With Sony and Canon dominating headlines, Nikon has been the forgotten company in the mirrorless race. That’s set to change, though, with the launch of the 45.7-megapixel Z9. It’s an innovative professional camera that offers awesome speed, incredible detail and outstanding 8K video capabilities.

Nikon is breaking some new ground with this camera, too. It’s the first mirrorless model with no mechanical shutter, relying instead on a fast sensor to minimize the jello effect that can happen with an electronic shutter.

The Z9 is not without some flaws, though. It’s quite heavy compared to rival models like Sony’s A1 and the lack of a fully articulating screen makes it a hard sell for videographers. At $5,500 for the body only, it’s quite expensive — let’s see how it measures up to the competition.

Design and handling

The Z9 is, honestly, a chunky camera. It weighs around 1,340 grams, or just 60 grams less than Nikon’s D6 DSLR. By contrast, the Canon EOS R3 weighs 1,015 grams and Sony’s A1 is a mere 737 grams. A camera that heavy can be fatiguing but then again, many professional photographers prefer big, heavy cameras for stability.

The large size meant Nikon could include a nice big grip that imparts a feeling of security, particularly with a big lens attached. Like Canon’s EOS R3, you can rotate it 90 degrees and get matching controls and a grip, allowing for easy operation in portrait mode.

The Z9 has a generous array of manual controls, but I’m not a big fan of the layout. The AF button is positioned on the left side next to the lens, which I find awkward. The top display also takes up space that could’ve been used for more buttons. On the plus side, it has a large number of programmable buttons, so you can set it up the way you like. If you’re a longtime Nikon user, you might enjoy the layout more than I did.

Compared to Sony and Panasonic’s latest cameras, the menu system is a bit too complicated. Rather than dividing things into subcategories, you have to scroll down for a while on certain menus to find what you need. Your best bet to avoid any hassle is to program the custom menus and buttons to do what you need.

Steve Dent/Engadget

With 2.09 million dots, the 3.2-inch touchscreen is relatively sharp and bright, but there are a few drawbacks. As mentioned, it only tilts up and out to the right, so self-shooting or vlogging is impossible unless you have an external display. Sony’s A1 also lacked a fully articulating display, but Canon has finally included one in its professional EOS R3 model.

The EVF’s resolution is a bit disappointing for a camera this expensive, at just 3.69 million dots, compared to 9.44 million on the Sony A1. Considering the high-res sensor, that seems like a miss. On the plus side, it uses “Dual-Stream” tech to send data to the sensor and EVF simultaneously. As a result, you get blackout-free performance during all shooting, and according to Nikon, that comes with zero frame skipping or repeating. That’s likely a light dig at some Sony cameras, which drop the EVF resolution under certain shooting conditions.

Nikon has gone all-in on CFexpress for the Z9, with two CFexpress and no SD card slots. This is smart for a couple of reasons. There’s no compromise on speed for burst shooting or video, which is crucial for high frame rates required for ProRes HQ and, down the road, RAW video. It also means that professional shooters can record to two cards at once for a backup, with no compromise on speed. The downside, of course, is that CFexpress and XQD cards are much more expensive than SD cards.

Other features include a huge EN-EL18D battery that delivers a solid 740 shots on a charge and over two hours of 4K video, as well as backward compatibility with D6 and other Nikon pro DSLR batteries. It also comes with a USB-C 3.2 slot for both transfers and storage, mic and headphone jacks, an Ethernet port for remote shooting and a full-sized HDMI jack. While the Z9 doesn’t have a mechanical shutter, it does have a cover to protect the sensor when you change lenses.

Performance

Steve Dent/Engadget

The Z9 is Nikon’s fastest mirrorless camera to date. You can shoot up to 20 RAW frames per second at full resolution, 30 JPEGs or a blindingly-fast 120 JPEGS per second at a lower11-MP resolution. Again, all of those speeds are in electronic shutter mode, because there’s no mechanical shutter. You can, of course, shoot in silent mode, or activate a simulated shutter sound if you really need to hear that “click.”

That’s fast, but I managed just 40 lossless RAW photos with a very fast CFexpress card before the buffer filled. If you’re fine using the lossy “high-efficiency” modes, however, you can shoot anywhere from 70 to nearly infinite photos without stopping.

The hybrid phase-detect autofocus is a big improvement over the last model, too. Nikon has brought 3D tracking over from its DSLRs, but it’s far better in the Z9 because it has 493 AF tracking points, far more than the others have.

It delivers very reliable subject tracking most of the time. Normally, you just set the tracking point on an athlete or other fast-moving subject, and the AF will keep them in focus. Occasionally, it will drop the subject and focus on the background or something else, particularly with birds or other fast objects. In general, it works as well or nearly as well as the subject-tracking AF on Sony’s latest A1 and A7 IV models.

Steve Dent/Engadget

The new face, eye, vehicle and animal detection is incredibly fast and accurate as well, even if the subject is a good distance away. In fact, the Z9 can recognize nine types of subjects automatically: human eyes, faces, heads and upper bodies; animal eyes, heads and bodies; and cars, planes, trains and motorbikes. I didn’t have a chance to test all of those, but for several different subjects, even at high burst speeds, most of my shots were in focus.

With no mechanical shutter, fast readout speeds are key to reducing skew and other issues. Happily, the Z9’s stacked, backside-illuminated CMOS sensor is up to the job. I never saw any rolling shutter unless I really whipped the camera around. So even for sports or bird photography, you’re not likely to see any issues.

At the same time, the 5-axis in-body stabilization reduces blurry shots due to camera movement or hand shake. The six stops of shake reduction are well below the Canon EOS R3’s 8 stops, however.

Image quality

With 45.7 megapixels on tap, the Z9 delivers sharp images with very high dynamic range. In fact, next to all rivals, it’s second only to Nikon’s own Z7 II, which is the current dynamic range champion and packs a similar sensor. All RAW images are now captured with 14-bit color depth, giving you plenty of room to tweak them in Adobe Lightroom or Nikon’s NX Studio.

Color accuracy is better than I’ve seen on previous models, possibly thanks to the Nikon Z9’s new Expeed 7 processor that brings improvements to white balance. However, skin tones aren’t quite as warm as with Canon’s R3 or R5. JPEG images look nice straight out of the camera, though they can be slightly over-sharpened with the default settings.

The Z9 has a wide ISO range of 64-25,600, expandable to 32-102,400. The low 64 setting is nice to have in bright sunlight if you want to avoid completely closing down the iris (which can hurt image quality). It performs pretty well in low-light, too. Grain is well-controlled up to about ISO 6,400, and shots are usable at ISO 12,800. Beyond that, noise starts to get distracting. The Z9 is about the same as Canon’s R5 in that regard, but not up to the level of Sony’s A7R IV or A1.

Video

The Z9 is Nikon’s most competent video camera to date, and by far. For starters, it can record 8K H.265 for up to two hours without any overheating or other issues. That’s something Canon’s smaller EOS R5 can’t do, though it is $1,600 cheaper. The Z9 can shoot very sharp 4K at up to 30 fps by oversampling the full width of the sensor, and up to 120 fps with continuous autofocus and sound.

Right now, you can record 4K using ProRes 4:2:2 HQ, a widely used format that creates large files, but is easy to edit. It also records in H.265 10-bit mode for 8K and 4K video. Meanwhile, Nikon’s N-Log mode lets you max out dynamic range, so you can easily adjust shadows and highlights when editing.

There’s more to come, though. In a future firmware update, Nikon will offer 8K 60p internal RAW video using a format called TicoRaw. When it arrives later this year, it’ll let you capture 11 minutes of lossless 12-bit 8K RAW video and 54 minutes of visually lossless 8K to a 1TB card.

As with photos, video autofocus is fast, reliable and can track a subject’s face even if they’re fairly far away. It also follows subjects smoothly when they’re moving toward the camera, though you may have to tweak the AF speed. Unfortunately, the only way to find or use that setting is to dive into the menus. Hopefully, Nikon will let you assign it to a button in a future upgrade.

Steve Dent/Engadget

Rolling shutter is more prominent in video than stills, particularly at the 8K and 4K oversampled resolutions. However, it’s still not nearly as bad on some cameras (looking at you, Sony’s A7 IV) thanks to the extremely fast sensor readout speeds. It’s not even really noticeable unless you whip-pan the camera or have a fast-moving subject.

8K and 4K oversampled video is extremely sharp. Quality drops a hair at the higher frame rate 4K settings, when pixel-binning kicks in, but it’s nothing you’d notice unless you’re looking for it. Colors are accurate, but again, skin tones aren’t quite as pretty as on Canon’s latest models.

Dynamic range is also top notch, particularly in the 400-800 ISO range in ProRes mode, making it easy to adjust shadows and highlights in post. If you love shooting ProRes footage, be sure to get some high-capacity CFexpress cards, because the file sizes can get huge; they take around 132GB for 10 minutes of 4K 60p ProRes HQ footage.

