http://www.wilderness-wales.co.uk/pg/pano/pano.html
Any comments, criticisms, suggestions etc. welcome.
Paul
--
Wherever I pitch my tent, that's my home.
http://www.wilderness-wales.co.uk/pg/pano/pano.html
Some nice fun photos but I find the "curved earth" images rather alarming
;-) If it bends that much I think we'll all fall off if we get anywhere near
Swansea!!
--
Carol Haynes
www.playoutdoors.org.uk
www.swaledaleoutdoorclub.org.uk
Group site for uk.rec.walking: http://groups.yahoo.com/group/ukrw/
To reply by email remove obvious spam trap from my address.
"Paul Saunders" <pv...@wildwales.fsnet.co.uk> wrote in message
news:amqugf$gvs$1...@news7.svr.pol.co.uk...
>
>Any comments, criticisms, suggestions etc. welcome.
>
They're all crap
I like the swing-lens panos. But don't tilt the camera <g>.
Bernard Hill
Selkirk, Scotland
> Someone has been spending a bit!! How much does something like that cost?
Relatively cheap actually. It costs £400 new from the Widescreen Centre but
I got mine for £260 second hand from a reputable source. Only a year old
and in very good condition.
Also the Epson 2450 scanner I got for £263, a bargain at that price. Total
cost for the two was less than you'd pay for a decent digital camera these
days, and the quality way better.
> Some nice fun photos but I find the "curved earth" images rather alarming
> ;-)
Yes, not the most subtle effect is it?
> If it bends that much I think we'll all fall off if we get anywhere near
> Swansea!!
Yeah, I think Worms Head is pretty much the end of the earth around these
parts...
> I like the swing-lens panos. But don't tilt the camera <g>.
My thoughts exactly. Those were test shots of course, so I needed to try
out the effect. Some people really like the curved distortion effects, but
they tend to be photographers rather than walkers. I think most
non-photographers prefer the normal looking shots, as do I. I'll try to
avoid tilting it whenever possible, but I may take the odd one now and
again.
I prefer the "distortion" effect because that's exactly how the real
world looks. You just have to look at each part of the picture
individually and not as a whole... just like you do with the world, in
fact.
To me, rectilinear lenses are the distorting ones. Have a look at
www.braeburn.co.uk/pix/trees.jpg (35K).
That's an 18mm lens so has a 90 degree aov and I really object to the
fact that the trees are growing out of the ground at an angle. Much
better to have curved trees.
Bernard Hill
Selkirk, Scotland
>To me, rectilinear lenses are the distorting ones. Have a look at
>
>www.braeburn.co.uk/pix/trees.jpg (35K).
>
>That's an 18mm lens so has a 90 degree aov and I really object to the
>fact that the trees are growing out of the ground at an angle. Much
>better to have curved trees.
That is perspective distortion caused by camera angle. Tilting the
camera up has caused converging vericals. If the camera had been level
the trees would have been straight, but you wouldn't get them all in
of course. Your eye and brain does a great trick when you see the
scene. It does the swinging lens thing *and* produces a straight
trunk. With a camera you have to make a compromise
--
Phil Cook looking north over the park to the "Westminster Gasworks"
It's caused by angle and camera distortion. The lenses are manufactured
as what are called "rectilinear" and introduce this distortion
intentionally to keep the straight line of the tree straight.
However the eye does *not* see straight lines as straight. I look at the
base of the tree and see parallel trunks. I look at the top and they are
closer together, showing that in real life the trees are curved to the
eye.
I prefer the true (spherical) perspective.
>Your eye and brain does a great trick when you see the
>scene. It does the swinging lens thing *and* produces a straight
>trunk. With a camera you have to make a compromise
That last is true. But for me the lens manufacturers have the wrong
compromise. I prefer the compromise of the swinging lens (cylindrical
projection) not the rectilinear.
Bernard Hill
Braeburn Software
Author of Music Publisher system
Music Software written by musicians for musicians
http://www.braeburn.co.uk
Selkirk, Scotland
01750 21854
+44 1750 21854
What does post processing in Photoshop entail when editing your images?
Regards
ChronicFatigue
> What does post processing in Photoshop entail when editing your images?
Er...taking the thread to photoshop newsgroup ?? ;-)
Chris
alt.graphics.photoshop
jt
Jhimmy
"Chris Gilbert" <Chris....@Consignia.com> wrote in message
news:3d93226e$1...@RGINF-S02.research-group.co.uk...
> >Some people really like the curved distortion effects, but
> >they tend to be photographers rather than walkers. I think most
> >non-photographers prefer the normal looking shots, as do I.
>
> I prefer the "distortion" effect because that's exactly how the real
> world looks.
Are you referring to the distortion on the "normal" shots, or the curved
earth effect from tilting the camera?
The "normal" shots (equirectangular) have a very realistic perspective IMO,
but not the curved earth shots, the world definitely doesn't look like that!
> You just have to look at each part of the picture
> individually and not as a whole... just like you do with the world, in
> fact.
I read an interesting article about this subject, our eyes only focus
sharply on the bit in the middle, most of the rest of what we see is blurred
and we don't take much notice of it.
http://www.ferrario.com/ruether/articles.html#perspective
> To me, rectilinear lenses are the distorting ones.
True, distances are stretched at the corners in order to keep straight lines
straight.
> Have a look at
>
> www.braeburn.co.uk/pix/trees.jpg (35K).
>
> That's an 18mm lens so has a 90 degree aov and I really object to the
> fact that the trees are growing out of the ground at an angle. Much
> better to have curved trees.
Well that's due to tilting the camera upwards, as Phil said. My panoramic
shot of the chimney didn't exactly look normal either. The 18mm lens tilt
is comparable to tilting my panoramic camera, both produce unrealistic
looking images.
> Your eye and brain does a great trick when you see the
> scene. It does the swinging lens thing *and* produces a straight
> trunk. With a camera you have to make a compromise
My shots of Tesco and Jessops are very good examples of the "cigar"
distortion you get with the swing lens, but in reality, it's not distorted
at all. Each end of the Tesco building gets smaller because it's further
away from the camera, hence the "cigar" effect, but this effect is real, we
just don't look at the whole scene at once, we concentrate on one small part
at a time and our brain corrects the rest.
> >Your eye and brain does a great trick when you see the
> >scene. It does the swinging lens thing *and* produces a straight
> >trunk. With a camera you have to make a compromise
>
> That last is true. But for me the lens manufacturers have the wrong
> compromise. I prefer the compromise of the swinging lens (cylindrical
> projection) not the rectilinear.
I prefer it too, at least as far as panoramas are concerned. Rectilinear is
necessary for architectural shots, but most landscapes don't have straight
lines so a bit of distortion, real or otherwise, is often not that
noticeable.
>The "normal" shots (equirectangular) have a very realistic perspective IMO,
>but not the curved earth shots, the world definitely doesn't look like that!
>
>> You just have to look at each part of the picture
>> individually and not as a whole... just like you do with the world, in
>> fact.
>
>I read an interesting article about this subject, our eyes only focus
>sharply on the bit in the middle, most of the rest of what we see is blurred
>and we don't take much notice of it.
>
>http://www.ferrario.com/ruether/articles.html#perspective
>
>> To me, rectilinear lenses are the distorting ones.
>
>True, distances are stretched at the corners in order to keep straight lines
>straight.
>
>> Have a look at
>>
>> www.braeburn.co.uk/pix/trees.jpg (35K).
>>
>> That's an 18mm lens so has a 90 degree aov and I really object to the
>> fact that the trees are growing out of the ground at an angle. Much
>> better to have curved trees.
>
>Well that's due to tilting the camera upwards, as Phil said. My panoramic
>shot of the chimney didn't exactly look normal either. The 18mm lens tilt
>is comparable to tilting my panoramic camera, both produce unrealistic
>looking images.
>
Indeed. But with your lens, even at the same 45 degree up angle the
trees would have been curved inwards at the top and parallel verticaly
at the bottom. The rectilinear lens is distorting what the eye actually
looks at just to keep straight lines straight.
Bernard Hill
Selkirk, Scotland
> What does post processing in Photoshop entail when editing your images?
Well since you asked, it's basically just correcting the images so that they
look like the original slides and optimising them for display on the web.
If I just scanned them and put them straight online they'd look terrible.
This is what I did with the panoramic images (slightly different procedure
with my other images because of using a different scanner).
First I scanned them with my new Epson 2450 scanner. This is much softer
than my Minolta scanner so the images required a lot more sharpening, but
the softness means that it doesn't pick up every tiny speck of dust so
hardly any time is spent removing them.
One curious thing about the Epson scanner is that I can't seem to switch the
automatic exposure off when it previews the image. Even with the colour
balance set to None in VueScan, the scanner still adjusts it automatically,
so I then have to "correct" the automatic corrections manually, which is a
bit silly.
Having said that, the Epson automatic settings do usually produce very good
colours, and usually quite accurate too when using the Epson software, but
not with the VueScan software for some reason. Due to the automatic
cropping on the Epson software being far too aggressive with my panoramic
shots, I had to use the VueScan software instead. The Epson is a very good
scanner for the price, but it has a few quirks, mostly in the software.
Anyway, after scanning each image I do this in Photoshop 7;
1. Crop image (using Marquee tool (M) - fixed aspect ratio - 29x12 - then
Alt-I-P to crop)
2. Levels (Control-L). Click on Options, then play about with various
values until I get the best result. Enhance Monochromatic Contrast tends to
work best with unusual colouring (like sunsets), whereas Enhance Per Channel
Contrast often produces good results but changes the colouring. Snap
Neutral Midtones tends to work well with the latter, but not the former. I
don't use Find Light & Dark Colours very often, but sometimes it works well.
The whole point of the above is to adjust the colour balance (and compensate
for what the scanner did wrong) to try to recreate the correct colours in
the picture. The new Photoshop options in the Levels tool are very good. I
find that different settings work best with different pictures so I never
apply the default auto settings, I fine tune them for each picture.
3. Curves (Control-M). This is mainly to adjust contrast in the picture,
but only if it needs it. Sometimes it's necessary to lighten the shadows,
or darken them, to get the most pleasing balance. The shadows in some of
the photos were much too dark, like the Afon Twrch Weir shot for example,
and sometimes too light, as in the heather shot. The last few shots had
very flat lighting so I needed to boost the contrast to put a bit of life
back into them.
It can also be used to compensate for under exposure, brightening the whole
image in a balanced way (scaling the dark and light points) unlike the
Brightness and Contrast tool which works like a sledgehammer and causes over
exposure in the highlights.
Also, if the Levels tool fails to sort the colours out, I use the Curves
tool on one or more colour channels to get rid of the colour cast, for
example adding a little yellow to reduce the excessive blue light on an
overcast day.
4. Resize (Alt-I-I). Resize the image to an appropriate viewing size for
the computer, in this case 2900x1200 pixels.
5. Image Mode. Change image mode from 16 bits per channel (48-bit TIFF) to
8 bits (24-bit image). It's important to do this after adjusting the colour
balance, contrast and brightness.
6. Contrast Masking (optional). I've only done this on a few shots, like
Swansea Beach Sunset and the last few overcast shots. It's mainly to darken
bright skies and lighten dark land. I've explained the whole procedure
previously so I won't repeat it here. Settings used on the pano scans were
typically Gaussian Blur radius 20, contrast mask layer opacity 50%. Layers
were then flattened and additional Levels or Curves applied to fine tune the
image.
It would be better to do contrast masking before changing the image mode,
but layers don't work in 16-bit mode unfortunately.
7. Dust and scratch removal. Very little if anything to remove with this
scanner, but what little there is I remove with the Healing Brush tool (J)
(works very well on skies) or the Rubber Stamp tool (S).
8. Sharpen image. Normally I'd use Unsharp Mask for delicate adjustments
but due to the softness of this Epson scanner I need to give it a full blast
of Sharpen More.
