RSS
 

Posts Tagged ‘analysis’

Analysis predicts drone Remote ID will cost 9X more than expected, DJI urges FAA to reconsider ruling

06 Mar

After numerous delays, the Federal Aviation Administration (FAA) released its Notice of Proposed Rulemaking (NPRM) for the Remote Identification of Unmanned Aircraft Systems at the end of last year. The 60-day public commenting period closed this past Monday, March 2nd, with over 52,000 comments submitted during that time.

DJI, the world’s leading drone manufacturer, has supported the need for Remote ID since 2017. In the interest of moving the industry forward, a proper ruling would allow flights at night, over people and beyond visual line of sight. When the NPRM was released, however, DJI publicly chastised the FAA for not incorporating recommendations submitted by the 74 stakeholders that make up the Aviation Rulemaking Committee.

In its 89-page comment to the FAA, DJI cites independent economic analysis that was prepared by Dr. Christian Dippon, Managing Director at NERA Economic Consulting. The study concludes that the societal costs associated with the Remote ID NPRM would total $ 5.6 billion. This makes it 9 times more costly than the $ 582 million the FAA predicts for the next decade.

“I worry about an impact on innovation, with fewer people interested in using drones,” – Brendan Schulman

The long-term ramifications, should the Remote ID NPRM pass in its current form, will extend beyond financial burdens. ‘I worry about an impact on innovation, with fewer people interested in using drones. Our economist’s survey found at least a 10% drop in drone activity if the proposal were implemented, but I think it could be much higher as the full impact is felt by operators,’ Brendan Schulman, DJI’s Vice President of Policy & Legal Affairs, tells DPReview.

Remote ID, simply put, is a digital license plate for drones. It allows authorities to identify the location, serial number, and a remote pilot’s identity in near real-time. The FAA is proposing that almost all drones should transmit that information over wireless networks to a service provider’s database. NERA’s study concludes that the monthly cost of a network-based service for a remote pilot would be $ 9.83 instead of the FAA’s $ 2.50 estimate.

A few vocal critics have suggested that DJI’s involvement in drafting Remote ID rulemaking has served their own interests, and that regulations will amount to a multi-billion dollar gain for the company. ‘The critics missed the context and history. Since 2017 we knew Remote ID was inevitable as a government mandate, and have been advocating for the best possible result for all drone users: low costs and burdens. Everything we have done on this topic has been focused on those goals. Keep costs low and respect drone user privacy. For example, in March 2017 we released a whitepaper strongly advocating for pilot privacy,’ Schulman explains.

DJI has advocated for a ‘drone-to-phone’ solution that provides Remote ID information on common smartphones without burdening drone operators with any extra costs or effort. DJI says that its solution is cheaper and easier than what the FAA is proposing. Any new ruling on Remote ID will not likely take effect until 2024.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Analysis predicts drone Remote ID will cost 9X more than expected, DJI urges FAA to reconsider ruling

Posted in Uncategorized

 

Analysis: With new AF firmware, Nikon’s Z6 & Z7 just got more competitive

21 Jun

$ (document).ready(function() { SampleGalleryV2({“containerId”:”embeddedSampleGallery_3673238986″,”galleryId”:”3673238986″,”isEmbeddedWidget”:true,”selectedImageIndex”:0,”isMobile”:false}) });

Nikon has released a much-anticipated firmware update intended to improve autofocus performance from it Z-series mirrorless cameras. Specifically, Nikon promised improvements to autofocus in low contrast and low light shooting situations, but perhaps the most anticipated feature this firmware brings is continuous Eye AF.

As soon as the new firmware became available we updated our Z6 and Z7, and we’ve shot extensively with the updated cameras over the past few weeks. I routinely use a broad spectrum of Eye AF implementations, particularly Sony’s best-in-class one, and I can state up-front that I’ve come away very impressed. But not without a fair share of caveats and suggestions for improvement.

