New Nikon J1/V1 sensors = half the surface area of micro 4/3rds!

Andrew Reilly wrote:
On Fri, 07 Oct 2011 02:33:28 -0400, Neil Harrington wrote:

One other interesting thing about the Coolpix 5700 that I haven't
seen anywhere else is that it doesn't use a Bayer-pattern sensor,
but instead a four-colour pattern.

That's interesting; I wasn't aware of that.

I recall reading years ago that Sony was developing a four-color
sensor, red/green/emerald/blue instead of RGGB. I wasn't aware that
it was ever actually used in a production camera. Maybe that's what
your 5700 uses?

No, NEF metadata (and the web sites) all claim that it is a CYGM
(Cyan, Yellow, Green, Magenta) pattern, and apparently there were

Wow, that's weird. Very interesting. I've always assumed the Bayer pattern
was pretty much universal.

several others from Sony and Canon that used similar, around that
time. According to Wikipedia
(http://en.wikipedia.org/wiki/CYGM_filter) the advantage is lower
luminance noise at the expense of colour accuracy. Irrespective of
that, I always liked its colour rendition.

I don't know why such things haven't been tried elsewhere, other
than that there was no particular advantage. I've always been very
impressed with the quality of the colours in its JPEGs (shooting
RAW/NEF was never a good idea: wait several minutes between shots
for the file to be stored...)

That's interesting too. I've never used NEF in such cameras myself.

It's not something that one does on purpose, after the first time...
My wife has selected it by accident a couple of times, while scrolling
through the image size options. The apparent result is that the
camera locks up completely and is handed to me with a comment along
the lines of "honey, I think I've killed it again: can you fix it?".

:-)


That sensor did spontaneously die in fairly spectacular fashion,
while on holiday once. Seemed as though every second and then fourth
and then... sensor row gave up, and images were produced that were
under exposed and weirdly stripy (and then stopped altogether, of
course.) Luckily it was a known problem and the subject of a Nikon
product recall, so they replaced it gratis.

I've heard that Nikon is very good in that way. Never yet had any Nikon
problems to repair, myself.

nospam wrote:
In article , Neil
Harrington wrote:

One other interesting thing about the Coolpix 5700 that I haven't
seen anywhere else is that it doesn't use a Bayer-pattern sensor,
but instead a four-colour pattern.

That's interesting; I wasn't aware of that.

it's still a bayer sensor.

I recall reading years ago that Sony was developing a four-color
sensor, red/green/emerald/blue instead of RGGB. I wasn't aware that
it was ever actually used in a production camera. Maybe that's what
your 5700 uses?

there are several variants, including two shades of g, rgbe, cmyg and
rgbw (white). sony did use rgbe a while back.

Thanks for the info. Obviously I didn't know as much about this as I thought
I did. :-)

Bruce wrote:
"Neil Harrington" wrote:
Bruce wrote:
The reason why the used GF1 is in demand is because its user
interface appeals to DSLR users whereas the GF2 and GF3 have very
little appeal to that market. I used a GF1 for a few months and
quite liked it, but what I really want to supplement my DSLRs is a
Leica M9-P, and the GF1 fell a little short. ;-)

Used GH1 bodies are also in demand because their firmware can be
hacked (in a similar way to CHDK) to improve some aspects of its
performance.

I've seen that mentioned, though I don't know the details. And
somewhere I've seen a hacked GH1 offered for sale.


The principal advantage of the hack is that improves the GH1's video
performance far beyond the camera's standard settings:
http://preview.tinyurl.com/2a92hok
or:
http://www.engadget.com/2010/06/16/p...-high-quality/

A search on "Panasonic GH1 Tester 13" will reveal a wealth of
information.

Very interesting, thanks.



I bought a job lot of GH1 kits with the 14-140mm lens,
split them and sold them separately at a very large profit margin.
My only regret is that I didn't wait until now, when brand new GH1
bodies (with a warranty) would be worth even more.

That's interesting. The price benefit doesn't seem to apply on this
side of the pond. I noticed a while ago that GH1 bodies had come
down in price, and checking Amazon just now I see a couple of
marketplace sellers have them at less than $570 including shipping.
If I'm not mistaken they had been over $1,000.
http://www.amazon.com/gp/offer-listi...&condition=new
I'm tempted.


If you shoot video and are prepared to hack the firmware to release
the camera's full potential, $570 is a steal. Sadly, Panasonic closed
the loophole and the GH2's firmware cannot be hacked.

Checking Amazon just now I see that some buyer reviews of this camera (at
the low price) are complaining that the camera is a grey market item, i.e.
made for the European market and not supported, guarantee-wise, by Panasonic
in the U.S.