Wrap-up

Steve Dent/Engadget

The Z9 is Nikon’s most versatile and powerful mirrorless camera yet, and the lack of a mechanical shutter is a non-issue. It’s particularly strong for video, and will be a powerhouse once the RAW video firmware update arrives. However, Nikon’s decision to not use a fully articulating screen is a shame, considering all this awesome video capability.

The Z9’s main rival is Sony’s $6,500 A1, which offers similar shooting speeds, resolution and video capabilities. The Z9 is $1,000 cheaper and will (eventually) have higher 8K frame rates with the upcoming firmware update. However, your choice may come down to whether you prefer Sony or Nikon’s camera systems.

Another option is Canon’s $6,000 EOS R3. With just a 24-MP sensor, though, it’s designed primarily for sports, so it’s not ideal for studio or landscape work. Overall, the Z9 is a great choice for pros who don’t want to compromise on any kind of shooting, whether that’s photos or video.

Last year’s OnePlus 9 Pro was one of the best phones of 2021 – which surprised some of us. The company substantially improved the cameras and packaged a high-spec phone in a premium body. Now, after several other OnePlus phones have launched with lower prices (and lesser specs), the $899 OnePlus 10 Pro, is finally here. The phone has long been available in China and the company has been teasing its new flagship since January.

With a cheaper price, a 6.7-inch 120Hz AMOLED screen, Snapdragon’s latest processor, a bigger battery and a gorgeous new green color option, the OnePlus 10 Pro sounds like it has everything it needs to repeat the success of its predecessor, but I’m left feeling underwhelmed.

Hardware

Mat Smith/Engadget

The camera unit, which includes three sensors and a flash, is surrounded by metal which bleeds over the edge to the frame of the 10 Pro. It’s an understated design touch, but I’m glad it looks different from older OnePlus models, its Oppo stepsiblings and other phones doing the same old camera sensor layout. Different is good.

The rest of the phone’s back is covered in a translucent finish that reminds me of the back of the iPhone 13 Pro. The OnePlus logo seems to be etched into this; I actually thought it was a sticker at first. There’s still some Hasselblad branding, too, along the side of the camera unit, but thankfully it's a little subtler than previous designs.

OnePlus’s Alert Slider has clung on for its eighth year. Once again, it’s located above the power button and switches between silent, vibrate and full volume modes, each of which can be manually adjusted to your preferred levels. There’s a USB-C port, supporting 80W SuperVOOC charging and stereo speaker grilles along the bottom edge. Finally, as usual, there’s a volume rocker on the left edge. 

OnePlus has upgraded the selfie camera to a 32-megapixel sensor, but it’s still a pinhole camera set in the top left corner of the display. The front-facing camera works with face unlock, while there’s also an in-screen fingerprint scanner. Both seem faster than OnePlus’ last-gen phone, and the fingerprint sensor has been shifted higher up the phone panel, making it easier to access. 

While the OnePlus 10 Pro looks different from last year’s 9 Pro, there’s an awful lot of spec overlap. Both models have the same size screen and run at 1,440 x 3,216 resolution, with adaptive refresh rates of up to 120Hz. This year’s phone does have an upgraded LTPO display, however, which OnePlus says is better optimized for dynamic changes in refresh rates. But you’d be hard-pressed to notice any difference between the 10 Pro and 9 Pro's screens; both are crisp, bright and colorful. While more and more phones are beginning to arrive with adaptive refresh rates, OnePlus does it better than most, dipping as low as 1Hz for static content on your phone screen, meaning less power drain. According to OnePlus, the upgraded screen should translate to 1.5 hours of additional use versus last year’s OnePlus 9 Pro. 

Cameras

Comparisons with last year’s OnePlus 9 Pro come up yet again with the cameras. The company heralds this as its second-generation Hasselblad camera, with improvements to the OnePlus Billion Color Solution (which still struggles to sell itself as a compelling feature) and an updated Hasselblad Pro mode, which I’ll explain later.

Glancing at the spec sheet, even if the camera array itself looks notably different, the OnePlus 10 Pro has very similar camera sensors – and in fact there’s actually one less than last year’s OnePlus 9 Pro. Yes, we’ve lost the monochrome sensor, which shouldn’t be a big deal. It was a low two-megapixel sensor and I didn’t miss it at all. Otherwise, all the numbers match: a 48-megapixel primary sensor, a 50-MP ultra-wide lens – this time capturing across 150-degree views, and – like the OnePlus 9 Pro – a middling 8-megapixel telephoto option that tops out at 3.3x optical zoom.

My experience matches what Chris Velazco said last year in his review of the 9 Pro. The primary sensor captures detailed images, especially in well-lit surroundings. OnePlus has further refined the sensor to improve dynamic range and noise reduction in images, and you’ll still get the best shots from the pixel-binned 12-megapixel mode. If you do want to capture all the detail you can, however, there’s a high-res shooting mode that’s easily accessible through most of the camera app’s modes. 

The new ultrawide camera gets a few tricks, too, including a new fish-eye capture mode that’s a bit of a gimmick, but it’s fun nonetheless. You can switch between a mild and strong fish-eye effect. I’m not going to win photography awards but the results are clean and it’s a harmless addition.

Mat Smith/Engadget

The telephoto camera remains the weakest part of the OnePlus camera setup. Given the 8MP resolution, shots seem blurry and low on detail. Sometimes my photos just lacked color and vitality, which is a shame because I use the telephoto cameras on phones a lot as it offers some degree of compositional freedom when framing my shots. (I can’t believe I just wrote the phrase “compositional freedom.”)

OnePlus’ flagship held its own against an iPhone 13 Pro and Pixel 6 Pro, but it wasn't the best. The cameras occasionally struggled with scenes with high dynamic range. 

When shooting this scene, with a plant in front of an open fire and a bright neon sign in the background, the 10 Pro couldn’t quite tame the aggressive lighting, blowing out the neon light a little too much. The iPhone captured the scene in slightly warmer tones, while the Pixel seemed cooler. When I compared the photos, however, it was the OnePlus 10 Pro, even if it lacked the detail and dynamic range, that had nailed the colors of the room. Perhaps that Hasselblad partnership is actually working? Even then, would I prefer a more accurately colored shot or a wider dynamic range? The latter.

I had to see if anything had truly changed between the two OnePlus flagships, so I compared a few sample photos. While many were indistinguishable from each other, the 10 Pro edged out last year’s 9 Pro when it came to nighttime photography, which could be due to computational improvements and a faster chip – or OnePlus’ own internal calibrations and tweaks. Having said that, OnePlus’ night photography and AI assistance are a little too heavy-handed for my liking; they made some low-light scenes look artificially bright.

Software-wise, the new Hasselblad Pro mode works across all three camera sensors, adding fine control like ISO levels as well as 12-bit RAW capture for those willing to dive deep into image editing. OnePlus’ RAW+ image format is also here, which attempts to combine all the information of a photo, but it’s augmented by the computational image capture we’ve seen in most phones over the last few years. Was I using RAW+ capture over the stripped-down auto mode? Most of the time, no. But it’s more flexible for users willing to push the 10 Pro’s cameras to their limit.

Performance and software

Mat Smith/Engadget

OnePlus has long been able to balance a streamlined Android experience close to stock while adding its own tweaks and features. The OnePlus 10 Pro continues that approach despite the closer collaboration with Oppo and a shared codebase. In fact, the latest version of OnePlus’ OxygenOS, version 12, was apparently one of the reasons that the 10 Pro took so long to arrive after its debut, with more time needed to tweak the software for regions beyond China.

I still appreciate the ability to easily switch off OS additions I don’t need. The OnePlus Shelf is a pop-up menu that can be pulled down from the top right corner of the phone. It groups together several adjustable tiles – like those widgets you’ve been able to add to your homescreen on Android phones for about a decade. 

In short, I don’t need it and would get frustrated when it pops up instead of the standard Android drop-down menu containing setting toggles and my notifications. Thankfully, I can just tap the settings cog inside the Shelf, and turn the entire feature off. 

One addition I won’t be disabling is a new AI adaptive brightness feature. The OnePlus 10 Pro can learn your display brightness preferences and make adjustments, hopefully before you do. Artificial intelligence features inside smartphones are often hard to notice in day-to-day use (think: battery optimization features that are meant to adapt to how you use your phone and reduce power consumption). But OnePlus’ adaptive brightness soon appeared when I would check the 10 Pro in the early hours of the morning, helping to shield me from an unnecessarily bright screen before I’d even had my coffe

There are also a few gaming improvements to make the most of the Snapdragon 8 Gen 1 chip. The HyperBoost game engine, courtesy of Oppo, tries to stabilize frame rates during gaming sessions, while also increasing the responsiveness of the touchscreen through a new feature called O-Sync. Both suffer from my issues with behind-the-scenes AI optimizations. It’s also not compatible with streaming games from Xbox Cloud Gaming or Stadia, which is how I game on smartphones most of the time.

The OnePlus 10 Pro wasn’t fazed by anything I threw at it – which has been true for most phones powered by Qualcomm’s Snapdragon 8 Gen 1. When playing a video on repeat, the 5,000mAh battery took over 14 hours to run down – which isn’t great compared to roughly 17 hours from Samsung’s Galaxy S22+.

