Reading mode: Light Dark

Serious G1 Deficiency Overlooked by Product Reviews

Go to last post
And here's another just in case the foreground object in the previous example wasn't low enough contrast. :)

The dark lettering in the background is black. The lettering in the foreground is 25% gray.

This shot is focused on the background lettering:



This shot is focused on the foreground lettering:

 
Peter, did you do all this in low light with focus assist on? If you did, that is the reason your G1 focuses on the center since all else is dark. I'm confused about this focusing thing as well, though I tend to believe in Kevin's statement but not so sure about that either.

Michael
Peter T said:
And here's another just in case the foreground object in the
previous example wasn't low enough contrast. :)

The dark lettering in the background is black. The lettering in the
foreground is 25% gray.

This shot is focused on the background lettering:



This shot is focused on the foreground lettering:

Click to expand...
 
It's not dark. But it is dim enough for the focusing light to come one. Here's another test done in bright light with no focusing assist light.



 
The original assertion was that the focus was influenced not so much by a region of the frame but by the overall area of highest contrast over the entire frame.

In both test images, there are two possible primary focus planes -- the foreground sheet of paper, and the background paper+striped couch. There's also the area of the floor, which is in neither of the two planes.

I saved both images to disk and made a simple grayscale cutout of the three areas in each -- the area of one focus plane in white, the other in black, and the floor area neutral gray. I could then just look at a histogram of the image to see the relative number of pixels in each flat-colored region.

In both images, the plane of focus covers more of the image than the other plane. So this test, if anything, indicates that the original assertion may indeed be correct.

kb

Peter T posted:
Peter T44178 said:


Click to expand...
 
Canon has not, and probably will never, provide the ability for their digital cameras to be PIM enabled. Sony, Olympus, Minolta, Nikon etc. etc. are all involved in this new technology that, according to the August 2001 issue of PC magazine test (page 50) makes a VERY significant improvement in prints that are outputted by the PIM enabled Epson printers. Sony has provided this capability to cameras ranging from the P50 to the S75/S85 with Nikon providing it for the 995 and Olympus to their tested camera as well. Canon probably never will because it would steer consumers to Epson's PIM enabled printers and away from Canon's own printers.
 
Bob,

I always use manual focus when shooting macros (even with my SLR's). It is the only way in which to control the desired image. Admittedly, it is more difficult to do with the G1 because of the lack of DOF preview but the zoom trick is definitely useful to ensure focus.
Bob Williams said:


If you look at the red rose shot I posted for Diane, you will see the
same effect there, but that image had a mild unsharp mask applied.
Bob H said:
Don T said:
I am embarrassed to admit that after owning my G1 or more than six
months it is only now that I have come to realize the G1’s
autofocus deficiency. This is no minor problem, unlike the magenta
overcast, which is easily fixed with an editing program. There is
no easy fix for a camera that has a very weak autofocus system.

Hi Don! Please forgive me if I am speaking "out of school" here, I am relatively new to photography (digital or otherwise) and may not really understand your concerns.
Click to expand...
Some previewers of the G1 have pointed out the problems with it's
metering system and Canon should address those concerns in future
upgrades. However, the G1 can now be had for under $600 US dollars
and I suspect under $500 very soon. Digital camera technology is
expanding almost exponentially it seems and yesterday's wonder
camera is viewed as mediocre today.
All that said, I think the G1 is an excellent camera for the price;
even with it's inherent flaws (chromatic aberration included).
I guess my point is, (and maybe it's because I havn't developed
enough of a discrimenating eye yet), but this may be "plenty" of
camera for those like myself and as the price continues to drop, it
may seem even more attractive in the future.
Click to expand...
Click to expand...
 
How about this test?




Kevin Bjorke said:
The original assertion was that the focus was influenced not so
much by a region of the frame but by the overall area of highest
contrast over the entire frame.

In both test images, there are two possible primary focus planes --
the foreground sheet of paper, and the background paper+striped
couch. There's also the area of the floor, which is in neither of
the two planes.

