OK. I was there for five hours!

(Well, I had to avoid the M25 rush hour traffic...).
It is
stunning - everything I have been waiting a
long time for.
I endorse all of Chris' (UrbanT's) comments. Here are some of my own:
1) The Gennum VXP processing is outstanding. Deinterlacing of SD (PAL) and HD video and film was superb. Scaling was equally impressive - on good quality DVD transfers (e.g. Star Wars Ep. IV, Phantom of the Opera, Robbie Williams Live at the Albert Hall), I detected no ringing.
2) The JVC was
considerably sharper than the Pearl. And Elliot considered the Pearl on demo to be amongst the best he has seen. Pretty certain this is primarily down to the lens. When viewing grid patterns, it was clear that the degree of Pearl misconvergence varied across the image. The HD1 misconvergence (1/4 pixel after a single red pixel correction horizonally) appeared to be linear across the panel. i.e. panel misconvergence rather than a lens effect.
3) The differences between Pearl and HD1 were much more apparent on HD-DVD than upscaled DVD. However, the JVC image simply had more depth to it for both types of material.
4) The simulaneous contrast was breathtaking. The droid scene inside the Jawa transported was a revelation in the shadow detail.
5) The colours were glorious and flesh tones were natural.
6) The JVC was simply more colourful and punchy than the Pearl.
7) When fed an 100IRE full screen signal from a Lumagen ProHDP, colour uniformity issues were obvious on the Pearl. The HD1 screen simply looked white!
On the technical side, I spent 20 mins with the HD1's designer, Alex Kobayashi, and can confirm the following:
1) 480i60 and 576i50 are accepted over HDMI.
2) 1080p24 is accepted, frame doubled to 1080p48 and displayed using 96Hz processing at the D-ILA panel.
3) 1080i50 and 1080p50 are accepted and displayed using 100Hz processing at the panel.
4) 1080i60 and 1080p60 are accepted and displayed using 120Hz processing at the panel.
5) 1080p48 is
not in the EDID table.
6) JVC are looking to see if the vertical stretch feature is possible.
7) Firmware upgrades are technically possible, but I got the feeling that they would require JVC to do them.
8) The blanked off panel is a factory connection for firmware upgrades (in conjunction with the RS232C).
9) Currently, 60i --> 24p processing for the removal of 3:2 cadence judder has not been implemented. I got the impression we wouldn't get this for the HD1. More of an HD2 feature.
10) User editing of the EDID table is not possible (for 576i out of SkyHD boxes).
11) Settings such as DNR (noise reduction) and Sharpness are global memories per physical input. That means that an HDMI switching solution prior to the HD1 precludes easy per-source calibration.
I put in my order for this on Monday morning. Tonight's demo completely justified my faith in all of the pre-production reports we've been getting. The single thing that prevents 10/10 from my perspective is that 480i60 and 1080i60 film material are displayed with 3:2 cadence judder. That's it and of course this is correctable with an external video processor.
I warmly congratulated Alex on his achievement. He and his team have achieved the "impossible" - a projector with virtually no flaws. This one will sell a lot.
Joel