Monitors are weird…..
Since building the Open Source Response Time Tool, I’ve tested a whole load of displays and… man.. Some of them do some really weird stuff… So I thought why not make a video talking about all the weird stuff I’ve seen in testing! A lot of this is also stuff I don’t fully understand, so if you do I’d love to learn more from you too. Without further ado… let’s talk OLEDs!
OLEDs are a weird one to test for response times. Being an organic light emitting display, the LEDs themselves can change brightness levels (and therefore colours) in nanoseconds if driven hard enough. Realistically it’s more like microseconds, but still that’s a lot, lot faster than a traditional LCD display that can take milliseconds – literally an order of magnitude longer. In fact, here is an LCD measurement and an OLED measurement overlaid. See the difference? OLEDs are FAST.
But let’s look at that OLED data a little more closely… See those dips every 6 milliseconds or so? That’s at the start of every new frame, and yeah, it dips to black. Every frame, quick dip to black, then instantly back on to display the new frame. I’ve captured that with highspeed cameras before too, and it’s something every OLED I’ve tested does and I’m honestly pretty confused. See most liquid crystal displays just change what voltage they are set to over time, so they meander from one state to the next, but OLEDs are instant, so the controller just says nope and switches off, then comes back to life with the new voltage to be set to. This is one of those things I’d love a more detailed explanation of, so if you know please do leave that in a comment below.
One of the other quirks of OLEDs is that thanks to their insistence on holding a given voltage until refreshed, sometimes the controller decides it fancies missing the target and overshooting. But then, unlike a traditional LCD panel which will just slowly work its way back down, on an OLED it holds that position for a full frame before correcting itself. This is another one of those really weird quirks that thanks to so many people having their hands on these OSRTT units, I’ve seen countless tests of various WRGB AND QD-OLEDs that still exhibit this same behaviour. Again, if you have any more information please do leave a comment.
Moving on from OLEDs – but not that far – we have MiniLEDs. Specifically, standard LCD monitors that use a MiniLED backlight. The two main reasons for that choice are maximum available brightness – MiniLEDs can get crazy bright – and for the fact you can switch some of the MiniLEDs off to give you local dimming. Basically because liquid crystal panels don’t block all the light the backlight tries to shine through them, what should look like black often looks like a dark grey, leading to a less than ideal viewing experience. But, if you can turn off the backlight too… well then it’s pure black, and that’s much better. The quirk here is that every MiniLED panel I’ve tested, be it on a £1000 gaming monitor or a £4000 gaming laptop, they all flicker like crazy.
That’s thanks to a thing called PWM or pulse width modulation. See, there are two ways of controlling the brightness of an LED, you can either just lower the voltage you run it at, or you can basically switch it on, then off, really fast so that the average light output works out to be what you want. PWM backlights generally suck, they cause eye strain, headaches and generally aren’t all that great for you – but the key thing is the frequency. AOC’s AG274QXM cycles at 500Hz, meaning it switches the backlight on and off 500 times per second. The Asus Zephyrus Duo 16 cycles 50,000 times per second, which is much, much better.
If I’m being honest, I’m not entirely sure why all MiniLED monitors with local dimming support need to use PWM for their backlight control – I assume it has something to do with the local dimming, and probably the cost, but I’m sure it’s possible to do local dimming, with MiniLEDs, using DC control instead. Might be costly though.
Jumping back to more traditional monitors, did you know that a whole load of them actually overshoot past RGB 255 – aka full white? You’d think that setting the screen to as bright as it can go would be the limit, but nope! My best guess for this is that the backlight is set in steps, say of 10, then they just shift how open the pixels are to fine-grain adjust the brightness, but that means when you set the overdrive too high and it swings past what it thinks is full brightness it ends up outputting more light than you’d otherwise expect from “full white”. That also helps compress how far the crystals have to move to achieve any given transition, so again that helps speed things up too.
One panel tech that just takes its sweet, sweet time though is VA or vertical alignment panels. Especially in the darker shades, these can take multiple times longer to transition than with brighter shades. This distinct lack of speed is what often creates a slow, smeared look to VA panel monitors. They are just really, really slow, and it shows.
Finally, I want to give you a little teaser of something really exciting coming soon. I talked a little about this in the last OSRTT video, but this little guy is OSRTT Pro. It’s a huge step forward in functionality, especially because it can now test at pretty much any brightness level you’d want. Anything from like 60 nits to 1,500 should work fine for this, and it should be even more accurate too. I’m really excited to get this out, so keep an eye on the channel for a full video on this soon!