Does the update address the autofocus usability and performance issues we detailed in our Nikon Z7 review? And how does it stack up under the microscope against the best Eye AF systems? Read on to learn more.

Table of contents:

  • Identifying eyes
  • Focusing on unintended subjects
  • Choosing your subject
  • Eye AF accuracy
  • Conclusion
  • Footnote: compared to Canon EOS R / RP
Eye AF on the updated Z6 is so effective that it even found the coach’s eye underneath his cap.

Nikon Z6 | Nikkor 70-200/2.8E FL ED VR

Identifying eyes

When it comes to finding eyes and tracking them, the updated Z cameras compare very favorably to the a7R III.

The Z6 and Z7 are able to routinely identify eyes behind glasses, and even when all you see is a person’s profile, as in the shot above, or this one. It then holds onto the eye nearly as tenaciously as the a7R III. As it turns out, it’s often even better than the Sony at initially identifying eyes in sideways-turned or downward-looking faces. That said, once the a7R III has identified the eye of your subject, it tends to be better at sticking with it, even if their face is turned to a profile. Faces and eyes have to be a bit larger in the frame before Eye AF kicks in on the Nikon, but in practical use this difference is negligible.

A situation where the Nikon Z6 found my daughter’s eye, but the Sony a7R III did not. Unfortunately, as you’ll see in this series of shots, accuracy of focus was comparatively poor, an issue that particularly manifests itself as light levels drop.

Nikon Z6 | Nikkor 35/1.8 S

But simply identifying eyes and faces isn’t all that useful, in and of itself. Accuracy of detection (i.e., not seeing faces where there aren’t any) and accuracy of focus are the more important factors. Let’s investigate.

Focusing on unintended subjects

A downside of Nikon’s eagerness to find faces and eyes means sometimes it finds them where they don’t exist, like in this featureless shag carpet or these trees in the background. When this happens, even if a real face re-appears the camera is often reticent to refocus on it, especially if it’s at a vastly different distance-to-camera. Often the camera will hang on to the mis-identified ‘face’. In comparison I’ve never found Sony cameras to identify a non-human subject as a face or eye.

A downside of just how easily Nikon Z cameras find faces and eyes is that quite often it finds faces in objects that are not faces. In this case, the camera detected a face in the ladder in the background. It was then slow to re-focus on my daughter when she revealed her face again, and in some instances remained on the ladder.

Nikon Z6 | Nikkor 35/1.8 S

This problem with false positives could be worked around if the Z cameras worked more like (recent) Sonys and Canons in which Eye AF works over the top of their subject tracking modes, meaning you can point your initial AF point at your subject to tell the camera what to target.

The omission of Nikon’s excellent 3D AF Tracking feature in the Z-series (and its replacement with an unintuitive, less effective subject tracking mode) was our biggest gripe about the Z-series when it launched. It would have been a perfect mode to layer Eye AF on top of.

Instead, Eye AF is only available in the Zs’ ‘Auto area’ AF mode, where the camera is left to decide what to focus on. While it makes sense that an auto area AF mode might assume that the nearest thing to the camera is your intended subject (see below) this is not always true.

Despite there being two clearly identifiable faces within this scene, in ‘Auto Area’ mode the camera chose the nearest subject. But can you blame it? Most ‘auto’ AF systems are designed to focus on near, central objects because It makes sense, most of the time.

Nikon Z6 | Nikkor 70-200/2.8E FL ED VR

The soccer ball was not my intended subject in the image above, my daughter was. And that’s the problem with Eye AF only being added onto an ‘auto’ mode: you have accept that sometimes the camera will focus with a mind of its own. Most ‘Auto’ area modes tend to prioritize central, nearer objects like the ball above, and that’s not always the thing you want.

There is a way to force the camera to refocus on your subject once its focused on an unintended one (more on that below), but this doesn’t always work reliably. If your subject is at a vastly different distance from the one the camera is currently focused on, the Z6 / Z7 can be particularly reticent to refocus. The AF system either flat out refuses to refocus, or does so hesitantly. This again highlights the need for a robust subject tracking autofocus mode, to which Eye AF should be coupled.