I saved both images to disk and made a simple grayscale cutout of
the three areas in each -- the area of one focus plane in white,
the other in black, and the floor area neutral gray. I could then
just look at a histogram of the image to see the relative number of
pixels in each flat-colored region.

In both images, the plane of focus covers more of the image than
the other plane. So this test, if anything, indicates that the
original assertion may indeed be correct.
Click to expand...
 
Bryan,

Please help me understand what you were trying to say? The rosebud actually displays very little contrast in comparison to the leaves. That is the reason the camera searched and focused in on the part of the scene with greatest contrast (leaves).

Recall, that contrast merely refers to the tonal gradation between highlights and shadows (this is not the same thing as brightness). The camera uses an algorithm to compare the pixel valuations in a region and calculates the contrast. That region with the greatest gradation becomes the area of focus.

Try this little experiment. Place a blank sheet of paper a few feet in front of a contrasty scene (a bookcase for example). Shot a photo and tell me what the likely area of focus will be?

If your G1 can focus in on the piece of paper, please send it to me. It must have some other sort of focusing system other than the contrast based one I own.
Bryan Siverly said:
Bob Williams said:
I will shoot some good examples tomorrow.
I can't believe that so many people are arguing whether or not
this happens. I would have shot examples toady, but I had not
read these answers yet.
Click to expand...
Gee, I can hardly wait to see more of your "samples." One wonders
why your G1 didn't focus on the most contrasty part of the scene...
the bud in the background, if it behaves the way you suggest. In
fact, the leaf in the foreground had to be the lowest contrast area
of the photo.

Bryan
Click to expand...
 
Bob Williams wrote:
. It is hard to understand also because
Member said:
it handles other reds very well in some cases, like the water slide
shot I posted. The focus in this shot applies to the other problem,
not to the red issue
Click to expand...
I wonder if it has something to do with the texture of the roses/flowers?? Don't have a clue about this, but they are more like 'textiles' in some ways and textiles 'suck' light and saturated colors are tough for even a pro to get right. This is just totally off my head, but roses are so velvetty. I'll be interested in seeing your final solution for the roses.

Diane
 
What you say is correct. But the autofocus sensor doesn't cover the entire image.

I suggest you modify your experiment. Place the blank piece of paper at the same plane as the contrasty scene. Make sure the center of the paper is in the center of the image. And place the camera 3 feet from the plane of focus. You will not get a focused image.
Adam said:
Bryan,

Please help me understand what you were trying to say? The rosebud
actually displays very little contrast in comparison to the leaves.
That is the reason the camera searched and focused in on the part
of the scene with greatest contrast (leaves).

Recall, that contrast merely refers to the tonal gradation between
highlights and shadows (this is not the same thing as brightness).
The camera uses an algorithm to compare the pixel valuations in a
region and calculates the contrast. That region with the greatest
gradation becomes the area of focus.

Try this little experiment. Place a blank sheet of paper a few
feet in front of a contrasty scene (a bookcase for example). Shot
a photo and tell me what the likely area of focus will be?

If your G1 can focus in on the piece of paper, please send it to
me. It must have some other sort of focusing system other than the
contrast based one I own.
Click to expand...
 
I forgot to mention to do this in non-macro mode. If you want to do it in macro mode make sure the plane of focus is less than 1 foot. If the camera can't find focus it focuses long, so you need to eliminate that variable.
 
Peter,

Thanks for doing this. Could you print a few of the variables so that we can understand the test better (distances, settings, etc.) Also, it strikes me that the two printed pieces of paper are different. If this is true, then it would necessarily affect the test results.
Peter T said:



Kevin Bjorke said:
The original assertion was that the focus was influenced not so
much by a region of the frame but by the overall area of highest
contrast over the entire frame.

In both test images, there are two possible primary focus planes --
the foreground sheet of paper, and the background paper+striped
couch. There's also the area of the floor, which is in neither of
the two planes.

I saved both images to disk and made a simple grayscale cutout of
the three areas in each -- the area of one focus plane in white,
the other in black, and the floor area neutral gray. I could then
just look at a histogram of the image to see the relative number of
pixels in each flat-colored region.

