RSS
 

Posts Tagged ‘Matches’

Sony a7R III dynamic range improved, nearly matches chart-topping Nikon D850

01 Nov

Sony has claimed 15 EV dynamic range from its newest ILC iteration: the a7R III. Is it true, or is it like Sony’s odd claim that the a7S had 15 EV dynamic range? Turns out: Sony has some strong grounds for its claim here.

The Sony a7R III retains its dynamic range even in bursts. That’s a big deal for a Sony ILC

At the launch event in NYC, we were able to gather enough data to measure the ‘engineering dynamic range’ of the a7R III.* And boy is it impressive. Possibly even more important: for the first time the a7R III retains this dynamic range even in continuous drive. That’s a big deal for sports and action photographers. But how true is Sony’s claim?

The Sony a7R II already had impressive Raw dynamic range, with the ability to expose short enough to keep highlights from blowing, but with low enough sensor noise to lift shadows without too much noise. The a7R III improves on this.

Oh and think this image is too dark? Wait till you view it on a HDR display, which is another can of worms altogether the stills industry should be discussing.

Photo: Rishi Sanyal

Sony has found a way to reduce shadow (or ‘read’) noise in its files such that the final output has higher dynamic range, and cleaner shadows if you need them, than files from its predecessor. To summarize it in a number at base ISO: 13.6 EV at the pixel, or for a 42.4MP file. Or 14.8 EV if you like to compare to DXO numbers (and only generate 8MP images from your 42.4MP camera). Either way, that’s a nearly half-stop improvement over its predecessor. See our table below, which also compares the a7R III to the full-frame chart-topping Nikon D850, ranking based on highest performer:

Pixel Dynamic Range 8MP ‘Print’ Dynamic Range
Nikon D850 (ISO 64) 13.78

15.03

Sony a7R III 13.63

14.84

Nikon D850 (ISO 100)

13.27 14.53
Sony a7R II 13.21 14.41

While the Nikon D850 is the top performer here, its important to note that this is only the case if you can give the D850 the extra ~2/3 EV light it needs at ISO 64 (which you often can if you’re shooting bright light or a landscape photographer on a tripod). At ISO 100, the a7R III dynamic range actually exceeds that of the D850, thanks to incredibly low read noise. That’s impressive for a camera constantly running its sensor in live view.

At ISO 100, the a7R III dynamic range actually exceeds that of the D850… impressive for a camera constantly running in live view

Keep in mind, though, that if you can give the D850 the extra exposure to take advantage of its ISO 64 dynamic range, all tones in your image benefit from the higher signal:noise ratio—even midtones and brighter tones will be more amenable to post-processing and sharpening thanks to being more ‘clean’ and less noisy to begin with. The D850 is able to tolerate as much total light as the medium format Fujifilm GFX 50S, as we showed here. That’s what allows one to get unbelievably crisp, ‘medium format-like’ like files from a Nikon D810 (just zoom in to 100% on that shot and tell me you’re not impressed).

But the Sony a7R III gets you nearly there. While in some circumstances the Nikon D810/D850, or medium format, may afford you slightly cleaner more malleable files, the a7R III takes a significant step at closing the gap. And that’s nothing short of impressive for a mirrorless ILC constantly running its sensor for a live feed (and all its benefits).

As for Sony’s marketing, it sounds like the claim of 15 EV is believable, but only technically if you consider how your images look when shrunk to 8MP files. To be fair, there’s some benefit to comparing dynamic range figures after resizing camera outputs to 8MP, since it’s a common basis for comparison that doesn’t penalize cameras for having higher resolution (and therefore smaller pixels).

In depth vs. a7R II

Let’s take a deeper dive. Here are our ‘engineering’ dynamic range measurements of the a7R III vs. the a7R II. ‘Engineering’ dynamic range means we are measuring the range of tones recorded between clipping and when the shadows reach an unacceptable noise threshold where signal is indistinguishable from noise (or when signal:noise ratio = 1). Have a look (blue: a7R III | red: a7R II):

The a7R III shows a 0.42 EV, or nearly a half a stop, improvement in base ISO dynamic range over the a7R II. That’s not insignificant: it will be visible in the deepest shadows of base ISO shots of high contrast scenes. How did Sony do this given the already low levels of read noise its known for? Possibly by going to better or higher native bit-depth ADCs, something Bill Claff had suggested based on our largely 12-bit findings of the Sony a9’s output. But let’s save that for the PST forums.

