That “+2 frames delay via emulation” was what I picked up from reading a few scattered Google results. But in my own just-now preliminary rough test, in JP SoulCalibur I got 5 frames of input delay on DC, and 3 frames of input delay in Flycast on my laptop.* Used HRAPV 2017 stick for both.
Well heck! That was with vsync off of course, so there’s some tearing, but jeez I’ll take that, that’s nuts. Stays the same in Flycast even if I turn on widescreen and crank up the internal resolution to the 9x max it allows me (7680x4320). This is mind-blowing. If I can get the games I could conceivably want to play dumped (SC, JoJo’s, CvS1&2, MvC2, GGX), Flycast will be my Dreamcast and my actually DC’ll be a laptop stand** and standby GD-ROM dumper.
*I might be adding a frame or so in my count; what really matters to me is the comparative score. I’m starting counting delay after the first full frame AFTER the button on the stick is fully depressed, 'cause not having a wired-up LED, that’s kind of the earliest sure sign I can read in the ~60 fps cell phone video I’m counting from: so button down, advance a frame, no more downward movement, that’s frame “zero,” advance a frame, and then this is the frame I count as 1 frame of delay, and so on.
**(The DC has been just about perfect as a laptop stand; even the slight convexity of the lid probably aids laptop cooling ^ _^ – I don’t use the actual laptop keyboard so the rounded wobble doesn’t matter.)
Oh shoot if it’s really able to drop the input delay below DC levels I’ll have to try DC Double Impact and 3rd Strike too; those have cool additional options on DC hardware but at least several frames more input lag than other platforms.