checking quality?

I have a Canon A630, and enjoy "pushing it".

I don't want a better (more recent and/or expensive) camera.

I want to MAX this one.

Can anyone recommend a test procedure to find
out what combination of lens length and aperture
gives me the best sharpness?

BugBear

On Thu, 22 Nov 2007 10:36:48 +0000, bugbear
wrote:

I have a Canon A630, and enjoy "pushing it".

I don't want a better (more recent and/or expensive) camera.

I want to MAX this one.

Can anyone recommend a test procedure to find
out what combination of lens length and aperture
gives me the best sharpness?

BugBear

A search at Google for this filename: "ISO_12233-reschart.pdf" pulls up 355
hits.

This is the same chart that dpreview uses (or used?) for all its lens and camera
resolution tests.

Print it up at the absolute highest quality and largest size that your printer
can do. 18"x13" works nicely if you have a wide-bed printer. Most any 8x10
300dpi printer can also resolve the smallest elements. Higher print resolution
and larger will always help. You can also print it up spanning several pages (4,
8x10s works well without a wide-bed printer), then trim & tape from the back for
a larger image at higher resolution.

Affix it at a distance relative to the focal lengths (FOV) that you will be
comparing. Test any combo of lenses and settings that you desire. Snap off your
photos. You'll have a relative comparison to find what works best with your
camera, what f-stops are best, or when comparing any two or more cameras or
lenses that you own.

I use this often to prove to myself that nearly all P&S cameras that I buy today
beat the DSLRs and their glass being sold. I don't need to depend on someone's
money-making biased review posted online.

I also use this to compare different post-processing methods, i.e. sharpening
methods, rotation/resizing methods, RAW vs. JPG originals, etc. A sharp image of
this being put through its paces to see what comes out the other side.

I've also used it to see if a camera with JPG-only output is using upsampling on
RAW sensor data before JPG conversion for its digital-zoom. Some JPG-only
cameras can provide slightly more detail when using digital-zoom than optical
zoom alone, depending on their in-camera work-flow. I never believe what I am
told online unless I have tested it myself.

A must-have file for any digital photography enthusiast that wants to know if
they are getting what they paid for, or in learning how to optimize the purchase
choice they made. Do you need a target to practice your hand-held shooting
technique? Works well for that too.

grant lexter wrote:
A search at Google for this filename: "ISO_12233-reschart.pdf" pulls up 355
hits.

This is the same chart that dpreview uses (or used?) for all its lens and camera
resolution tests.

Print it up at the absolute highest quality and largest size that your printer
can do. 18"x13" works nicely if you have a wide-bed printer. Most any 8x10
300dpi printer can also resolve the smallest elements. Higher print resolution
and larger will always help. You can also print it up spanning several pages (4,
8x10s works well without a wide-bed printer), then trim & tape from the back for
a larger image at higher resolution.

Affix it at a distance relative to the focal lengths (FOV) that you will be
comparing. Test any combo of lenses and settings that you desire. Snap off your
photos. You'll have a relative comparison to find what works best with your
camera, what f-stops are best, or when comparing any two or more cameras or
lenses that you own.

http://www.dpreview.com/reviews/spec...canon_a630.asp

Any advice on your to manipulate/control focus? My camera
obviously has auto-focus, and also has a hard-to-use
manual focus.

Clearly, if I can't get accurate, consistent focus,
all else is academic.

BugBear

On Thu, 22 Nov 2007 11:27:19 GMT, grant lexter wrote:

On Thu, 22 Nov 2007 10:36:48 +0000, bugbear
wrote:

I have a Canon A630, and enjoy "pushing it".

I don't want a better (more recent and/or expensive) camera.

I want to MAX this one.

Can anyone recommend a test procedure to find
out what combination of lens length and aperture
gives me the best sharpness?

BugBear

A search at Google for this filename: "ISO_12233-reschart.pdf" pulls up 355
hits.

This is the same chart that dpreview uses (or used?) for all its lens and camera
resolution tests.