Other sellers are selling the GH1 at about $1350 to $1680, presumably
officially imported ones. None of these are sold by Amazon, but other
sellers "fulfilled by Amazon" or whatever.

Others are selling the GH2 from $948 up.

Neil Harrington wrote:
Wolfgang Weisselberg wrote:

Canon does have cross sensors that are capable of f/8, but work
with increased quality at f/2.8 or faster ... and since they
work the same way, this means it's not physically impossible to
build split-image focus helpers that can do both --- i.e. without
sacrificing focussing accuracy at larger apertures.

Not a *single* split-image rangefinder prism, though.

There are varifocal spectacles, so why should that be impossible?

I've never heard of such an arrangement, though. The
Canon system you mention of course will only give f/8 accuracy at anything
less than f/2.8.

Still enough to stay within DOF on the target.

Since a split-image rangefinder prism works by using the distance between
opposite sides of the aperture as a rangefinder base, if you don't have the
required aperture (whatever that may be) you just don't get any rangefinding
at all -- either the top or the bottom will go black. That's why TSANSTAAFL.

You assume it *has* to be two simple prisms per rangefinder.
There are, for example, quad-rangefinders (for horizontal and
vertical) available commercially. I do not see that it *has*
to be impossible to make one that's two-step or even gradual in
their accuracy.

-Wolfgang

Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:

Canon does have cross sensors that are capable of f/8, but work
with increased quality at f/2.8 or faster ... and since they
work the same way, this means it's not physically impossible to
build split-image focus helpers that can do both --- i.e. without
sacrificing focussing accuracy at larger apertures.

Not a *single* split-image rangefinder prism, though.

There are varifocal spectacles, so why should that be impossible?

Progressives. How could they have anything to do with this?


I've never heard of such an arrangement, though. The
Canon system you mention of course will only give f/8 accuracy at
anything less than f/2.8.

Still enough to stay within DOF on the target.

Maybe, maybe not. If you really needed that feature at f/2.8 you'd need it
at f/3 as well, and presumably wouldn't get it -- you'd be back to f/8
focusing accuracy, I should think.


Since a split-image rangefinder prism works by using the distance
between opposite sides of the aperture as a rangefinder base, if you
don't have the required aperture (whatever that may be) you just
don't get any rangefinding at all -- either the top or the bottom
will go black. That's why TSANSTAAFL.

You assume it *has* to be two simple prisms per rangefinder.
There are, for example, quad-rangefinders (for horizontal and
vertical) available commercially. I do not see that it *has*
to be impossible to make one that's two-step or even gradual in
their accuracy.

I'd like to see how they would do that.

Neil Harrington wrote:
Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:

Canon does have cross sensors that are capable of f/8, but work
with increased quality at f/2.8 or faster ... and since they
work the same way, this means it's not physically impossible to
build split-image focus helpers that can do both --- i.e. without
sacrificing focussing accuracy at larger apertures.

Not a *single* split-image rangefinder prism, though.

There are varifocal spectacles, so why should that be impossible?

Progressives. How could they have anything to do with this?

They demonstrate that such things are possible.


I've never heard of such an arrangement, though. The
Canon system you mention of course will only give f/8 accuracy at
anything less than f/2.8.

Still enough to stay within DOF on the target.

Maybe, maybe not. If you really needed that feature at f/2.8 you'd need it
at f/3 as well,

You've got an f/3 lens for your camera? Or does your AF not
focus at wide open?

and presumably wouldn't get it -- you'd be back to f/8
focusing accuracy, I should think.

The f/8 focussing accuracy is enough to place the plane of
focus so that the focussed object is within the DOF.
That's the spec Canon says the AF will bring. And with f/2.8
or faster it'll be within 1/3rd of the DOF.


Since a split-image rangefinder prism works by using the distance
between opposite sides of the aperture as a rangefinder base, if you
don't have the required aperture (whatever that may be) you just
don't get any rangefinding at all -- either the top or the bottom
will go black. That's why TSANSTAAFL.

http://www.katzeyeoptics.com/page--K...lus--plus.html

You assume it *has* to be two simple prisms per rangefinder.
There are, for example, quad-rangefinders (for horizontal and
vertical) available commercially. I do not see that it *has*
to be impossible to make one that's two-step or even gradual in
their accuracy.

I'd like to see how they would do that.

And would you belive them if they told you they did?

-Wolfgang

Bruce wrote:
Bruce wrote:
If you shoot video and are prepared to hack the firmware to release
the camera's full potential, $570 is a steal. Sadly, Panasonic
closed the loophole and the GH2's firmware cannot be hacked.

