"Given this, I think low-light might still be a long way off XLH1/XHA1 low-light performance IMO. (happy to be proved wrong!)."

Here are the specs of the Canon XLH1: Minimum Illumination: 60i, 1/60 shutter speed = 7 lux; 30F, 1/30 shutter speed = 4 lux; 24F, 1/48 shutter speed = 6 lux

considering the Hv20 will be 6 lux effectively at 60i, they should be comparable.

Yup, it does. I said in another post that it's that plus the noise reduction processing that's supposedly 'on-chip'. Those two together could help the low-light.

I think from the get-go Canon handicapped the Hv10's low light on purpose. Considering they are using the same CMOS sensor, that already had the progressive capabilities in the HV10. Progressive already requires more light than interlaced (about a stop) and with more noise +6db. So it seems to me to get better light out of the HV20, buy improving the lens slightly, and cleaning up the noise through the DSP seems rediculously hard to do. They would first have to fix the light loss from progressive and clean up the noise, then improve up the HV10, seems impossible to me.

Initially, they must have dumbed down the HV10, to not steal from the A1 too much. But as the HV10's sold faster than they could make, they decided on utilizing all of the performance available for the HV20.

Just a guess.

The focal length and maximum aperture is exactly the same as the HV10. Chances are it is the exact same unit from the HV10. So there is no more light coming to the HV20 than there is to the HV10. The filter ring is larger but that's just what the outside housing of the camcorder happens to be. It doesn't necessarily mean the lens is larger.

How Canon achieved the better low-light performance is a mystery right now. According to Camcorder Info:

If I had to speculate, I'd say they are enabling some noise reduction circuitry from the A1 cameras that they disabled in the HV10.

I dont think that this is true. I have noticed in low light on the HV10 it seems to use the NR1 and NR2 settings (like the A has) I see much ghosting and trails when the HV10 runs out of light and I also noticed that the HV10 uses gain up to +24db in darkness.

I think they dumbed it down, intentionally.

from what we know about the HV20 so far. and Canon's foray of CMOS implementation on consumer grade models, i don't think my optimism is too enthusiastic when looking at the XHA1 footage thinking that could be (or pretty darn close) to what im getting at a sub $1k camcorder!

check out the following link posted by Elton at DVXuser...

)>> One shot with a few simple cc's in FCP just to show different looks. All handheld in 24F mode with OIS engaged.

150 MB h.264 (2 min.)
http://www.realm.cc/upload/Elton/FacesInTheCrowd.mov <<(

its a big file but worth it to at least have a sneak peak of what the HV20 could very well be. (or pretty darn close!)

full link here... http://www.dvxuser.com/V6/showthread.php?t=77926

Excuse me, but you don't need to leave DV Info Net to see Barlow's clip.

It's hosted here as well: http://media.dvinfo.net/canonxh/FacesInTheCrowd.mov

For more details, see http://www.dvinfo.net/conf/showthread.php?t=79651