9. Save the image once as Maximum quality JPEG, Baseline Optimised. This
is the "master" image that I keep on the computer for personal viewing.
Then I need to prepare it for the web.
10. Resize (Alt-I-I) to 1208x500. Adjust Canvas Size to 1200x500 (slight
cropping).
12. Sharpen. This time I use Unsharp Mask, settings - Radius 1, Threshold
4, Amount 50% to 100%, depending on image. Most shots use 100%.
13. Save as Maximum quality JPEG, Baseline Optimised.
Then I need to create the thumbnail.
14. Resize to 360x150.
15. Sharpen. Unsharp Mask, same settings.
Finally, once I've got all the web images and thumbnails in the Web folder,
I throw the whole lot into JPEG Optimizer and bulk optimise them all at a
quality setting of 70.
> Nope not at all .. just wondered how much of what we see in those images
> were shot on film and not manipulated in a graphics package.
All of the shots except Swansea Bay Sunset are pretty much identical to the
original slides. So what you see on my website is very close to what I got
on my film.
The problem is that most scanned shots *don't* look like the originals, so
most of the processing is done to recreate the original photo (or original
scene, they're not always the same), not to change the image into something
fake.
Computer monitors are quite a different medium to film, so scanned images
need to be post processed to make them look comparable. Straight scans
don't usually look very good, which explains many of the sub-standard shots
you see on the Internet. A lot of people expect a perfect image straight
out of the scanner, but that doesn't happen. You need to work on them a bit
first.
As for the Swansea Bay Sunset shot, I really don't know what happened to the
colours in that one. The blue wasn't that shade of blue and the magenta was
more orangey. It was much worse originally, but even after considerable
fiddling about it's still not right. I got fed up with fiddling in the end.
It's not that different to the original, but it was different enough for me
to make a note on the web page that the colours weren't correct.
I'm primarily a photographer, so I don't go out and take crap photos then
try to turn them into something wonderful on the computer. I try to take
the best photos I can then use digital processing to recreate them as
accurately as possible (or the original scene, if the camera didn't capture
it properly).
If I do digitally manipulate an image to create something unreal, then I'll
state that specifically. I don't see anything wrong with artistic image
manipulation in principle, provided that you're honest about it.
Speaking of which, I have to admit that there is one digitally manipulated
image on my site;
http://www.wilderness-wales.co.uk/pg/nw/ow75.jpg
I removed a telegraph pole and the associated wires.
> I've always liked wide-angles lens'. I liked the Swansea docks photos
best.
Thanks. Very nice light that day.
Photoshop Elements makes very good panoramas from multiple shots. The
newly released version 2 actually does 360 degree panoramas too.
Current price is about 80 quid but you can often find deals. The
other thing it lets you do is optimise for web and email so you end up
with tiny files - 20kb or thereabouts.
My "Rydal Water" at
http://www.themos.org/lakes/pages/Rydal%20Water.htm was compiled from
2 separate photographs
Grasmere at http://www.themos.org/lakes/pages/grasmere.htm was
compiled from 3 separate photos.
I omitted to turn autoexposure off while taking the photos and it is
possible to see the join in the sky. Will try harder next time.
Tom
> The one of Tesco made my pulse quicken in time in time with the
> mighty retailer itself.
Oh I see! Only just got that joke... :-)
Paul
--
http://www.wilderness-wales.co.uk/
> Photoshop Elements makes very good panoramas from multiple shots. The
> newly released version 2 actually does 360 degree panoramas too.
Actually I just got that bundled with my new scanner, I thought it was just
a cut down version of Photoshop. Does it do much else that Photoshop
doesn't?
> Current price is about 80 quid but you can often find deals. The
> other thing it lets you do is optimise for web and email so you end up
> with tiny files - 20kb or thereabouts.
I use JPEG Optimizer for that. Maybe I should make some comparisons.
> My "Rydal Water" at
> http://www.themos.org/lakes/pages/Rydal%20Water.htm was compiled from
> 2 separate photographs
>
> Grasmere at http://www.themos.org/lakes/pages/grasmere.htm was
> compiled from 3 separate photos.
>
> I omitted to turn autoexposure off while taking the photos
Easy to forget that if you're used to using automatic. Were you using a
digital camera?
> and it is
> possible to see the join in the sky. Will try harder next time.
I wouldn't have noticed in the Rydal pic if I hadn't been looking for it,
just looks like the clouds are brighter on one side. With Grasmere the
joins are noticeable on the water but the sky looks fine.
Paul
--
http://www.wilderness-wales.co.uk/
Nice results.
I think that some of the pictures show that the camera does have to be
carefully levelled if it's full potential is to be realised. Tilting up
or down doesn't really work (e.g. the Marina shot, Tesco shot) ...
rather like using a fish-eye but a bit weirder :-)
I'm not a great fan of 'super-wide' panoramas as you need to
continuously scan the image with the eyes rather than be able to
appreciate the composition. So ...'Swansea Docks' ..... interesting, but
not my 'cup of tea'.
I think that you should be able to take some really great landscapes
with the camera.
(It would be nice to see some of the pictures displayed full-screen)
One last thing. One or two of the pictures show a light band between
hills and sky. Over-sharpening in the post-processing?
regards
<lurk>
--
Jake
I liked composition and display of both photographs .... especially as
they're easy to look at on a 800x600 screen.
My only comments would be:
Rydale ........ needs a little additional sharpening?
Grasmere ..... suffering badly from JPEG artifacts. I took a closer look
at the image and it appears to have been compressed to a 53.18:1 ratio.
A typical digital camera would usually have fine/normal/basic settings
(or something of that type). The compression used here is well below
'basic'.
The greater the compression - the lower the quality.
regards
<lurk>
--
Jake
> Grasmere ..... suffering badly from JPEG artifacts.
It's all those tourists.
--
===============================================
Chris Game <chrisgameATyahoo!DOTcoDOTuk>
===============================================
Why don't you just use Save for Web in PS ?
Carol
> Why don't you just use Save for Web in PS ?
>
> Carol
>
Or you can actually customise your own if your dead anal like me by using
Image/Image Size and editing the variables in that control panel.
You can see the results of me reducing 1600x1200 images to 640x480 (all 72
res) here:
I wrote it up in about an hour last night and it is running off my computer
(fixed IP address) via IIS so nobody please start trying to exploit it thank
you!
Sundays were made for learning :)
Ands
Oh aye the wee doggy is on Ben Wyvis there, or at least on the path up to
the foot on An Cabar, though they could pretty much be anywhere!
Ahhh ...
Nice pics - I like the coulors of the grass and bracken.
Carol
> Nice results.
Thanks.
> I think that some of the pictures show that the camera does have to be
> carefully levelled if it's full potential is to be realised.
Definitely.
> Tilting up
> or down doesn't really work (e.g. the Marina shot, Tesco shot) ...
> rather like using a fish-eye but a bit weirder :-)
No, unless you specifically want a special effect. I won't be doing that
too often, unless it's unavoidable, and even then I'm hoping to sort out
some software solutions to correct the distortion.
Actually the Tesco shot wasn't tilted, that was perfectly levelled on a
tripod. The distortion in that shot is called cigar distortion and it's
characteristic of swing lens cameras. The reason is that each end of Tesco
is further away from the camera, so gets progressively smaller, resulting in
apparent distortion.
It's not actually distorted though, it looks like that to the eye, but our
brains don't perceive it that way, mainly because we have to turn our heads
to see the whole thing and we only look at one small part of it at any one
time, so we imaging that it looks straight, but it actually doesn't. The
camera has captured a true representation in this case. "Normal" lenses are
corrected to make the straight lines appear straight in the photo, but this
is not an accurate representation of reality, it's how we imagine it looks.
Of course, normal lenses don't have this amount of coverage anyway, so the
distortion is not usually noticeable, except in very wide angle lenses.
> I'm not a great fan of 'super-wide' panoramas as you need to
> continuously scan the image with the eyes rather than be able to
> appreciate the composition. So ...'Swansea Docks' ..... interesting, but
> not my 'cup of tea'.
I'd have to agree with you there. Now if only my monitor was 12 feet wide
it would look great! Actually, this sort of panorama would look great if it
were printed at a large size and hung on a very long wall, but you just
can't appreciate it on a computer screen.
> I think that you should be able to take some really great landscapes
> with the camera.
That's the idea.
> (It would be nice to see some of the pictures displayed full-screen)
Not sure what you mean. You do know that you can click on the thumbnails to
see a larger image, don't you? Most of the images are 1200x500 pixels in
size.
> One last thing. One or two of the pictures show a light band between
> hills and sky. Over-sharpening in the post-processing?
I presume you mean the last few of the Afon Twrch? That was due to contrast
masking, not sharpening. The sky was much brighter than the land so I used
contrast masking to balance them out. Tricky to do a perfect job without a
lot of fiddling about, but I still think it looks better than dark grey land
and a bleached out sky.
Thanks for the comments.
Paul
--
http://www.wilderness-wales.co.uk/
> > I use JPEG Optimizer for that. Maybe I should make some comparisons.
>
> Why don't you just use Save for Web in PS ?
Tried that once, wasn't impressed, it just reduced the image quality. JPEG
Optimizer reduces the file size whilst maintaining image quality, by
discarding less important information like colour information. Colour
quality can be reduced quite a lot without being very noticeable.
But maybe Save for Web has been improved since I first looked at it. Maybe
I should look at it again.
Paul
--
http://www.wilderness-wales.co.uk/
"Carol Haynes" <lairdswoodDO...@yahoo.co.uk> wrote in message
news:apglsp$19201$1...@ID-104261.news.dfncis.de...
>
> Ahhh ...
>
> Nice pics - I like the coulors of the grass and bracken.
>
> Carol
>
Aye well it is a Sony 2MP which to me suggests quality and they were
captured at 1600x1200 and reduced to 640x480 as previously said though this
does not affect the colouration of the images. They ARE very vibrant
colours indeed but I am not so sure they are the same colour as they would
have been on the day! :P Maybe that is where the extra 3MP of real stills
photography (real stills being 5MP or the equivalent?) is being missed by my
digital camera, not enough dots to make up the real colours? Then again it
may be the "real colours" and I should just go back to bed for another extra
hour sometime this afternoon ;P
Ands
I do agree that JPEG Optimizer is pretty well a great app yeah and resizing
your images and dropping the res manually in PS will not quite do the same
as this cracking wee app as it has been specifically written to take into
consideration all the factors and creata the smallest image at the quality
level you choose. Then again saving for web should really do the same
thing, too lazy to go and try it although I think I have done before and it
did not really reduce the files size as you say.
Cheers,
Ands
> Maybe that is where the extra 3MP of real stills
> photography (real stills being 5MP or the equivalent?)
Excuse me while I pick myself up off the floor... :-)
I don't want to start a big debate about this, but a Nikon 4000 scanner will
produce images of roughly 6000 x 4000 pixels from a 35mm slide. That's
24MP. My scanner does about 7MP.
According to scientific tests, film can record even more detail than a 4000
dpi scanner can extract, but in practice few people will be using sharp
enough lenses and there'll be some camera shake unless you use a rock solid
tripod.
Various sources disagree on exactly how many MP 35mm film is equivalent to,
and one of the reasons is because it's impossible to make a precise
comparison for a variety of technical reasons, like grain size and errors
introduced at the scanning stage. I think a number of websites agree that
the MP equivalent is roughly 20MP, but even then you need to be using the
highest quality lenses and film and your technique needs to be perfect.
Your average snapshooter isn't going to get that sort of quality from a £30
compact using cheap film.
Apparently when used professionally, film scans are reduced in size before
printing, to minimise grain I believe, so a 24MP scan can end up as a 6MP -
12MP image at the printing stage, so the practical quality difference is not
so far off as you suggested.
