DLSR and Pixels
Eugene
DLSR is supposed to be good because the pixel
is bigger. In the Canon 300D. The pixel size is
7.4 micron, 6+ megapixels. Now with the Canon 1000D,
pixel size is 5.7 micron, 10+megapixel. Both has
the same CCD size of 22.5 x 15mm. Notice the pixel
size has gone smaller. Wouldn't this defeat the purpose
of DSLR (Bigger Pixel Rocks)? Has anyone compared
the quality of the 300D vs 1000D? Which is cleaner?
Using the same 22.5 x 15mm grid, how small can
the pixel get (or how large can the megapixels be,
as they are inversely proportional) before noise would become bad
enough that it won't be far from
point&shoot CCDs?
E
Mr. Strat
<4a28adb4-b6fd-4b79...@z8g2000prd.googlegroups.com>,
Eugene <eugen...@gmail.com> wrote:
Do you have any knowledge of or experience with photography?
It doesn't look like it.
Jürgen Exner
Actually no. dSLRs are good because you can customize them for your
personal needs with different lenses and other accessories.
Better light gathering because of larger pixels is a nice added bonus.
>In the Canon 300D. The pixel size is
>7.4 micron, 6+ megapixels. Now with the Canon 1000D,
>pixel size is 5.7 micron, 10+megapixel. Both has
>the same CCD size of 22.5 x 15mm. Notice the pixel
>size has gone smaller.
Surprise, surprise. It's called mathematics, you know.
>Wouldn't this defeat the purpose
>of DSLR (Bigger Pixel Rocks)?
Not at all, because larger pixel size is nice added bonus, nothing more.
jue
Derge
Well, yes and no. The point of DSLRs isn't that they have big pixels.
The point is that they have big sensors.
The notion of "pixel level" noise is a bit of a red herring, since
ultimately we, as human beings, need to view an image taken with any
camera at fixed dimensions, whether in print or on the web. The camera
usually doesn't determine what those dimensions are: We do! If you're
sending snapshots to your grandmother, for instance, they're only
going to be 800x600 JPEGs, no matter what camera took them. A
comparison at that size would give you similar pictures. If anything,
the images from the EOS 1000D would be *less* noisy *because* it has
smaller pixels, which means lower read noise per unit area.
You obviously understand the relationship between pixel size and
noise. The other edge of the sword is this: Anything "lost" by using
smaller pixels can be reclaimed by resizing the larger image in
Photoshop. This only works in one direction, obviously. Smaller pixels
give us a choice.
Eugene
> On Apr 22, 8:18 pm, Eugene <eugenhug...@gmail.com> wrote:
>
> > DLSR is supposed to be good because the pixel
> > is bigger. In the Canon 300D. The pixel size is
> > 7.4 micron, 6+ megapixels. Now with the Canon 1000D,
> > pixel size is 5.7 micron, 10+megapixel. Both has
> > the same CCD size of 22.5 x 15mm. Notice the pixel
> > size has gone smaller. Wouldn't this defeat the purpose
> > of DSLR (Bigger Pixel Rocks)? Has anyone compared
> > the quality of the 300D vs 1000D? Which is cleaner?
> > Using the same 22.5 x 15mm grid, how small can
> > the pixel get (or how large can the megapixels be,
> > as they are inversely proportional) before noise would become bad
> > enough that it won't be far from
> > point&shoot CCDs?
>
> > E
>
> Well, yes and no. The point of DSLRs isn't that they have big pixels.
> The point is that they have big sensors.
>
What? But pixels are sensors. What you talking about?
> The notion of "pixel level" noise is a bit of a red herring, since
> ultimately we, as human beings, need to view an image taken with any
> camera at fixed dimensions, whether in print or on the web. The camera
> usually doesn't determine what those dimensions are: We do! If you're
> sending snapshots to your grandmother, for instance, they're only
> going to be 800x600 JPEGs, no matter what camera took them. A
> comparison at that size would give you similar pictures. If anything,
> the images from the EOS 1000D would be *less* noisy *because* it has
> smaller pixels, which means lower read noise per unit area.
Or you mean bigger pixels? How can smaller pixels be less noisy?
E
David J Taylor
[]
> Or you mean bigger pixels? How can smaller pixels be less noisy?
>
> E
If you are viewing the complete image, is it better that it be made up of
5MP or 10MP, given that each of the 10MP pixels will be smaller, slightly
noisier, but that the 10MP image will be a little sharper?
Perhaps the optimum varies for different images, and some shots will
benefit more from the greater sharpness, and others from the lower noise?