Benefiting from Oppo’s R&D, the OnePlus 10 Pro supports 80W SUPERVOOC charging. Yes, fast charging isn’t anything new for OnePlus, but damn, this is genuinely fast. With the appropriate charger (don’t worry, there’s one included with the phone), the 10 Pro can be fully charged in a mere 32 minutes. If you want even faster charging, OnePlus says 80W SUPERVOOC can charge the phone from 0 to 61 percent in just 15 minutes; I got around the same figures when recharging the phone myself. It’s a strong feature, and importantly, something that the OnePlus 9 Pro doesn’t have. 

Wrap-up

Mat Smith/Engadget

OnePlus was on a roll. Last year, it proved it could deliver a true high-end smartphone rival to the likes of the iPhone and the Galaxy S series with the OnePlus 9 Pro. But with the OnePlus 10 Pro, the company has struggled to push the envelope further.

The OnePlus 10 Pro has impressive fast-charge capabilities, a gorgeous screen that’s incrementally better than last year’s model, while OxygenOS continues to add more to the Android experience without derailing it. I especially like how the green version looks, but at the same time, the smartphone competition moves fast. 

(While I tested the covetable green version, in the US, OnePlus is selling the black model with 8GB of RAM and 128GB of storage for $899 in the US. The company says the green version, with 12GB of RAM and 256GB of storage, will arrive later. The price is still TBC. )

OnePlus might have nailed its cameras last year, but it’s not at the same level of the best phone cameras in 2022. Its computational photography seems overly aggressive and the telephoto camera, again, disappoints compared to the competition. When the Pixel 6 Pro costs the same, the OnePlus 10 Pro is hard to recommend over Google’s own flagship, even to die-hard OnePlus fans looking to upgrade from the OnePlus 9 Pro. At least it’s slightly cheaper.

Xiaomi has come a long way since the launch of its budget Redmi line back in 2013, and while the latter has since extended its coverage across a wider price range, it isn't about to abandon the budget market. Ahead of its 12th anniversary on April 6th, Xiaomi announced a new Redmi 10 5G which, starting from $199, happens to be the company's — and some markets' — cheapest 5G smartphone to date. To put things into perspective, the OnePlus Nord N200 5G costs $240, and Samsung's recently-launched Galaxy A13 5G is priced at $250.

The Redmi 10 5G — not to be confused with the 4G-only Redmi 10 or 10A — is powered by an octa-core MediaTek Dimensity 700 processor with dual 5G standby, and it runs on MIUI 13 based on Android 11. You get a reasonably-sized 6.58-inch 2,408 x 1,080 LCD with up to 90Hz refresh rate, along with a generous 5,000mAh battery which supports 18W fast charging (interestingly, the phone comes with a 22.5W charger), a side-mounted fingerprint sensor and a 3.5mm headphone jack. 

Xiaomi

There's not much going on in terms of design — just a plastic body with a grooved concentric circle texture and flat edges. Likewise with photography: just a 50-megapixel main camera paired with a 2-megapixel depth camera on the back, plus a 5-megapixel "dot drop" selfie camera above the screen.

As with all Xiaomi phones these days, the Redmi 10 5G won't be headed to the US, but a spokesperson confirmed that we can expect it to land in most of Xiaomi's global markets, including the UK, Europe, South East Asia, Middle East and more. You'll have two options: the $199 model comes with 4GB of RAM and 64GB of internal storage, whereas the $229 version bumps the storage up to 128GB. These will come in "Graphite Gray," "Chrome Silver" and "Aurora Green."

Xiaomi

Also announced for the international markets are the higher-end Redmi Note 11 Pro+ 5G and Redmi Note 11S 5G, both of which also pack a MediaTek processor (Dimensity 920 and Dimensity 810, respectively) with dual 5G standby, a side-mounted fingerprint reader, an IR blaster, a 3.5mm headphone jack and dual speakers (with JBL tuning on the Pro+).

The Redmi Note 11 Pro+ 5G has been available in China since October, and it's all about its 15-minute fast charging on its 4,500mAh dual-cell battery, thanks to Xiaomi's 120W HyperCharge tech. You also get a 6.67-inch 2,400 x 1,080 AMOLED screen, which delivers up to 120Hz refresh rate and up to 360Hz touch sampling rate for some slick scrolling, as well as a high brightness mode of up to 700 nits for outdoor usage. With its 108-megapixel main camera (Samsung HM2 sensor), 8-megapixel ultra-wide camera, 2-megapixel telemacro camera and 16-megapixel punch-hole selfie camera, you should be able to get some nice photos out of this device. Prices range from $369 (6GB + 128GB) to $449 (8GB + 256GB), and you can pick "Graphite Gray," "Star Blue" or "Forest Green."

Xiaomi

The marginally cheaper Redmi Note 11S 5G is a slight variant of China's earlier Redmi Note 11 5G, with the main difference being the extra 2-megapixel macro camera on the back. The rest appears to be identical: 6.6-inch 2,400 x 1,080 LCD (90Hz refresh rate, 240Hz touch sampling rate), 5,000mAh battery with 33W fast charging, 50-megapixel main camera, 8-megapixel ultra-wide camera and 13-megapixel punch-hole selfie camera. Prices range from $249 (4GB + 64GB) to $299 (6GB + 128GB), with color options including "Midnight Black," "Twilight Blue" and a sparkly "Star Blue."

Oppo’s latest phone looks and sounds like an expensive, powerful flagship. With a 6.7-inch WQHD+ screen that can hit 1,300 nits of brightness, adaptive refresh rates, and a new AI chip, made in-house to amp up low light camera performance – even at 4K. There are also two 50-megapixel primary camera sensors, a new (but familiar-sounding) Hasselblad collaboration and, well, just look at this beautiful thing.

So far this year, Oppo already announced its MariSilicon X neural processing unit (NPU), its intriguing attempt at foldables, but not a new flagship phone. But that’s where MWC comes in, which so far is shaping up to be a relatively quiet mobile show. It’s good news for Oppo. Barring any leftfield surprises (Samsung and Huawei both have press events this week) it can take center stage. Fortunately, Oppo has provided a device ahead of the public reveal, so I’ve got some early impressions and thoughts to add to proceedings.

The Find X5 Pro borrows heavily from the Find X3 Pro’s shape and style from last year, but I think it looks even better. The camera unit now has a diagonal slope at the bottom, so it should fit in your hand better. There’s no space-age, hyper-reflective silver color option this time, just glossy black or ceramic white. It was a bit much.

Mat Smith/Engadget

This black one picks up fingerprints and smudges with a little too much ease, but I didn’t fear scratching it before I had to film this preview or grab photos for my story. Oppo says the ceramic body is actually nanometer microcrystalline ceramics. The material takes 168 hours to produce but Oppo says results in a phone that can resist scratches over 1200 HV – Vickers Pyramid number that registers hardness. For comparison, stainless steel scores 200 HV, while diamond reaches around 10,000 HV. In summary, you might not need the protective case that also comes in the box. There’s also Gorilla Glass Victus on the screen.

I think it looks great. And while it might not be quite as unique as the Pixel 6’s cyclops look, it at least looks different to the army of phones with their squarish camera arrays. There’s a little originality to the design here. There are also some (reliably?) familiar specs. It’s another phone arriving with Qualcomm’s Snapdragon 8 Gen 1 chip – arguably the most powerful Android phone chip available in phones at this moment. There’s also a 5,000mAh battery – bigger than the one found in its predecessor. Bigger batteries are always better in my opinion.

Cameras seem to be Oppo’s focus with the Find X5 Pro. The company has made changes across camera hardware, processing, user interfaces and everything in between. Like its last Find X device, the X5 Pro has dual 50-megapixel primary sensors, but each has a few different specs, lens layouts and even materials.

The ultrawide camera, at the top of the Find 5X, includes a freeform lens. Oppo says this should reduce image distortion – that weird stretched effect at the edges of an image – when putting those wide angles to use. It has a f/2.2 aperture and can cover a 110-degree field of view. It’s a 50-megapixel sensor, which uses pixel binning techniques to offer what Oppo says is “2um binned pixel size”.

Just below it, the second 50MP camera taps into the same pixel binning techniques but has an f/1.7 aperture with a glass and resin combination lens. Using glass in just a part of the lens array should improve color accuracy up to 77 percent, according to Oppo. This camera also features five-axis optical image stabilization to boost your chances of a clear shot, and improve imaging performance across low-light settings. All of those axes of stabilization mean this sensor takes up a lot more space inside the Find X5 Pro, but there was still room for a 13-megapixel telephoto camera with f/2.4 aperture and a new 13-channel spectral sensor. Last year’s Oppo flagship had a mere 5-channel sensor, but this upgrade will improve color accuracy and the precision of white balance settings.

Mat Smith/Engadget

On the front, there’s a 32-megapixel selfie cam, with a new RGBW sensor that should offer, again, more accurate color capture and improved dynamic range. The selfie camera will also monitor whether you’re looking at your phone, keeping the screen lit if you’re still watching or reading.