Suffice it to say the a7R III improves on low ISO dynamic range, without sacrificing anything on the high end

It’s worth noting our a7R II figures are higher than DXO’s published 12.69 EV (13.9 EV ‘Print’) figures, possibly because they tested an older unit prior to uncompressed Raw and improvements to Sony’s compression curve. We retested it literally today with the latest firmware, and get figures of 13.2 EV or 14.4 EV normalized for ‘Print’ (Bill measures 13.3 EV, which you can see by clicking the camera name in the legend). See our 8MP, or ‘Print’ normalized, dynamic range figures below. These are more comparable to what DXO might report, for the benefit of your own comparative efforts (blue: a7R III | red: a7R II):

You can see the Sony a7R III encroaching on the ~15 EV rating of the Nikon D850 at ISO 64, but achieved at ISO 100 on the Sony, thanks to lower read noise. Impressive, though keep in mind again that the overall image quality improvement of an ISO 64 file from a D850 is due to total captured light (and it’s all about total captured light, which you can read about here).

Independently, our friend Bill Claff has tested the a7R III and also shows a similar 0.3 EV improvement over the Mark II (you can see the dynamic range numbers by clicking on the relevant camera in the legend at the upper right). He also shows the slight advantage of the Nikon D850 over the a7R III, which comes in at 13.7 EV vs. the a7R III’s 13.6 EV at the pixel level.**

Sony: a job well done. And all this at no cost to high ISO performance (we have comparisons coming showing parity between high ISO a7R III and a9 performance). Now please offer us visually lossless compressed Raw so we don’t have to deal with >80MB files for no reason. 🙂

ISO-Invariance

A camera with such great dynamic range performance suggests it’s probably fairly ISO-invariant, but is it?

Well, yes and no. It’s ISO-invariant in exactly the way it should be, but not so in the ways it shouldn’t be. Confused? Read on.

The a7R III, like many Sony predecessors, has a second gain step at the pixel level that amplifies signal, at the cost of higher tones, to preserve higher signal, and less noise, in dark tones. But it does so at a higher ISO—640 to be exact. At this point, the camera has amplified its signal in the analog domain so much that any remaining noise barely affects it.

That’s why the camera shows no difference between amplifying that ISO 640-amplified signal digitally (in-post) or in the analog domain in-camera. While we’ll have a more rigorous and controlled ISO-invariance test coming soon, you can see even in our cursory test at the launch event below that comparing ISO 6400 vs ISO 640 shot at the same exposure but raised 3.3 EV in-post to maintain the same brightness as ISO 6400 shows no difference at all in noise performance.

$ (document).ready(function() { ImageComparisonWidget({“containerId”:”reviewImageComparisonWidget-56009185″,”widgetId”:560,”initialStateId”:null}) })

What’s the advantage to the latter? 3.3 EV of highlights you otherwise lose by amplifying to ISO 6400 levels in-camera, but that you don’t lose if you ask ACR to digitally brighten 3.3 EV in post (anything that gets blown from that 3.3 EV push can easily be recovered in ACR since it’s there in the Raw file).

Below ISO 640 there’s some extra noise to, say, shooting ISO 100 and boosting 6 EV in post as opposed to shooting ISO 640 and boosting 3.3 EV. But there’s simply no excuse to the camera’s traditional ISO 6400 method of shooting ISO 6400-appropriate exposure and then amplifying the analog signal 6 EV in post to get ISO 6400 levels of brightness; instead, 2.7 EV of that push could be done in the analog domain by switching dual gain to ISO 640 levels, but the remaining 3.3 EV push should be saved for Raw conversion in order to retain 3.3 EV (or more) of highlight detail. Indeed, this is easily seen in Bill Claff’s ‘Shadow Improvement’ graphs that show little to no benefit to analog amplification above ISO 640 on even the Sony a7R II (or ISO 400 on the Nikon D850). And only a highlight cost of stops, upon stops, upon stops, since tones get amplified above the clipping point of the ADC at higher ISOs.

