INebulaUK
I certainly don't disagree with you. I just think that same problem exists even when calibrating with LD enabled, particularly on TCL where the dimming algorithm can be pretty aggressive and unpredictable
I'm sure you've seen it yourself. Run an SDR grayscale sweep with LD enabled using a 10% window on black, then a raised background, then a larger window or full-field. You'll see Y, xy, and often the EOTF shift with each geometry. Constant-APL patterns can certainly help, but you're still calibrating against a particular set of conditions. Change the APL and the response changes again. Now disable LD and repeat those measurements. At least on my QM7K, the response is considerably more consistent. That's the argument for calibrating the underlying panel response first and letting the dimming algorithm do its thing afterward.
Interestingly, I take the opposite approach with HDR/DV because LD is so integral to the luminance response and tone mapping. But I think that's part of both the challenge and the fun of calibration. There isn't necessarily one correct approach. I don't have a professional reference monitor at home, so ultimately I'm judging how well the results translate to real content and how closely they align with my other, more stable calibrated displays. It's a compromise either way, and I think there's a reasonable argument for both methods. |