The Z7 initially chose to focus on the background in its ‘Auto area’ AF mode. This is unfortunately common behavior for the Z-series cameras. The bigger problem is that after the camera has focused on the background, it’s reticent to refocus on a foreground subject if the subject is so far from the background that it’s extremely blurred.

Nikon Z7 | Nikkor 50/1.8 S

Choosing your subject

As you can see, most of the problems come from the camera choosing the wrong subject. You are given some tools to rectify this, but these have their own drawbacks. Let first explain why this matters.

Event, wedding, and many other photographers trying to capture candid portraits have a very specific requirement that many manufacturers tended to overlook in the past: the desire to choose which face, among many, to focus on. On the latest Sony cameras you can simply place the AF point over your desired face and half-press the shutter button.

Nikon offers two ways to choose your subject: tap on a detected face or eye on the rear LCD or use the joystick or four-way controller to jump between eyes and faces.

I wasn’t able to select my daughter (in the background) as the target because the Z6’s AF system did not detect her face. If Eye AF were combined with a proper subject tracking mode, as it is for Sony’s ‘Real Time Tracking AF’ or Canon’s Face+Tracking on its EOS R/RP, then I could always rely on the camera at least tracking my daughter, whether or not it has detected her face or eyes.

Nikon Z6 | Nikkor 70-200/2.8E FL ED VR

Normally, tapping on the LCD from ‘auto’ area mode is one way of engaging subject tracking. But if you tap on a detected face, the camera will automatically enter Eye AF mode instead of subject tracking mode. Note though, if you are a bit sloppy with your tap such that the camera ends up tracking a part of your subject’s shoulders or body, the camera may enter subject tracking mode and won’t engage Eye AF until you manually cancel tracking and try again.

Using the joystick to select your subject works quite well on the Z-series cameras

Either way, repeatedly having to take your eye away from the EVF just to tap on the LCD isn’t a great user experience, especially if you’ve just moved from a DSLR and are used to always having your eye to the viewfinder.

Thankfully, using the joystick to select your subject works quite well on the Z-series cameras, even allowing you to quickly toggle between the left and right eye of your subject. There’s a caveat, though: in order for you to switch to a different person, the camera has to have identified that person’s face. In candid shooting, if your subject is looking away, or for some reason isn’t identified as a face, you’ll find yourself repeatedly hitting the joystick or tapping the screen with no result.

Thankfully, a moment later, the Z6 did detect my daughter’s face, and I was able to hit ‘left’ on the joystick to switch from the blonde child to my daughter (in-focus). But this way of choosing your subject can be a gamble, dependent upon whether or not the camera has detected your intended subject as a face / eye to switch to.

Nikon Z6 | Nikkor 70-200/2.8E FL ED VR

In practice, some of these concerns were allayed because the Z6 and Z7 are just so darn good at identifying eyes and faces that more often than not, I was able to select my daughter. And once I’d selected her, it stuck to her like glue. Even if she looked away, the Z6 continued to track her head, ready to jump back to her eye when she faced the camera again. Now, if your subject completely turns its back to you or looks away for long enough, the camera will eventually jump off to another eye in the scene.

So it’s not bad, by any means. However, the very best current system don’t have these problems. The systems introduced in the Sony a9 and a6400 will track a subject and automatically switch in and out of Eye AF as necessary, as it finds or loses them on the subject being tracked. And that’s the bar the Nikons have to live up to.

Eye AF accuracy

Now that we’ve talked about the ability of the cameras to detect eyes and select the right one, let’s get to an arguably more important question: how often are the resultant photos actually focused on the eyes?

In good light, for the most part, the Nikon Z6 and Z7 do a great job of focusing on the eye. It’s a good deal better than Canon, but it doesn’t quite compete with the repeated pinpoint accuracy of Sony’s system, particularly in challenging light as we’ll see below.

