Dpreview G1 Preview Revisited

[GregGory]

Not sure what you refer to, hopefully not DPR's "DR tests."

Actaully, I did... Are they flawed?

Except for the industry-leading K10D with its amplifyer-free readout,
I don't think Pentax and Sony are known for low read noise performance
either. Data is scarce, but I'd say they are closer to 4/3rds
performance,
particularly normalised to sensor area, than they are to the best APS-C
cameras, 40D and D300, in this regard.

Canon CMOS has been the market leader for some time, the D300 is expensive. But all in all, I don't think there's evidence that any difference is down to the SSIS. But please prove me wrong if you have some info!

Now I'm not following you at all. I was talking about DR and read noise.
Yes I agree it has very little to do with stabilisation.

What I meant was, in the future it might become evident that the DR cannot be improved further with SSIS hitting a wall. Then LSIS might become the right/ the only viable path.

The G10 is not so interesting to me because of 28mm and its weight.

It's a (small) brick! 400g (yuck) but it does a lot of stuff right, had Canon chosen a lower MP sensor, I would probably have picked this one on beforehand, but I like 'aperture priority' mode - not 'diffraction-level' mode...

The FZ28 not so interesting because of size.

I was traveling 4-5 weeks this summer, the FZ8 was an excellent companion, but I did miss the WA and HD movies. I had my FZ8 in my belt case ALL the time, always ready for action (reaction time around 2-2.5 sec), but not disturbing at all when not in use. I think the FZ28 would still be 'non-disruptive' on my side. A G1 + 14-140 less so... But I'll have to try that to find out.

BTW, I find a belt case many many times more comfortable than a jacket pocket, but then you would have to remove the TC every time you picked up the cam.

The LX3 is interesting

but the number of copies with a soft left side is disturbing me.

Yes, that's a potential problem. The Danish prices are still 'loco', the Germans are OK-is, around € 405-410 delivered, but if I get a bad copy, then I'm in trouble. I hope the introduction of the G10 will lower the price a bit. The LX3 is really tempting, since it has RAW burs, f2, and 24mm. What I learned this summer is, that the long end is completely useless from dusk til dawn, just adds bulk. It's much nicer to stroll around in the evening with the LX3 than the FZx8. AND one more thing, the LX3 is the undisputed low-light champ. The m43 system won't change that until a SSIS body turns up, sadly. A f2.8 OIS would actually be better, but still, a lot more bulk for 1-2 extra stops.

--

 

[GregGory]

Still looking for the swivel LCD and FF sized VF on those, and the
E420 doesn't have any kind of IS... In this respect, the G1 compares
to the E3, maybe the Tweener if the VF greatly improved.

The G1 does not compare well to the E3. It is much more
appropriately compared to cameras like the DMC-L10, E420 and E520 -
perhaps also the upcoming E-A1

Concerning the VF size and the swivel LCD, yes they are comparable, the sensor might even be better. The L10 has the swivel LCD, but puny VF like the E520, and the E420 has a pinhole.

And the E-420 does IS if you simply purchase a Panasonic IS lens for it.

Great, slap on the L10-kit and your E420 would be MUCH larger than the G1 kit, with a sub-puny VF, and no swivel LCD... The E4x0 is simply not competition for the G1, there might (probably will) be a smaller G1 with smaller VF and no swivel LCD.

--

 

[GregGory]

Almost all DSLR lenses have to be stopped down one or two
stops the get good results,

This is simply not true with the typical Olympus ZD or
Panasonic/Leica lens. They typically are quite good wide open.

3 lenses have been tested by DPR, stopping down is almost not required with the 14-42 kit lens, but it's slow, so the the stopping down is almost a 'built-in' feature.

The other two are better stopped down.

--
Jay Turberville
http://www.jayandwanda.com

--

 

[GregGory]

In general, I would hesitate to buy a G1 for birding... With today's
EVF technology, the m43 system excels in most respect, but the
extreme long end for max DR and incredibly fast moving subjects not
being one of them. It would still be a lot better than any FZ, but at
the foreseeable pricepoint it better be right up there with the DSLR
counterparts.

