFT camera has any advantage in terms of subject isolation is if the subject is too close for the FF camera to focus.
Trollshavethebestcandy: First!
Does this do better with 5 axis ibis or the lens stabilization?
@deep7: "The advantage of having it in the lens is it keeps the viewfinder image steadier." This is a key advantage in SLRs, certainly, but it isn't necessarily any more with electronic viewfinder cameras. In fact the E-M5's in-body system can be set up to give a stabilized viewfinder image on half-press.
Andy Crowe: Will the new lens test data be directly comparable to the older lens test data, or will the new lab affect the sharpness scores?
It'll be roughly comparable, in principle, but it won't be presented in the same widget and we won't be allowing direct comparison. Instead the aim is to include properly-comparable data for similar lenses to whatever's being reviewed.
jkrumm: One strange thing about Dxo lens tests, unlike lenstip and slrgear, they rate 4/3 and m43 lenses from poor to middling, at least in resolution. Not sure what the reason is. Maybe they are not taking sensor size into account? Will be interesting to see what they get out of the new Oly 75 1.8, which appears to be about as flawless as lenses get.
DxOMark's lens scoring gives lower marks to Micro Four Thirds simply because it's a system score that *does* effectively take sensor size into account (and considers larger sensors to be better). The explanation is on their website.
Trollshavethebestcandy: First!
Does this do better with 5 axis ibis or the lens stabilization?
By definition, 5-axis, as in-lens can't correct rotation.
ProfHankD: Gordon Bell at Microsoft Research has been pushing this concept for years, and there have been various prototypes -- including SenseCam out of Microsoft Cambridge (which claims to have started in 2003):
http://research.microsoft.com/en-us/um/cambridge/projects/sensecam/media.htm
Many of us have done smarter automatic capture control in existing cameras -- I've even done it using CHDK inside of Canon PowerShots.
If you want a camera as an event recorder, rather than a medium for producing art, this is a commercial step in the right direction.
Autographer is in fact directly based on Microsoft SenseCam technology.
Eugene CH: What about Battery???
It uses a built-in battery (with USB charging) that's supposed to last an entire day - not impossible given that there's no rear LCD screen to power.
Timmbits: Hey DPR: shouldn't the the image for the Panasonic 1/1.7" crop be the same height as the other 1/1.7"? whereas, in 16:9 is is shorter but as wide?
Why is this, all of a sudden, looking very biased towards canon? advertising revenue, or kickbacks for the article?
The LX7 uses an unusual sensor - it's 1/1.7" type, but 3:2 aspect ratio (so lower but wider than a conventional 4:3 1/1.7" sensor). The 4:3 frame that we're showing here is then cropped-in from the edges of sensor, leading to a smaller active area than Canon's 1/1.7" sensor.
(The giveaway here, BTW, lies in the two lens's focal lengths. Both offer 24mm equivalent wideangle, but the S110's is 5.2mm while the LX7's is 4.7mm. This means that the LX7 must be using a smaller active area of the sensor.)
So no, this isn't biased towards Canon, it's just the way the two cameras work. Your allegation that we're unfairly manipulating the information presented for 'advertising revenue' or 'kickbacks' is therefore utterly unfounded, and simply reflects your own lack of understanding. Your apologies are accepted in advance, BTW.