Film and digital images have quite different characteristics. It's not
simply a case that one is better than the other. As I understand it, the
big advantage of film is that it's much sharper. However, digital has no
grain (let's forget the noise issue for now). So although film is sharper,
it's also grainier, so the big question is how much can you enlarge film
before you see the grain? Does the grain matter? Some reckon it doesn't.
If grain doesn't bother you then you can enlarge film much more than
digital, but if you limit yourself to not enlarging past the point of
visible grain, then the difference is much smaller.
For practical purposes by average users I'd say that 6MP is quite
sufficient, as much as most people are ever likely to need. Coincidentally
that's exactly the same size as the largest PhotoCD image (not counting Pro
PhotoCD).
Canon have now released an 11MP SLR for £5,500. Kodak have released a 13MP
camera, so clearly the pros want more pixels.
> is being missed by my
> digital camera, not enough dots to make up the real colours?
No, that's not it. Each pixel contains all the necessary colour information
(24-bits - 16 million colours) although the colours are interpolated in
cheap (sub £1000) digital cameras.
The colours look fine to me though, but then I use Velvia...
Paul
--
http://www.wilderness-wales.co.uk/
Me neither but I don't want to let the side down so...
I like my £190 (cheap off ebay) 2MP Sony camera as it lets me take over 100
images at 1600x1200 that I can reduce for the web or burn on CD on one 128MB
Sony memory stick. I have no aspirations to be a professional photographer
but like to capture a good day out so I can look back on it/record it for
the website (www.highlandhiking.co.uk) I am one day to manage to get my act
together to work on! :P For the weight and size of it I am well happy,
having to carry extra films, lenses etc does not appeal to me when I find
carrying maps, water, compasses etc enough of a pain in the behind already
;)
Horses for courses, interpolated ones or not! :P
Ands
>
>> I'm not a great fan of 'super-wide' panoramas as you need to
>> continuously scan the image with the eyes rather than be able to
>> appreciate the composition. So ...'Swansea Docks' ..... interesting, but
>> not my 'cup of tea'.
>
>I'd have to agree with you there. Now if only my monitor was 12 feet wide
>it would look great! Actually, this sort of panorama would look great if it
>were printed at a large size and hung on a very long wall, but you just
>can't appreciate it on a computer screen.
In which case I'm not sure how you would view it. You could walk along
it and study each section?
>
[snip]
>
>> (It would be nice to see some of the pictures displayed full-screen)
>
>Not sure what you mean. You do know that you can click on the thumbnails to
>see a larger image, don't you? Most of the images are 1200x500 pixels in
>size.
>
O.K. I didn't know they were thumbnails.
>> One last thing. One or two of the pictures show a light band between
>> hills and sky. Over-sharpening in the post-processing?
>
>I presume you mean the last few of the Afon Twrch? That was due to contrast
>masking, not sharpening. The sky was much brighter than the land so I used
>contrast masking to balance them out. Tricky to do a perfect job without a
>lot of fiddling about, but I still think it looks better than dark grey land
>and a bleached out sky.
"Swansea from Townhill" is a good example of what I was looking at.
>
>Thanks for the comments.
>
You're welcome :-)
--
Jake
I beg to differ....
If you project onto a flat surface, then straight lines remain straight.
Only when you project onto a curved surface do they bend.
Now the eye projects onto a curved surface (the retina) and has the
the required transformation engine (wetware) to convert the curved lines
back to straight one.
A 'normal' film camera projects onto flat film. So provided there are
no rectilinear distortions in the lense, lines remain straight (but not,
of course, necessarily parrallel).
The panoramic camera projects onto curved film - so bends straight
lines. The amount of curvature will depend (I think) on the distance
to the object being photgraphed. Near items suffereing from the
greatest distortion.
David
Yes, but is the detail there on the original to make 24Mp worth while? I
can always scan a 10x8 photo at any source at 600dpi on my scanner to
make a picture of 29Mp but it proves nothing about the resolution
quality of the picture - maybe I took it at 640x480 and printed it
enlarged...
>
>According to scientific tests, film can record even more detail than a 4000
>dpi scanner can extract, but in practice few people will be using sharp
>enough lenses and there'll be some camera shake unless you use a rock solid
>tripod.
Agreed.
>
>Various sources disagree on exactly how many MP 35mm film is equivalent to,
>and one of the reasons is because it's impossible to make a precise
>comparison for a variety of technical reasons, like grain size and errors
>introduced at the scanning stage. I think a number of websites agree that
>the MP equivalent is roughly 20MP, but even then you need to be using the
>highest quality lenses and film and your technique needs to be perfect.
>Your average snapshooter isn't going to get that sort of quality from a £30
>compact using cheap film.
I think 20Mp is optomistic: certainly for good quality medium or large
format this would be a minimum, but all I can say for 35mm is that my
10x8 from my 4Mp digital are as good as the similar size printed from my
Pentax with zoom lens.
Printing quality, lens quality are as you say actually more important
than pixel count for picture quality.
However one point of confusion is that a scanned picture contains all
the 24M pixels in full colour. A digital camera does not have all its
pixels in full colour: half are green light sensitive, a quarter are
blue, quarter red so the best estimate of actual pixel count is to
divide the Mp quoted by the manufacturer by 3.
That brings my Olympus E10 down to about 1.3 actual sampled full-colour
points.
>
>Apparently when used professionally, film scans are reduced in size before
>printing, to minimise grain I believe, so a 24MP scan can end up as a 6MP -
>12MP image at the printing stage, so the practical quality difference is not
>so far off as you suggested.
>
>Film and digital images have quite different characteristics. It's not
>simply a case that one is better than the other. As I understand it, the
>big advantage of film is that it's much sharper. However, digital has no
>grain (let's forget the noise issue for now). So although film is sharper,
>it's also grainier, so the big question is how much can you enlarge film
>before you see the grain? Does the grain matter? Some reckon it doesn't.
>
>If grain doesn't bother you then you can enlarge film much more than
>digital, but if you limit yourself to not enlarging past the point of
>visible grain, then the difference is much smaller.
>
>For practical purposes by average users I'd say that 6MP is quite
>sufficient, as much as most people are ever likely to need. Coincidentally
>that's exactly the same size as the largest PhotoCD image (not counting Pro
>PhotoCD).
I tend to agree. Of course the crucial question is "how large do you
want to print?" and for up to A4 4Mp is OK. At A3 then it's 5 or 6
ideally. Some say that 10x8 or A4 is about the worst case as with larger
pictures you stand further back to view them.
>
>Canon have now released an 11MP SLR for £5,500. Kodak have released a 13MP
>camera, so clearly the pros want more pixels.
It's always good to have more, especially if you are cropping before
printing.
>
>> is being missed by my
>> digital camera, not enough dots to make up the real colours?
>
>No, that's not it. Each pixel contains all the necessary colour information
>(24-bits - 16 million colours) although the colours are interpolated in
>cheap (sub £1000) digital cameras.
Actually the colours are interpolated in all digital cameras (except the
new Foveon sensor ones which have yet to prove themselves).
Each pixel of the *output file* contains the full colour information,
but at the sensor stage you only have 1 byte (8 bits) of intensity at
each pixel: before outputting the file to the memory card the camera
invents information to make it 3 bytes per pixel. This is the "Bayer
Interpolation" process.
The more expensive cameras can supply 10 or 12 bits of intensity at each
point and give it to the user in RAW format. But whether you can
sensibly use that in your photo editor is a bit doubtful and depends on
the manufacturer's software supplied with the camera. Certainly when
converting to a format for printing (Tiff or jpg) you will drop straight
down to 8 bits.
Maybe that's what you meant about the sub £1000 cameras.
Bernard Hill
Braeburn Software
Author of Music Publisher system
Music Software written by musicians for musicians
http://www.braeburn.co.uk
Selkirk, Scotland
01750 21854
+44 1750 21854
PMFJI
You're assuming the rays go through the camera without deviation. Your
statement is correct for a pinhole camera.
However wide-angle lenses compress the viewing area of maybe 90 degrees
into less on the film, so unless there is some attempt by the camera
lenses not to do so, they will give curves, like a fish-eye lens.
Look at this photo:
http://www.braeburn.co.uk/pix/trees.jpg
The angle of view vertically is about 90 degrees (rectilinear 18mm
lens). But when you view this picture you are not looking at 90 degrees
top to bottom and so you get the distortion of the trees growing out of
the ground at an angle because so-called rectilinear lenses keep
straight lines straight at the price of altering angles.
>Now the eye projects onto a curved surface (the retina) and has the
>the required transformation engine (wetware) to convert the curved lines
>back to straight one.
Or perhaps I would argue that reality is spherical as we look around us,
so that straight lines are of course curved, and should be, avoiding the
bad effects of the photo above.
>
>A 'normal' film camera projects onto flat film. So provided there are
>no rectilinear distortions in the lense, lines remain straight (but not,
>of course, necessarily parrallel).
>
>The panoramic camera projects onto curved film - so bends straight
>lines. The amount of curvature will depend (I think) on the distance
>to the object being photgraphed. Near items suffereing from the
>greatest distortion.
>
> David
>
Perhaps I don't disagree with you... it depends what you are saying. To
me, reality is spherical with me at the centre. So straight lines should
not appear straight on photos either. That's what I call rectilinear
distortion and is illustrated in the picture above.
Cheers for all that information. It is a geeks dream to digest such things!
;) Not sure I will remember it all but it was helluva interesting to read.
Ands
>In article <apgqeq$nbu$1...@news6.svr.pol.co.uk>, Paul Saunders
><pv...@wildwales.fsnet.co.uk> writes
>>"Carol Haynes" <lairdswoodDO...@yahoo.co.uk> wrote
>>
>>> > I use JPEG Optimizer for that. Maybe I should make some comparisons.
>>>
>>> Why don't you just use Save for Web in PS ?
>>
>>Tried that once, wasn't impressed, it just reduced the image quality. JPEG
>>Optimizer reduces the file size whilst maintaining image quality, by
>>discarding less important information like colour information. Colour
>>quality can be reduced quite a lot without being very noticeable.
>>
>>But maybe Save for Web has been improved since I first looked at it. Maybe
>>I should look at it again.
>Maybe I'm doing something wrong, but "Save for web" is not entirely
>satisfactory for me. Some of my original images are so large, that to
>get them down to a decent size for viewing means the file size is still
>way too large for web work if I want to keep as good a quality as I'd
>like.
I haven't looked at it too much but I think you are supposed to resize
the image first and then Save for Web. That's the way I work. Decide
the pixel dimensions you want the web image to be and process in PS to
get it looking how you want and then go to JPEG optimiser or save for
Web to knock the thing down to a small enough file size.
--
Phil Cook looking north over the park to the "Westminster Gasworks"
Thanks Phil, This is just the way I do it. I usually size my images to a
decent size - say 800 x 600 or a bit smaller then go into .jpeg
optimizer. I like my web files to be about 30kb but they usually end up
at around 70 to 80 kb!
What is your idea of a good file size for the web please?
> I like my web files to be about 30kb
I think that's too small.
> but they usually end up
> at around 70 to 80 kb!
Much better.
> What is your idea of a good file size for the web please?
I'd say around 100k, that's perfectly reasonable for a decent quality photo
I think. Most of my pics are between 50k and 120k, 70-80k is about the
average, that's for 750x500.
My panos vary from 40k to 230k (only one that big), the average is around
110k, which is not bad for a 1200x500 image.
Paul
--
http://www.wilderness-wales.co.uk/
I'd agree with Paul. Much bigger than 120kb is probably going to try
the patience of anyone on a 56k modem. Even if you are expecting it
having to wait for a large file to load is infuriating. having a
decent thumbnail to preview is important., make them too small and you
can't see what you will be looking at.