I am interested in the G1 specifically for birding. No vibration
from the shutter is a big plus when mounted on a tripod and shooting
low shutter speeds. It may not be such a great option for birds in
flight though.

Yes, I should have been more specific, my statement goes for BIF! The G1 is quiet, but unfortunately not completely silent, but a heck of a lot better than any DSLR. AND Luis has also pointed out a few things that would make his birding experience better with the G1 (EVF reivew being one), but he's more into BOB (birds on branches)

I think it would take less than an afternoon for the Pany R&D
department to design a macro tube. Basically it's two mounts and some
wire.

It might be more than that if you do it "right." The tube changes
the effective f-numbers of the lens. But you are correct that it
should be trivial for Panasonic to design an extension tube.
Further, I see no reason not to use one of the two Olympus macros. A
lot of macro work is done using manual focus, not AF. In fact, a lot
of the focus fine-tuning is done by moving the camera back and forth
a tiny bit.

If the m43 => 4/3" converter costs the same as the macro tube, and if you don't already have the oly macro lens, then.... But you're right, AF is not a necessity for macro most of the time. Bellows or a good rail-mount can be just as good if you have decent live view.

--

 

[Ehrik]

Not sure what you refer to, hopefully not DPR's "DR tests."

Actaully, I did... Are they flawed?

The JPEG tests are fooled by NR, look at the A350 review where the
DR goes up to 10.5 stops at ISO 3200! A quantitative JPEG DR test will
always be bogus. Better shoot a standardised scene with barely clipped
highlights and with shadow detail, push the detail a standard amount and
let the reader inspect and compare NR/detail tradeoff between cameras.

The "RAW headroom" tests have the problem that they include
highlight recovery which depends on how the relative strengths of the
Bayer filters interact with the colour temperature of DPR's studio lights.
If they mismatch a lot, some channel with clip a lot sooner, and you'll
have more room for recovery. But change the light and the camera
that was best might become worst. A P&S with RAW should have the
same basic potential for recovery as a 1Ds3 or D3.

Maybe using the "JPEG" or "ACR default" instead will give a useful upper end
but then there is still the risk ACR does different NR and sharpening
on raw files from different cameras (John Sheehy claims it does), so
again the dark end clip off won't be consistent.

But all in all, I don't think there's evidence that any
difference is down to the SSIS.

Ah! Now I get what you were talking about. No I don't think
SSIS has to add any noise. A sensor that moves may have a harder
time to dissipate heat than one that's fixed. But sensors don't draw
much power, particularly not CMOS/NMOS designs.

The antishake motors could also generate heat in the vicinity of the
sensor. But with current implementations they are only run during the
exposure.

In any case, heat only causes noticeable noise with long (tripod) exposures.

The other problem could be interference from the SSIS motors using the
same battery and causing voltage changes from the power draw to affect
the readout circuitry. But LSIS should have the same problem.

Induction of currents in the readout circuitry could also be a problem
(thinking of Pentax' electromagnets) I suppose.

But none of this has been demonstrated.

I think the better performance with e.g. the 40D and D300 is just because
of more evolved technology or more expensive components.

The G10 is not so interesting to me because of 28mm and its weight.

It's a (small) brick! 400g (yuck) but it does a lot of stuff right,
had Canon chosen a lower MP sensor, I would probably have picked this
one on beforehand, but I like 'aperture priority' mode - not
'diffraction-level' mode...

Diffraction never gets worse from smaller pixels, they are just better at
recording it.

BTW, I find a belt case many many times more comfortable than a
jacket pocket, but then you would have to remove the TC every time
you picked up the cam.

I've had a belt case for my Panny camcorder, and I've also carried the
Kodak P880 in it. But I feel it gets in the way. A belt case for the FZ28
could probably be slightly smaller but I just don't imagine it could
be small enough to be convenient.

