Chuck Norcutt wrote:
> I recently attended my granddaughters' dance recital.... We were also warned
> on entering that no video cameras would be allowed. But no one made any
> comment about my 5D.
>
My record of my granddaughters' latest dance/drama exploits are on
video. Although I like video better overall, I would also like some good
stills like the excellent ones you have posted. That's one reason I've
been following the cross-over products with some interest.
The Canyon S3 IS, for example, is quite an accomplished video camera
(and probably would have passed muster like the 5D at your event). But
the EVF is too slow and grainy for me and the LCD too small and low res.
So I've been waiting on reviews of the S5 and the TX-1. Well, the TX-1
seems to be a failure, but I'm still hoping for the S5. The idea of
being able to video a performance and occasionally take a high res
still, even though it loses a few frames, is quite attractive.
My Sony Videocam will do that, recording the still to a memory stick,
but it's at video resolution, 640x480, not really what I want. I will
say that 20x optical zoom, excellent IS and quite good low light
performance are pluses to the true videocam. It's really nice to be able
to do slow zoom and pan to catch my special subjects among all the
others and add some variety to the video.
> ......... Although the stage was
> reasonably lit for human vision it still took ISO 3200 and f/3.5-4.5
> (mostly) to get a shutter speed of 1/160 to try to stop some motion.
> This was the first time I've ever taken more than a couple test shots at
> 3200.
>
> Some surprises. Although there is modest noise visible on the CRT there
> is essentially zero noise visible on 4x6 prints... either those from
> nearly full size images to those which are cropped to about 25%.
Yeah, pixel peeping and modest size prints are quite different animals.
Unlike the high isos on some cameras, that on the 5D, and I assume, on
the 1D(s) series, is really usable for serious work.
> ............ I've yet to see what the lab prints will
> look like. Since I'm new to CS3 which I used for the conversion it
> occurred to me afterwards that CS3 might be doing some noise reduction.
> Sure enough, I checked and saw that everything has been converted with
> a default value of 0 for luminance noise reduction and 25 for color.
That's been the default in ACR for a long time.
>
> However, moving the sliders around I couldn't see any actual change in
> noise level. I also sharpened the images in PhotoShop so whatever noise
> is there has also been a bit accentuated from the sharpening.
>
I don't have CS3 yet. It will probably wait for a great deal or a new
camera that needs a newer version of ACR. To date, I don't find the
noise reduction in ACR or PS to come close to what independent products
like NeatImage and Noise Ninja can do.
My famous(infamous?), oft posted jellyfish full pixel shot benefits from
judicious application of NeatImage.
The other thing to keep in mind is that both above products come as
plug-ins for PS, et.al. So you can apply them to mask layers and be
selective about where and how much noise reduction is applied.
Elimination of noise in blank areas, where it is most obvious, but not
in areas of fine detail, where it isn't obvious, but reduction would
soften the image, can give something like the best of both worlds.
> Although I shot in raw mode I set the camera to auto white balance so
> that an as-shot conversion would give me a starting point. The next
> surprise was that I made almost no changes in color temperature after
> processing more than 100 images. I don't know how the camera manages to
> do *white balance* amidst a profusion of colored lights but it was rare
> that I could manually come up with a rendition that I liked better.
> This was especially surprising since I have seen it make some not so
> flattering choices in light which I'd consider much less challenging.
>
A couple more thoughts here (surprise, surprise!). Although good stage
lighting uses colors to create mood and drama, it has to blend on the
people in such a way as to make their skin look reasonably natural. So,
for example, RGB spots may be offset on a singer to add color, but they
overlap on the face to give white. So it doesn't surprise me that you
got decent AWB; the overall balance of the spots was likely pretty neutral.
I also think that Canyon is probably technically able to create "better"
AWB in incandescent light than their cameras do. I know Phil and Simon
will keep bashing them, but I think it's intentional. I'll bet that
their research shows that when people take pics in warm light, they
expect a warm image. You know all these manufacturers are showing lots
of pics to focus groups, etc. and finding out what they like, even
giving them cameras and seeing how they like what they get, whether or
not we would like it.
Ever tried taking a candlelit portrait of a couple and balancing it so
it looks like daylight? How did they like it? :-) The pros know how
to control WB, but the punters who actually buy most of the cameras
can't/won't, but when they make a pictbridge print, print over the net
or print at a kiosk, they get the result they expect and want.
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|