Print it up at the absolute highest quality and largest size that your printer
can do. 18"x13" works nicely if you have a wide-bed printer. Most any 8x10
300dpi printer can also resolve the smallest elements. Higher print resolution
and larger will always help. You can also print it up spanning several pages (4,
8x10s works well without a wide-bed printer), then trim & tape from the back for
a larger image at higher resolution.

P.S.

To save you some time:

The only program I found that allows you to poster-print any PDF file at the
highest possible resolution of your printer, to any number of page-panels or
degree of enlargement that you want is PhotoLine 32 (www.pl32.net). Using its
advanced page setup options. File Print Options Mode Poster.

After loading, do a "Mark All Layers", "Group Marked Layers". This is necessary
so that all your resizing is applied to all vector graphic data layers, not just
the one or sub-group selected. Resize the PDF document to match whatever final
poster-print resolution needed (Eg. Scale Document 800%). As the original
document resolution is only 4724 px / 2953 px, or therabouts, depending on which
version you download. When output to the printer every poster-panel page will
retain that higher resolution detail. It might also also help to choose the
Layout options of antialias and pixel-mode options if you are using a lower
resolution printer. Ultimately, you will always be limited by your printer's
highest available resolution per page. If you are fortunate enough to have a
high-resolution wide-bed printer, making a wall-chart for studio testing
purposes is easy.

Adobe's "Reader" nor any other program that I tested will work. Or at least I
couldn't find a way to make them work for this purpose.
..
If anyone knows of an easier way to get high-resolution poster-print options
from vector-graphic data PDF files do share. Freeware is nice.


You might also find a "eia1956.jpg" (or "eia1956.pdf" if you can find it)
test-chart online. The JPG when upsampled with a good S-Spline utility, it too
can work as a test target for other purposes.

I just found this one http://www.bealecorner.com/trv900/respat/EIA1956.pdf

It's been a while since I hunted down test-charts online, if anyone has more
good ones, care to share ...

On Thu, 22 Nov 2007 12:24:14 +0000, bugbear
wrote:

grant lexter wrote:
A search at Google for this filename: "ISO_12233-reschart.pdf" pulls up 355
hits.

This is the same chart that dpreview uses (or used?) for all its lens and camera
resolution tests.

Print it up at the absolute highest quality and largest size that your printer
can do. 18"x13" works nicely if you have a wide-bed printer. Most any 8x10
300dpi printer can also resolve the smallest elements. Higher print resolution
and larger will always help. You can also print it up spanning several pages (4,
8x10s works well without a wide-bed printer), then trim & tape from the back for
a larger image at higher resolution.

Affix it at a distance relative to the focal lengths (FOV) that you will be
comparing. Test any combo of lenses and settings that you desire. Snap off your
photos. You'll have a relative comparison to find what works best with your
camera, what f-stops are best, or when comparing any two or more cameras or
lenses that you own.

http://www.dpreview.com/reviews/spec...canon_a630.asp

Any advice on your to manipulate/control focus? My camera
obviously has auto-focus, and also has a hard-to-use
manual focus.

Clearly, if I can't get accurate, consistent focus,
all else is academic.

BugBear

Not being familiar with that camera, your methods that you can find by trail &
error experimentation until you get the hang of it, will be far better than
anything I could suggest by playing out the scenario in my mind. This is one of
those hands-on things that only time and experience will bring you.

Obviously, as with any auto-focus cameras, the steadier you can hold it the
faster the camera can latch-onto/analyze the image details to find proper focus.
This is especially true in dim-light and long-zoom scenarios. Engaging IS into
continuous mode can help with that (if your camera has IS), but then you can't
see how much you are shaking the camera. So while it may latch onto proper focus
faster, you might be using longer shutter-speeds than the IS can compensate for.
A catch-22 situation if your hand-held camera techniques aren't up to par. I
prefer IS in single-shot mode so I can see how much I am shaking the camera and
dampen it out myself. Anything that I can't prevent the camera can then easily
deal with, even during long exposures up to 1 second. Knowing how and why your
camera does what it does can make or break how well you can use it.