While few would complain about the above result (this is a 50% crop), if you roll over to the a7R III example, you’ll see that the eye itself is in better focus, while the Nikon Z7 shot is slightly front-focused. Interestingly, on those occasions when Eye AF accuracy struggled on the Nikon, I almost always found the result to be fractionally front-focused.

Nikon’s Eye AF reticule tends to be rather large, especially compared to Sony’s, and I wonder if this is at least part of the reason. After all, contrasty features around the eye like eyelashes and eyebrows fall in front of the eye. In the image below, focus seems to lie on my daughter’s eyebrow, something that happened with quite some frequency:

I found Eye AF on the Z6 to front-focus more than I’d expect, sometimes significantly so. Here the camera appears to have focused on my daughter’s eyebrow.

Nikon Z6 | Nikkor 35/1.8 S

The extent to which this inaccuracy might affect you will vary from lens to lens. With many lenses and subject distances, the slight difference in distance between an eyebrow and an eye won’t matter. But for shallow depth-of-field applications, and particularly with fast wide primes like the 35/1.8 where the difference in camera-subject distance between the eye and the eyebrow can be significant, I found the inaccuracy to be problematic. Furthermore, if you’re shooting a model with prominent eyelashes, you may end up experiencing more ‘eyelash AF’ than ‘eye AF’.

As light levels drop, accuracy suffers even further. The shot below is severely front-focused, despite light levels being at a modest 4 EV (approx). Interestingly, the camera fared better under similarly dark daylight conditions (as opposed to artificial light). The a7R III did not struggle with accuracy under these conditions.

Even at 4 EV, Eye AF autofocus accuracy can take a dive with artificial lighting. Interestingly, I had better results at similar daylight levels.

Nikon Z6 | Nikkor 35/1.8 S

That’s not to say the Z6 is incapable of capturing accurately focused shots in these lighting conditions. Quite the contrary: I achieved a number of successful shots, like the ones below.

In darker conditions though, down to around -1EV, I was only getting a hit-rate of around one shots in three from the Z6, compared to the Sony a7R III, which typically missed only one or zero shots in a comparable sequence in repeated testing. The Z7 in particular would often hunt, and in Release Priority take completely out-of-focus shots. The Z6 fared better.

Furthermore, the Nikon 35 S and 50 S primes tend to sometimes hesitate to refocus, and this problem was only made more obvious in low light as the AF system tended to slow down. I’d often depress the shutter button and shoot a number of out-of-focus shots before the focus element finally even started to move to the correct position. This seems to be a problem not with Eye AF, but with the default behavior of the AF system, particularly with native lenses (often I achieved better results with adapted F-mount lenses).

In ~1 EV lighting, at best only two of these six shots are critically focused (click image to view 1:1). I had the same hit-rate with a different subject under 1 EV artificial lighting, where only one or two of six shots were critically focused. In all these instances, the eye was successfully detected by the camera.

Nikon Z6 | Nikkor 35/1.8 S

Another (related) area where Eye AF, and focus in general, can struggle is under severely backlit conditions. Here, like many DSLRs, the Z6 can resort to hunting. While I was still able to get the perfectly focused shot on the left, the camera struggled, and many of the shots were misfocused like the one on the right.

In focus one moment…

Nikon Z6 | Nikkor 70-200/2.8E FL ED VR

Out of focus the next.

Nikon Z6 | Nikkor 70-200/2.8E FL ED VR

Conclusion

My goal in this article was to determine when and where the newly introduced Nikon Z6 / 7 Eye AF worked well, and when it didn’t. I tested the performance side-by-side with the Sony a7R III but also considered the best implementation from a usability point-of-view, which is the system in the Sony a9 and a6400.

Because this is a complicated thing to test (thanks for reading this far), I’ve broken our findings up according to four criteria, with winners indicated in bold.