In both images, the plane of focus covers more of the image than
the other plane. So this test, if anything, indicates that the
original assertion may indeed be correct.
Click to expand...
Click to expand...
 
Regarding powerpoint presentations:

You're not really limited by the size of the screen, but the ability of your projector. Many projectors top out at 1024X768, but in any case, the maximum you're going to get with a high-end projector is something similar to what you'd get with a large computer monitor and a good video card -- use that as your basis -- then, use the appropriate sized screen for your audience (size, distance from screen, etc.). I have projected digital images on much bigger screens than that, and they look fine if the projector is good. Just check the resolution and color depth of your projector.

Jeff
Basia said:
Greetings.
Click to expand...
Basia said:
The main use will be in powerpoint presentations. Is the resolution
good enough to blow up on a 5X7 foot screen
Click to expand...
 
Both pieces of paper are the same. What difference do you think it would make if they weren't the same?

Here's EXIF data for shot1:

ExposureTime : 1.00Sec
FNumber : F2.5
ExifVersion : 0210
DateTimeOriginal : 2001:08:08 07:13:47
DateTimeDigitized : 2001:08:08 07:13:47
ComponentConfiguration : YCbCr
CompressedBitsPerPixel : 5/1 (bit/pixel)
ShutterSpeedValue : 1.00Sec
ApertureValue : F2.5
ExposureBiasValue : EV0.7
MaxApertureValue : F2.0
SubjectDistance : 0.80(m)
MeteringMode : CenterWeightedAverage
Flash : Not fire
FocalLength : 14.59(mm)

Shot 2:

ExposureTime : 1/1Sec
FNumber : F2.5
ExifVersion : 0210
DateTimeOriginal : 2001:08:08 07:14:21
DateTimeDigitized : 2001:08:08 07:14:21
ComponentConfiguration : YCbCr
CompressedBitsPerPixel : 5/1 (bit/pixel)
ShutterSpeedValue : 1/1Sec
ApertureValue : F2.5
ExposureBiasValue : EV0.7
MaxApertureValue : F2.0
SubjectDistance : 0.40(m)
MeteringMode : CenterWeightedAverage
Flash : Not fire
FocalLength : 14.59(mm)
Adam said:
Peter,

Thanks for doing this. Could you print a few of the variables so
that we can understand the test better (distances, settings, etc.)
Also, it strikes me that the two printed pieces of paper are
different. If this is true, then it would necessarily affect the
test results.
Click to expand...
 
Jim Rose said:
Canon probably never will because it would steer consumers
to Epson's PIM enabled printers and away from Canon's own printers.
Click to expand...
PIM is one gimmick that I don't get. Why would I want to rely on some technology that attempts calibration for me?

It's not that complicated a task to print a couple of photos and then to calibrate your monitor to match. And I've never had a situation where my HP printer failed to properly render the colors of my photos.

How in the world did we get by without PIM for so long? Gee, what a struggle it's been. :)

Bryan
 
I don't have a photo printer yet. Since I'm technically inclined I'm sure I could handle the calibration. But the impression I have is that for a lot of people it is quite a struggle and PIM will be good for them.

And if I remember correctly the article said the results were better than what they got with calibration.
Bryan Siverly said:
How in the world did we get by without PIM for so long? Gee, what
a struggle it's been. :)
Click to expand...
 
The reason I asked this is that if the two pieces of paper are of similar contrast, the results would be dependent on the algorithm of the camera as to how it deals with items of similar contrast at different distances. Perhaps a better test would be to place a large item of low contrast in the background and move a contrasty object around the frame (center, sides, etc). See what the camera focuses on and let us know.
Peter T said:
Here's EXIF data for shot1:

