Hi Lynn,
I think there are some controversies below that I would like to
address:
On May 24, 11:27 am, l_d_allan <
lynn.d.al...@gmail.com> wrote:
> My understanding is that more and more Photoshop book authors are
> recommending to apply some sharpening early in the workflow. This is
> based on writings by Bruce Fraser (now deceased) and others. He called
> this "creative sharpening", which tended to be a relatively low to
> moderate amount of sharpening. Subsequent sharpening would be done at
> or near the very end, based on output considerations.
> It seems like this might be an emerging concensus ...
Most photoshop book authors don't do panoramas, and don't take the
very special environment of stitching not into consideration. while
their suggestions might work fine in a classical environment (I love
sharpening in ACR for my normal photos), they are counter-productive
for panos:
I for example do not sharpen my RAW->tiff files because they go into a
exposure blending process, where initial sharpening would create nasty
color fringes. in addition, the individual tiles get a lot of warping
with various kinds of interpolations, and since sharpening is a lossy
process (information gets lost in this process), giving the
interpolators the maximum amount of information rather than less info
is beneficial (see all the lengthy discussions about jpg vs. tiff as
input files). but if you like it, go for it, but please don't call it
"emerging consensus".
> ********** a semi-related issue *********
> is that the noise reduction (NR) of LR and ACR is much less effective
> than plug-ins like Noise Ninja, NeatImage, etc. imo. Every book I've
> read about Photoshop has described Photoshop/LR noise reduction as
> less than effective. LR3 and ACR6 are supposed to be better, but that
> wasn't my experience, based on the LR betas. My experience is that the
> NeatImage NR plug-in is much better than ACR/LR's NR. My evaluation of
> commercial NR plug-in trials was that they were all much better than
> ACR/LR.
this is unfortunately very wrong, for various reasons:
- the NR in ACR6 is spectacular (yes, i am an adobe fanboy), and beats
about any other NR (IMO). it is so good in fact, that I can afford to
shoot panos at ISO5000 now. but whatever you use, you want to remove
the noise before warping.
- noise is a regional statistical phenomenon, in other words, the
noise is more or less equally spaced. every NR application is relying
on this fact. if you apply your NR AFTER warping, the noise is not
equally spaced anymore, and NR apps can't remove the noise anymore,
because they can't detect it properly (especially in areas of really
large distortions near the Z/N). in addition, noise can appear in
patterns (i.e. vertical streaks), which after warping can be very
visible, and almost impossible to blend nicely (at least in cases with
heavy noise).
> Ideally, it would be better for ACR/LR to allow integration of NR plug-
> ins, so higher quality NR was done earlier in the processing pipeline.
there is no need for it at all. first, CS4 is outdated, and adobe will
not work on the code ever again (besides minor security fixes).
second, adobe has already really good NR in ACR6, and they are
tremendously proud of it (rightfully so), that they don't see a reason
why to offer third-party NR in ACR... if you still don't like the
results of ACR6, you can use plugins in the app environment (LR/PS),
or use a different RAW developer.
> This is (yet another) aside, but NR plug-ins can be harder to use with
> PTGui because they like to have EXIF info, especially ISO, which PTGui
> doesn't seem to keep in the generated pano.
I don't see this a mayor flaw of ptgui not to include the ISO... if
you really need it, use exiftool to add that piece of info to your
file.
joergen