Many of the Canon cameras have an option where you can press the "Set" button
when in manual-focus mode. Once you are near to proper focus and then pressing
the "Set" button engages an auto-focus on a micro-step level. Limiting its
focus-search range to whatever is nearest to what you already focused on. If
your camera has that learn to use it. It's a remarkably effective way to combine
manual focus with an auto-focus assist. I think it's even called "Manual-focus
Assist" if I remember right. It does the job and does it well. If you learn to
use it.

If your camera can run CHDK and you use that, then there's a CHDK menu option in
the Misc. section to re-map your manual-focus button presses to your zoom-toggle
lever. This greatly improves the speed and accuracy of manual focusing on those
cameras. If your camera uses button-presses instead of a zoom-lever for zooming
then there's no advantage to that unless zoom has easier to control buttons than
your manual focus buttons.

That's about all the help I can offer without the camera in-hand.

Practice and hands-on experience with instant feedback will always be your best
teacher. You have to train your mind and muscles to work effectively in
conjunction with the camera in your hands. Get those neural pathways engrained.

On Nov 22, 4:36 am, bugbear wrote:
I have a Canon A630, and enjoy "pushing it".

I don't want a better (more recent and/or expensive) camera.

I want to MAX this one.

Can anyone recommend a test procedure to find
out what combination of lens length and aperture
gives me the best sharpness?

BugBear

As others suggest, a lens test chart is certainly the BEST way. But a
quick and dirty test is to shoot a brick wall from varying distances.
Make sure you note which order you vary the aperture/FL combos. Shoot
the longer FLs from longer distances, of course. You want to shoot
from distances so that the bricks appear pretty small in the image at
that FL.

This is an old trick described in several old photography books,
before the net gave us such easy access to lens test charts.

bugbear wrote:
I have a Canon A630, and enjoy "pushing it".

I don't want a better (more recent and/or expensive) camera.

I want to MAX this one.

Can anyone recommend a test procedure to find
out what combination of lens length and aperture
gives me the best sharpness?

Theoretically, an aperture number larger than f/3.3 will suffer from
diffraction on that 1/1.8" sensor with 10MP:
http://www.cambridgeincolour.com/tut...ensor-size.htm
....so f/4 won't help and usually a wider aperture like f/2.8 is softer.

Usually the middle of a given zoom range is sharpest with the least
distortions & aberrations.

I discovered an interesting method for objectively evaluating sharpness
in real-world photos: use unsharp mask at different radii with a large
amount so you can see the effect. Zoom in to 200% to make it easy to
assess. If a radius of 0.3 pixels improves image detail, that's darn
sharp. If you have to go to a radius of 1 or 2, there's a problem. Look
for actual image detail, not just grain, noise & jpeg artifacts although
it could work on noise reduced edges so work with converted unsharpened
raw files. Here's one (from raw) that shows some improvement at 0.25
pixels radius:
http://edgehill.net/California/Bay-Area/San-Francisco/edgehill-garden/Nursery/plants/11-16-07/raw

grant lexter wrote:
Any advice on your to manipulate/control focus? My camera
obviously has auto-focus, and also has a hard-to-use
manual focus.

Clearly, if I can't get accurate, consistent focus,
all else is academic.

BugBear

Not being familiar with that camera, your methods that you can find by trail &
error experimentation until you get the hang of it, will be far better than
anything I could suggest by playing out the scenario in my mind. This is one of
those hands-on things that only time and experience will bring you.

Obviously, as with any auto-focus cameras, the steadier you can hold it the
faster the camera can latch-onto/analyze the image details to find proper focus.
This is especially true in dim-light and long-zoom scenarios.

Any testing of the kind that I'm concerned with will be done
with the camera on a tripod, and using the delayed shutter release.


Many of the Canon cameras have an option where you can press the "Set" button
when in manual-focus mode. Once you are near to proper focus and then pressing
the "Set" button engages an auto-focus on a micro-step level. Limiting its
focus-search range to whatever is nearest to what you already focused on. If
your camera has that learn to use it. It's a remarkably effective way to combine
manual focus with an auto-focus assist. I think it's even called "Manual-focus
Assist" if I remember right. It does the job and does it well. If you learn to
use it.

Yep, already do, when approrpriate.