  • Detecting eyes: Nikon Z
  • Avoiding false positives: Sony
  • Eye AF ease-of-use: Sony
  • Accuracy of Eye AF: Sony

Considering the updates in the Z6 and Z7 represent Nikon’s first stab at Eye AF, this is really impressive.

Our main outstanding concerns are twofold: The first is false positives, where the camera will misidentify something as a face and stick to it, and a general tendency of ‘Auto area’ AF to focus and stay focused on the background. Second, and perhaps more important: we’d like to see Eye AF integrated with a reliable and easy-to-use subject tracking mode, like 3D Tracking on the company’s own DSLRs. The only reason we might consider ‘ergonomics of Eye AF’ to be comparable is if we compare the updated Z cameras to the older Sony a7R III. The a9 and a6400, on the other hand, have the best Eye AF implementation to-date, both from a performance and usability standpoint (and as such should be a model for others to follow).

Nikon’s introduction of an effective Eye AF system is a huge step in the right direction

For now, let’s take a step back and give Nikon kudos for the fact that its improved face and eye detection allows me to get shots like the one below without having to think, whereas previously Auto Area mode may well have focused on my daughter’s arm in the foreground.

Face detect worked beautifully here to quickly nail this shot.

Nikon Z7 | Nikkor 50/1.8 S

Nikon’s introduction of an effective Eye AF system is a huge step in the right direction for its mirrorless line of cameras, which even before this update, were already in my opinion the most enjoyable and photographer-friendly of today’s ILCs to shoot with.

If this firmware update is at all indicative of what’s to come from the Nikon Z-series, then we’re keen to see what’s next. The rebirth of 3D Tracking, pretty please?

$ (document).ready(function() { SampleGalleryStripV2({“galleryId”:”3673238986″,”isMobile”:false}) })

Sample photoSample photoSample photoSample photoSample photo

Footnote: For the sake of completeness, we also compared the updated Z6 / Z7 to Canon’s EOS R and RP. The Nikon Z cameras are considerably better at finding eyes than either of the Canons, where the eye has to be pretty big in the frame to even be identified. The updated Z-series cameras are also much more responsive when it comes to tracking the eye: like the Sony a7/R III they follow eyes around the frame quickly, whereas the EOS R / RP lag and frequently have to play ‘catch up’ with moving subjects.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Analysis: With new AF firmware, Nikon’s Z6 & Z7 just got more competitive

Posted in Uncategorized

 

Analysis: The Sony a7R III is still a star eater

21 Nov

We sent some files to our friend Jim Kasson for analysis, and he confirms that the Sony a7R III is definitely still a Star Eater, despite several claims to the contrary that have been published online over the past week.

Looking at Kasson’s graphs, one can clearly see the noise reduction kick in near Nyquist in Kasson’s energy plots. Indeed, in our own shots of the stars with the a7R III and latest a7R II (firmware v3.00 and above), our final files only show stars that are larger than one pixel with a few neighboring pixels: suggesting that smaller stars are indeed ‘eaten’ or dimmed due to a spatial filtering algorithm.

At a 3.2-second exposure, the ‘spacial filtering’ (Star Eater) is very mild, and won’t affect your stars.
But as soon as you hit 4-seconds, spacial filtering kicks in big time, causing the same Star Eater problems that was seen in the a7R II

This is a missed opportunity for Sony, and something dedicated astrophotographers will want to consider when deciding between the a7R III and other options that don’t have this same issue (a Nikon D850 for example). Other photographers happy with the number of stars still in their shots simply won’t care.