That’s half the hardware story. Oppo is debuting its Marisilicon X NPU chip in its new phone. During my briefing, the company walked through how the chip enables a new “night-mode” video recording, even at 4K resolution with HDR enabled. There’s a barrage of specs and numbers to explain what Oppo’s homemade NPU is adding, but a lot of it distills down to better and faster signal processing and even improved power efficiency, which could be important if you’re capturing a lot of 4K content.

I briefly tested out 4K Ultra Night Video, Oppo’s take on low-light video recording. While it does certainly help to amplify detail in the dimmest situations, at least when I’ve used it so far, the NPU hasn’t translated to 4K video notably better than its competition.

I tried filming last night with the Find X5 Pro and Pixel 6 Pro, and Google’s footage seemed more natural and less harsh than what Oppo’s processing produced. Foliage detail was a little too excessively blown-out to the point of turning gray – it appeared to tune out the green.

There’s also the new collaboration with Hasselblad. Which isn’t quite so new. We’ve already written about it when OnePlus revealed its Hasselblad-infused camera software in the 10 Pro.

But for the sake of consistency, let’s go over what Hasselblad is adding, aside from cachet.

The phone’s Pro mode offers users some Hasselblad-seasoned fine-grain controls. That means a big orange shutter button and three Hasselblad approved filters for stills and videos. I can’t explain why, but I really liked what each of these ‘special’ filters did to my video and photos. You even get Xpan Mode, where you can shoot video in 65:24 ratio. If you don’t care to do the math, it’s just a very wide, cinematic aspect ratio. For reasons I can’t fully explain, I loved this too.

Mat Smith/Engadget

Now OnePlus and Oppo are more tightly aligned, it’s not surprising to see software overlap like this. Yes, you also get Hasselblad’s natural color solution, where the company has tried to replicate the color accuracy of its very expensive cameras into these smartphone camera sensors.

Oppo is saving a little bit of spectacle for its MWC show, and said spectacle, unfortunately, includes pricing and availability. I’ll update this story once Oppo spills those all-important details. However, you’re unlikely to see the Find X5 Pro in the US – something you may hear several times during Europe’s big mobile show.

DJI’s announcement of the Ronin 4D last year shook things up with features never seen on a cinema camera. It had a full-frame sensor, up to 8K RAW video, one of the first vertical stabilization arms and a LiDAR system that can handle not only focus but gimbal control.

However, things have changed a bit recently. Most crucially, the Ronin 4D is no longer available with Apple ProRes RAW video now that it has gone on sale. Because of that, DJI has also lowered the base price of the 6K model from $7,200 to $6,800.

The Ronin 4D is not without other flaws like excessive weight and rolling shutter, too. Still, there’s nothing else out there that can remotely rival its tech and design. DJI might be best known for drones, but is it actually the most innovative camera company in the world?

Design

Steve Dent/Engadget

The Ronin 4D is a complex product, so let’s take a look at the features and design. First up is the “Zenmuse X9” mirrorless camera designed exclusively for the R4D. I’m testing the 6K version, but DJI plans to release an 8K model later on.

It has a full-frame 24-megapixel sensor and choice of three mounts: DJI’s own DL mount, a Sony E-Mount, and the M-mount used for Leica rangefinder/mirrorless cameras.

DJI has three native lenses including 24mm, 35mm and 50mm models, all with f/2.8 maximum apertures. They’re all pretty lightweight at 180 grams each as they’re designed for gimbal use, and cost $1,199 each – very expensive for f/2.8 prime lenses.

You can also mount Sony and M-mount glass, but lenses have to be relatively small or you can’t balance them correctly. That’s not a huge limitation as it would be pretty unusual to use a telephoto lens on a gimbal.

You can even mount manual lenses and get autofocus via DJI’s optional focus module controlled by the LiDAR system. Finally, it has an incredible nine ND filter levels that give you maximum exposure control, letting you crank open the iris even in bright sunlight.

The main camera gimbal has three axes of movement with quick-release locks for swift balancing. The standout feature, though, is the Z-arm that gives it the “4D” name – one of the first on a gimbal, according to DJI. It uses an internal balancing system combined with multiple motion detection sensors on the front and buttons that help it smooth out walking, running or other jarring vertical movements.

On the right side of the body you’ll find HDMI, time code and DC power ports. On the left are switches to control power, recording, the main gimbal and the Z-axis, along with a multi-dial that controls focus by default. However, it can also be programmed for other functions like gimbal control.

Steve Dent/Engadget

For storage, you can use either a CF express card, external SSD over the USB-C port (much like recent Blackmagic cameras), or DJI’s ProSSD module. The latter is a paid $799 upgrade, but required for high video frame rates. It’s nice to have the three options, particularly the external SSD, but you can only record on one of those at a time – so there’s no way to create a backup while recording.

The Ronin 4D has multiple removable modules, starting with the strong top handle. It’s a good way to hold the camera if you want a low angle, so it’s equipped with a rocker switch to adjust the gimbal (height only) and a button to put it in “sport” mode for quick movements. What it really needs is a record button, though, so you don’t have to reach around to the side.

Attached to that handle is the 5.5-inch, 1,000 nit display that’s the primary way to change camera and gimbal settings. It’s easy to adjust the viewing angle thanks to the nice friction mount, and has a super-logical menu system similar to the one on Blackmagic’s cameras.

It can be operated either via touch or with the 14 manual buttons and dial. Camera settings like f-stop and shutter speed are found along the top and bottom of the display. Power, media, gimbal and other controls are on the left.

Steve Dent/Engadget

Next up are the camera grips. Both attach via a quick-release lever, and you can change the angle to adapt to your working style with another clever lever – even while you’re holding on to it.

The left grip has a joystick to move the main gimbal up and down or side to side, along with buttons for focus and gimbal tracking. The exposure button brings up the waveform, zebra and other tools depending on what you have selected, while the trigger button locks the gimbal or centers it with a double press. Finally, there’s another sport mode button on the inside of the handle.

The right handle has a large wheel that’s primarily used for focus, but also changes camera settings like the ND filter, shutter speed and iris when you hit the “mode” button. Doing that also ingeniously changes the wheel’s operation from smooth to “clicky” with distinct detents. It also has a focus assist button and AF lock button, while the trigger either locks the Z arm or activates/deactivates it with a double press.

Steve Dent/Engadget

You also have the option to mount an optional $399 transmitter at the back that relays video to the High-Bright 7-inch, 1,500 nit remote monitor that’s also optional ($1,699). It lets you not only view video but control many camera and gimbal functions from up to 20,000 away. To do that, you can remove the left and right grips normally used on the camera and attach them to the screen.

It can also record proxy video at up to 1080/60p via the built-in microSD card slot, though that function was not enabled in my review unit. However, I was able to record video to an Atomos Ninja V recorder via the HDMI port.

Also at back is the battery plate that uses DJI’s TB-50 batteries. Those deliver 1-2 hours on a charge depending on operation, according to DJI, but I usually found they were good for just over an hour with active gimbal use.

Finally, the LiDAR focus module is attached above the camera and connected via a locking USB cable. With supported lenses, you get both autofocus and tracking – more on that shortly.

Operation

Steve Dent/Engadget

I was pretty intimidated by all these features to start with, thinking it would be a bear to learn the Ronin 4D. However, it was surprisingly easy to pick up the basic operation and then gradually learn the more granular controls.

Balancing the camera is easier than on any gimbal I’ve tried, as everything is right there in front of you. You just unlock each axis one at a time, unlock it using the quick release controls and slide the gimbal until it’s balanced. Once that’s set, DJI’s “autotune” feature takes care of the fine balance.

You can then turn on the 4D Z-arm for walking or other situations where the gimbal may bounce, but it’s by no means required for simple handheld operation. If you don’t need it, keeping it shut off will save you some battery life. If you do use it, you’ll need to balance for the smoothest operation, but that’s easily done via an on-screen guide.

The gimbal can be set so that it always points the same way or (“locked”) or set in “follow” mode to smoothly follow the operator’s movement. In the latter mode, can set it follow only pan, pan and tilt, or pan, tilt and roll (FPV). The camera can be manually tilted or panned using the joystick, then centered using the C1 button or double pressing the left trigger.

The Z-arm has similar lock and follow modes, with “lock” keeping the camera in the same position. If you move the camera more than the arm’s range of movement, though, you risk smacking the arm against its top or bottom stops. “Follow,” again, allows the Z-arm to smoothly follow your movements.

Steve Dent/Engadget

The LiDAR focus system is impressive, and could be something we’ll see on future cameras if they can shrink it down small enough. It works at a distance of around 33 feet by generating a real-time 3D point cloud.

You can choose between either manual focus, autofocus or assisted manual focus. AMF uses autofocus but also moves the manual focus dial. That way, you can take control by just grabbing the wheel, and give control back by releasing it.

There are three focus areas, or regions of interest as DJI calls them. The first is spot focus, the second is a wide box and the third is “smart” that detects faces or lets you draw a box to track objects.