ExposureTime : 1.00Sec
FNumber : F2.5
ExifVersion : 0210
DateTimeOriginal : 2001:08:08 07:13:47
DateTimeDigitized : 2001:08:08 07:13:47
ComponentConfiguration : YCbCr
CompressedBitsPerPixel : 5/1 (bit/pixel)
ShutterSpeedValue : 1.00Sec
ApertureValue : F2.5
ExposureBiasValue : EV0.7
MaxApertureValue : F2.0
SubjectDistance : 0.80(m)
MeteringMode : CenterWeightedAverage
Flash : Not fire
FocalLength : 14.59(mm)

Shot 2:

ExposureTime : 1/1Sec
FNumber : F2.5
ExifVersion : 0210
DateTimeOriginal : 2001:08:08 07:14:21
DateTimeDigitized : 2001:08:08 07:14:21
ComponentConfiguration : YCbCr
CompressedBitsPerPixel : 5/1 (bit/pixel)
ShutterSpeedValue : 1/1Sec
ApertureValue : F2.5
ExposureBiasValue : EV0.7
MaxApertureValue : F2.0
SubjectDistance : 0.40(m)
MeteringMode : CenterWeightedAverage
Flash : Not fire
FocalLength : 14.59(mm)
Adam said:
Peter,

Thanks for doing this. Could you print a few of the variables so
that we can understand the test better (distances, settings, etc.)
Also, it strikes me that the two printed pieces of paper are
different. If this is true, then it would necessarily affect the
test results.
Click to expand...
Click to expand...
 
Name calling and insults are the necessary tools of those who cannot express their opinion either intelligently or convincingly enough otherwise. This clearly demonstrates a certain mean-spiritedness and perhaps repressed anger.
Neal said:
My advise: get glasses or just shut-up and live with it. You are
the one with the problem. Don't bother us geniuses with your
whining and complaints about an innocent piece of metal. Boy, what
a 'tard you are.
Don T said:
I am embarrassed to admit that after owning my G1 or more than six
months it is only now that I have come to realize the G1’s
autofocus deficiency. This is no minor problem, unlike the magenta
overcast, which is easily fixed with an editing program. There is
no easy fix for a camera that has a very weak autofocus system.

Rather than explain the problem in my own words, please read the
exert that I have extracted from Kevin Bjorkes discussion on G1
autofocus (botzilla.com):
Autofocus
“The G1 appears to use a contrast-based AF system. It also
appears to use the entire frame to measure contrast. In other
words, the contrast of pixels along the edges and in the corners
are given just as much importance by the AF system as pixels near
the center.
This uncentered approach has serious consequences for autofocus use
on real scenes. While it's fine for landscapes and may even do well
for group shots, what about a head-and-shoulders portrait?

Canon's G1 manual mentions "autofocus crosshairs" on page 38, but
this is simply wrong they're in the optical finder and clearly have
nothing to do with the Autofocus. What's more, they may give you
the impression that the AF is concentrating on the area of the
crosshairs (like the "target marks" did on the previous Canon
design, the Pro 70) and that's simply not true

The Pro 90 also has a "target bracket" area, which might actually
do something (it's part of the EVF and LCD, not the optical
finder). But both the G1 and Pro 90 manuals claim weakness to
"Subjects with extremely low contrast to the surroundings." Even
the A20 has AF brackets why does the G1 have this crippled system?
Is it some failed prototype of the A20's AiAF ?”

If Kevin is correct in his evaluation of the G1 system, and I for
one have no reason to doubt him, the G1 has one major flaw that has
not been picked up by the various experts who have published
extensive reviews of the G1.
Had I known this at the time I was ready to buy I would have bought
a different camera. (Probably a Nikon 990/995)
One might ask why didn’t I notice this “flaw”
earlier? I have taken many in-focus pictures with the G1 because
the autofocus works adequately for many conditions considering the
very generous depth of fields associated with short focal length
lenses. In fact the focal length of most digital camera lenses are
so short, you could get by with fixed focus for many of your
routine shots. It is in those difficult shots where you most need
the autofocus as for example macros or views with both foreground
and distant objects in the field of view.
There are workarounds using contrast targets or switching between
digital zoom and manual focus (see discussion on botzilla.com ) but
these shouldn’t be needed with a camera of this quality.

