The Sony A7R V is a full-frame mirrorless camera with 61 Megapixels, 8k video and a brand new autofocus system which uses AI to better-detect and track subjects.
Announced in October 2022, it arrives just over three years after the Mark IV version and costs around $3900 or £4000 pounds. At the time of making this review, the previous R IV A model could be picked up for around $3000 or pounds, so the big question is whether it’s worth spending an extra grand on the new version or bagging a bargain on the old one.
To find out I spent some quality time testing a final production A7R V and in this review I’ll show you what’s new, how it performs in practice, and ultimately if it’s the right camera for you. As always, my complete review is in the video below, but if you prefer to read a written version of the highlights, keep scrolling!
Let’s start with the weather-sealed body, which shares a similar control layout to the earlier R IV, with the minor refinements of subsequent models. As before the R V continues to miss out on any controls on the upper left surface, while on the right, dedicates one of the two function buttons to movie recording, converts the earlier EV dial to a blank customisable control, and simplifies the main mode dial.
The movie and S&Q positions from the older R IV dial now find themselves on a separate and more sensible collar control around the base, allowing you to switch between them and still photos, while maintaining the main exposure mode. This is the same approach as the A7 IV.
Round the back, the previous record button to the right of the viewfinder has now become a custom function button, while the finger, thumb and wheel controls, along with the joystick remain present.
But while the R V may look similar to its predecessor at first glance, Sony’s made some big changes on the composition. The EVF inherits the 9.44 million dot OLED panel of the Alpha 1 with 0.9x magnification, making it both bigger and potentially more detailed than the 5.76 million dot panels of either the R IV or Canon EOS R5.
It may lack the optional 240fps mode of the Alpha 1, but can still deliver a tremendous view when composing or playing back images. Note to exploit the full detail though you will need to set the Display Quality to High and accept a temporary reduction in detail as the camera focuses; obviously this becomes limiting if you’re continuously autofocusing on a moving subject, but it’s actually a common restriction across multiple systems, and again in Single AF or playback it looks great. Note the EVF resolution will also reduce when you choose 120fps refresh.
The screen now sports a larger and more detailed 3.2in 2095k dot panel (although the interim R IV A was also better than the original R IV in this regard), but what makes it really special is the articulation, switching the vertical tilt of the R IV’s screen for a new four-axis mechanism. This attempts to satisfy the desires of both tilt and flip fans, by effectively mounting a tilting mechanism atop one that flips.
As such you can perform a quick, simple and discreet vertical tilt up by 98 degrees or down by 40, or angle the entire unit out using the side-hinge to flip forward for framing pieces to camera or fold back on itself for protection. You could even tilt the unit and then flip the screen out to the side, thereby avoiding getting in the way of the side ports.
It’s such a good idea you wonder why no-one’s thought of it before. Oh hang on, they have. Panasonic offered something similar on the Lumix S1H three and a half years earlier, and again on the GH6 after that.
In practice it really does offer the best of both Worlds on the R V, although with a hinge upon another hinge, it can’t help but lack the solidity of a simpler solution as you wrangle it to the desired angle. That said I liked using it and assume it’ll end-up on future Alpha bodies.
Sticking with the displays, Sony’s also redesigned its user interface with new menus and wide support for touch control too. It’s interesting to remember the earlier R IV had barely any touch control beyond positioning the AF area, but thankfully Sony has since caught up with rivals in this respect.
In addition the R V also inherits the Main menu views from the FX30, showing a bunch of settings at a glance with the chance to adjust some by touch as well as traditional controls. It’s a useful view, but to me feels a lot like an expanded version of the existing Fn menu, which also remains available here. Do we need both going forward?
Before moving onto the sensor and quality, a quick update on connectivity. The R V keeps the microphone and headphone jacks of the R IV, along with both USB C and Micro USB ports for simultaneous power and tethering for example, but in a welcome upgrade, switches the old Micro HDMI port for a full-size version. Take that Canon!
Oh and the USB C port has also been updated to support the faster charging speeds of Power Delivery sources with Sony claiming it’s now three to four times faster than before, and you can also run the camera on USB power if you like.
The USB C port supports 3.2 Gen 2 for 10 Gigabit/s speeds as well as optional Gigabit ethernet adapters, making it quicker for tethered shooting. The built-in Wifi is also faster than before.