Focus needs to be calibrated for each supported lens, but once set up, it’s very accurate. Unlike standard autofocus systems, it can work in the dark. The biggest drawback compared to phase or contrast detect AF is that it focuses on the front of a scene, so someone’s nose might be in focus but not their eyes.

You can even set up LiDAR autofocus with manual focus lenses using the LiDAR motor kit. That would normally require a geared cinema lens, but the kit comes with a geared belt you can wrap around other lenses that don’t have that. I tested it briefly with a Sony 28mm f/2.0 lens (with the AF turned off) and found it was reasonably accurate, though it required multiple calibrations.

Steve Dent/Engadget

DJI has plenty of tools for manual focus, too, like peaking. But the best one is the LiDAR focus waveform that shows an object’s shape and distance. After some practice, I was wondering how I ever managed to do manual focus without it.

One of the coolest features enabled by the LiDAR module is ActiveTrack Pro, similar to that found on DJI’s Mavic 3 and other drones. That lets you lock onto a subject and have the gimbal track it while also maintaining autofocus. It automatically detects faces when in “smart” mode, and from there, you just press the track button on the right handle. That turns the yellow box into a green box, indicating that ActiveTrack is alive. You can track other types of subjects simply by drawing a box on them.

The feature works very well, keeping your subject framed exactly as you set it up. It tracks human subjects well even if you move behind or to the side of them. If something goes wrong, the framing can also be tweaked manually using the left trigger. It can get thrown if your subject moves too fast, though, as the camera can only pivot and tilt so quickly.

On the whole, the Ronin 4D delivers smoother footage than any comparable product I’ve tried – even over challenging terrain. The ActiveTrack Pro feature is particularly welcome, as it allowed me to take my eyes off the subject to see where I was going while still maintaining the desired framing.

Steve Dent/Engadget

Finally, let’s look at the video specs and quality. Originally, the Ronin 4D was announced with ProRes RAW along with ProRes HQ and H.264 10-bit recording. However, DJI announced that the RAW option would not be available at launch, even though it worked perfectly when I tested it. The company didn’t say why, but the cinema camera company RED has a patent on that technology that all users must license.

Instead, it’s shipping with Apple ProRes 4444 XQ. That’s a very high quality codec, but it doesn’t offer the advantages of ProRes RAW – particularly the ability to keep the original, pristine sensor data.

Both ProRes options do deliver good quality, but also gobble up storage quickly. The H.264 10-bit option is more efficient, but not very edit friendly – so it would be nice to have something in between those.

You can record up to 6K at 48 fps or 4K at 96 fps using the full width of the sensor, in all codecs. Anything above that, including 6K at up to 60 fps and 4K at 120 fps will introduce a Super 35 crop.

The Ronin 4D’s 24.1-megapixel sensor has dual native ISOs at ISO 800 and 5000, with a claimed 14 stops of dynamic range. Image quality is right up there with other cameras I’ve tested lately, like the Panasonic S1H, Sony A7S III and Canon EOS R3.

Steve Dent/Engadget

In RAW mode with Log settings enabled, it delivers crisp, color accurate video that’s easy to adjust in post. It’s not quite as good in low-light as the EOS R3 or other recent mirrorless cameras, but video is usable up to around ISO 6400, up to ISO 12,800 if you’re careful with exposure. Beyond that, noise can get excessive, particularly with highly saturated reds and blues.

The biggest drawback with the sensor is rolling shutter. While it’s not worse than models with similar sensors like Panasonic’s S1, it’s excessive for a camera that’s designed for movement. Another current issue is audio, as the 3.5mm mic jack isn’t up to snuff for production needs. However, DJI is addressing that with an upcoming module with more production-friendly XLR inputs.

Considering that the Zenmuse X9 is DJI’s first full-frame camera, the image quality is shockingly good overall. It might even make sense for DJI to release the camera separately in some form, as the Ronin 4D form factor isn’t practical for non-gimbal shooting. That way, productions could more easily match up footage.

Wrap-up

Steve Dent/Engadget

In sum, the Ronin 4D is an impressive piece of technology that could change the way producers think about handheld shooting. DJI has created a new product class and invented technology like LiDAR-powered ActiveTrack Pro and assisted manual focus. At the same time, it delivers full-frame image quality with no compromises, apart from the rolling shutter.

So is DJI the most innovative camera company right now? I haven’t seen anything else that can match the level of technical innovation on the Ronin 4D. It’s even hard to compare the Ronin 4D to other products, because it essentially has no peers. And for a first-generation product, the design and build quality is astonishingly good.

With a $6,800 price tag (and more like at least $8,000 properly kitted out), the Ronin 4D is a pretty niche product – though it’s easily worth that price. Even if it doesn’t sell widely, though, it’s a technical tour de force that has put other camera manufacturers, cinema and otherwise, on notice. If DJI keeps innovating at its current pace, that sleepy world could get a lot more interesting.

After a bit of a delay, Panasonic has finally revealed the 25.2-megapixel GH6, the highest-resolution Micro Four Thirds mirrorless camera to date. One can think of it as a mini version of its S1H "Netflix camera," as it's loaded with pro video features like 5.7K ProRes V-Log video, a tilt-swivel display and CFexpress capture. However, it's still stuck with contrast detect rather than phase detect autofocus like most modern cameras. Luckily, I was able to get my hands on a preproduction unit to check out the new features and get an early feel for it. 

First, let's talk about the most interesting changes. As mentioned, the GH6 has a 25.2-megapixel sensor with a high-speed readout, though it isn't stacked like its Micro Four Thirds rival, the OM-1. That obviously gives a much-needed image resolution boost, though photo burst performance is actually down a touch from the GH5 II (8 fps with continuous AF compared to 9 fps). 

The GH6 still uses a contrast detect autofocus system (what it calls "depth from defocus" or DFD) rather than phase detect like nearly all its rivals. As with the GH5 II, it also includes face/eye/head/body/animal AI recognition. While Panasonic has pretty much perfected its DFD system, it's still not as fast or reliable as phase detect systems. 

When shooting people pictures in relatively low light, it sometimes failed to get focus quickly enough to nail the shot, for instance. By comparison, Sony and Canon's latest cameras (the A7 IV and EOS R6) have no such difficulty, even in low light. And I still saw some wobble or hunting with video AF. 

In terms of video, things are more positive. The GH6 can now shoot 5.7K video with V-Log at up to 30 fps with internal capture to either ProRes or ProRes HQ, or 5.7K up to 60 fps to MP4 10-bit H.264/H.265 recording (all-intra or longGOP). You can also shoot 10-bit 4:2:0 4K at up to 120 fps using the longGOP codec. All of these modes use the full sensor width with super-sampled video, though the 4K 120p setting introduces some pixel binning. 

It will also be able to record ProRes RAW to an external Atomos recorder via a future firmware update. More interestingly, the GH6 is one of the first mirrorless cameras with HDMI 2.1 output, so it'll eventually let you record 4K RAW externally at up to 120 fps via a future update. When it first ships, it'll support 4K at up to 60 fps, with simultaneous external and internal recording. 

Steve Dent/Engadget

In terms of video quality, Panasonic is promising 13+ stops of dynamic range using ProRes and V-Log recording. It also claims you'll see less noise at higher ISO settings (ISO 6400 and up). For both video and stills, it wasn't quite as good in low light as the GH5s (which has less than half the resolution) but it wasn't far off. 

Capturing 5.7K video at up to 30 fps in ProRes HQ uses internal data rates of 1,903 Mbps, far beyond SD UHS II capability. As such, the GH6 is the first Micro Four Thirds mirrorless camera with a CFexpress Type B card slot. That's required for ProRes recording and many of the MP4 high-frame-rate codecs. 

The drawback right now is that there's only one CFexpress slot, so you won't be able to get a backup for certain video resolutions unless you use an external recorder. However, the GH6 will eventually support recording to an external SSD recorder via the USB-C 3.2 port, which also allows you to power the camera while shooting.

If you're concerned about rolling shutter with the extra resolution, the high-speed sensor readout is supposed to reduce that more than ever. I confirmed that rolling shutter was very manageable, even at the highest 5.7K and 60p resolution. However, it is present if the camera or subject move too quickly.

As with past GH models, video quality is outstanding, with the 5.7K video particularly sharp and useful for editing. The ProRes option has also saved me an extra step compared to the GH5s, as I no longer need to transcode video. The drawback is that the video files are 4-5 times the size — so you'll want to invest in high capacity CFexpress Type B cards, which can get very expensive.

The GH6 has five-axis in-body stabilization like the other models, but now offers up to 7.5 stops of shake reduction with supported lenses, just behind the 8 stops on Canon's EOS R6. With boosted IS and E-stabilization, it really smooths out video well even for situations like walking, better than most mirrorless cameras I've used. 

The GH6 kind of looks like the GH5 II, but it's quite a bit heavier (823 grams compared to 727 grams with a battery and memory card), and the body has changed in several notable ways. It's much thicker, to start with, thanks in part to an active cooling fan that ensures the GH6 doesn't suffer overheating issues like Canon's EOS R5 and R6. To that end, Panasonic is promising unlimited recording in all video modes, including 5.7K 60p and 4K 120p.