Conclusion:
For most of us who have invested so much in the G1 it is a little
too late to do much about it. We will just have to learn how to
live with it and hope Canon will fix the problem on a subsequent
update. I doubt if it can be fixed without some hardware changes.
First Canon has to acknowledge that a problem even exists.
For those considering a G1, give this some thought before you buy.
There are cameras with better auto focus systems.
Don T

--
Donald T
Click to expand...
Click to expand...
 
Ted B. said:
Name calling and insults are the necessary tools of those who
cannot express their opinion either intelligently or convincingly
enough otherwise. This clearly demonstrates a certain
mean-spiritedness and perhaps repressed anger.
Click to expand...
Ted, I think this is a tongue and cheek satirical comment. He rerfers to himself and all the other G1 advocates as " geniuses".

Anyway I took it to be kidding about anyone who has the nerve to complain about a camera that is revered by so many.

Just ignore it. I hope my post has value in getting people to think about the subject. As of the moment there are about 100 replies which is a lot.

Don T
Ted B. said:
Neal said:
My advise: get glasses or just shut-up and live with it. You are
the one with the problem. Don't bother us geniuses with your
whining and complaints about an innocent piece of metal. Boy, what
a 'tard you are.
Don T said:
I am embarrassed to admit that after owning my G1 or more than six
months it is only now that I have come to realize the G1’s
autofocus deficiency. This is no minor problem, unlike the magenta
overcast, which is easily fixed with an editing program. There is
no easy fix for a camera that has a very weak autofocus system.

Rather than explain the problem in my own words, please read the
exert that I have extracted from Kevin Bjorkes discussion on G1
autofocus (botzilla.com):
Autofocus
“The G1 appears to use a contrast-based AF system. It also
appears to use the entire frame to measure contrast. In other
words, the contrast of pixels along the edges and in the corners
are given just as much importance by the AF system as pixels near
the center.
This uncentered approach has serious consequences for autofocus use
on real scenes. While it's fine for landscapes and may even do well
for group shots, what about a head-and-shoulders portrait?

Canon's G1 manual mentions "autofocus crosshairs" on page 38, but
this is simply wrong they're in the optical finder and clearly have
nothing to do with the Autofocus. What's more, they may give you
the impression that the AF is concentrating on the area of the
crosshairs (like the "target marks" did on the previous Canon
design, the Pro 70) and that's simply not true

The Pro 90 also has a "target bracket" area, which might actually
do something (it's part of the EVF and LCD, not the optical
finder). But both the G1 and Pro 90 manuals claim weakness to
"Subjects with extremely low contrast to the surroundings." Even
the A20 has AF brackets why does the G1 have this crippled system?
Is it some failed prototype of the A20's AiAF ?”

If Kevin is correct in his evaluation of the G1 system, and I for
one have no reason to doubt him, the G1 has one major flaw that has
not been picked up by the various experts who have published
extensive reviews of the G1.
Had I known this at the time I was ready to buy I would have bought
a different camera. (Probably a Nikon 990/995)
One might ask why didn’t I notice this “flaw”
earlier? I have taken many in-focus pictures with the G1 because
the autofocus works adequately for many conditions considering the
very generous depth of fields associated with short focal length
lenses. In fact the focal length of most digital camera lenses are
so short, you could get by with fixed focus for many of your
routine shots. It is in those difficult shots where you most need
the autofocus as for example macros or views with both foreground
and distant objects in the field of view.
There are workarounds using contrast targets or switching between
digital zoom and manual focus (see discussion on botzilla.com ) but
these shouldn’t be needed with a camera of this quality.

Conclusion:
For most of us who have invested so much in the G1 it is a little
too late to do much about it. We will just have to learn how to
live with it and hope Canon will fix the problem on a subsequent
update. I doubt if it can be fixed without some hardware changes.
First Canon has to acknowledge that a problem even exists.
For those considering a G1, give this some thought before you buy.
There are cameras with better auto focus systems.
Don T

--
Donald T
Click to expand...
Click to expand...
Click to expand...
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top