Also, bear in mind that a more recent camera may have slight improvements,
resutling in a better signal-to-noise ratio for each pixel.
There isn't a single answer, you see.
David
Doug Jewell
>
> DLSR is supposed to be good because the pixel
> is bigger. In the Canon 300D. The pixel size is
> 7.4 micron, 6+ megapixels. Now with the Canon 1000D,
> pixel size is 5.7 micron, 10+megapixel. Both has
> the same CCD size of 22.5 x 15mm. Notice the pixel
> size has gone smaller. Wouldn't this defeat the purpose
> of DSLR (Bigger Pixel Rocks)? Has anyone compared
> the quality of the 300D vs 1000D? Which is cleaner?
noise. But like many things, all else isn't equal. Modern
cameras have improved sensor design, including less sensor
area dedicated to support electronics, so the actual
sensitive area hasn't shrunk by the same proportion that the
MP increase would indicate. Additionally support circuitry
has improved, post-processing (both in and out of camera)
has improved, so once again the performance doesn't scale
linearly with MP. While I haven't done a direct comparison,
and I have a 450D not a 1000D, the "vibe" I get from my 12MP
450D is that it is only slightly noisier than the older
lower MP cameras. In some cases maybe less noisy. The
increase in noise is also offset by a dramatic increase in
resolution. I'm confident that a 450D image downsampled to
6MP would be significantly better than a 300D image.
> Using the same 22.5 x 15mm grid, how small can
> the pixel get (or how large can the megapixels be,
> as they are inversely proportional) before noise would become bad
> enough that it won't be far from
> point&shoot CCDs?
have a pixel density of 43MP/cm^2. the 22.3x14.9mm sensor
used in the cropped canon cameras has an area of 3.3227cm^2,
so for the same pixel density an SLR could have a 142MP
sensor. ie, a 142MP 1.6x crop DSLR would deliver similar
noise levels to a 12MP P&S.
The current 10-15MP cameras are obviously a long long way
short of the pixel densities in P&S cameras, and despite
concern that the higher pixel counts are increasing noise,
they are still significantly better than P&S.
>
> E
Derge
>
> What? But pixels are sensors. What you talking about?
>
A sensor is a silicon chip. Its surface is divided into an array of
photosites. Each photosite contributes to the number of pixels.
>
> Or you mean bigger pixels? How can smaller pixels be less noisy?
>
What I said was that the *image* made *from* the smaller pixels would
be less noisy.
The EOS 1000D captures images at 3888x2592.
The EOS 300D captures images at 3072x2048.
If you resize an EOS 1000D image to 3072x2048, it will be less noisy
than the *same* image from an EOS 300D.
When you resize the image, the same principles that make the smaller
pixels noisier are working in reverse.
The fact is that the more samples you take, the less accurate any of
them need to be. This is why there is no "sweet spot" for APS-C
sensors. The only reason we don't have 400 megapixels DSLRs is because
our cameras and our personal computers aren't powerful enough yet. If
they were, we would.
Don Stauffer
never considered pixel size one of them. One can use ANY focal plane in
a DSLR, in a compact, a rangefinder, or any other type of camera.
To me the big advantages of an SLR are easily interchangable lenses, and
the ability to see the actual optical focus when using manual focus.
Also, for macro work there is no parallax error.
I do not consider electronic viewfinders acceptable until the display
has as many pixels as the primary sensor.
John McWilliams
> I do not consider electronic viewfinders acceptable until the display
> has as many pixels as the primary sensor.
That may be quite some time, or quite a large camera back.
What's the maximum number of ppi can an LCD properly display? Most
monitors are less than 100 ppi.
What can the human eye resolve- and when does extra resolution become
undetectable?
--
John McWilliams
John Passaneau
when I change something to when it changes on the LCD. Until LCD become
as fast as an optical viewfinder it's a no sale for me.
As it stands now, sports photography with electronic viewfinders would
be impossible.
John Passaneau
Chris Malcolm
> Eugene wrote:
>>
>> DLSR is supposed to be good because the pixel
>> is bigger. In the Canon 300D. The pixel size is
>> 7.4 micron, 6+ megapixels. Now with the Canon 1000D,
>> pixel size is 5.7 micron, 10+megapixel. Both has
>> the same CCD size of 22.5 x 15mm. Notice the pixel
>> size has gone smaller. Wouldn't this defeat the purpose
>> of DSLR (Bigger Pixel Rocks)? Has anyone compared
>> the quality of the 300D vs 1000D? Which is cleaner?