Steve Dent/Engadget

The other major change is with the grip that's considerably deeper than on the GH5 II (or GH5 and GH5s. With that, the GH6 is considerably more comfortable to use than any Micro Four Thirds camera I've tried, particularly with large lenses. Panasonic has made a number of other changes to the button and dial layout as well, most notably with the addition of a dedicated audio button that let you change things like gain levels and quality. It also has an extra record button on the front of the camera that's extremely handy for vlogging .

The GH6 now has a tilt-swivel display like the SH1, with the same 1,840K dot resolution as the one on the GH5 II. That makes the camera thicker, but more practical for low angle and other types of shooting, and the tilt option keeps the screen from being blocked by cables if you use an external recorder. It's also Panasonic's brightest display yet, and was highly visible even in sunlight.

The 3.68 million dot OLED electronic viewfinder (EVF) is unfortunately unchanged from the GH5 lineup. While it's true that the GH6 is more of a video than a stills camera, a lot of shooters prefer to use an EVF in bright sunlight, so a sharper image would have been nice. Also unchanged is the BLK-22 battery that delivers fewer shots than the GH5 II (380 compared to 410) and slightly less shooting time — likely due to the fan and extra resolution.

Steve Dent/Engadget

The GH6 has a couple of other serious downsides compared to rivals. One of them is the relatively small sensor size compared to a full-frame camera. That offers fewer creative possibilities, though it's also more forgiving in terms of focus and makes for cheaper, lighter lenses. 

The biggest flaw, however, is the contrast-detect autofocus. The good news is that Panasonic seems to have finally realized that buyers want phase detect and hinted that it's not out of the question for future models. 

The GH6 was rumored to be priced at $2,500, but there's some good news on that front. It will go on sale in March for $2,200 (body only), making it relatively cheap to models that can match its video specs. It looks a solid choice for that, with video quality on par with any rivals. We'll take a closer look at everything in our upcoming review. 

Logitech’s Mevo Start cameras were built for live streaming. I don’t mean webcam conference calls or video that you can record and edit (though you can do both of those things with the Start). These are live streaming cameras, first and foremost, aimed at the growing number of creators on YouTube, Twitch and elsewhere.

The Mevo Start 3-pack kit ($1,000) consists of three streaming cameras that connect together wirelessly through the companion Multicam app. With the ability to instantly switch and stream between three different camera angles, it has the potential to overhaul how streamers broadcast. Until now, that has usually demanded thousands of dollars of equipment, along with significant expertise.

If you have a strong internet connection, some practice live streaming and, to be honest, a lot of patience during the setup, this could change how you stream. It could also lower the barrier to entry – so I thought I would test it out. Am I willing to make the leap from my camera, PC and OBS setup to three tiny wireless cameras?

Mat Smith/ Engadget

Let’s start with the Mevo Start itself. It’s a solid camera small enough to slip into a trouser or jacket pocket. On the back, it has inputs for mics, USB-C for both charging devices and being charged itself, and a microSD slot too. The unit has a nice matte finish (sadly, yes it’s a fingerprint magnet), and a rubberized power button at the back that’s easy to find without having to look. The 3-pack consists of, yes, three of them, with USB-C charging cables. From there, it’s all down to you to figure things out, which for many, will mean tripods, plural, and microSD cards if you want to record footage of each individual camera. Oh and also a standalone mic if you want to improve the audio quality.

I’m not going to dig too deep into the minutiae of sound and image quality, but the camera mics will pick up everything around you, so most users will need a mic for decent audio talking to the camera – especially in busy environments.

Unless you’re stepping up from a webcam to the Mevo Start, you’ll probably notice a dip in image quality compared to streaming from a flagship phone or digital camera. Streaming resolution on the Start tops out at 1080p at 30 frames per second, which is a little disappointing when we’re starting to get comfortable with 4K content, but also because that means any digital zooming (well, cropping – more on that later), is going to lead to even grainier video. Conversely, this means less heavy lifting for your internet connection, which is a good thing.

While it is not the most high-fidelity streaming option, the Start aims for versatility over specialization. And let’s not forget: the eye-catching feature is this multi-camera setup. There’s nothing quite at this price (and level of accessibility) for a multi-camera streaming setup. Normally it would cost far more than $1,000 to offer a comparable camera ‘studio’ capable of live broadcasting online.

Logitech has made some sensible choices with the Start specifications, too. Each camera can record and stream for up to six hours without the need to be plugged into a power source. This means less cable calculus when packing the cameras and also offers you greater freedom when you’re rigging up the cameras. I balanced one on top of my kitchen cabinet, pointing down as I prepared vegetables for a meal.

The Start also has a very slight light hood around the camera lens to help reduce light flare, while the thread for tripods fits two different types of stands. Logitech has thought about these cameras from the perspective of a stream.

But maybe it should have thought a little more: My first afternoons with these cameras were so frustrating, I wanted to give up.

Once you’ve downloaded the Mevo MultiCam app, you’ll then need to individually download firmware updates to each camera… if your phone or tablet deigns to register the existence of each camera. For me, it led to a frustrating loop of a camera being detected and trying to corral it into a firmware update before it suddenly disappeared from the Bluetooth ether. This part ended up taking hours, turning into days. (Well, a couple of afternoons.)

It was the first sign that consistency is the biggest issue with the Mevo camera setup. And while I could give the product the benefit of the doubt, live streams are fraught at the best of times. I’ve done countless streams for Engadget, flanked by professionals who make sure I’m not screwing things up. I’ve never disappeared from an Engadget livestream, but I can’t say the same of my week of stream attempts with Mevo Start. But before I relive that horror, what is the Mevo Start capable of? I was pleasantly surprised to discover: a lot.

The Mevo app has an auto-director mode that will switch between the cameras without you having to tap on your phone. Better still, you can choose to prioritize one camera and adjust how long the robo-director holds a shot before switching. You can even pick the kind of transition used, from immediate cuts through to slide swipes and other gentler changes. Alas, no star swipe.

It’s these elegant touches that elevate a plain stream into something a little slicker. You can also crop the video on each camera, helping to cut out, say, a messy corner of your apartment, or just offer up a closer look at whatever you’re cooking or crafting.

As the camera maxes out at 1080p, the more you crop, the more fidelity you will lose on your stream, which really limits the usefulness of the feature. While no one streams in 4K, if these cameras had 4K you’d at least be able to crop in without losing too much quality. In the same settings menu, you can adjust the degree of latency you want to use. High latency means that your feed will be held up a little longer before it’s streamed to viewers, ensuring a more consistent connection and – hopefully – video feed. There’s also a low-latency option if you’re extremely confident in your web connection, and this will ensure the video stays closer to real-time.

You can fine-tune the cameras further, with profiles for outdoor lighting and even an HDR toggle. There are also some stylistic profiles for vivid high-contrast, grayscale and more, adding more flexibility depending on your own lighting setup, location or simply artistic tastes.

There’s even a degree of editing available through overlays, where you can add lower third text (like you’ll see most TV news shows, and a lot of YouTube streams) or even introductory full-screen text to your video stream. It’s simplistic but it’s built into the app – it’s easy.

So there is a lot to praise in the Start kit, but then you, well, start using it. The main test for me was to see how easy it was to stream directly to YouTube. The first issue I encountered is one every new streamer will struggle with: you need over 1,000 subscribers to your YouTube channel in order to enable a simple connection. As I didn’t want to mess up Engadget’s official channel cooking up a storm with tempeh (with hindsight, I could have made an unlisted stream), instead I had to use the Real-Time Messaging Protocol (RTMP). And go a little mad.

Once your camera angles are in place and your connection looks stable, you can tap the live button on the app. This will take you through to all the livestream options, including platforms YouTube, Facebook and Vimeo. The RTMP option is all the way to the right of these options, which is tough to find on a phone.

Mat Smith/ Engadget

While you can set up everything on a single device, but I’d suggest having a laptop or something else internet-connected nearby for setting up your streams on YouTube, as you may find yourself flitting between a web browser (needed to tinker with YouTube Studio settings not available on the app) and Mevo multicam app a little too much.

Tap on RTMP, and you can set up the hooks for YouTube to pick up at its end. Other notes that might help you: Server URL and Stream URL are different things, don’t mix them up. Also don’t expect your stream to go instantly live. My Start camera stream typically took a minute to buffer or settle into a feed that the public could watch. What’s frustrating is that YouTube Studio’s interface shows a buffering window, regardless of whether your stream is live or not. I would have to check the shareable link to the stream to check if I was even live. The “live” labels on the Mevo app itself are not accurate.

The multicam app was mostly responsive and did what I expected it to do, but it was also fussy. Swap between one too many apps (remember how I initially tried to adjust my Youtube settings on Safari on the same phone?), and the connection will cut out. Absent-mindedly lock your screen, and again, the stream is lost. I think it’s common sense that you probably didn’t want to cut a livestream connection if you locked your phone. Or maybe I’m making excuses for myself. It’s probably a little of both.

Despite having a healthy internet connection and keeping in close proximity to the cameras, I lost my stream a few times for reasons I’ll never know. And that’s a shame, as there’s so much potential in this Mevo camera trio.