In another welcome, albeit unsurprising upgrade, Sony’s equipped the R V with the twin dual-format card slots of recent Alphas, with each slot able to accommodate either SD or faster CF Express Type-A cards.
Sadly the latter are still slower and more expensive than the CF Express Type B cards used by everyone else, but in Sony’s favour you can at least still use a pair of SD cards if you like, whereas those with Type B support typically sacrifice one of their SD slots.
There’s pros and cons to both approaches, although I should note Sony’s dual format slots do require SD to be inserted with their tops facing away which is opposite to normal and constantly catches me out.
Meanwhile, power is still delivered by the NP-FZ100 Lithium Ion pack, good for between 450 and 550 shots under CIPA conditions, and which again can be quickly charged over USB using Power Delivery. The R V is also compatible with the optional VG-C4EM grip which can accommodate two battery packs.
Moving inside, the R V employs what appears to be the same 61 Megapixel Back-Illuminated sensor as its predecessor, but now paired with the latest BIONZ XR image processor and the brand new AI Processing Unit.
Sony’s also added a visible light and IR sensor from recent bodies to improve auto exposure, which, coupled with the AI Processor should also make auto white balance more accurate.
The latest image processor means the R V now gains HEIF as well as JPEG options, and for the RAW shooters there’s now the choice of three Lossless Compressed sizes in addition to lossy Compressed and Uncompressed formats. There’s also now four compression levels for JPEG or HEIF files from Super Fine to Light.
As before, the maximum resolution remains 9504×6336 pixels and two lower resolutions are also available, with all three sizes available in a choice of four aspect ratios. Like the R IV before it, the sensor has so many pixels to start with that even when cropped to the APSC frame, there’s still a respectable 26 Megapixels remaining.
Before going any further, it’s interesting no-one’s challenged Sony on resolution in the full-frame format. The earlier R IV introduced the 61 Megapixel sensor back in July 2019, but today both Canon and Nikon still offer no more than 45 Megapixels on their most detailed mirrorless cameras. So if you’re after the highest resolution in a single frame in the full-frame format, the Alpha R IV and V will be on your shortlist. But which one to get?
Let’s start with pure image quality and this landscape view I shot with the R V at 100 ISO fitted with the FE 24-70 G Master Mark II lens at f8; in my review of that lens, it proved to be one of Sony’s sharpest. Zooming-in for a closer look reveals a tremendous amount of real-life detail with huge potential for big prints or tight crops. But in terms of pure resolution, no different from the R IV.
If you desire even more detail, the R V inherits the Pixel Shift mode of the R IV, capturing four or 16 frames with the electronic shutter and using the stabiliser to subtly shift the sensor between each one.
These are then combined in Sony’s Imaging Edge software later to generate an image with reduced colour moire artefacts and, in the case of the 16-frame version, up to 240 Megapixels of detail. At the time of making this review, Canon did not offer a Pixel Shift mode.
I’ve tested multiple pixel-shifting modes across a variety of cameras and while they all boosted the detail on static portions on the frame, anything in motion resulted in undesirable artefacts. This is why they’re great for archive or studio-based product photography, but less suitable for landscapes or architecture where a surprising amount on the frame can be in motion.
However Sony’s now released a new version of Imaging Edge which supports Motion Correction to better deal with parts of the composition which move during Pixel Shift capture.
This is billed as a new feature for the R V, but since it’s applied in software running on your computer, I wondered if it could also benefit older models. Well I’ve got good news and bad news: Motion Compensation is available for the Alpha 1, but sadly no earlier models. I’m not sure why or how this restriction works, but the bottom line is you can apply Motion Compensation to Pixel Shift files captured by the R V, but not by the R IV.
To see how well it works in practice I reshot my Central Park landscape with the 16-frame Pixel Shift mode, selected the files in the latest version of Imaging Edge and made two new versions, one without Motion Compensation to represent the older R IV, and one with Motion Compensation to see what, if any benefit the R V could enjoy over it.
So on the left we have the single 61 Megapixel frame which you saw earlier, and on the right the 16-frame Pixel Shift version processed without Motion Compensation. When examining a static portion of the frame, the Pixel Shift version on the right definitely has greater real-life detail, but if I move to a different portion of the frame, you’ll see the undesirable artefacts around areas that were in motion – and with water, foliage and people all moving during the capture, that’s a big issue for this shot.