It is. And broadband is not even evailable in many parts of the UK.
Neither ADSL nor cable. ISDN is, of course but it's very little faster
than dial-up.
And some of us run internet-based businesses :-|
Interesting Paul!
I assume that 70 to 100 k are for pictures on their own with a link from
another page. If such pictures were included in a subject page with text
it would cause the page to take too long to down load for viewer
comfort.
A TR for example would have thumbnails linking to a page with bigger
single pictures.
Thanks Phil - I read Paul's reply first, and I agree with you both on
this one.
On one occasion I did a report on a fishing trip and included pictures
in the report. It took ages to download fully, so I split it into three
pages. Thumbnails would have been another option.
You may have visited my site, so perhaps a critique would be possible
please.
>You may have visited my site, so perhaps a critique would be possible
>please.
Whilst folk are discussing Web authoring, I have been thinking it was
about time I put some of that free web space to use myself. So what do
most of you use as an editor?
I have Front Page 2000, but people tend to be rude about it, and
Freeserve doesn't support Front Page Extensions. My plan was to learn
the basics with FP, and then perhaps try and get to grips with
Dreamweaver.
BTW I have books on both editors, and a couple on HTML.
--
Trevor Dennis
Remove s-p-a-m to email
I wouldn't recommend learning the basics of HTML from Front Page, since
the HTML it produces is pretty difficult to read and full of a lot of
tags you don't need.
I'd personally suggest Dreamweaver as a good starting point (though it's
not cheap), or perhaps just use a free editor such as Netscape Composer.
I personally tend to use Bluefish, running on Linux (
http://bluefish.openoffice.nl/) but I don't think a Windows version exists.
Jon.
I use Dreamweaver, you can download a 30 day trial to see if you get
on with it.
IMHO, HTML is the best way to go if you can be bothered learning it -
but I can't, hence DW :-)
Any fool can learn the basics of HTML. Have a look at my site
http://www.p-t-cook.freeserve.co.uk all written by my own fair hand
after looking at the code of four pages generated by Front Page Expess
and a few other sites. You only need to know half a dozen tags. Take a
look here http://www.jonstorm.com/html/
Yes, I used the same strategy : I bought a very low-level book first, to
get a general hint of what html looks like ; then I used the precise
specifications of html and of css which can be found at
http://www.w3.org (with their useful validator to check the astounding
amount of mistakes you can do which are accepted by your browser, but
perhaps not by others).
http://www.blooberry.com/indexdot/ was also a very useful reference.
And I also needed http://css.nu/pointers/bugs.html for the exact
opposite of the html validator : this shows what is correct in your code
but does not work on bugged browsers (mainly Netscape 4, which is an
absolute crap). Hint : work out your site under Netscape 4. If it works
there, there are reasonable chances that it works elsewhere too.
--
My walking website : http://www.lavaurs.com
Last update : GR 10 in the Hautes-Pyrénées
http://www.lavaurs.com/en/sentiers/P_003
>I use Dreamweaver, you can download a 30 day trial to see if you get
>on with it.
>
>IMHO, HTML is the best way to go if you can be bothered learning it -
>but I can't, hence DW :-)
Cheers Simon. It had been my intention to get going with Front Page, and
possibly switch to Dreamweaver. But Freeserve doesn't seem to like FP,
so there doesn't seem much point in learning it. So I'm currently
struggling through the absolute beginners tutorials in DWMX help. It's a
bugger getting old.
Carol
"Trevor Dennis" <tre...@tdennis36.s-p-a-m.fsnet.co.uk> wrote in message
news:cuhWeVFF...@tdennis36.fsnet.co.uk...
>
>Cheers Simon. It had been my intention to get going with Front Page, and
>possibly switch to Dreamweaver. But Freeserve doesn't seem to like FP,
>so there doesn't seem much point in learning it. So I'm currently
>struggling through the absolute beginners tutorials in DWMX help. It's a
>bugger getting old.
Good luck! Whatever you end up going for, make sure it's not FP,
there's far to much MS-only stuff in there (which is why so many ISPs
won't support it).
S.
--
This space left blank for new .sig
> I like my Ł190 (cheap off ebay) 2MP Sony camera as it lets me take over
100
> images at 1600x1200 that I can reduce for the web or burn on CD on one
128MB
> Sony memory stick. I have no aspirations to be a professional
photographer
> but like to capture a good day out so I can look back on it/record it for
> the website (www.highlandhiking.co.uk) I am one day to manage to get my
act
> together to work on! :P For the weight and size of it I am well happy,
> having to carry extra films, lenses etc does not appeal to me when I find
> carrying maps, water, compasses etc enough of a pain in the behind already
> ;)
That's simple, just leave the maps, water and compasses at home. Only take
the essential stuff!
I agree about the lenses though. Aside from the extra weight, I hate the
time consuming fiddliness of changing the damn things. Not only does it
disrupt the flow of the walk, but it can be a problem in difficult
conditions. I prefer the simplicity of a good zoom lens, I generally just
carry two, but often can't be bothered to change at all. Two cameras with a
zoom on each works better, a wide zoom and a telephoto zoom, both mounted on
a belt so they can be whipped out quickly without ruining the flow of the
walk too much.
Of course, if I'm really in the mood to take serious photos, then fiddling
about with lenses, tripods, filters and exposure meters is all part of the
fun, but I don't get much walking done in those situations.
While we're on the subject of digital cameras, I still can't get hold of a
Canon G3 anywhere, still not on general release in Britain yet. Perhaps a
week or two, but they've been saying that for six weeks now.
Paul
--
http://www.wilderness-wales.co.uk/
> >> I'm not a great fan of 'super-wide' panoramas
> >Actually, this sort of panorama would look great if it
> >were printed at a large size and hung on a very long wall, but you just
> >can't appreciate it on a computer screen.
>
> In which case I'm not sure how you would view it. You could walk along
> it and study each section?
Just scroll left and right. Not ideal but what else can you do?
> >> One last thing. One or two of the pictures show a light band between
> >> hills and sky. Over-sharpening in the post-processing?
> "Swansea from Townhill" is a good example of what I was looking at.
Yeah, probably over sharpened.
Paul
--
http://www.wilderness-wales.co.uk/
> To me, reality is spherical with me at the centre.
> So straight lines should not appear straight on photos either.
Absolutely.
Paul
--
http://www.wilderness-wales.co.uk/
I wrote;
> > It's not actually distorted though, it looks like that to the eye, but
our
> > brains don't perceive it that way, mainly because we have to turn our
heads
> > to see the whole thing and we only look at one small part of it at any
one
> > time, so we imaging that it looks straight, but it actually doesn't.
The
> > camera has captured a true representation in this case.
David Laight replied;
> I beg to differ....
>
> If you project onto a flat surface, then straight lines remain straight.
> Only when you project onto a curved surface do they bend.
Not true. With wide angle lenses straight lines would appear curved were it
not for the fact that the lenses are corrected to straighten the lines.
This is most apparent with extremely wide angle lenses, and the distortion
due to correction shows itself as objects being larger toward the edges and
corners of the photo than in the centre. This is particularly noticeable if
you photograph a person's face close up with a very wide angle lens, and
their face is offset from the centre. You'll see the person's face being
stretched toward the corner of the photo.
With standard lenses and longer, straight lines are not much of a problem,
because the field of view is quite narrow and projects onto a flat surface
quite well, but a wide angle lens does not, since the field of view is
actually curved.
One way of thinking of it is rather like the way that the ancients perceived
the night sky, as a large black sphere encompassing the earth dotted with
lights on the inside. In effect, focusing on infinity is rather like
focusing on the inside of that imaginary sphere, and projecting the image
onto a flat film distorts the shape in much the same way that map
projections of the curved earth's surface do.
Multiple elements inside the lens serve to correct the distortions just like
certain map projections. In the case of a 180 degree fish eye lens, there
is no correction, so half of the inside of the imaginary sphere is captured
on a circle within the film. If you were to lie flat on the ground and
point such a lens directly above you, it would capture the entire sky, so
the edge of the circle would correspond to the horizon, which we usually
think of as a straight line, but it's only straight if you look at in a
particular way (the way we usually look at it).
All straight lines are curved in a fish-eye photo (except those that pass
through the centre), the result of projecting a semi-sphere onto flat film.
The only way to avoid that would be to take the photo on spherical film, but
no-one's yet invented that so far as I know, and I have no idea how you'd
view the result!
With wide angle lenses the result is the same, but less pronounced, and
extra lens elements are inserted to correct it. It's very easy to see the
distortions of the lens corrections though, just look through a wide angle
lens (with an SLR) and pan the camera left and right. You'll see the
landscape distort as you pan.
> Now the eye projects onto a curved surface (the retina) and has the
> the required transformation engine (wetware) to convert the curved lines
> back to straight one.
This would be rather like projecting onto the spherical film I mentioned, so
perhaps little transformation would be necessary since we're receiving a
spherical image on a spherical surface.
In fact, you can see curved lines with your eyes if you look hard enough.
Look at some long straight lines, but without moving your eyes. That's the
hard part. Only the centre part of our vision is in sharp focus, so we tend
to pan our eyes left and right, only concentrating on the central part of
the image. We tend not to look very closely at the blurry areas in our
peripheral vision, but if you try hard enough you'll see that straight lines
really do look curved.
In the case of a panoramic camera, it captures a field of view roughly as
wide as we can see with our peripheral vision, but we don't look at real
panoramas like that, we turn our eyes and our heads to scan the whole scene.
When we look at the whole scene, we're not really looking at the whole
scene, just the central part. The blurry edges just give context.
A panoramic camera shows the whole scene in sharp focus, which is not how we
see it. It also compresses the field of view into something much narrower
than reality, so the perspective looks wrong to us. The only true way to
view a panorama would be to display it on the inside of a sphere, and to
look at it from inside the sphere, at the correct distance so that the field
of view exactly matches the original field of view. In that situation it
wouldn't look distorted, because you'd have to turn your head to see the
whole thing.
The nearest practical way of doing this would be to display the panorama at
a very large size in an arc on the inside of a large cylinder that you could
stand inside, viewing it from the exact centre of the cylinder.
> A 'normal' film camera projects onto flat film. So provided there are
> no rectilinear distortions in the lense, lines remain straight (but not,
> of course, necessarily parrallel).
But only because of lens corrections.
> The panoramic camera projects onto curved film - so bends straight
> lines.
No, it doesn't work like that. In a true panoramic camera the lens rotates
as it takes the photo, projecting a narrow vertical strip of light onto the
curved film as it rotates. At all times the distance of the lens from the
film is the same, so the film is effectively flat relative to the lens. In
fact the reverse is true for flat film, because the edges of a curved lens
are further away from the film.
The rotation of the lens approximates the turning of one's head to look at a
panorama, and matches the shape of the imaginary sphere much better than an
extreme wide angle or fish-eye lens. In fact the only lens distortion in a
panoramic photo is that which occurs vertically, the top and bottom of the
image being slightly stretched. But horizontally there is no stretching.
> The amount of curvature will depend (I think) on the distance
> to the object being photgraphed. Near items suffereing from the
> greatest distortion.
This is true anyway. Near objects look larger than more distant objects,
and if part of an object is more distant than another part it will look
smaller, so the object will appear distorted, *except* that it doesn't
appear distorted in real life because our brains compensate for what we see.
For some reason our brains like to imagine that straight lines look
straight. This is probably because we know that straight lines are
straight, so we can't accept them looking curved, even though they really do
look that way. Since we usually only pay close attention to the sharp
centre of our field of vision, we conveniently fail to notice the curvature
in the scene as a whole.
Photographs show this distortion far too clearly though, which is why lenses
are corrected to make straight lines appear as we imagine them to look.