Checking Amazon just now I see that some buyer reviews of this
camera (at the low price) are complaining that the camera is a grey
market item, i.e. made for the European market and not supported,
guarantee-wise, by Panasonic in the U.S.

Other sellers are selling the GH1 at about $1350 to $1680,
presumably officially imported ones. None of these are sold by
Amazon, but other sellers "fulfilled by Amazon" or whatever.

Others are selling the GH2 from $948 up.


As I said, it's a steal at $570, but at $948 you would have to be
quite keen on videography. It's something I might take an interest in
when I retire, but that's nine years away. ;-)

Yes. I have little interest in it myself.

I am perfectly happy with my G1 and G2 anyway.

Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:

Canon does have cross sensors that are capable of f/8, but work
with increased quality at f/2.8 or faster ... and since they
work the same way, this means it's not physically impossible to
build split-image focus helpers that can do both --- i.e. without
sacrificing focussing accuracy at larger apertures.

Not a *single* split-image rangefinder prism, though.

There are varifocal spectacles, so why should that be impossible?

Progressives. How could they have anything to do with this?

They demonstrate that such things are possible.

Why would there be any question about that? And what does that have to do
with split-image rangefinder prisms?

A progressive lens for eyeglasses is not varifocal in the ordinary meaning
of that word. At any point on the lens it is single focal length, and that
focal length does not vary. You could make a wedge prism progressive in the
same way, probably, but I don't see how you could make it do what you're
asking it to do.



I've never heard of such an arrangement, though. The
Canon system you mention of course will only give f/8 accuracy at
anything less than f/2.8.

Still enough to stay within DOF on the target.

Maybe, maybe not. If you really needed that feature at f/2.8 you'd
need it at f/3 as well,

You've got an f/3 lens for your camera? Or does your AF not
focus at wide open?

I have one f/2.8-4 zoom lens, yes. At some focal length it is f/3,
obviously.


and presumably wouldn't get it -- you'd be back to f/8
focusing accuracy, I should think.

The f/8 focussing accuracy is enough to place the plane of
focus so that the focussed object is within the DOF.
That's the spec Canon says the AF will bring. And with f/2.8
or faster it'll be within 1/3rd of the DOF.


Since a split-image rangefinder prism works by using the distance
between opposite sides of the aperture as a rangefinder base, if
you don't have the required aperture (whatever that may be) you
just don't get any rangefinding at all -- either the top or the
bottom will go black. That's why TSANSTAAFL.

http://www.katzeyeoptics.com/page--K...lus--plus.html

You assume it *has* to be two simple prisms per rangefinder.
There are, for example, quad-rangefinders (for horizontal and
vertical) available commercially. I do not see that it *has*
to be impossible to make one that's two-step or even gradual in
their accuracy.

I'd like to see how they would do that.

And would you belive them if they told you they did?

Probably, if they explained how they did it. The Katzeye site you've given
above makes no such explanation, nor does it make the claim that you're
implying. It looks as though they are just using shallow prisms to reduce
the rangefinder base (i.e., that work at a smaller aperture) and thus, of
course, lose some rangefinder accuracy. If I'm missing something in their
description, please point it out to me.

Neil Harrington wrote:
Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:

Canon does have cross sensors that are capable of f/8, but work
with increased quality at f/2.8 or faster ... and since they
work the same way, this means it's not physically impossible to
build split-image focus helpers that can do both --- i.e. without
sacrificing focussing accuracy at larger apertures.

Not a *single* split-image rangefinder prism, though.

There are varifocal spectacles, so why should that be impossible?

Progressives. How could they have anything to do with this?

They demonstrate that such things are possible.

Why would there be any question about that? And what does that have to do
with split-image rangefinder prisms?

If progressives are possible, variable opening aperture
spit-image rangefinders are probably possible, too.


A progressive lens for eyeglasses is not varifocal in the ordinary meaning
of that word. At any point on the lens it is single focal length, and that
focal length does not vary. You could make a wedge prism progressive in the
same way, probably, but I don't see how you could make it do what you're
asking it to do.

So the range finder would have a single max opening at every
point, too, but that could vary over the prism.


I've never heard of such an arrangement, though. The
Canon system you mention of course will only give f/8 accuracy at
anything less than f/2.8.

Still enough to stay within DOF on the target.

Maybe, maybe not. If you really needed that feature at f/2.8 you'd
need it at f/3 as well,

You've got an f/3 lens for your camera? Or does your AF not
focus at wide open?

I have one f/2.8-4 zoom lens, yes. At some focal length it is f/3,
obviously.

Not necessarily. The lens could be built to jump the
aperture at certain focal lengts by actuating the aperture a
bit when turning the focus ring.

Anyway, you're arguing you need that feature. You don't.
Not at f/2.8. But it helps at f/1.4.


