HV20 - used to get Still images
 

Hi,
I am receiving my HV20 tomorrow and I wish to enquire about which is the best way to get still images and which will result in better quality. As I understand it I have 3 options...

1) Use on Still mode. Picture will be 3.1 megapixels.
2) Take pictures while recording. Pictures will be 2.0 megapixels
3) Take frame grabs from video afterwards. Pictures will be 2.0 megapixels.

Please feel free to correct me if I wrong.
Would I see a noticeable difference in quality between the 2.0 and 3.1 pictures?
Also, is there a stills camera that the HV20 is comparable to in terms of image quality?
Many thanks.

I used my HV20 to take stills.

Used it as still camera and taking pics during video recording. (have not tried post grabs)

Both pics came out looking great.

One key difference that I noticed...

The "still camera mode" the pictures were in 4x3 aspect ratio (or close to it)

During the "video mode" the pics came out in 16x9 aspect ratio (or close to it) (Note: I was shooting video in HDV 24P widescreen)

Both looked great though.

THe HV10/20 take surprisingly good stills. I have both still and video-grab set to 2MP (16:9) although you can up it a bit to 3MP for stills, but I don't like the aspect ratio (I view photos on my widescreen monitor/TV, less 4x6 prints.)

I haven't noticed too much quality difference for still-mode vs video grab, except for low-light ones; video-grabs there come out very grainy, almost like it had to split the signal w/ tape (tape video was unaffected), maybe something w/ digital gain occuring after the still part of the pipeline?
some thoughts:
- still mode gives better manual control like regular camera eg flash/focus/drive modes etc, along with the same video controls like A/V priority, white balance, exposure lock etc.; so you may prefer still modes for those. Of course, there is also NO flash in video grab :)

- video grab mode does VERY well at taking pretty good shots while you're taking regular video (although you might move the camera a little, or linger/zoom on a shot differently than a regular video scene);

one great advantage about that, esp for fast-moving/continuous events, is that video is ALWAYS constantly exposed/focussed properly, so there is NO focus/exposure/shutter delay like w/ still mode.
eg if you're videoing some sports or other action, and you want to grab a still of a perfect shot or jump or something, just press the still button and you got it (or can get it per-frame on playback later.)
With still modes, you'd probably lose the shot due to delay, and you won't have the video either.

- the LCD viewfinder behaves differently in video vs still; dark scenes are not very visible in still mode (eg flash or any gain is after the shot, not during framing.) Video mode is WYSIWYG basically (w/ the grain problem for low light as noted earlier.)

I mix and match both, mostly taking video grabs unless I have time to frame & compose a still shot (and switch to camera mode.)

Note also that my HV10 w/ 1GB sandisk card, takes some time (10-30sec?) to "set up" working on the SD card... so video "boots up" faster to take videos than still shots do.
If you take a video grab in those initial 30secs or so, video will continue, but further stills (and even cam controls like zoom/exposure?) appear "hung" until the card finishes booting/writing.

Pictures taken at 3.1mp recorded to a MiniSD card should look better. But considering the sensor is still only 1920x1080, to end up with a 2048x1440 sized image, it's obviously being enlarged somehow. I'm still yet to get myself a MiniSD card, so all photos from this camera to date have been frame grabs. They don't look too bad though.

Also remember that with Bayer style sensors such as those used in nearly cameras the RGB are interpolated, so you're only really getting one colour per pixel. Either red green or blue. Only a few still cameras with Foveon sensors such as Sigma's SD14 - http://www.sigma-sd14.com, are able to capture images with all 3 colours per pixel. This has the advantage of enabling a much smaller file size.

I have created 4x6 pictures from stills in card mode and camera mode (frame grabs). Both look great.

As far as the card mode goes however, it is 3.1 megapixels using 4x3 and I think 2.9 megapixels using 16x9. Feel free to correct me if I am wrong here.

I use the stills to post on the web, and hence I do not need a lot of megapixels. However, if you are looking to print the stills, card mode would be a better option.

It has been my observation that taking stills in video mode while recording have a larger file size than taking the images from frame grabs. Mind you this is just a general observation. I have not gone back to the specific frame to test.

Note that the human eye also has many more rod cells, which are not sensitive to color, than cone cells which are. Color perception is reconstructed in the brain. The Bayer matrix color works pretty much in the same manner.