With the panoramic camera, there is no such correction, it looks as it
really appears, but you can't really appreciate a panoramic photograph
properly unless it is enlarged considerably and displayed in an arc, and
viewed from the focal point of the arc so that the relative field of view
would be correct.
In the case of the Tesco photograph, which exhibits "cigar-distortion",
http://www.wilderness-wales.co.uk/pg/pano/tesco1200.jpg
the image is not distorted because the film is curved, but because the
building is straight. The apparent curvature is real, but we do not notice
this in real life because we only see part of the scene at one time. If you
look at just a small part of the photo it looks fine, it's only when you
look at the whole scene in it's entirety that it looks wrong, because we
can't see the whole scene with such clarity in real life.
The reason for the curvature is simple. Each end of the building is further
away from the camera than the centre of the building. Distant objects are
smaller, so the ends of the building appear smaller than the centre. As
each end of the building increases in distance, it becomes smaller,
gradually tapering away in a gentle curve, not a straight line as you might
expect. If they did taper away in straight lines in both directions then
those straight lines would meet at an angle in the centre of the image.
Clearly this is impossible, so they cannot be straight, they must be curved.
The standard rules of perspective invented by artists are not correct, they
fail to take the curvature into account. In art, one side of a building is
often shown as tapering away, but with straight lines, not curved. This is
fine when you are looking at a building from an angle, but not when looking
at it head on. In the latter case, artists depict the straight lines as
straight, yet this is not so.
In most cases, with a narrow field of view, the distance of each part of the
building away from the observer is quite similar, so slight curvature isn't
noticed, but in the Tesco case the building is very wide and the observer is
very close, necessitating turning of the head to see the whole scene
properly. This is what the panoramic camera is doing, turning it's "head"
to capture the whole scene. The reason it looks so odd is that it's
displayed as a flat image with a narrow field of view, which doesn't match
the circumstances in which it was taken.
In contrast if the building were curved in an arc, and I were to photograph
the building from the focal point of the arc, such that all parts of the
building were an equal distance from the camera, the result would be
perfectly straight lines, exactly as it would appear to an observer. Thus
this camera is ideal for photographing the curved interiors of circular
shaped buildings, provided you stand at the focal point of the arc.
It would have been much easier to simply draw a few diagrams rather than
trying to explain this in text. The important thing to be aware of when
using a panoramic camera is that the camera photographs an arc, rather than
a flat plane as is the case with a "normal" corrected lens.
Because of this it's best to try to find subject matter which is equidistant
from the lens in all directions, so subject matter that forms straight lines
is far from ideal. What you need to look for is a natural curve in the
landscape, then photograph it from within the curve. A good example would
be the Glyders, which curves around (actually it's more of an "L" shape but
it's close enough). The focal point of this curve is in the vicinity of Pen
yr Ole Wen.
Pen yr Ole Wen is known as a great viewpoint for the Glyders, but it would
particularly suit the panoramic camera. The only problem is that it faces
south-west, so the light would be in the wrong place for much of the day,
especially in the afternoon. The ideal time would be in the very early
morning during the summer months when the sun is high in the sky.
The sun is quite a problem with this panoramic camera. Since the field of
view is so wide it tends to intrude into many desirable scenes. The best
solution is good planning, but you can't have the sun in the ideal spot for
every scene encountered throughout the day. Paradoxically therefore, I've
been finding that days with sun and lots of clouds around can be ideal for
panoramic shots taken into the light.
Simply compose the picture, wait until a cloud covers the sun, then take it,
which is the exact opposite of what I'd normally do! The immediate
foreground will be in shade of course, but since the scene is so wide, many
other parts of the picture will be sunlit, and the clouds covering the sun
can be a very interesting part of the photo in themselves, especially if
there are rays bursting through. I've already had some success with this
technique.
Paul
--
http://www.wilderness-wales.co.uk/
> >I don't want to start a big debate about this, but a Nikon 4000 scanner
will
> >produce images of roughly 6000 x 4000 pixels from a 35mm slide. That's
> >24MP. My scanner does about 7MP.
>
> Yes, but is the detail there on the original to make 24Mp worth while?
In most cases, probably not. In some cases, yes. If you've taken a photo
with sufficient care with a good quality lens on a sturdy tripod it may be
worthwhile, but for most people most of the time it just isn't necessary.
However, if you plan on selling scanned images to publishers it's worth
buying, because they consider this to be a professional quality scanner.
At the moment there is some snobbery in the digital publishing business
where they expect scanned photos to be between 50 and 75 meg in file size.
The Nikon produces 65 meg scans. This may just be elitism, but they
apparently don't want to consider purchasing scans made with the "cheap"
scanners that most of us own. Having said that, the Nikon does offer some
extra features which can enhance scan quality, like digital ICE, grain
reduction and restoring faded colours.
> I
> can always scan a 10x8 photo at any source at 600dpi on my scanner to
> make a picture of 29Mp but it proves nothing about the resolution
> quality of the picture - maybe I took it at 640x480 and printed it
> enlarged...
True, but even if you hadn't the quality wouldn't be as good. Any print is
of inferior quality compared the original neg or slide. Anyone who owns a
decent slide scanner has seen the evidence of that.
> >I think a number of websites agree that
> >the MP equivalent is roughly 20MP,
>
> I think 20Mp is optomistic:
Perhaps. One site said that it's only a third of that, giving the reason
that colour film is only a third as sharp as black and white film.
Presumably the quality estimate was for black and white film in that case,
but is there really that much difference between black and white and colour
film or is this just naive speculation based on the fact that colour images
are made out of three separate colours? Sounds suspiciously like the colour
interpolation that digital cameras use, see later in this post.
> all I can say for 35mm is that my
> 10x8 from my 4Mp digital are as good as the similar size printed from my
> Pentax with zoom lens.
I imagine they would be. I've seen A3 photos printed from a 4MP digital
camera (Canon S40) and was amazed at the sharpness, and I was looking
closely! I won't really believe this until I get one and do my own
comparative tests.
> Printing quality, lens quality are as you say actually more important
> than pixel count for picture quality.
Yes.
> However one point of confusion is that a scanned picture contains all
> the 24M pixels in full colour. A digital camera does not have all its
> pixels in full colour: half are green light sensitive, a quarter are
> blue, quarter red so the best estimate of actual pixel count is to
> divide the Mp quoted by the manufacturer by 3.
No, I've heard this argument before and it doesn't pan out.
> That brings my Olympus E10 down to about 1.3 actual sampled full-colour
> points.
Colour resolution yes, but not luminosity resolution. If you take a 1.3MP
full colour scan and interpolate it up to 4MP, the result will be much less
sharp than an equivalent 4MP image. I suggested an experiment along these
lines to someone who suggested the same thing, and after trying it he agreed
that it was far worse.
The fact is that although the colour resolution is only 1.4MP, the
luminosity resolution is 4MP, and our eyes are more sensitive to luminosity
than colour. Luminosity is primarily responsible for our perception of
sharpness in an image. Colour is not very important to sharpness. I've
done experiments in Photoshop where I've blurred only the colour in a photo,
and the result still looked just as sharp, but with an odd colour smoothing.
I tried this as a way of reducing grain, and it worked to a degree, it got
rid of the "colour grain" (the multi-coloured pixels) but not the
"luminosity grain".
In effect, a digital camera is taking a black and white photo at a high
resolution, then superimposing colour over it at a lower resolution. Since
luminosity is responsible for showing shapes, the lower resolution colour is
not important to the perception of sharpness. This is why black and white
photography is viable, colour is not very important in showing shape. In
fact, if you removed the luminosity from an image and left only colour (hue
and saturation), you'd hardly recognise what you were looking at.
I've just tried this again for a bit of fun and verified that blurring the
colour only results in no loss of sharpness. Even with maximum blur (which
results in the original image looking practically like a grey card) the
image still looks just as sharp when the blurring is applied only to the
colour (it looks just like a black and white photo in fact), but when the
luminosity is blurred (effectively removed) the result looks like
meaningless splodges of colour. You can barely make out what the image is,
and it doesn't look sharp at all.
Slight blurring of the colour results in an image which is almost
indistinguishable from the original, except for very small brightly coloured
details (like people on a beach).
In practical terms, for digital cameras, the colour interpolation is only a
problem when there are sharp divisions between bright and dark objects, or
very different colours. Then the colour can "spill over" into adjacent
pixels, but with higher resolution cameras this isn't so noticeable. For
typical landscape work it would rarely be a problem and you could always fix
it in Photoshop if it were sufficiently objectionable.
Coming back to the argument that colour film is three times less sharp than
black and white because of the three different colour layers, I don't see
why this would make the photo any less sharp, for the reasons stated above.
The only reason I can see why black and white would be sharper than colour
is if the grains are that much finer, which may well be the case, but three
times sharper?
> >For practical purposes by average users I'd say that 6MP is quite
> >sufficient, as much as most people are ever likely to need.
>
> I tend to agree. Of course the crucial question is "how large do you
> want to print?" and for up to A4 4Mp is OK. At A3 then it's 5 or 6
> ideally.
Yet I've seen a very impressive A3 print from a quality 4MP camera.
> Some say that 10x8 or A4 is about the worst case as with larger
> pictures you stand further back to view them.
I looked very close.
> >Each pixel contains all the necessary colour information
> >(24-bits - 16 million colours) although the colours are interpolated in
> >cheap (sub £1000) digital cameras.
>
> Actually the colours are interpolated in all digital cameras (except the
> new Foveon sensor ones which have yet to prove themselves).
Not all. You're thinking primarily of consumer cameras. Some expensive
digital cameras use a triple colour filter system, where the photo is taken
three times, once through each of the three colour filters. Naturally such
cameras are useless for action photos and have to be used on a tripod.
Also there are the large format scanning backs which literally scan a scene
through the lens and can take up to half an hour to do it. The quality and
dynamic range are supposed to be phenomenal (about 11 stops I believe) and
so is the price (about £30,000 last time I checked).
> Each pixel of the *output file* contains the full colour information,
> but at the sensor stage you only have 1 byte (8 bits) of intensity at
> each pixel: before outputting the file to the memory card the camera
> invents information to make it 3 bytes per pixel. This is the "Bayer
> Interpolation" process.
Yes.
> The more expensive cameras can supply 10 or 12 bits of intensity at each
> point and give it to the user in RAW format.
The Canon G3 can do 12 bits, the G2 is 10 bits.
> But whether you can
> sensibly use that in your photo editor is a bit doubtful and depends on
> the manufacturer's software supplied with the camera.
No, it's not doubtful, it's quite possible, with the Canons at least.
> Certainly when
> converting to a format for printing (Tiff or jpg) you will drop straight
> down to 8 bits.
Tiff can support 48 bit images (3x24 bits), so no loss of data there.
> Maybe that's what you meant about the sub £1000 cameras.
Paul
--
http://www.wilderness-wales.co.uk/
> I assume that 70 to 100 k are for pictures on their own with a link from
> another page.
Yes.
> If such pictures were included in a subject page with text
> it would cause the page to take too long to down load for viewer
> comfort.
Well only if there were lots of other things on the page or enormous
quantities of text. Text is pretty small and loads in very quickly. It
also gives you something to read while you're waiting.
> A TR for example would have thumbnails linking to a page with bigger
> single pictures.
Yeah.
Paul
--
http://www.wilderness-wales.co.uk/
Why? It's a fact about the Bayer Interpolation. What "doesn't pan out"?
THe new foveon chip in the Sigma SD9 is the only one with a 3million
sensors which register in all colours. And a quick look at 1:1 on the
sample pictures with any other pictures from any other camera shows the
difference.
>
>> That brings my Olympus E10 down to about 1.3 actual sampled full-colour
>> points.
>
>Colour resolution yes, but not luminosity resolution. If you take a 1.3MP
>full colour scan and interpolate it up to 4MP, the result will be much less
>sharp than an equivalent 4MP image. I suggested an experiment along these
>lines to someone who suggested the same thing, and after trying it he agreed
>that it was far worse.