Just my two oere
Erik from Sweden

 

[GregGory]

The JPEG tests are fooled by NR, look at the A350 review where the
DR goes up to 10.5 stops at ISO 3200! A quantitative JPEG DR test will
always be bogus. Better shoot a standardised scene with barely clipped
highlights and with shadow detail, push the detail a standard amount and
let the reader inspect and compare NR/detail tradeoff between cameras.

The "RAW headroom" tests have the problem that they include
highlight recovery which depends on how the relative strengths of the
Bayer filters interact with the colour temperature of DPR's studio
lights.
If they mismatch a lot, some channel with clip a lot sooner, and you'll
have more room for recovery. But change the light and the camera
that was best might become worst. A P&S with RAW should have the
same basic potential for recovery as a 1Ds3 or D3.

Maybe using the "JPEG" or "ACR default" instead will give a useful
upper end
but then there is still the risk ACR does different NR and sharpening
on raw files from different cameras (John Sheehy claims it does), so
again the dark end clip off won't be consistent.

Thanks for the 'highlights'! Saved me a few hours of reading... Yes, I remember a dude in the Oly forum who claimed he could make a test that would match the DR of some Oly with any Canon (suppose not a FF).

In any case, heat only causes noticeable noise with long (tripod)
exposures.

BTW, if heat was a problem, suppose they would have some kind of heat sinks in the body, but never seen that.

The other problem could be interference from the SSIS motors using the
same battery and causing voltage changes from the power draw to affect
the readout circuitry. But LSIS should have the same problem.

Induction of currents in the readout circuitry could also be a problem
(thinking of Pentax' electromagnets) I suppose.

But none of this has been demonstrated.

I think the better performance with e.g. the 40D and D300 is just
because
of more evolved technology or more expensive components.

Probably

The G10 is not so interesting to me because of 28mm and its weight.

It's a (small) brick! 400g (yuck) but it does a lot of stuff right,
had Canon chosen a lower MP sensor, I would probably have picked this
one on beforehand, but I like 'aperture priority' mode - not
'diffraction-level' mode...

Diffraction never gets worse from smaller pixels, they are just
better at
recording it.

That's the same thing, isn't it? Or am I missing something? Basically the discs start to overlap => resolution doesn't improve, despite more pixels. 15mp is insane IMO. I'm really interested in a DR comparison between the LX3 and the G10, the winner probably gets my money. There was an LX3 review by a blogger, I really liked the way he handles the noise test, focusing on the saturation and colours in general. It's much much more important, especially for prints.

BTW, I find a belt case many many times more comfortable than a
jacket pocket, but then you would have to remove the TC every time
you picked up the cam.

I've had a belt case for my Panny camcorder, and I've also carried the
Kodak P880 in it. But I feel it gets in the way. A belt case for the
FZ28
could probably be slightly smaller but I just don't imagine it could
be small enough to be convenient.

The really great thing is the 'reaction time'. I've timed it, it really only takes 2-2.5 sec to turn it on, by the time the lens has extended, it's already before my eyes. I can pick it up with one hand, and even put it back. I went horseback riding, and I could still take a bunch of pictures and videos with strictly one hand. This is the reason why I haven't migrated up to the Drebel, I would simply miss tons of opportunities because of the slower reaction time.

I really hope there will be a m43 cam, smaller & lighter than the G1, but still some kind of grip (eg. E420), and a SMALL versatile lens like the 14-140 but no OIS and smaller, or even the 12-60 oly m43 version. I imagine I would Ebay every Canon DSLR bit I have, and say goodbye to the FZs.

--

 

[Jay Turberville]

Almost all DSLR lenses have to be stopped down one or two
stops the get good results,

This is simply not true with the typical Olympus ZD or
Panasonic/Leica lens. They typically are quite good wide open.

3 lenses have been tested by DPR, stopping down is almost not
required with the 14-42 kit lens, but it's slow, so the the stopping
down is almost a 'built-in' feature.

The other two are better stopped down.

Now you are shifting the issue. The point you made wasn't that lens performance improves if you stop down. The point you made was that, "Almost all DSLR lenses have to be stopped down one or two stops to get good results."