While I’ve touched on some of the settings you can access on the cameras, there’s further flexibility inside the app. You can tweak frame rate, suppress echo and even choose the bitrate and recording format for your backup recording, all of which will be stored on your iOS or Android device. I like that these backup files are neatly stored in one place.

Mat Smith/ Engadget

For beginners or streamers looking to upgrade their streaming style, there are a few bumps on the road to a flawless multicam setup. Mevo does offer a glimpse at an effortless solution – it’s definitely streamlined many of the processes, and the jack-of-all-trades specs mean that this could be the ideal setup for upgraded live music streams, community projects or simply to add polish to existing livestreams. You will have to consider consistency though. My experience has been mixed, but others have fared better. There’s also a vibrant user community across both Facebook and Reddit that was able to help me when things went wrong. Hopefully, there’s a future Mevo camera in the works that’s a little more stable and polished. And maybe 4K capable.

OM Digital Solutions has unveiled the Micro Four Thirds OM-1, its first mirrorless camera since purchasing Olympus. It packs a new sensor and technology that gives it a big leap in performance over the last flagship Olympus models and could be the last Olympus-branded model you'll see. 

Marking the 50th anniversary of the original OM film SLR, the OM-1 offers some big improvements over the E-M1 Mark III. That starts with the stacked, backside illuminate 20.4-megapixel Live MOS sensor with a "quad pixel" AF arrangement using four photodiodes per pixel. With the higher-speed readout and updated TruePic X processor, it can shoot RAW photos at speeds of up to 50 fps in electronic shutter mode with phase-detect autofocus enabled (with supported PRO series lenses), or 120 fps with AF locked. 

OLYMPUS CORP.

With the stacked design, you'll also see far less rolling shutter when shooting in electronic mode. That's a good thing, as speeds go down considerably to 10 fps with the mechanical shutter. Unfortunately, you can only capture up to 139 RAW images at a time with the mechanical shutter, or just 96 RAW images at 50 fps — less than 2 seconds of shooting. Those speeds are no doubt limited by OM's choice to use dual SD UHS-II card slots, rather than something faster like CFexpress. 

Though the OM-1 is mainly aimed at wildlife photographers, it can handle more detailed images as well thanks to the High Res Shot mode. That allows you to capture 50-megapixel images handheld in just five seconds, or 80-megapixel images on a tripod. Other modes include live ND to simulate long exposures, HDR, and focus stacking, which delivers an image where everything is in focus.

OLYMPUS CORP.

The AF is considerably improved in terms of AI smarts, as well. OM says that its latest face and eye-detect algorithms are much more responsive, and it now detects cars, airplane and trains on top of people, birds and animals. It also has a new subject tracking feature that's completely separate from face and eye detection. 

The OM-1 has big improvements in terms of video, as well. You can now shoot 4K (UHD or DCI) at up to 60 fps, in either 8-bit H.264 or 10-bit H.265 modes. The latter works either in OM-Log mode for improved dynamic range, or HLG mode for HDR work. Better still, the OM-1 can output 12-bit Apple ProRes RAW video at up to 60 fps to an Atomos Ninja V or Ninja V+ external recorder, supposedly without any subsampling (pixel binning). 

OM Digital Systems

Other feature include 8 stops of in-body image stabilization with supported lenses (7 stops without), high-end build quality with an outstanding IP53 camera rating against dust, water, and freezing conditions, a redesigned menu system, 5.76 million dot 120 HZ electronic viewfinder and battery that delivers 520 shots on a charge (CIPA rated) and 90 minutes of video recording. It has a fully articulating 1.62 million dot rear display that should make this camera ideal for vlogging.  

The OM-1 looks to be a strong rival to Panasonic's GH5-II Micro Four Thirds camera, offering faster shooting speeds and superior phase-detect autofocus, while matching up well in terms of video. It remains to be seen how it will stack up against the upcoming Panasonic GH6, however, which is set to be launched on February 21st at 8PM ET. 

The OM Digital's OM-1 is now available on pre-order for $2,200, with delivery set for early March. Olympus also unveiled a pair of new lenses, the M.Zuiko Digital ED 12-40mm f/2.8 Pro II ($1,000) and the M.Zuiko Digital ED 40-150mm f/4 PRO ($900). 

After Sony released the A7 III way back in 2018, I wondered if Canon and Nikon could ever catch up to its autofocus and other technology. With the launch of the 24-megapixel EOS R3, however, it’s Canon that’s on the cutting edge with its “eye control” AF that lets you focus on a subject just by looking at it.

The R3 is also Canon’s first camera with a backside illuminated, stacked sensor. That gives it burst shooting speeds in silent mode of up to 30 fps with autofocus and auto-exposure enabled. Plus, it has top flight video specs with RAW capture at 6K up to 60 fps.

Here’s the rub, though: For $6,000 it has relatively low resolution, while Canon’s own 45-megapixel EOS R5 costs more than $2,000 less. So, who is this camera good for exactly? And is the eye control AF a useful feature, or just a toy? With some help from my pro photographer friends, I spent some time with a final production model to find out.

Body and handling

The first thing you’ll notice on the R3 is the big camera grip body that makes it look like a DSLR and mirrorless camera mashup. Much like the 1DX Mark III that inspired it, it has dual controls for both landscape and portrait shooting modes, including a matching shutter button, dials and joystick.

Despite the big body, the EOS R3 is relatively light at 2.3 pounds including a battery and memory card. The 1DX Mark III, by contrast, tips the scales at 3.17 pounds, while the EOS R5 is 1.6 pounds.

As you’d hope for a big camera, it has a big grip. That gives a feeling of security when you’re holding it, and it handles great even with big telephoto lenses like Canon’s RF 70-200mm f/2.8.

The R3 also packs plenty of dials and buttons so that you can operate the camera without diving into the menus, for the most part. The control layout is similar to the 1DX Mark III and uses the same infrared control button. That lets you set the focus point just by sliding your thumb over it – faster than a joystick, but easier to activate inadvertently until you get used to it.

Steve Dent/Engadget

A dedicated button lets you switch easily between photo and video modes, while giving you separate settings and menus for each mode. However, it uses the same old Canon menus, which aren’t quite as intuitive as on the latest models from Sony, Nikon and Panasonic. You can control it in a variety of ways, though, using the dials, joystick or touchscreen. That makes it quick to get to a setting, once you figure out where it is.

The big 3.2-inch touch display has a crisp 4.15-million dot resolution, nearly double that of the R5. For scrolling through menus, browsing photos and more, it’s extremely responsive – much more so than Sony’s A1 or any other camera I’ve tried for that matter. The display also flips out for low-angle shooting or vlogging, though the R3 is a bit heavy for the latter.

The 5.76-million dot OLED EVF is similarly sharp and offers a smooth 120 Hz refresh rate. It doesn’t stack up on paper to the 9.4-million dot, 240Hz EVF found on the A1, but I couldn’t see much difference between them, to be honest.

It uses the same huge LP-E19 battery from the 1DX Mark III that delivers up to 620 shots on a charge or about two and a quarter hours of 4K 30p video shooting. Depending on how you shoot, you may easily be able to exceed those figures, however.

For storage, you get one fast CFexpress type B and one SD UHS II slot. It’s nice to have the SD option if you prefer to shoot with those cards. However, if you want to back up the CFexpress card with SD, it’s going to slow performance. Since this is mostly a pro camera, Canon should have offered dual slots for both like Sony does, or perhaps two CFexpress slots.

Steve Dent/Engadget

It has a new hotshoe interface that supports Speedlite and other flashes along with accessories like Tascam’s new XLR 4-channel microphone interface. That finally matches what Sony and Panasonic have offered for years now.

The fiddly micro-HDMI port is not ideal for video shooters, and it’s an odd decision considering Canon had space with the R3’s big body. It also comes with USB-C for data transfers and in-camera charging, though you’ll need another optional accessory if you want to power the camera externally while you shoot.

Now, let’s talk about Canon’s new eye control feature. It’s only really useful if you calibrate it, but luckily that’s easy to do – you just stare at five dots. You might need to calibrate it multiple times depending on whether you’re wearing glasses and contacts, or even for different environments. Luckily, you can save up to six different settings.

Once it was calibrated, I could select an object to focus on just by looking at it. Even if the eye control circle wasn’t quite on a subject, the autofocus would usually select it if it was close. From there, the face, eye or subject tracking would kick in as needed to track the subject.

It worked even in tricky environments with lots of subjects or movement, though it turns off once you hit the focus button and start shooting. It worked fine for me, but didn’t function at all for my photographer friend with light blue eyes and an astigmatism. So if you’re interested in the feature, you may want to test it out before making a purchase, as the functionality seems to depend on your eye color and other factors.

Performance

Steve Dent/Engadget

First and foremost, the EOS R3 is a speed demon. That starts with the shutter, which goes all the way up to 1/64,000th of a second in electronic mode, faster than any other consumer camera. It also supports some of the fastest bursts we’ve seen, up to 30 fps in silent mode or 12 fps with the mechanical shutter. Unlike some recent Sony models, you get those full speeds with uncompressed and not just lossy RAW files.