That's no surprise! I'm not sure what your point is.
>
>The fact is that although the colour resolution is only 1.4MP, the
>luminosity resolution is 4MP, and our eyes are more sensitive to luminosity
>than colour. Luminosity is primarily responsible for our perception of
>sharpness in an image. Colour is not very important to sharpness. I've
>done experiments in Photoshop where I've blurred only the colour in a photo,
>and the result still looked just as sharp, but with an odd colour smoothing.
>I tried this as a way of reducing grain, and it worked to a degree, it got
>rid of the "colour grain" (the multi-coloured pixels) but not the
>"luminosity grain".
I don't think you are accurate there. Remember that half the elements
only register red and blue light, and we don't register luminosity on
those pixels in any great sensitivity. We rely pretty much on the 50% of
pixels which are green-sensitive, so the luminosity is based on only
half the pixels, and full colour on 1/3 of them.
As I said, visit the sample images of the foveon on
and look at them at 1:1 compared to any picture from any other camera at
1:1.
>
>In effect, a digital camera is taking a black and white photo at a high
>resolution,
High-ish. 50% of the pixel count :-)
> then superimposing colour over it at a lower resolution. Since
>luminosity is responsible for showing shapes, the lower resolution colour is
>not important to the perception of sharpness. This is why black and white
>photography is viable, colour is not very important in showing shape. In
>fact, if you removed the luminosity from an image and left only colour (hue
>and saturation), you'd hardly recognise what you were looking at.
>
>>
[snip]
>> I tend to agree. Of course the crucial question is "how large do you
>> want to print?" and for up to A4 4Mp is OK. At A3 then it's 5 or 6
>> ideally.
>
>Yet I've seen a very impressive A3 print from a quality 4MP camera.
Yes, but that's because we don't really need 300dpi resolution for a
good picture. At 4Mp, A3 size is about 130dpi.
>
[snip]
>> Actually the colours are interpolated in all digital cameras (except the
>> new Foveon sensor ones which have yet to prove themselves).
>
>Not all. You're thinking primarily of consumer cameras. Some expensive
>digital cameras use a triple colour filter system, where the photo is taken
>three times, once through each of the three colour filters. Naturally such
>cameras are useless for action photos and have to be used on a tripod.
OK, you must allow me to discount the £10,000+ cameras from any
discussion as I have no experience of them. Not even seen a picture.
All I know is that the Bayer interpoation is used in cameras right up to
£6000 except the SD9 which has just been released.
>> The more expensive cameras can supply 10 or 12 bits of intensity at each
>> point and give it to the user in RAW format.
>
>The Canon G3 can do 12 bits, the G2 is 10 bits.
>
>> But whether you can
>> sensibly use that in your photo editor is a bit doubtful and depends on
>> the manufacturer's software supplied with the camera.
>
>No, it's not doubtful, it's quite possible, with the Canons at least.
You would have to use the manufacturer's software of course. The screen
only ever shows you 8 bits. Does Photoshop work in 12 bits per colour???
Bernard Hill
Selkirk, Scotland
>While we're on the subject of digital cameras, I still can't get hold of a
>Canon G3 anywhere, still not on general release in Britain yet. Perhaps a
>week or two, but they've been saying that for six weeks now.
Paul, I've bitten the bullet and ordered an S45 today. Waiting for the
G3 was very tempting, but I wanted to get the hang of it before going on
holiday, and I really wanted to get away from lugging heavyweight camera
equipment up hills. I shall be very interested to swap notes when you
get your G3.
BTW www.equipment-express.co.uk had the best deal on the S45.
Can't think I when I was this excited about getting a new toy. ;-)
> >> so the best estimate of actual pixel count is to
> >> divide the Mp quoted by the manufacturer by 3.
> >
> >No, I've heard this argument before and it doesn't pan out.
>
> Why? It's a fact about the Bayer Interpolation. What "doesn't pan out"?
The "actual pixel count" being divided by three. The resolution is what it
is, it's not a third of the image size. The *colour* resolution is only a
third, agreed, but that doesn't reduce the luminosity resolution.
> THe new foveon chip in the Sigma SD9 is the only one with a 3million
> sensors which register in all colours. And a quick look at 1:1 on the
> sample pictures with any other pictures from any other camera shows the
> difference.
I never said there wasn't a difference. Obviously a full colour sensor will
produce a better result than an interpolated monochromatic one, but in that
case it's the colour resolution which has improved, not the luminosity
resolution.
> >> That brings my Olympus E10 down to about 1.3 actual sampled full-colour
> >> points.
> >
> >Colour resolution yes, but not luminosity resolution. If you take a
1.3MP
> >full colour scan and interpolate it up to 4MP, the result will be much
less
> >sharp than an equivalent 4MP image. I suggested an experiment along
these
> >lines to someone who suggested the same thing, and after trying it he
agreed
> >that it was far worse.
>
> That's no surprise! I'm not sure what your point is.
My point is that even though the colours are interpreted, that doesn't
reduce the resolution of the image. A 4MP image is still a 4MP image even
if the colours are interpolated. If you take a full colour 1.3MP image and
interpolate it up to 4MP then it won't be as sharp as the 4MP image with
interpolated colours. But obviously a 4MP full colour image will be better,
since the colours will be sharper.
> >I've
> >done experiments in Photoshop where I've blurred only the colour in a
photo,
> >and the result still looked just as sharp, but with an odd colour
smoothing.
> >I tried this as a way of reducing grain, and it worked to a degree, it
got
> >rid of the "colour grain" (the multi-coloured pixels) but not the
> >"luminosity grain".
>
> I don't think you are accurate there.
Yes I am, but I did my experiment on full colour scans from film, not images
from a digital camera.
> Remember that half the elements
> only register red and blue light, and we don't register luminosity on
> those pixels in any great sensitivity. We rely pretty much on the 50% of
> pixels which are green-sensitive, so the luminosity is based on only
> half the pixels, and full colour on 1/3 of them.
No, I disagree. I know that we are less sensitive to red and blue light,
but we can still see them. Look at a full colour image in Photoshop and
compare the three different colour channels (viewed in monochrome). There
is plenty of "luminosity" in the red and blue channels, both show images
which are clearly recognisable as what they are. The green channel gives
the most natural looking result, but all three colour channels provide
clearly recognisable images, so all three colours must contribute to the
luminosity resolution of a digital camera.
The colour "smoothing" caused by the Bayer interpolation will naturally
degrade the image quality, but it doesn't reduce the actual resolution.
> >In effect, a digital camera is taking a black and white photo at a high
> >resolution,
>
> High-ish. 50% of the pixel count :-)
Nah, 100%. Black and white photographers often take photos through red
filters. Are these photos invisible to our eyes? Of course not.
> >The Canon G3 can do 12 bits, the G2 is 10 bits.
> You would have to use the manufacturer's software of course. The screen
> only ever shows you 8 bits. Does Photoshop work in 12 bits per colour???
Yep, it works in 16 bits per channel. I always make the initial adjustments
to my raw scans in this mode , then reduce to 8 bits per channel later.
Paul
--
http://www.wilderness-wales.co.uk/
> what are the differences between the G2 and G3 ?
There are lots of small improvements, like extra modes and rearrangements of
controls and menus to make it more ergonomic. Apparently there's now a
wheel you can turn to adjust things like focus. You can do time-lapse
photography with it, it can do automatic exposure bracketing and even focus
bracketing. You can save custom settings. Video size is larger and maximum
time is longer. Processing speed is increased. Infra-red remote is
included. Any photo can optionally be saved in RAW format after being
taken, even if the camera isn't in that mode.
The biggest differences though are these;
1. 12-bit sensor rather than 10-bit, so RAW images will produce 36-bit
pictures. This gives more latitude for manipulating images later and I
think increases the dynamic range also.
2. Higher quality sensor. It's a new design. Same number of pixels but a
better quality image.
3. Zoom lens increased from 3x optical to 4x, so it's now equivalent to
35-140mm rather than 35-105mm.
> Also, when you do manage
> to get one, could you send me an email with your thoughts on it ? I was
> thinking about changing from my A20 after Xmas.
Certainly, but I'll probably forget. Email me sometime to remind me.
Once I get one you'll certainly hear about it on this newsgroup, and of
course I'll be putting plenty of pics onto my website. I'm not in a
position to compare with other digital cameras, since I've never used one
before, but I will be doing some test comparisons between the G3 and my film
cameras to see how the quality compares. I'll also be interested to see if
I can find any applications that digital can do better than film, aside from
the obvious instant results. Blended exposures is one area that digital
seems well suited to.
Paul
--
http://www.wilderness-wales.co.uk/
> Paul, I've bitten the bullet and ordered an S45 today. Waiting for the
> G3 was very tempting, but I wanted to get the hang of it before going on
> holiday, and I really wanted to get away from lugging heavyweight camera
> equipment up hills.
Heavyweight? 1 lb?
I've just had a look at the features on the S45 and they pretty much match
the G3. The main difference is the lens, the G3 has a better, faster lens
with a greater zoom range, and of course is bigger and heavier (and more
expensive?).
> I shall be very interested to swap notes when you
> get your G3.
Certainly.
Paul
--
http://www.wilderness-wales.co.uk/
OK. Not massive. But my EOS 630 with its F1:2-2:8 28-70mm zoom weighs a
ton, and as it's more than I care to lose, I carry it on multi-day walks
rather than leave it in the car. With the S45 I can leave it at home and
rely on my Ricoh GR1 for film. I am oh so looking forward to NOT
carrying the EOS. :)
>I've just had a look at the features on the S45 and they pretty much match
>the G3. The main difference is the lens, the G3 has a better, faster lens
>with a greater zoom range, and of course is bigger and heavier (and more
>expensive?).
Yep. Plus the S45 has no flash shoe, (and an inadequate onboard flash).
>> I shall be very interested to swap notes when you
>> get your G3.
>
>Certainly.
Well I picked it up this morning in glorious sunshine, but by the time I
had charged the battery it grey flat and boring. Still very pleased with
the results, but they needed some perking up in PS to get the best out
of them. Thinking about it though, the end result was much better than I
could have achieved with scanned film.
Still struggling with all the functions. Fingers crossed for a nice
weekend.
> I'll also be interested to see if
>I can find any applications that digital can do better than film, aside from
>the obvious instant results. Blended exposures is one area that digital
>seems well suited to.
The obvious answer for us photographer mortals is that we can be more
prolific, but with the amount of film you get through, digital probably
won't make that much difference.
I am going to take loads and loads of sky pictures for perking up less
than inspiring landscapes. Bryce 5 and the KTP6 Sky Effects filter do
reasonable jobs with skies, but can never compete with the real thing.
Then plenty of experimental shots with deliberate camera movement,
and/or very long exposures. The instant feedback would help fine tune
technique and hopefully get some interesting shots.
Multi exposures from a fixed viewpoint - as I suspect you mean with the
blended exposures. Perhaps a cityscape at dusk, through sunset and
beyond, with the classic car light trails and maybe a few neon signs.
You could surely blend such a sequence for a nice effect in PS.
BTW The software that comes with the G3 has a panorama stitching
facility. No idea how well it works yet.
>
>I am going to take loads and loads of sky pictures for perking up less
>than inspiring landscapes. Bryce 5 and the KTP6 Sky Effects filter do
>reasonable jobs with skies, but can never compete with the real thing.
Good luck! I always find it hard if not impossible to get a pasted in
sky to look good. The light needs to match. The best sky is the one
that was there at the time. Or rather the best landscape includes a
good sky.
--
Phil Cook back looking north over the park to the "Westminster
Gasworks" instead of gum trees and brown grass.
> >Heavyweight? 1 lb?
