Testing Sony’s X95L: countering OLED with mini-LEDs and Bravia Core

Luca Fontana

26.9.2023

Translation: Patrik Stainbrook

Sony wants to shake up the LCD market with its mini-LED TV – just like Samsung did last year with Neo QLED. But it’s actually Sony’s Bravia Core, a new kind of streaming service, that really gets me excited.

Full disclosure: the TV, a 65-inch version of the X95L, was provided to me by Sony for testing.

Not everything that glitters is OLED. Not with Sony. True, Sony is definitely one of the better OLED TV builders in the industry. Regardless, the X95L isn’t an OLED TV, but a so-called LCD TV with mini-LEDs: thousands of closely spaced LEDs provide the background light for your TV. Unlike OLED pixels, LCD pixels don’t shine by themselves. That’s why the TV game in 2023 will be primarily split between these two technologies.

Right off the bat, the central question: which technology is better? The answer isn’t always obvious. OLED TVs are generally considered superior in contrast, colour fidelity and volume. But in bright living rooms, LCD TVs come up trumps as their tech makes them even more radiant. If you mainly watch TV in the evening, you’re more likely to go for an OLED. If you use it more during the day, when the living room’s bright, an LCD TV is more worthwhile.

To avoid comparing apples with oranges, I’ll primarily contrast Sony’s X95L with the Samsung Neo QLED QN95B in this review – that’s Samsung’s mini-LED TV. You can find the review of that model here. In short, I was thrilled. The bar that Sony’s X95L has to clear in this detailed review is pretty high as a result.

Let’s get stuck in.

Design: a convincingly clever stand system

There it proudly stands, on its two black metal duck feet. It isn’t really my style; I’ve never really liked Sony’s industrial look. But it’s practical. Firstly, the feet aren’t placed centrally, like on most competitors. Secondly, there’s a gap of 8.5 centimetres between the panel and the TV cabinet. This leaves enough room for most large soundbars without them protruding into the picture.

This gap would only be unsightly if you don’t have a soundbar. But Sony has thought of this too. The two duck feet can also be attached to the panel so that the gap is «closed». The feet are then barely visible from the front and back, while the panel sits directly on the TV unit.

Very smart indeed. Looks nice too. If I didn’t have a soundbar, I’d definitely go for this configuration. Here it is in the Zurich digitec store.

Otherwise, Sony stays true to what most other manufacturers imagine a TV to be. Modern. Slim, with narrow edges. Nothing extraordinary – and that’s a good thing. TVs should be TVs, in my opinion.

Viewed from the side, Sony’s X95L is quite thick at 6 centimetres. This is down to the additional LED layer that lights up the LCD pixels in the panel. There you go, mini-LEDs.

Now for the specs. Sony’s X95L offers the following:

And a word about weight. Without the foot stand, the TV weighs in at 32.2 kilos. So if you want to mount your TV to the wall, you’ll need a VESA 300 × 300 mm mount. You can find them in the shop here. With the two stands, the TV weighs 33.7 kilogrammes.

Measurements: good, but no top values for Sony’s X95L

The following is a deep dive into the subject matter. If you’re not into charts and diagrams, just skip it all and scroll straight to the section «The picture: mini-LED worthy material with the usual strong processor». From that point onwards, you can expect a lot of my subjective impressions and quite a bit of video material.

Cinema mode achieved the best scores. However, since almost all streaming services automatically switch to Dolby Vision for HDR content, the measurements below refer to the TV’s Dolby Vision Bright mode.

Maximum brightness

Right, let’s take a look at the brightness of the X95L. In the chart, I compare directly with Samsung’s QN95B, the mini-LED competitor, and LG’s G3, currently the best and, according to my measurements, the brightest OLED TV in the world. There are two axes: the vertical one stands for brightness, the horizontal one for the section in which the brightness is measured.

Huh? Sony’s X95L results in an outcome that I’ve never seen before in my measurements:

The X95L doesn’t score well in either brightness or black level. Spoiler: this is exactly what will affect later measurements.