Now let’s return to the static area, again with the single 61 Megapixel frame on the left, but now the Pixel Shift version on the right has Motion Compensation enabled. This unsurprisingly has no impact on the building details, but moving to the areas with portions in motion show how much better the software has dealt with the compositing problem.
Sony wouldn’t reveal how this works, but I’m assuming it’s comparing the frames in the burst and only taking detail from, say, one frame for areas in motion.
Thing is, this would result in a high-res image with holes of lower resolution portions where motion had occurred, but comparing the single frame with the composite on the right shows that’s not the case.
Some additional processing is clearly taking place so that the areas of motion at least lack the jagged edges of the single frame on the left.
Here’s another example, starting with a single 61 Megapixel frame, before zooming-in on a detailed area, where again the image looks great. But now here’s the single frame on the left and the 16-frame Pixel Shift version on the right, processed without Motion Compensation where you can see a boost in detail for the static regions, but again undesirable artefacts for those areas which moved.
And now switching the crop on the right for one with Motion Compensation enabled, which has again dealt with the areas of motion pretty effectively.
Here’s another portion of the frame showing pedestrians walking along, obviously captured just fine on the single frame on the left and looking very odd on the right for Pixel Shift without Compensation. But now here’s the version with Compensation and it’s looking much better.
Obviously your mileage will vary, but from my tests, I’d say the latest version of Imaging Edge has greatly expanded the subjects and situations where you can exploit the benefits of Pixel Shift. It’s allowed it to move out of the studio and more into the real World and become a tool I would certainly consider for landscape and architectural photography.
So while the R V is no better than the R IV in terms of pure resolution on a single frame, or indeed when capturing static subjects with Pixel Shift, the ability to deploy Motion Compensation to effectively capture more dynamic scenes gives the R V a quality edge over its predecessor. Plus you’re also getting more file options including HEIF and three Lossless RAW sizes.
Next for a run through of the extended sensitivities from the lowest 50 ISO to the highest 102400 ISO, and in the absence of RAW support from Adobe at the time I made this video, I’m showing you out-of-camera JPEGs.
Like most cameras I’ve tested, the extended low sensitivities of 50, 64 and 80 ISO are really only there to achieve specific exposures and actually suffer from reduced dynamic range. The best quality, as always, starts at the base sensitivity which on the R V is 100 ISO.
In this test, the R V maintains most of its detail right up to 800 ISO and only loses a little to noise and processing at 1600 ISO. But from 3200 ISO onwards, there’s a significant drop in detail due to the smearing effect of noise reduction. You can really see it here on the leaves which steadily lose their surface detail up to 12800 ISO, after which the R V lets noise speckles become apparent on the image.
To be fair the R V isn’t designed to be used at high ISOs. Most of the landscape, architecture and portrait photographers it’s primarily aimed at will be able to deploy the lower sensitivities for the best results, but if you need them, the high ISOs are there.
The R V also becomes a more practical camera to shoot handheld thanks to improved sensor-shift stabilisation, or IBIS. Sony claims the five-axis system is now good for up to eight stops of compensation, versus the 5.5 on the R IV.
You’re looking at two photos I took handheld with the R V at 70mm, both at a shutter speed of 1/10. On the left without stabilisation, and on the right with IBIS alone enabled; there wasn’t any additional optical stabilisation here.
A shutter of 1/10 was the slowest I could handhold a sharp image at 70mm when IBIS was enabled, and it’s clearly far superior to the result without. Indeed thanks to the unforgivably high resolution of the sensor, the slowest I could handhold a sharp result on the R V at 70mm without IBIS was 1/320. So on the day, IBIS on the R V gave me five stops of compensation.
But Sony claims up to 8 stops, so let’s now put my unstabilised result at 1/320 on the left with my stabilised one at 1/10 on the right. This is the five stops I achieved on the day. Now let’s switch the crop on the right for one taken a stop slower at 1/ 5, so this is a six stop difference and some wobbling has become visible. Next one stop slower at 0.4 seconds for a seven stop difference and the motion is more apparent, and finally at 0.8 seconds for an eight stop difference, where the shake is now quite apparent.
Obviously like all stabilisation systems, your mileage will vary, not just influenced by the lens you’re using, but also your technique and environment on the day. I may not have achieved more than five stops at 70mm in this test, but that’s still more than I’ve previously measured from the R IV and it allowed me to handhold the R V at respectably slow shutter speeds while exploiting its full resolution.