>
> OK. Not massive. But my EOS 630 with its F1:2-2:8 28-70mm zoom weighs a
> ton,
I still think of 35mm SLR gear as lightweight. Now MF and LF gear is heavy,
that's serious gear. I was very pleased when I got my panoramic camera to
discover that it only weighed 2lb.
> and as it's more than I care to lose, I carry it on multi-day walks
> rather than leave it in the car.
I don't blame you, my cameras are worth more than my car. But I take your
point about light weight when backpacking. 35mm gear can break the camel's
back.
> With the S45 I can leave it at home and
> rely on my Ricoh GR1 for film. I am oh so looking forward to NOT
> carrying the EOS. :)
There are definitely times when it's a great relief to leave all the SLR
gear behind, but not on clear sunny days!
> Yep. Plus the S45 has no flash shoe, (and an inadequate onboard flash).
Yeah, the G3 seems to have quite a few flash features and accessories.
Rarely ever use flash myself, but maybe I'll make the effort to experiment a
bit more since it won't cost anything. Using flash to light foreground
subjects in twilight with a colourful sky behind has a lot of potential.
> Well I picked it up this morning in glorious sunshine, but by the time I
> had charged the battery it grey flat and boring.
Still sunny here, well starry anyway...
> Still very pleased with
> the results, but they needed some perking up in PS to get the best out
> of them.
I'm convinced that all photos benefit from perking up in PS.
> Thinking about it though, the end result was much better than I
> could have achieved with scanned film.
Good to hear you say that.
> Still struggling with all the functions.
It will no doubt take a bit of practice. Free practice, that is!
> Fingers crossed for a nice weekend.
Yes. I've just had some good news. Jessops are expecting a delivery of a
G3 with my name on it. Probably tomorrow, I mean today... Must get to
sleep soon, so that I can wake up before they close... ;-)
Paul
--
http://www.wilderness-wales.co.uk/pg/mg/mg.html
> > I'll also be interested to see if
> >I can find any applications that digital can do better than film,
> The obvious answer for us photographer mortals is that we can be more
> prolific,
Well that goes without saying. What I'm particularly interested in is
applications that it can do better than film, or at least easier. Panoramas
is probably one area. Not only are they very expensive with film, but a
digital panorama can be of much larger resolution and hence compete with
film. And shots don't always have to be joined in a line, they can be
arranged in rows, giving the equivalent of an extremely wide-angle lens.
> but with the amount of film you get through, digital probably
> won't make that much difference.
Wanna bet? I've been holding back because of the cost! Just wait until the
floodgates open! (Once I get enough memory cards.)
> I am going to take loads and loads of sky pictures for perking up less
> than inspiring landscapes.
I've been photographing skies since I first started photography, long before
I ever realised what I'd one day be able to do with them. The only problem
is searching through my collection and finding them. I could build up a
great sky library if I took the trouble to find and scan them all.
> Bryce 5 and the KTP6 Sky Effects filter do
> reasonable jobs with skies, but can never compete with the real thing.
No. I've only just seen examples of that on the web and I must say that I'm
surprised that anyone serious about photography would consider adding a
computerised sky to a real photo. Adding a real sky is the obvious choice.
> Then plenty of experimental shots with deliberate camera movement,
> and/or very long exposures. The instant feedback would help fine tune
> technique and hopefully get some interesting shots.
Yes, the experimentation is probably the most appealing thing. I've always
been loathe to do more experimenting than necessary because of the cost of
film, so many aspects of photography I've failed to master because I just
couldn't afford it. I'm looking forward to being able to play with a camera
without worrying about the cost.
What may be particularly useful is the ability to test exposures in
difficult lighting conditions before taking a serious shot on film. Pros
who use MF typically take a polaroid to test the exposure, a digital camera
would be far simpler, allowing for the differences of course.
> Multi exposures from a fixed viewpoint - as I suspect you mean with the
> blended exposures.
Yes, the G3 would be ideal for this. In particular, it has the ability to
auto-bracket, so it can take three different exposures in quick succession,
and you can set how wide you want the bracketing to be. Apparently, I've
read that it's possible to take such shots hand held if you use a fast
enough shutter speed. Obviously a tripod would be better to ensure that the
framing is identical for each shot, but it seems that this technique could
even be used for action photography.
The problem with auto-bracketing on a film camera is that you have to wind
the film on in-between shots, and with a less than rock solid tripod, doing
that might alter the framing by a tiny amount, making it difficult to line
them up later. In fact, lining up film scans is difficult anyway because
the frames may not be scanned in identical positions.
> Perhaps a cityscape at dusk, through sunset and
> beyond, with the classic car light trails and maybe a few neon signs.
> You could surely blend such a sequence for a nice effect in PS.
I hadn't thought of that, that's a very interesting idea. The G3 can do
time lapse photography so you could set it up on a tripod and set it to take
one shot every five minutes for say an hour or so.
I was primarily thinking of doing the blended exposures of subjects with
extremely high contrast ranges, like for example the full moon at night.
Ever seen the halo around the moon when there's a thin cloud in front of it?
There is absolutely no way you can capture that on a single exposure, the
moon is far too bright and the sky around far too dark. The G3 only goes
down to 15 seconds I think, which may not be good enough for night
exposures, although using it at f2 at ISO 400 would help a lot.
> BTW The software that comes with the G3 has a panorama stitching
> facility. No idea how well it works yet.
Yes I know, it sounds ideal for accurate stitching. My brother has stitched
many seamless panoramas from his Nikon digital camera and says that it's way
easier than stitching film scans together. I don't know why this is, but if
his experience is anything to go by, it seems to be the case. I hope so,
because I've never been entirely happy with film scan stitching (maybe
taking photos with a 24mm lens introduces too much distortion to be able to
easily compensate).
Paul
--
http://www.wilderness-wales.co.uk/pg/mg/mg.html
> > I'll also be interested to see if
> >I can find any applications that digital can do better than film,
> The obvious answer for us photographer mortals is that we can be more
> prolific,
Well that goes without saying. What I'm particularly interested in is
applications that it can do better than film, or at least easier. Panoramas
is probably one area. Not only are they very expensive with film, but a
digital panorama can be of much larger resolution and hence compete with
film. And shots don't always have to be joined in a line, they can be
arranged in rows, giving the equivalent of an extremely wide-angle lens.
> but with the amount of film you get through, digital probably
> won't make that much difference.
Wanna bet? I've been holding back because of the cost! Just wait until the
floodgates open! (Once I get enough memory cards.)
> I am going to take loads and loads of sky pictures for perking up less
> than inspiring landscapes.
I've been photographing skies since I first started photography, long before
I ever realised what I'd one day be able to do with them. The only problem
is searching through my collection and finding them. I could build up a
great sky library if I took the trouble to find and scan them all.
> Bryce 5 and the KTP6 Sky Effects filter do
> reasonable jobs with skies, but can never compete with the real thing.
No. I've only just seen examples of that on the web and I must say that I'm
surprised that anyone serious about photography would consider adding a
computerised sky to a real photo. Adding a real sky is the obvious choice.
> Then plenty of experimental shots with deliberate camera movement,
> and/or very long exposures. The instant feedback would help fine tune
> technique and hopefully get some interesting shots.
Yes, the experimentation is probably the most appealing thing. I've always
been loathe to do more experimenting than necessary because of the cost of
film, so many aspects of photography I've failed to master because I just
couldn't afford it. I'm looking forward to being able to play with a camera
without worrying about the cost.
What may be particularly useful is the ability to test exposures in
difficult lighting conditions before taking a serious shot on film. Pros
who use MF typically take a polaroid to test the exposure, a digital camera
would be far simpler, allowing for the differences of course.
> Multi exposures from a fixed viewpoint - as I suspect you mean with the
> blended exposures.
Yes, the G3 would be ideal for this. In particular, it has the ability to
auto-bracket, so it can take three different exposures in quick succession,
and you can set how wide you want the bracketing to be. Apparently, I've
read that it's possible to take such shots hand held if you use a fast
enough shutter speed. Obviously a tripod would be better to ensure that the
framing is identical for each shot, but it seems that this technique could
even be used for action photography.
The problem with auto-bracketing on a film camera is that you have to wind
the film on in-between shots, and with a less than rock solid tripod, doing
that might alter the framing by a tiny amount, making it difficult to line
them up later. In fact, lining up film scans is difficult anyway because
the frames may not be scanned in identical positions.
> Perhaps a cityscape at dusk, through sunset and
> beyond, with the classic car light trails and maybe a few neon signs.
> You could surely blend such a sequence for a nice effect in PS.
I hadn't thought of that, that's a very interesting idea. The G3 can do
time lapse photography so you could set it up on a tripod and set it to take
one shot every five minutes for say an hour or so.
I was primarily thinking of doing the blended exposures of subjects with
extremely high contrast ranges, like for example the full moon at night.
Ever seen the halo around the moon when there's a thin cloud in front of it?
There is absolutely no way you can capture that on a single exposure, the
moon is far too bright and the sky around far too dark. The G3 only goes
down to 15 seconds I think, which may not be good enough for night
exposures, although using it at f2 at ISO 400 would help a lot.
> BTW The software that comes with the G3 has a panorama stitching
> facility. No idea how well it works yet.
Yes I know, it sounds ideal for accurate stitching. My brother has stitched
> Good luck! I always find it hard if not impossible to get a pasted in
> sky to look good. The light needs to match.
True, so a range of skies taken from many different angles and focal lengths
is necessary.
> The best sky is the one that was there at the time.
Absolutely, if possible.
> Or rather the best landscape includes a good sky.
Yes, exactly.
I'm tempted to think that the best reason for pasting a sky would be to
create a larger than life image, one which is obviously fake, rather than
trying to make it believable. For example, Hollywood sometimes uses the
gigantic moon effect in the background of some shots, so you could paste a
300mm telephoto sky onto a wide angle shot, or vice-versa! It would look
very odd, but that might be the whole point.
I think if you want a good normal photograph, it's best to just take a good
normal photograph in the first place. Probably less effort and less hassle.
If you want to do collages, then why not do something really unusual that
would be impossible to take anyway?
Speaking of skies, check this one out;
http://www.wilderness-wales.co.uk/pg/mg/mumblesA.jpg
> Phil Cook back looking north over the park to the "Westminster
> Gasworks" instead of gum trees and brown grass.
Ah, that's so sad. Having lived in London for a short while, I can
sympathise with you... :-(
Paul
--
http://www.wilderness-wales.co.uk/pg/mg/mg.html
>Well that goes without saying. What I'm particularly interested in is
>applications that it can do better than film, or at least easier. Panoramas
>is probably one area. Not only are they very expensive with film, but a
>digital panorama can be of much larger resolution and hence compete with
>film. And shots don't always have to be joined in a line, they can be
>arranged in rows, giving the equivalent of an extremely wide-angle lens.
You'll find the G3 has a 'Stitch assist' mode, which, together with the
software that comes with the camera, makes panoramas. And as the camera
and software are working in harmony, I suspect it does a reasonable job
of it.
>No. I've only just seen examples of that on the web and I must say that I'm
>surprised that anyone serious about photography would consider adding a
>computerised sky to a real photo. Adding a real sky is the obvious choice.
Sky substitutions, or enhancement in PS can often make all the
difference to a picture. If anyone was interested I could put together a
sort of PhotoShop sky enhancement FAQ. That would range from easy to do
accurate masks, to enhancing the original, to sky replacement.
>What may be particularly useful is the ability to test exposures in
>difficult lighting conditions before taking a serious shot on film. Pros
>who use MF typically take a polaroid to test the exposure, a digital camera
>would be far simpler, allowing for the differences of course.
Well after using the S45 over the weekend I am convinced digital is the
way forward. The camera produced much better results from some very
dreary conditions than I'd expect from film. The lack of grain and
scanning artefacts produced remarkably clean and punchy images, that
respond very well to adjustment in PS.