Sony’s X95L is quite decent in terms of brightness in the 100 per cent window – that is, an area as large as the entire display. There, the X95L comes in with 648 nits. This is almost as bright as Samsung’s QN95B. LG’s G3, on the other hand, «only» has 250 nits, still a very good value for OLED TVs.

We’ll get to that.

The white balance

To measure the accuracy of the white balance, I need two tables:

1. Greyscale delta E (dE)
2. RGB balance

The greyscale dE shows how much the greyscale generated by the TV deviates from the reference value. The RGB balance indicates in which direction the greyscale levels produced by the TV deviate from the reference value. Why is this important? Let’s take a look at the concrete X95L example:

If you were to put the TV right next to a reference monitor, that would mean:

• Value is 5 or higher: most people will see the difference to the reference monitor.
• Value between 3 and 5: only experts and enthusiasts will be able to tell the difference.
• Value between 1 and 3: only experts will see the difference, enthusiasts won’t.
• Value below 1: even experts will not see any difference.

The colour gamut

Now for measuring the colour gamut; the coverage of the most common colour spaces. These are:

• Rec. 709: 16.7 million colours, standard colour space for SDR content like live TV and Blu-rays
• DCI-P3 uv: 1.07 billion colours, standard colour space for HDR content, from HDR10 to Dolby Vision
• Rec. 2020 / BT.2020 uv: 69 billion colours, still barely used in the movie and TV industries

The large blob of colour, including the darkened areas, shows the full range of colours detectable by the human eye. The lightened area on the left shows the BT.2020 colour space. On the right, you’re seeing the same, only the smaller DCI-P3 colour space. The white boxes show the actual boundaries of the respective colour spaces. The black circles, on the other hand, represent the limits actually identified during the measurement.

The measurement showed the following colour space coverage:

• Rec. 709: 100% (good = 100%)
• DCI-P3 uv: 89.35% (good = >90%)
• Rec. 2020 / BT.2020 uv: 63.48% (good = >90%)

Remember what I said above about the relationship between peak brightness and colour space coverage? Exactly this can be seen here: the X95L achieves only 89.35 per cent coverage in the important DCI-P3 colour space. Samsung’s QN95B achieved 92.49 per cent coverage in the same test – and thus more than the target 90 per cent that a good TV should have.

For comparison: Sony’s and Samsung’s QD OLED TVs as well as LG’s OLED TVs all come to around 99 per cent coverage in this test.

Now for the BT.2020 colour space. Sony’s X95L covers this with «only» 68.48 per cent. Admittedly, currently only Sony’s and Samsung’s QD OLED TVs achieve a value of slightly over 90 per cent. It’s exactly why the BT.2020 colour space is still hardly used in the industry (see info box above). But even Samsung’s mini-LED TV still manages 71.27 per cent coverage. I’d have thought Sony’s mini-LED TV could manage more here.

The colour error

Now for the colour error. What it does is describe how accurately colours are represented. As with greyscale above, the deviation from the TV to the reference value is referred to as dE. The white boxes indicate the reference colours sent to the TV by the test pattern generator. The black circles, on the other hand, represent the colours actually measured. Again, dE values below 5 are good for non-calibrated TVs.

Here, too, I’m somewhat disappointed – although things are generally much better. The measurements show an inherently good colour fidelity in Dolby Vision Bright mode. But one that «only» achieves an average dE of 5.79 with a total of 40 measured values. This is only marginally above the target value of 5. In the end, it’s still above. It’s why I said things were much better. Samsung’s QN95B scored a very solid 2.97 in this test.

Reflections

You can’t measure reflections on your screen as such. But some of you requested I take a look at it in my tests. Good idea. For testing, I recreated a standard situation in a living room: first, an evening photo. There’s an oven behind me and a standard lamp next to the TV. The light from the standard lamp is reflected in the glass door of the oven behind me and thrown back onto the TV.

Here’s the result:

When testing Samsung’s mini-LED TV, I hadn’t used the mirroring test yet. Therefore, you can see the comparison with LG’s G3 here, which is visibly better with reflections. Much better. Again, the X95L disappoints me.