If your camera can run CHDK and you use that, then there's a CHDK menu option in
the Misc. section to re-map your manual-focus button presses to your zoom-toggle
lever. This greatly improves the speed and accuracy of manual focusing on those
cameras. If your camera uses button-presses instead of a zoom-lever for zooming
then there's no advantage to that unless zoom has easier to control buttons than
your manual focus buttons.

That's about all the help I can offer without the camera in-hand.

OK, all understood.

The problem is that in manual focus, either via rear buttons
or the zoom (I do have CHDK) the only feedback is the little square
in the middle of the display, optionally zoomed X2.

If I'm trying to test the limits of the resolution of the lens,
I need (AFAIK) "perfect" focusing, and I remain unclear
of the best way to get this, or failing perfection,
the best approach to perfection avaialable.

BugBear

On Fri, 23 Nov 2007 10:52:41 +0000, bugbear
wrote:

The problem is that in manual focus, either via rear buttons
or the zoom (I do have CHDK) the only feedback is the little square
in the middle of the display, optionally zoomed X2.

If I'm trying to test the limits of the resolution of the lens,
I need (AFAIK) "perfect" focusing, and I remain unclear
of the best way to get this, or failing perfection,
the best approach to perfection avaialable.


The quickest way that I know on these cameras, for best-case focus, now that I
know you have that MF + Set-button assist.

Auto-focus on the chart with a half-shutter press. While holding in on the
half-press after focus is locked, hit the MF button to lock focus. Then hit Set
once (while in MF mode) to let it micro-focus. These charts being high-contrast
graphics with sharp edges makes it easy for most cameras to latch onto them at
the right spot 99+% of time. Using this auto-focus / MF-lock /
micro-manual-focus-assist seems to be good when I have the time, it's like a
fail-safe double check of auto-focus. Also, once you do it a few times you could
do it blindfolded and fast.

There's only one slight drawback to this, with the camera I was testing. I found
if it already auto-focused to near the best spot, then hitting the Set button
while in MF mode might make it micro-increment just one tiny bit past best
focus. This may have just been an idiosyncrasy of the camera I was testing. It
also didn't happen often enough for me to pin down what caused it exactly. It
seemed to happen when auto-focus locked onto slightly past best focus, then the
"Set" focus-assist press would throw it one small notch past even further,
rather than back in closer for better focus. Almost like it was saying, "Well,
we're already out of focus, what's the sense of me trying to help!" This didn't
happen often enough to really concern me. I do recall I was using a make-shift
lighting setup that day, and it wasn't any too bright. Might have had something
to do with it. You may have to experiment with your camera to see if yours does
this any or at all.

You might find doing it manually will be best of all. Did you read those threads
in this news-group about watching for edge and texture scintillation in EVF
pixels to help with manual focus? That can help immensely if depending on manual
focus alone. You can probably find the threads at a Google groups search with
the words "scintillation", and "kangaroo", and "orchids". As those were
mentioned in the discussion as subjects that are also hard to see unless you
train your eye to see them. Then once you do they are easy to spot.


If using CHDK you could always use one of those focus-bracketing scripts too (or
write one). Then let it take half a dozen or more images in micro focus
increments so that one will have to be in best focus. But I would only do this
if I needed to compare one camera against another and wanted to see which had
the best lens. Otherwise it would not be a fair test of hands-on performance,
nor will it teach you the best way to do it on your own when not running
focus-bracketing script.

But if you write a good camera-test CHDK script using a resolution chart I hope
you share it. A well-written CHDK script and one of those charts could
automatically test a camera. Using that uBASIC "print to textfile"
script-command it could also record what it was doing for each test shot.

Good luck with the tests. I hope you reach that moment quickly, where you say to
yourself, "This is one excellent camera, I still can't find any important
negative things about it. That's enough tests, it's a keeper!" I like when that
happens. My two favorites are like that.

bugbear wrote:

If I'm trying to test the limits of the resolution of the lens,
I need (AFAIK) "perfect" focusing, and I remain unclear
of the best way to get this, or failing perfection,
the best approach to perfection avaialable.

Shoot a brick wall at a slight diagonal, then you get the full range of
focus.