There seems to be a definite learning curve however. Some shots I took
using the camera's 'landscape' setting resulted in some very blurred
members of the family against a pin sharp background. Incidentally, a
nice feature of digital is that the camera settings are stored as
metadata in the image file. I can look at Photoshop 7's File Browser and
see that the afore mentioned shots were taken at f4.9 @ 1/10th.
>Yes, the G3 would be ideal for this. In particular, it has the ability to
>auto-bracket, so it can take three different exposures in quick succession,
>and you can set how wide you want the bracketing to be. Apparently, I've
>read that it's possible to take such shots hand held if you use a fast
>enough shutter speed. Obviously a tripod would be better to ensure that the
>framing is identical for each shot, but it seems that this technique could
>even be used for action photography.
>
>The problem with auto-bracketing on a film camera is that you have to wind
>the film on in-between shots, and with a less than rock solid tripod, doing
>that might alter the framing by a tiny amount, making it difficult to line
>them up later. In fact, lining up film scans is difficult anyway because
>the frames may not be scanned in identical positions.
It had occurred to me that a problem with auto bracketing would be that
you couldn't accurately assess the results until you downloaded the
images. (This was from the point of view of discarding worst result to
save Flash Card capacity). But you can zoom in and pan to actual pixel
level on playback, which gives an excellent indication of image
sharpness, exposure etc.
BTW Paul. You can upload and change the start-up image as displayed on
the LCD when you turn the camera on. If the spec is the same as the S45,
then you'll need a 320 * 240 pixel jpeg of up to 20Kb. A Wilderness
Wales logo would look very cool on your G3. ;-)
>Good luck! I always find it hard if not impossible to get a pasted in
>sky to look good. The light needs to match. The best sky is the one
>that was there at the time. Or rather the best landscape includes a
>good sky.
There are several techniques which can be used to enhance skies in
PhotoShop. As I mentioned elsewhere in this thread, I could put together
a mini FAQ on the subject if anyone was interested.
> You'll find the G3 has a 'Stitch assist' mode, which, together with the
> software that comes with the camera, makes panoramas.
Yes, with which I imagine you could get them lined up very precisely, even
hand hand, although I'd prefer to use a monopod at least.
> And as the camera
> and software are working in harmony,
> I suspect it does a reasonable job of it.
This is my assumption also. I've never entirely trusted other programs like
Panavue because the focal length is always wrong. My 24mm shots do not
match up when I use the 24mm setting, instead I let the software detect the
lens size and it always comes up with something quite different, like maybe
30mm. Even then I've had to tweak the numbers by hand until the pictures
matched. I assume this is because of corrective elements inside the lens.
With the Canon, the software should understand how the lens is designed.
Does yours have stitch assist too?
> Sky substitutions, or enhancement in PS can often make all the
> difference to a picture. If anyone was interested I could put together a
> sort of PhotoShop sky enhancement FAQ. That would range from easy to do
> accurate masks, to enhancing the original, to sky replacement.
That sounds good. I'd be interested. The trickiest thing is when you have
a complex object in the sky, like a tree. There are functions in Photoshop
for dealing with dark or light fringing, I imagine they'd be useful in such
a situation.
> Well after using the S45 over the weekend I am convinced digital is the
> way forward.
Glad to hear you say that.
> The camera produced much better results from some very
> dreary conditions than I'd expect from film.
A lot depends on the film you use and the techniques. In dull light you can
use 400 film hand held with a fast shutter speed and produce awful pictures,
or you could use Velvia on a tripod with a small aperture and produce
excellent quality pictures. The latter technique is no good for snapshots
or action shots though.
The bright lenses and range of film speeds on digital cameras should produce
decent results in bad light though.
> The lack of grain and
> scanning artefacts produced remarkably clean and punchy images, that
> respond very well to adjustment in PS.
Yes, the lack of grain is the single biggest advantage IMO.
> There seems to be a definite learning curve however. Some shots I took
> using the camera's 'landscape' setting resulted in some very blurred
> members of the family against a pin sharp background.
Well that's because "landscape" means "slow shutter speed, focused on
infinity". Use with tripod! :-)
> Incidentally, a
> nice feature of digital is that the camera settings are stored as
> metadata in the image file. I can look at Photoshop 7's File Browser and
> see that the afore mentioned shots were taken at f4.9 @ 1/10th.
Yes, that's very handy, can you check that in-camera?
> It had occurred to me that a problem with auto bracketing would be that
> you couldn't accurately assess the results until you downloaded the
> images. (This was from the point of view of discarding worst result to
> save Flash Card capacity). But you can zoom in and pan to actual pixel
> level on playback, which gives an excellent indication of image
> sharpness, exposure etc.
Well I wouldn't auto bracket very often, only in special cases (it would
waste too much memory). If it was a special case then I'd take some time
over it, maybe even taking some test shots first to determine the optimum
exposures, then delete them and take them again with auto-bracketing.
> BTW Paul. You can upload and change the start-up image as displayed on
> the LCD when you turn the camera on. If the spec is the same as the S45,
> then you'll need a 320 * 240 pixel jpeg of up to 20Kb. A Wilderness
> Wales logo would look very cool on your G3. ;-)
:-)
And an easy way to identify it if I lost it...
Paul
--
http://www.wilderness-wales.co.uk/pg/mg/mg.html
>"Trevor Dennis" <tre...@tdennis36.s-p-a-m.fsnet.co.uk> wrote
>
>> You'll find the G3 has a 'Stitch assist' mode, which, together with the
>> software that comes with the camera, makes panoramas.
>
>Yes, with which I imagine you could get them lined up very precisely, even
>hand hand, although I'd prefer to use a monopod at least.
Oh yes. Use a monopod. Even with the help of the edge of the last shot
to help line up it is rather hard to get them to line up properly if
doing it hand held.
>
>> And as the camera
>> and software are working in harmony,
>> I suspect it does a reasonable job of it.
>
>This is my assumption also. I've never entirely trusted other programs like
>Panavue because the focal length is always wrong. My 24mm shots do not
>match up when I use the 24mm setting, instead I let the software detect the
>lens size and it always comes up with something quite different, like maybe
>30mm. Even then I've had to tweak the numbers by hand until the pictures
>matched. I assume this is because of corrective elements inside the lens.
>With the Canon, the software should understand how the lens is designed.
Does it get the focal length from the files?
<snip>
>> Incidentally, a
>> nice feature of digital is that the camera settings are stored as
>> metadata in the image file. I can look at Photoshop 7's File Browser and
>> see that the afore mentioned shots were taken at f4.9 @ 1/10th.
>
>Yes, that's very handy, can you check that in-camera?
Not on my A40. You might get better info out of the G3
>
>> It had occurred to me that a problem with auto bracketing would be that
>> you couldn't accurately assess the results until you downloaded the
>> images. (This was from the point of view of discarding worst result to
>> save Flash Card capacity). But you can zoom in and pan to actual pixel
>> level on playback, which gives an excellent indication of image
>> sharpness, exposure etc.
>
>Well I wouldn't auto bracket very often, only in special cases (it would
>waste too much memory). If it was a special case then I'd take some time
>over it, maybe even taking some test shots first to determine the optimum
>exposures, then delete them and take them again with auto-bracketing.
>
>> BTW Paul. You can upload and change the start-up image as displayed on
>> the LCD when you turn the camera on. If the spec is the same as the S45,
>> then you'll need a 320 * 240 pixel jpeg of up to 20Kb. A Wilderness
>> Wales logo would look very cool on your G3. ;-)
>
>:-)
>
>And an easy way to identify it if I lost it...
You can do the same thing (change the startup screen) on some GPS's so
if somebody finds it and turns it on they will get your details.
My A40 doesn't have that feature but I have just thought that I could
put a pic on the CF card with my details rather like the recomendation
to photograph a card with your address on it on film, which I never
seem to do.
--
>My A40 doesn't have that feature but I have just thought that I could
>put a pic on the CF card with my details rather like the recomendation
>to photograph a card with your address on it on film, which I never
>seem to do.
Blimey. That's a good idea.
>If you are in the habit of deleting all pics on the card once they have
>been downloaded and saved, it means you either have to take the I/D
>picture again, or else delete each pic one by one. Tedious if you have
>taken a lot. :-)
Or how about using one of those little plastic containers designed for
dog leads? You write your details on a tiny roll of paper, insert it in
the capsule, and hang it from the wrist strap loop.
I can select all then use Ctrl+click to deselect the details image
before deleting all the others. Since the file will always be at the
beginning it won't be hard to find.
>Yeah, the G3 seems to have quite a few flash features and accessories.
>Rarely ever use flash myself, but maybe I'll make the effort to experiment a
>bit more since it won't cost anything. Using flash to light foreground
>subjects in twilight with a colourful sky behind has a lot of potential.
A problem with the S45 is the proximity of the flash to lens axis. It
produces the worst red I've ever come across. It does have an RE
reduction feature, but the onboard flash is too weak to really help.
Perhaps I am not doing it right.
The G3 should be a lot better in this respect, and RE isn't a problem
with landscape photography anyway. ;-)
>Yes. I've just had some good news. Jessops are expecting a delivery of a
>G3 with my name on it. Probably tomorrow, I mean today... Must get to
>sleep soon, so that I can wake up before they close... ;-)
Have you got it yet?
Ahhh! I usually delete on the camera, rather than using the softyware.
Must try that wizard wheeze.
Adny-ism left in, which raises the question where is he? Not seen a
post from him for a while.
--
Gordon
> The G3 should be a lot better in this respect, and RE isn't a problem
> with landscape photography anyway. ;-)
Will it be strong enough for foreground lighting though?
> >Yes. I've just had some good news. Jessops are expecting a delivery of
a
> >G3 with my name on it. Probably tomorrow, I mean today...
> Have you got it yet?
No, it still hasn't come in. It's due anytime, it's on their lists, but
maybe there's a UK shortage at the moment? Loads of people are on waiting
lists for it. I believe I'm first on the list in my local Jessops though.
I ordered it two months ago!
Paul
--
http://www.wilderness-wales.co.uk/pg/mg/mg.html
Gone to see his Merkin galfiend I believe.
--
Dominic Sexton
http://www.dscs.demon.co.uk/
Ahh the lovely Shirley
http://www.ziplink.net/~holm/ngg/nggshirley.html
> Oh yes. Use a monopod.
I plan to. It's actually easier than using a tripod because that's a pig to
get aligned properly, even with a spirit level on the tripod. I'd only take
a pano hand held if I was in a big rush, like if I was about to lose the
light or something.
> Even with the help of the edge of the last shot
> to help line up it is rather hard to get them to line up properly if
> doing it hand held.
The main problem is sideways rotation, that's one reason why it can be very
tricky to join them up later, if your pano software doesn't allow you to
adjust that. I've become acutely aware of the sideways rotation problem
since getting my panoramic camera. At first it was very difficult to get
the spirit level centred, until I realised it was the sideways rotation that
was causing the problem. Now I find myself adjusting it quite quickly.
Because of this I think I'll find it easier to line up shots for panos now,
because I'm aware of the problem and am getting used to adjusting it. It's
strange that what looks level to your eye often isn't, and when the spirit
level tells you it's correct it sometimes doesn't look it, but it is, as the
photos demonstrate. Perspective is a peculiar thing.
> >With the Canon, the software should understand how the lens is designed.
>
> Does it get the focal length from the files?
I would hope so.
> Phil Cook back looking north over the park to the "Westminster
> Gasworks" instead of gum trees and brown grass.
Sad to be back?
Paul
--
http://www.wilderness-wales.co.uk/pg/mg/mg.html
Tedious certainly - but well worth it. Great idea.
--
Bill Grey
http://www.billboy.co.uk
<quote>
>softyware
</quote>
More an "Unwinism" surely?