But reflections are much more glaring during the day, without closed curtains, blinds or shutters, when light also hits the panel from the side. I mean, look at this:

Fortunately, Sony’s X95L shines bright, as is typical for LCDs. Annoying reflections are rarely noticed during the day. At least not in bright scenes. In dark scenes, however, the reflections were annoying from time to time.

Interim verdict after measuring

The measurements speak loud and clear: Sony’s X95L certainly isn’t a bad mini-LED TV. But it doesn’t achieve top values: the moderate peak brightness unfortunately has a visibly negative effect on colour space coverage. True, the colour fidelity is good, but not outstanding. And I always struggle with reflections, especially during the day, in dark scenes. Let’s see what the practical test has to say about it.

The picture: mini-LED worthy material with the usual strong processor

That’s the theory done with. But how are things in practice? Reminder: to avoid comparing apples with oranges, I’ll primarily contrast Sony’s X95L with the Samsung Neo QLED QN95B in this review – that’s Samsung’s mini-LED TV from last year.

Colour rendering

Let’s circle back a bit. I want to see what colours look like outside of computer-generated worlds. Take James Bond – Skyfall. James and young quartermaster Q are at an art museum, looking at a painting of a proud old battleship being unceremoniously towed to the scrapyard. The scene is obviously an allusion to the ageing secret agent.

Here, Sony’s picture already convinces me more. Pay special attention to the natural skin tones. A slight red tone creeps into the Samsung picture. I already criticised this in the Neo-QLED review. Otherwise, however, both TVs are operating at a very high level.

Black crush and shadow details

How does the latest Sony mini-LED perform in dark scenes? For this test, the first scene from Blade Runner 2049 comes into play. But first, Sony’s X95L seemed to shine particularly bright with an orange background. There’s no other way I can explain why my camera delivers such an overblown image. I always fix the aperture in place so that the test video doesn’t flicker. Therefore, ignore those few seconds on the Sony TV.

Brightness gradations

And here’s for a final image test: brightness gradations. I want to see how well Sony’s X95L renders bright areas of the image in particular. Again, I notice that Sony obviously puts more power into the LEDs in bright image areas than Samsung. Pay attention to the sun, which is barely visible on the Japanese TV and also «swallows» the clouds around it. It seems Samsung controls its LEDs in a more balanced way, especially in terms of brightness.

Processor: the usual strong level

The processor is the TV’s brain. Its main task is to receive, process and then display image signals. In this context, processing means recognising poor image quality and enhancing it. It does this by removing noise, enhancing colours, smoothening edges, making movements more fluid and adding any missing pixel information.

Motion processing and judder

For this test, I want to make the processor sweat. Concretely, by taking a look at judder. Something all TVs are affected by. Especially with long camera pans. Sam Mendes’ 1917 is full of such steady, slow-flowing camera movements, making it perfect for the judder test. In my comparison with models by other manufacturers, I pay particular attention to the vertical beams in the barn, checking whether they run smoothly through the image or judder.

Let’s skip to the next scene from 1917. Again, Mendes’ camerawork constitutes an immense challenge for most processors. Especially hard edges in front of a blurred background – take the helmets of the two soldiers below whenever they cross in front of branches and shrubbery. Both the processor and pixels have to react incredibly fast.

Sony’s processor doesn’t show any weakness in this discipline. Only the judder is – as always – a bit visible as long as you’re making a direct comparison.

Pixel response time

Next up, the Apple original For All Mankind. I want to see how long a single pixel takes to change colour. You can tell if the pixels aren’t changing colour fast enough when the image looks smeared – an effect known as «ghosting». When the camera pans over the surface of the moon, pay attention to the superimposed text.

Upscaling

Now for one of the most difficult tests: upscaling. I want to see how well the processor upscales lower quality material. For example, Blu-rays or good old live television shows. Or The Walking Dead. The series was deliberately shot on 16 mm film to preserve the old-fashioned grain that creates the feeling of a broken, post-apocalyptic world.

Gaming: input lag and game mode

When measuring the colour accuracy in Game Mode, I get a splendid average Delta E of 3.01 (read the section on colour error further up if you’re interested and want to learn more). This is no reference image level, to be sure. But it’s one of the best values I’ve ever measured for a TV in Game Mode.

