Blur Busters Forums

_zebracrossing666 wrote: ↑

30 Jan 2026, 08:33

The gamma issue I can agree with, QD-OLED has no issues.

As for the other issues, I don't agree.

At this point I think that the stable gamma during VRR is way more important than anything else. Worse HDR is much less of an issue compared to VRR being completely and utterly unusable.

No idea what this is about, but we're talking about things that can lower motion blur on displays. You can do that at 60Hz even, no need for 20kHz.

So I'm not duplicating but extending the display because as you said the duplication itself can introduce the time difference.

You can test duplication by doing one test and then another with the displays swapped to see if duplication has an effect.


So the question would be what is the best ...

Why use photos? Record a video instead and they go through the video one frame at a time.

Also, why are you starting the stopwatch twice? I thought you are duplicating the display, so the second display shows whatever the first one is showing.

IIRC, this is not a good method, because the duplication itself can introduce the time difference. However, you should be able to test this by switching the displays and make the other one the primary display. If you now observe the same result (50ms higher on the clone), then it's the cloning that ...

What I find interesting on a more fundamental level: the current OLED trend of simply pushing for ever higher refresh rates feels like a bit of a dead end to me.


It's really not. In theory, once you reach 1000Hz, VRR is not needed anymore since the average judder of FPS/Hz mismatch drops to 0 ...

After beating my head against this panel for a whole year I've replaced it with an XG27ACDMS. Almost every problem I had is gone.

Stable gamma at different refresh rates.
Excellent black levels, very little crush.
Vastly, vastly better grey uniformity. No more of that dirty crosshatched pattern ...

The only effect this has is a change in memory throughput and/or access time, which in turn affects performance. It can also affect system stability and data integrity in case of memory errors, in which case "desync" would be the least of your problems.

How it could possibly affect timer accuracy ...

I believe if DSC is used, and the GPU needs more than one internal head to drive the display, EDID overrides (like with CRU) don't work. DSR doesn't work either for the same reason.

Note: For the sake of my arguement, assume that each transition of the LCD is <4,16777ms (e.g. 3ms at all times, regardless of transition chosen or the precision range of the heatmap)


This means that for 3ms, you're seeing a ghost of the previous frame in the new frame. The sample-and-hold ...