We’ll drop in one of our sample photos shortly for your pixel-peeping pleasure. But for now, we can say this with confidence: while a lot of stars still survive ‘Star Eater’, the a7R III continues the trend of noise reduction that dims or erases small stars at exposure longer than 3.2s.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Analysis: The Sony a7R III is still a star eater

Posted in Uncategorized

 

Financial statement analysis and valuation easton filetype pdf

02 Sep

You light surfaces only, our experts create writing masterpieces that earn our customers not only high grades but also a solid reputation from demanding professors. Earth of departed sunset, my course runs below the financial statement analysis and valuation easton filetype pdf of plummets. The past and present wilt – have you practis’d so long […]
BooksChantcdCom

 
Comments Off on Financial statement analysis and valuation easton filetype pdf

Posted in Equipment

 

Digging into the Fujifilm X-Pro2: Studio analysis and full-production sample gallery added

25 Jan

Our Fujifilm X-Pro2 First Impressions just got a significant update. With a production-ready camera in hand, we’ve been able to send it through our studio scene and dynamic range tests, make some judgements about image quality from its new 24MP X-Trans image sensor and put together a fresh sample gallery. Compare the X-Pro2 to its peers and take another look at what Fujifilm’s newest flagship can do.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Digging into the Fujifilm X-Pro2: Studio analysis and full-production sample gallery added

Posted in Uncategorized

 

Sony Alpha 7R II dynamic range analysis published

06 Oct

Our in-depth testing of the Sony Alpha 7R II continues, and we’ve just published 3 pages of Raw dynamic range analysis. Along with AF, the a7R II’s dynamic range capabilities have been the subject of much attention and could help put Sony’s latest full frame camera at the top of its field. Take a look and see how the a7R II’s 42MP BSI CMOS sensor performs when its Raw files are pushed to the limit. Read more

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Sony Alpha 7R II dynamic range analysis published

Posted in Uncategorized

 

Analysis: DxO ONE – the World’s ‘First Truly Connected Camera’

18 Jun

DxO’s new camera, the ONE, offers an appealing combination of features and some highly impressive technology. Designed to be used in combination with Apple smartphones, the ONE is built around a capable 1 inch type sensor and bright 32mm equivalent lens. DPReview’s technical editor has been digging into the specifications, and talking to DxO’s resident scientists to find out what makes the ONE tick. Click through for his analysis

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Analysis: DxO ONE – the World’s ‘First Truly Connected Camera’

Posted in Uncategorized

 

Canon EOS 7D Mark II studio analysis added to first impressions

02 Dec

The Canon EOS 7D Mark II is the long-awaited replacement to the 7D, which was launched in 2009. It features a 20.2MP APS-C CMOS sensor and dual DIGIC 6 image processors. It has a new 65-point, all cross-type autofocus system as well as an updated version of Canon’s Dual Pixel CMOS AF system that provides continuous phase detect focusing during video recording. We’ve just updated our first impressions review with analysis of the camera’s performance in our studio scene. See how it compares

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Canon EOS 7D Mark II studio analysis added to first impressions

Posted in Uncategorized

 

iWatch | High Velocity Bowling PlayStation Move Analysis

25 Oct

www.iwaggle3d.com Released more than two years ago on PlayStation Store, Sony’s bowling simulator High Velocity Bowling got recently updated (for free) to support stereoscopic 3D visuals and, most importantly, the PlayStation Move motion controller. So, how does it play now? Check the video to find out!

 
Comments Off on iWatch | High Velocity Bowling PlayStation Move Analysis

Posted in 3D Videos

 

DxO launches DxOMark Mobile – device IQ analysis as used on Connect

16 Oct

DxOLogo.png

DxO Labs has announced ‘DxOMark Mobile’ – its assessment of mobile phone image quality that will support mobile device reviews on connect.dpreview.com. In its first group of tests, DxO has found that the best contemporary mobile devices (in this case the Nokia 808 PureView, with its large sensor) will out-perform a 5-year-old high-end compact camera. And when it comes to video, the Samsung Galaxy SIII will trump last year’s Canon PowerShot S100. DxOMark Mobile will analyze 14 aspects of mobile imaging and produce a final score that will be comparable to existing DxOMark figures. We’ll be reporting DxOMark Mobile figures alongside our overall scores in connect reviews.

News: Digital Photography Review (dpreview.com)

 
Comments Off on DxO launches DxOMark Mobile – device IQ analysis as used on Connect

Posted in Uncategorized