• 4x HDMI 2.1 ports (4K120Hz)
• Auto Low-Latency Mode (ALLM)
• Variable frame rates (HDMI Forum VRR)

Sony – like Samsung, LG, Philips, TLC and Panasonic – has also entered into a partnership with many large gaming studios. As part of the HGIG, or HDR gaming interest group. According to the manufacturer, this should ensure that HDR is displayed as the game developers intended – like when playing Spider-Man: Miles Morales on the PlayStation 5.

What Sony conjures up here is a picture with absolutely perfect colours. In addition, I notice that black is really black, the edges look sharp and the image stays sharp even during fast and jerky camera pans. Look out for Miles’ dark silhouette against the light, the detailed textures of a snow-covered New York and the extreme detail in the clouds during fights. This is what a good game mode looks like.

Just a shame it doesn’t always feel as good. For example, when playing FIFA23, I failed many a dribble or perfectly timed shot because my inputs weren’t processed and displayed by the TV as quickly as I’m used to on other TVs. LG’s G3, for example, has an input lag of only 10.1 milliseconds.

Smart OS: Google TV

A little treat: Sony itself has added a kind of art mode. Instead of turning off the TV, you can display some nice pictures with the date and time. This is meant to liven up the rectangular black hole that a TV otherwise is when off, at low power and low brightness.

Bonus round: BRAVIA Core’s insane!

80 megabits per second!

This is also reflected in the fabulous picture quality, which can even easily compete with that of UHD Bluray, sporting a data rate between 108 and 128 Mbit/s. Also because Bravia Core adds that most movies are stored with IMAX Enhanced metadata. That’s important. It’s the only way the TV can get the most out of IMAX Enhanced picture mode.

As an example, I’ll show you a comparison between Blade Runner 2049 via Bravia Core and Blade Runner 2049 on my UHD Bluray.

IMAX Enhanced wins, in my opinion. Blacks are even richer than on UHD Blu-ray, and colours are a bit more vibrant without seeming artificially bright. I don’t spot annoying artefacts, picture noise in dark scenes or unsightly streaks on the UHD Blu-ray – as expected – or in the streamed version.

Purchasing via credits and the Bravia library looks like this:

What’s missing now for Bravia Core to become the ultimate streaming service is primarily content from other movie studios. The challenge, however, lies in the technology: for Pure Stream to work, content must be stored on special Sony servers. This is one big mess, both technically and for licensing reasons, that’s difficult to resolve with other studios, according to Sony executives. But they’re working on it.

Verdict: not the top dog of the mini-LED game – but close

Initially, I said I wanted to compare Sony’s top-of-the-line mini-LED primarily to Samsung’s mini-LED TV from last year. The result: Samsung wins. Sony’s X95L falls behind, especially in terms of brightness and colour fidelity. In the case of the former, by a surprisingly wide margin. Even LG’s G3 OLED TV manages a better peak brightness – tech-wise, a right hook from LG that must have landed hard with Sony.

Outside of the direct comparison, however, Sony shows why the Japanese company is still one of the best TV manufacturers. The picture sports great colour reproduction, always looks natural and is supported by the usual excellent if for my taste much too restrained Sony processor. Even in tandem with a Sony PlayStation 5, the X95L doesn’t show any weakness.

In addition, there’s the ace up the sleeve that other manufacturers don’t have: Bravia Core. This makes even movies that I previously found to be less exhilarating than UHD Blu-ray competition look like complete remasters. This is down to Pure Stream and the IMAX Enhanced metadata. A gamechanger in the streaming world. If only Sony could get other film studios involved as well…

Luca Fontana

Senior Editor

Luca.Fontana@digitecgalaxus.ch

I'm an outdoorsy guy and enjoy sports that push me to the limit – now that’s what I call comfort zone! But I'm also about curling up in an armchair with books about ugly intrigue and sinister kingkillers. Being an avid cinema-goer, I’ve been known to rave about film scores for hours on end. I’ve always wanted to say: «I am Groot.» more

---

Product test

Our experts test products and their applications. Independently and neutrally.

Show all

These articles might also interest you