You assume it *has* to be two simple prisms per rangefinder.
There are, for example, quad-rangefinders (for horizontal and
vertical) available commercially. I do not see that it *has*
to be impossible to make one that's two-step or even gradual in
their accuracy.

I'd like to see how they would do that.

And would you belive them if they told you they did?

Probably, if they explained how they did it. The Katzeye site you've given
above makes no such explanation,

99% of the buyers wouldn't be able to follow such an explanation,
(many probably don't even grasp how a split prism really works ---
and they don't care!) no matter how simple it was, and I guess
it's not simple.

nor does it make the claim that you're
implying.

What part of
| The Katz Eye™ “Plus� series of focusing screens has a unique
| split prism design which overcomes the "blackout" problem,
| without sacrificing large aperture sensitivity.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
didn't you find?


It looks as though they are just using shallow prisms to reduce
the rangefinder base (i.e., that work at a smaller aperture) and thus, of
course, lose some rangefinder accuracy.

Not according to their claim.

If I'm missing something in their
description, please point it out to me.

See above.

-Wolfgang

Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:
Neil Harrington wrote:
Wolfgang Weisselberg wrote:

Canon does have cross sensors that are capable of f/8, but work
with increased quality at f/2.8 or faster ... and since they
work the same way, this means it's not physically impossible to
build split-image focus helpers that can do both --- i.e.
without sacrificing focussing accuracy at larger apertures.

Not a *single* split-image rangefinder prism, though.

There are varifocal spectacles, so why should that be impossible?

Progressives. How could they have anything to do with this?

They demonstrate that such things are possible.

Why would there be any question about that? And what does that have
to do with split-image rangefinder prisms?

If progressives are possible, variable opening aperture
spit-image rangefinders are probably possible, too.

It's an interesting idea, but I'd sure like to see it diagrammed.



A progressive lens for eyeglasses is not varifocal in the ordinary
meaning of that word. At any point on the lens it is single focal
length, and that focal length does not vary. You could make a wedge
prism progressive in the same way, probably, but I don't see how you
could make it do what you're asking it to do.

So the range finder would have a single max opening at every
point, too, but that could vary over the prism.

I'd still like to see that diagrammed.



I've never heard of such an arrangement, though. The
Canon system you mention of course will only give f/8 accuracy at
anything less than f/2.8.

Still enough to stay within DOF on the target.

Maybe, maybe not. If you really needed that feature at f/2.8 you'd
need it at f/3 as well,

You've got an f/3 lens for your camera? Or does your AF not
focus at wide open?

I have one f/2.8-4 zoom lens, yes. At some focal length it is f/3,
obviously.

Not necessarily. The lens could be built to jump the
aperture at certain focal lengts by actuating the aperture a
bit when turning the focus ring.

guffaw!
"jump the aperture"?! Come on.


Anyway, you're arguing you need that feature. You don't.
Not at f/2.8. But it helps at f/1.4.


You assume it *has* to be two simple prisms per rangefinder.
There are, for example, quad-rangefinders (for horizontal and
vertical) available commercially. I do not see that it *has*
to be impossible to make one that's two-step or even gradual in
their accuracy.

I'd like to see how they would do that.

And would you belive them if they told you they did?

Probably, if they explained how they did it. The Katzeye site you've
given above makes no such explanation,

99% of the buyers wouldn't be able to follow such an explanation,
(many probably don't even grasp how a split prism really works ---
and they don't care!) no matter how simple it was, and I guess
it's not simple.

It sure doesn't look simple.


nor does it make the claim that you're
implying.

What part of
The Katz EyeT "Plus" series of focusing screens has a unique
split prism design which overcomes the "blackout" problem,
without sacrificing large aperture sensitivity.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
didn't you find?

I don't know, it was about a couple of weeks ago. I've read a million things
since then.



It looks as though they are just using shallow prisms to reduce
the rangefinder base (i.e., that work at a smaller aperture) and
thus, of course, lose some rangefinder accuracy.

Not according to their claim.

If I'm missing something in their
description, please point it out to me.

See above.

A split-image rangefinder prism works by using the distance between opposite
sides of the available aperture as the rangefinder base. In order to do what
is claimed above for KatzEye's product, it would have to be able to this at
different maximum apertures, such that even the much shorter RF base of,
say, f/9 would be usable (albeit at reduced RF accuracy of course) without
blacking out either half. That seems to imply that the RF prisms could not
be the usual shallow angled surfaces, but at least one of the long surfaces
would have to be curved in some way rather than flat.

How could this be done and still keep the image parts within the RF section
sharp all over, as they normally are with such a device? I'm trying to
imagine what such a progressive "prism" might look like in the viewfinder.