Ok, now for arguably the biggest new feature, certainly a headliner on Sony’s own marketing: the new AI Processing Unit which allows the R V to support AI-based autofocus with improved subject recognition.
This now adds insects, cars, trains and planes to the basic detection list, while refining existing bird and animal eye detection by adding grazing animals like sheep as well as head and body recognition for them. This should in turn allow the camera to identify and keep tracking a subject when an eye isn’t always in view.
But most importantly it can now have a go at estimating human poses, which should allow it to better track people when they’re turned away, maybe wearing a mask or helmet, or when they’re not the dominant subject. Even face-on with eye-detection, the camera will now try to prioritise on the surface of the actual eyeball rather than eye-lashes.
And while I believe the imaging sensor is the same as before, the R V employs a different shaped array of phase-detect autofocus points covering 93% of the frame vertically and 86% horizontally. Compare that to the array on the R IV which covered an area 99% vertically and 74% horizontally.
So while the R V still doesn’t offer the 100% (or thereabouts) coverage of other models in the range, I’m happy to trade a little vertical coverage for a slightly wider array, plus you’re still getting contrast-detect in the borders, just try to keep any anything in motion away from the far left and right edges if possible.
Starting with humans, I filmed the back of the screen to show it in action. Notice how distant subjects are now identified by their torso, which then switches to head, then face, and then eye when closer. If the subject turns away, the R V switches back to head or torso. This is the AI part in action, using what it’s learned about human bodies and poses to better track them in a variety of situations.
The on-screen tracking boxes certainly stuck with the desired subject better than the previous R IV, which may have had good eye-detection but lost human recognition as soon as subjects turned away. Here you can see the R V tracking the eye as the subject turns to profile and back.
But it’s important to remember Sony’s rivals have also become pretty good at this. Here’s Fujifilm’s latest X-T5 also tracking me pretty well in profile, and briefly switching to a box around my head when turned away. Canon’s face and eye detection also works well in profile.
Let’s take a closer look at some portraits I shot with the R V. When using a sharp lens like the FE 50mm f1.2 G Master, the potential for detail is again amazing, at least when the AF system delivers the goods.
While the AF box during composition stuck fast to the subject’s eye, I still captured a few images where the eye was a fraction out of focus or the eye-lashes were sharper than the eye-ball. To be fair this was with the lens at f1.2 with its unforgiveably shallow depth-of-field, and it was also no worse than the best of its rivals, but equally not the jump in performance I’d expected.
Moving onto animals, here’s a brief test at a Sony event showing how the AF system can now recognise the bodies and heads of dogs, as well as other animals. So even if an eye isn’t visible, you should still get an image of the animal itself in focus. This seemed to work well in my brief tests.
New to the R V is insect detection, and here it is again at a Sony event, sometimes placing a box around the subject, but at other times not. I found its success varied hugely depending on the insect in question, and to be fair, they certainly come in an enormous variety of shapes and sizes.
Later I tried it outdoors with bees on flowers, where it successfully identified and followed the subject. I expect this to improve over time.
Also new to the R V’s subject detection list are cars, trains and planes, and here you can see the system recognising cars with ease, concentrating on their noses at the front. This also worked well with trains and planes in my tests, concentrating on their front windows when sufficiently close. Again useful, but present on many rivals too.
As before though, to help the AI system know what you’d like it to prioritise, the R V still asks you to select the desired subject type from a list, and while animals and birds now have an additional joint option, you know, for times when you might photograph animals and birds, what I really want is a fully auto option.
I realise this would greatly slow-down a recognition system, plus of course there’s the potential for confusion if there’s someone driving a vehicle, or a human with a pet for example.
But Canon’s already offered an Auto subject option on the R6 II which I look forward to trying out, and they also have a system on the EOS R3 which can track your eye to see what you’re actually looking at, allowing you to directly tell the camera what to focus on. Right now, it’s expensive and demands a chunky viewfinder head, but it actually works.
I guess what I’m saying is when Sony says it now has AI-based autofocus with a dedicated processor, I hoped for greater actual automation and results that were significantly better than the competition. Better than the R IV, definitely, but not better overall than say the EOS R5.
I was also bewildered by the number of sub-menu options available, some of which appeared slightly contradictory or conflicting. Again I understand they all allow you to drill-down to specifics to improve the recognition, but I just found them all confusing.