>> Using the same 22.5 x 15mm grid, how small can
>> the pixel get (or how large can the megapixels be,
>> as they are inversely proportional) before noise would become bad
>> enough that it won't be far from
>> point&shoot CCDs?
> The DSLR has a number of advantages over other types of cameras. I have
> never considered pixel size one of them. One can use ANY focal plane in
> a DSLR, in a compact, a rangefinder, or any other type of camera.
> To me the big advantages of an SLR are easily interchangable lenses, and
> the ability to see the actual optical focus when using manual focus.
> Also, for macro work there is no parallax error.
> I do not consider electronic viewfinders acceptable until the display
> has as many pixels as the primary sensor.
Why do you want it be so much better than an optical viewfinder, and
so much better than you can see?
--
Chris Malcolm
Don Stauffer
res of focal plane. I can see the most minute changes in the lens focus
on my D40. I can see MUCH more detail than I can on any LCD screen I
have used on other cameras, like my P & S.
I have not run the numbers on visual acuity, but I certainly can tell I
can see FAR better in my SLRs than I can on an LCD screen.
bugbear
>>
>>> I do not consider electronic viewfinders acceptable until the display
>>> has as many pixels as the primary sensor.
.
.
.
> I don't think I do. The optical viewfinder should be pretty close to
> res of focal plane. I can see the most minute changes in the lens focus
> on my D40. I can see MUCH more detail than I can on any LCD screen I
> have used on other cameras, like my P & S.
Yes - but that merely means you want/need more pixels
than present LCDs. It does not (neccessarily) mean
you need the same res as the sensor.
The latter would clearly be "sufficient", but it
the question is "is it neccessary".
BugBear
Don Stauffer
Some SLRs have optical focusing aids- microprisms and such. These
increase the accuracy of focus and rely on optical "tricks." I doubt if
an LCD viewfinder would be able to use such techniques.
J. Clarke
No, but it can magnify the image.
Bob Larter
It's enough of a pain to magnify an image you've just taken, the last
thing in the world that I want is to have to do it while focusing.
--
W
. | ,. w , "Some people are alive only because
\|/ \|/ it is illegal to kill them." Perna condita delenda est
---^----^---------------------------------------------------------------
Chris Malcolm
> J. Clarke wrote:
>> Don Stauffer wrote:
>>> bugbear wrote:
>>>> Don Stauffer wrote:
>>>>> Chris Malcolm wrote:
>>>>>> Don Stauffer <stau...@usfamily.net> wrote:
>>>>
>>>>>>
>>>>>>> I do not consider electronic viewfinders acceptable until the
>>>>>>> display has as many pixels as the primary sensor.
>>>>
>>>>> to res of focal plane. I can see the most minute changes in the lens
>>>>> focus on my D40. I can see MUCH more detail than I can on any LCD
>>>>> screen I have used on other cameras, like my P & S.
>>>>
>>>> Yes - but that merely means you want/need more pixels
>>>> than present LCDs. It does not (neccessarily) mean
>>>> you need the same res as the sensor.
>>>>
>>>> The latter would clearly be "sufficient", but it
>>>> the question is "is it neccessary".
>>>>
>>>> BugBear
>>> True. Probably half or even a quarter would do.
>>>
>>> Some SLRs have optical focusing aids- microprisms and such. These
>>> increase the accuracy of focus and rely on optical "tricks." I doubt
>>> if an LCD viewfinder would be able to use such techniques.
>>
>> No, but it can magnify the image.
> It's enough of a pain to magnify an image you've just taken, the last
> thing in the world that I want is to have to do it while focusing.
That's just an interface issue. Some cameras have the option of
instantly magnifying the image as soon as you move the manual focus
ring.
--
Chris Malcolm
J. Clarke
Yep, the problem with the ones I've seen is that they don't let you specify
what portion of the image is to be magnified, they just magnify the center.
Seems to me that done right it could be more convenient under some
circumstances than using an eyepiece magnifier.
Doug Jewell
image. This is the area that it will use for contrast-detect
auto-focus, or that it will enlarge for manual focus. You
can use a corner of the image for focus if you really want
to. It works ok, but personally I prefer the optical viewfinder.
Bob Larter
I'm sure that'd be fine if I was only interested in what was happening
in the center of the image, but I think I'll stick with my 45 AF points
on a nice sharp optical viewfinder.
J. Clarke
Hmm. I need to get down to B&H and play with cameras. Maybe when they've
got the 500d in stock . . .