You can take a lot of shots in those modes, too. At 30 fps with the electronic shutter, you can shoot 150 shots to an SD UHS II or CFexpress card before the buffer fills, according to Canon. However, I was able to shoot many more than that with a fast CFexpress card, with just a slight reduction in shooting speeds after the 150 frame mark. It’ll handle 1,000 RAW uncompressed photos or more with the mechanical shutter before stopping.

The Dual Pixel autofocus can keep up with those speeds too, so I had very few photos that weren’t sharp. Face and eye detection is fast and fluid for people, though a bit less reliable for animals or birds. The EOS R has a car tracking feature that’s mainly designed for race cars, and unfortunately I didn’t have access to a Bugatti Chiron during testing.

Shooting sports is this camera’s forte, and at an indoor soccer game with decent lighting, Samuel, the pro photographer I was working with, had only a few out-of-focus shots. It didn’t perform quite as well as the A1 did for birds, but it was still better than most cameras I’ve tried. Overall, the EOS R3 has a very powerful AF system that puts Canon right up there with Sony.

The IBIS system can deliver 8 stops of shake reduction with supported lenses, more than any rival camera. That allowed me to get sharp shots handheld at low shutter speeds when shooting in low light. And thanks to the sensor’s fast readout speeds, rolling shutter is well controlled and only noticeable on fast-moving subjects or quick pans.

Image Quality

The EOS R3 may be Canon’s best mirrorless camera yet for image quality, particularly when it comes to dynamic range. The new 24-megapixel sensor offers at least a stop more dynamic range than the EOS R5, giving you more room for adjustment with RAW images. JPEGs also look great straight out of the camera, with sharpness and noise reduction well balanced.

Canon’s color science is still the best out there, delivering stellar color accuracy and natural skin tones. As usual with Canon, it has a slight bias toward warm tones.

The R3 really shines in low light, too. Noise is nearly non-existent up to ISO 3200, with hardly any noticeable drop in dynamic range. It remains well controlled up to ISO 12,800, and images are usable at ISO 25,600 and even higher if you expose them perfectly.

The biggest downside is the lowish 24-megapixel resolution. With the 45-megapixel EOS R5 or 50-megapixel Sony A1, you can shoot birds or wildlife at a longer distance and still have room to crop in. The R3 is far more limited if you want to retain detail.

Along with my own testing, I loaned the EOS R3 to professional photographer Samuel Dejours from Studio NathSam in Gien France, to get his opinion. "The EOS R3 really handles great, though it's quite heavy. The Sony A1 is our usual camera and it's much lighter," he told me. "However, it does feel better balanced when you use a large telephoto lens. I really liked the infrared autofocus control button; I found it to be much quicker than using the joystick."

"I was excited to try the eye autofocus. However, despite all attempts of calibration, etc., it simply didn't work for me. Perhaps with my light-colored eyes, it didn't work. Hopefully that's something they can fix with a firmware update."

"It was nearly at the same level of the A1 for autofocus, though the A1 was perhaps more rapid for certain types of shooting. It was great for sports, with pretty astonishing speed and AF that almost always nailed focus. In terms of image quality, I'm not sure I'd be able to tell the A1 and EOS R3 apart if I didn't know which was which. That makes other things become more important, like the resolution, speed, etc. Overall, it was an excellent experience, and it shows that Canon is really close to Sony for speed, autofocus, video and image quality."

Video

For a sports camera, the R3 is a surprisingly strong for video. You can shoot DCI 6K RAW at up to 60 fps using the full width of the sensor. It also supports full-frame 4K shooting at up to 120 fps using All-I capture, albeit with subsampling that can reduce detail. Both of these modes also require a CFexpress card as SD-UHS II is too slow.

All other 4K modes at 60p, 30p or under support full sensor oversampling, allowing for extremely sharp video with fine detail. You can also use an APS-C crop for natively sampled (pixel-for-pixel) 4K footage that’s just a touch less sharp, if you need to zoom in slightly for example.

What about overheating? Luckily, the EOS R3 has far fewer restrictions in that regard than the R5 or R6. Regular, oversampled 4K is not temperature limited, and 6K RAW or 4K 60p is good for at least an hour. 4K at 120 fps does have a 12-minute limit, but that’s an exotic use case that would affect very few users. While using the camera, albeit in coldish weather, I received no temperature warnings – even during long takes.

Steve Dent/Engadget

The camera’s excellent high ISO performance is handy for video, allowing more flexibility for indoor shooting. Skin tones are natural and colors accurate, even in relatively low light. The R3 also supports Canon Log 3 or PQ, along with RAW or RAW light, letting you max out the dynamic range. That in turn gives you options to create HDR videos or adjust images in post.

Complimenting all that is the Dual Pixel, AI-powered AF. As we’ve seen before with Canon’s system, it reliably nails focus without any hunting or other annoying issues. Eye tracking is generally fluid and reliable as well.

It’s not quite as sticky as the AF on Sony’s A1, however. It occasionally focused on the background rather than foreground subjects or missed focus altogether. This might be fixable in a future update, though. Keep in mind that the eye control AF function only works with photos and not video.

As for the in-body stabilization with video, it’s only really good for stable handheld shooting and you’ll need a gimbal for anything else. If you try any sudden moves, the system has a tendency to jolt the image, potentially ruining shots. The electronic IBIS mode helps, but it’s not quite as smooth as I’ve seen on cameras like the A1. Meanwhile, rolling shutter is present, but it’s better controlled than on the R5.

Wrap-up

Steve Dent/Engadget

The EOS R3 is a speed demon with lowish resolution, solid video chops and a high $6,000 price tag. With that odd mix, it’s mainly designed for news or sports journalists. Other users would do better with higher megapixel cameras like the $6,500 Sony A1 or Canon’s $3,900 EOS R5.

It’s actually Canon’s best mirrorless camera yet for video, which is something I didn’t expect. Though it doesn’t have 8K like the R5, it does offer 6K 60p and has far fewer limitations. It’s hard to justify the price for video alone, however, considering other options out there like the A1, $2,500 EOS R6 or Canon’s all-new $4,500 EOS R5C cinema camera.

Still, it could be a hit in the pro market if Canon can pry the 1DX Mark III and other DSLRs out of photojournalists’ hands. For the rest of us, the EOS R3’s awesome speeds and video capabilities are hopefully a preview of what’s to come in more affordable models down the road. Most importantly, the innovative eye control AF feature, while limited, shows us that Canon can beat Sony at the technical innovation game when it tries.

Canon has revealed the EOS R5C, a new member of its cinema camera line that looks like a good choice for photos, too. It strongly resembles Canon's EOS R5 mirrorless camera, but has a big hump in back to accommodate an active cooling system. As such, it offers most of the benefits of the EOS R5 with none of the overheating issues when shooting 8K or high frame rate 4K video. 

On the video side, the EOS R5C can shoot 8K at up to 60fps in Canon's 12-bit RAW LT format using the entire width of the sensor, a big step up from the 8K 30p RAW available on the R5. Better still, it can work at that setting "indefinitely," while the R5 is limited to just 20 minutes at 8K 30p due to overheating issues. It can also handle RAW 6K/60p with a Super 35mm crop and shoot 4K at up to 120 fps with no crop and full AF capability. The latter could use line-skipping like the EOS R5, however.

It also supports ProRes RAW output to an external recorder via the HDMI port at up to 8K/30P. "Proxy data can also be simultaneously recorded to an SD card in-camera, helping to provide efficient post-production operations," Canon said. Unfortunately, it uses micro HDMI rather than a full-sized port — not ideal for a dedicated cinema camera. 

Unlike the R5, however, the R5C doesn't have in-body stabilization — so any optical shake reduction for stills or video will only be available via supported lenses. However, the optical lens stabilization can work in concert with Canon's electronic stabilization, with a 1.1x image crop.

On the photography side, the full-frame 45-megapixel sensor is a big plus, as is the 20 fps shooting speed in electronic shutter mode (12 fps with the mechanical shutter). It uses Canon's Dual Pixel autofocus system for both stills and video, along with eye/face detection and subject tracking, so it should offer the same excellent AF performance as the R5. When you flip the camera over to photo mode, all the menus and button settings change accordingly. 

It comes with a 3.2-inch variable-angle flip out LCD monitor and 5.76 million-dot viewfinder, the same as the EOS R5. While you get Canon Log 3 for improved dynamic range, the Log 2 option found on other Canon cinema cameras isn't available. Other features include dual card slots (one CFexpress and one SD UHS-II), animal eye detection (cats, dogs, or birds), vehicles detection, a multifunction shoe for microphones and other accessories, a timecode terminal for multi-cam shoots and a DC coupler to provide continuous power.

The EOS R5C will arrive in March for $4,499, a $600 premium over the R5. It's a pretty interesting model as it can do more than some Canon cinema camera models for a lot less money and is much smaller, to boot. At the same time, it might give photo/video hybrid shooters serious pause if they're looking to buy Sony's $6,500 A1 hybrid.