I also wonder whether the R V sensor is the best mate for this new technology, with its modest readout and burst speeds. I suspect once the same AI system is deployed on a body with a stacked sensor, faster readout and quicker bursts, like a future Alpha 1 or 9, it will really come into its own. But for now while I can say the R V autofocus is a big step-up from the R IV in terms of subject recognition, it’s not hugely different to many of its rivals.
Moving onto bursts, the top speed remains the same as before: 10fps with the mechanical shutter, slowing to 8fps if you want auto exposure and autofocus, or 7fps when you’re using the electronic shutter. But thanks to support for faster CF Express cards, the buffer can now be cleared much quicker than the R IV, allowing up to 583 compressed RAW, 547 lossless RAW or over 1000 JPEGs.
I confirmed this in my tests with a Compact Flash Type A card, where I could effectively shoot unlimited bursts at 10fps in any format with little to no delay for the buffer to clear afterwards.
But like other cameras that support dual card formats, shooting bursts with SD will incur penalties: even with one of Sony’s quickest TOUGH SD cards, I only managed a burst of 39 Lossless RAWs at the full resolution before the R V slowed down, and once I let go of the shutter, the camera took about 20 seconds to fully write its buffer to the card. Of course at a top speed of 10fps, that still gets you almost four seconds of action in RAW, and if you’re happy to shoot in JPEG, you’ll enjoy longer bursts with SD cards.
Here’s some bursts of bikes cycling towards me – or driving as many were powered – and as you can see, the R V is quite capable of tracking and grabbing some action shots.
These were taken with the mechanical shutter, and if 10fps is sufficient for you, the R V certainly gives you plenty of resolution for cropping-in if desired.
This sounds tempting for some wildlife photographers, but I personally found the Alpha 1 far more effective at capturing action thanks to its much faster sensor, which in turn was reflected across the overall handling. Of course it’s also a much more expensive camera, but if you were already investing in big pricey lenses, that’s the body I’d pair them with.
To illustrate one of the differences, here’s a pan I made with the R V using its mechanical shutter where, as you’d expect, the buildings are vertical.
But for comparison, here’s the R V using its electronic shutter where the modest readout speed has resulted in visible skewing.
To be fair, this plagues most cameras without stacked sensors to some degree, but definitely limits the usefulness of their electronic shutters. In Sony’s own range, this effect is greatly reduced on the stacked sensors of the Alpha 1 and 9 models.
Moving on, in a long-overdue upgrade, the Mark V becomes the first Alpha ILC to sport focus bracketing and a Bulb timer. I’ve been requesting both for years now, so I’m relieved they’ve finally made an appearance here.
Focus bracketing allows up to 299 images to be captured with finely shifted focus, although you will need separate software to stack them afterwards on your computer. I use Helicon Focus for stacking.
To show what you can do, here’s a single frame taken at f4 where the lens is delivering optimal sharpness, but obviously with too shallow a depth of field to get it all in focus.
Now here’s a version at the minimum lens aperture of f22, which has certainly increased the depth of field, but it’s still not sufficient to have the whole note in focus, and the small aperture has also resulted in diffraction, softening the detail.
And now here’s a version where I used the new focus bracketing feature to capture 100 frames at intervals which would cover the entire note during the sequence. I then stacked these in Helicon Focus software using the default settings.
Since each frame was taken at f4, the lens was working close to its sharpest, while the stack has taken care of the large depth of field. To be fair, Sony should have had focus bracketing a long time ago, but at least it’s now here and provides the R V with a new string in its bow for high-res macro work.
Moving onto the Bulb Timer, the R V now lets you preset an exposure between two seconds and 900 seconds, which allows you to easily make long exposures beyond the usual 30 second limit without the need for a cable release. Just put the camera into Bulb, choose the exposure time in the menu, set the self-timer to avoid touching the camera and you’re all set.
Here’s a shot I took using the Bulb Timer on the R V for a 60 second exposure. It works in the same way as the Bulb Timer on Canon cameras, and again it’s a feature I feel Sony should have had a while ago, but again at least it’s here now.Check prices on the Sony A7R V at B&H, Adorama, WEX UK or Calumet.de. Alternatively get yourself a copy of my In Camera book, an official Cameralabs T-shirt or mug, or treat me to a coffee! Thanks!