I'd say you're better off with the XRGB-3. You can use the Mini along with a HDFury to do this, but the Mini's 480p output (which you'd have to use to feed an Emotia is less sharp than the XRGB-3's B1 output. I also don't think that scaling a PSP signal to 240p a good idea to begin with, since you're forcing the 272 active lines into a 240p signal. With or without scanlines, that's clearly degrading the signal.
I also don't trust the Mini's lag display on the info screen. I believe the Mini's lag is closer to the XRGB-3's B0 mode than to it's B1 mode. I've done some tests on this and I will do more and I will post the results once I'm done.
I leave the Super Emotia GX on the interlaced option for PSP games. I was under the assumption that this was the equivelant of 480i, and that it would be the same as the native interlaced S-Video output from a PSP-3000 unit. If I am off base, any extra insight would be appreciated.
The last time I tested the XRGB-3's B0 mode for lag, I had 26-30ms lag for DC 480p 60hz content, 34ms lag for 240p content, and 60ms lag for 480i content. This was done by splitting the VGA out to a CRT monitor and the XRGB-3/Emotia/SDTV rig. The XRGB-3 was the only device adding any lag in the chain. I took the average of 20 samples for each result set.
Has anyone gone to these lengths to identify the true amount of lag on the XRGB-Mini?
What are people using to determine the Mini's actual lag? Are people relying on an onscreen built in lag identifier on the mini, or just guessing based on advertised specs? I have been cautious ever since people had stated that the Gefen VGA-to-DVI Scaler Plus was lag free. Though when I purchased and tested it, it has 16ms of lag.
I should do a retest with the XRGB-3 actually in Zoom mode for the PSP. Depending on what the XRGB-3 hardware is doing, it may yield slight differences in lag. The Dreamcast actually had a 30ms average on the 1024 option, versus 26ms average on the 1280 option. I couldn't test the 1600 option since I didn't have an LCD at the time of testing. Using 1600 would flag an "out of range" error on the CRT. I am not sure whether the 4ms discrepancy has to do with deviations in the scaling hardware/algorithms of the XRGB-3, or that my sample size of 20 was too small, and that I would have to do 50-100 samples under each resolution to get a more accurate result.