I’m going to use this as an opportunity to ask manufacturers like Sony, Nikon and the like: please accept the digital revolution that even your video departments have accepted (in their ‘E.I.’ modes). Please stop throwing away highlight data for almost no shadow benefit to ostensibly stick to poor antiquated ‘film’ analogies, or to work around CCD/CMOS read noise limitations that no longer exist. We’ve been singing this tune since 2014 when we designed our ISO-invariance test, and it’s even more relevant today with dual-gain architectures. ACR understands digital ‘push’ tags and you can brighten the image preview (and JPEG) as necessary. This is not to single out Sony: Nikon, Olympus and Panasonic are just as easy to blame, if not Canon of late after having modernized its sensor architecture to catch up with the rest.


Footnotes:

* Sony’s claim that the a7S had 15 EV dynamic range was patently false, as even the a7R II which has been measured to have less than 15 EV dynamic range performs better. But since there’s no standard for dynamic range measurement, it’s hard to say whether or not anyone’s claim is right or wrong – manufacturers can claim whatever they wish.

** But again, that’s not the whole story until you consider the higher signal:noise ratio of all tones at ISO 64 on a D850 compared to ISO 100 on any other full-frame at ISO 100.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on Sony a7R III dynamic range improved, nearly matches chart-topping Nikon D850

Posted in Uncategorized

 

This is what happens when your camera’s frame rate matches a bird’s wing flap

19 Jul

Here’s a neat trick you might not have realized is possible. If your camera’s frame rate matches the flapping rate of a bird’s wings, you can create a video where it seems like the bird is floating ‘magically’ on frozen wings.

The video above is going viral today after YouTuber Ginger Beard shared it on his nascent channel. As he explains on Reddit, the video was captured with a Hikvision DS-2CD2342WD-I security camera set to “max resolution” at 2688×1520 and 20 frames per second. Apparently, the bird’s wings were also set to 20fps, because this ‘magic’ video is what popped out the other side.

And if you like this, check out this similar video shared on the DPReview forums a few months ago. In that one, the camera’s frame rate is perfectly synced to a helicopter’s rotor.

Articles: Digital Photography Review (dpreview.com)

 
Comments Off on This is what happens when your camera’s frame rate matches a bird’s wing flap

Posted in Uncategorized

 

Terrapattern: Satellite Image Search Engine Matches Similar Places

31 May

terrapattern shipyard

A powerful tool for artists, designers and researchers, Terrapattern lets users seek out similar-looking locations from an aerial perspective, finding connections and patterns between disparate landscapes and built environments.

terrapattern culs de sac

The premise is simple: start with a single place, be it a park or street, stadium or shipyard, then let the tool work its magic. The results are uncanny: colors, textures and shapes tied together by computer vision and clever algorithms. The broader use cases are infinite, but specific ones are possible too, like: a user could look for abandoned ships floating around the island of Manhattan.

terrapattern abandoned ships

The system works by looking at its subjects in layers, looking for identifying features like curves, edges and shadows that indicate height. In a way, its task is similar than some pattern recognition software since it is not called upon to identify the subject, just match it.

terrapattern street grids

“For our purposes,” explain the creators, “‘interesting’ features are anthropogenic or natural phenomena that are not only socially or scientifically meaningful, but also visually distinctive—thus lending themselves ideally to machine recognition. Examples could include things like animal herds, methane blowholes, factories, destroyed homes, or logging roads. Many other patterns await discovery.”

terrapattern buses

The system draws on data from OpenStreetMap, combing through hundreds of thousands of images looking for something like whatever you submitted. Researchers can use tools like this to monitor natural habitats or make archaeological finds, but ordinary people can employ this tool to create art or make inquiries about the cities they live in. Even a quick tour around the engine reveals emergent macro-patterns from individual tiles, some worthy of wall art treatment.

terrapattern golf courses

Terrapattern’s creators are indeed excited for more non-standard and unexpected uses: “Terrapattern is ideal for discovering, locating and labeling typologies that aren’t customarily indicated on maps. These might include ephemeral or temporally-contingent features (such as vehicles or construction sites), or the sorts of banal infrastructure (like fracking wells or smokestacks) that only appear on specialist blueprints, if they appear at all.”


WebUrbanist

 
Comments Off on Terrapattern: Satellite Image Search Engine Matches Similar Places

Posted in Creativity