so-called "convolutional method" or whatever it's called, activated, which according to the game's description, is the most modern DLSS technology that avoids loss of quality in movement.
The loss of quality in movement you're referring to is called temporal blur - it's a side effect of temporal rendering techniques which includes the convolutional method of DLSS upscaling. A 1080p temporal image might look 1080p when there's no movement, but it'll start looking 720p when there's movement. This doesn't just happen with DLSS, it happens to TAA, happens to FSR, happens to XESS, regardless of upscaling or native.
But the most recent transformer-based versions of FSR and DLSS actually minimize the temporal blur. It's not completely eliminated but near enough. See pic related, a wall in the distance in Space Marine 2 cropped and zoomed for extra emphasis. Sharpening is at 0%. The 2160p native TAA looks like 2160p TAA when you're frozen, but starts looking like it's not even 1440p when you move diagonally. Wheras 2160p DLSS Performance upscaling (so 1080p input resolution) looks like 2100p when you're frozen and approximately 2050p when you move, which is way better. FSR4 behaves similarly.
If you truly hate how movement destroys detail in newer games, then you should be using DLSS4/FSR4 whenever you can. Because those have better motion clarity than standard rendering on top of giving you better performance. Technically the best motion clarity would be achieved by not using any temporal rendering at all, but at this point that causes other headaches more disruptive than any temporal blur.