I just reviewed the three DPReview lens tests and each review said that the lens tested delivered either "good" or in the case of the 12-60mm "very good" results wide open. So in the opinion of the reviewer, you don't need to stop down to get good results with those lenses.

While I haven't formally tested my 50mm ZD macro, 17.5-45mm ZD, 14-50mm Leica, 40-150mm ZD, 14-45mm ZD, or 70-300mm ZD, my impression from using them is that they all give "good" results wide open. I never find myself thinking that I'd better stop down in order to get good results.

--
Jay Turberville
http://www.jayandwanda.com

 

[Jay Turberville]

The G1 does not compare well to the E3. It is much more
appropriately compared to cameras like the DMC-L10, E420 and E520 -
perhaps also the upcoming E-A1

Concerning the VF size and the swivel LCD, yes they are comparable,

Actually, I think the EVF in the G1 and the optical finder in the E-3 are significantly different. For instance the G1 image is much larger. In this area, we see more differences than similarities when comparing the G1 to any 4/3rds DSLR.

the sensor might even be better. The L10 has the swivel LCD, but puny
VF like the E520, and the E420 has a pinhole.

The E-420 has the same finder as the E-520 and L10.

And the E-420 does IS if you simply purchase a Panasonic IS lens for it.

Great, slap on the L10-kit and your E420 would be MUCH larger than
the G1 kit, with a sub-puny VF, and no swivel LCD... The E4x0 is
simply not competition for the G1, there might (probably will) be a
smaller G1 with smaller VF and no swivel LCD.

Fair enough. The E-520 is a better target for comparison. The DMC-L10 is better still. But the E-3 is such a different kind of camera - large, fast AF, waterproof and designed and marketed toward a pro or semi-pro market that such a comparison really doesn't fit.

Getting back to the issue of appropriate price, the G1 could reasonably sell for a bit more than the E-520, but it really needs stay somewhat close to that price point if they want to sell a lot of them.

--
Jay Turberville
http://www.jayandwanda.com

 

[John Carson]

If Panasonic is aiming the G1 - as their biggest market- the P&S
market upgrading to a DSLR, having a camera that costs near twice the
price of an e-420 system or Nikon D40x is very hard to justify in
this economy.

New technology always sells at a premium. Think of the first CD players, the first DVD players, the first Blu-ray players, the first HD television sets. The early adopters pay premium prices and then prices come down. The G1 may not as revolutionary as those examples (well...maybe it is close to HD television), but I don't think "overpricing" it by $200 to $400 at first release will do Panasonic much harm.

--
john carson

 

[Raval Online]

The G1 is meant as a DSLR replacement for those who do not want to
upgrade to a DSLR for various reasons. It is not meant for those who
already have DSLR equipment and look for a large-sensor pocketable
camera to complement their larger equipment. That's where a lot of
the misunderstanding comes from.
The G1 will compete directly with the entry level DSLRs from Canon,
Nikon, Sony, etc.
Personally I really enjoy live view and shooting from waist level
with a fold out LCD, that makes the G1 very interesting for me.
--

G1 is better than the FZ50 which is all I need to know and in 2yrs time when I'm looking for a replacement for my FZ50 I'll be looking at the latest version of the G1.

To me the negativity on this forum towards the G1 seems to be like an infection, someone said a few negative things about the camera and then everyone regurgitated what they read.

The camera hasn't launched yet and already some apparently know the ins and outs of the camera.

We should stop reading reviews from sites like DPreview and start reading these forum members post to get the skinny.
--

playing photographer since 2006 | DMC-LZ2 & DMC-FZ50 | http://www.ravalonline.com

 

[Ehrik]

The JPEG tests are fooled by NR, ...
The "RAW headroom" tests have the problem that they include
highlight recovery which depends on how the relative strengths of the
Bayer filters ...
there is still the risk ACR does different NR and sharpening
on raw files from different cameras ...

Thanks for the 'highlights'! Saved me a few hours of reading...

Welcome. I'm not sure this is written down anywhere; it's my own "analysis",
so it may be wrong. My understanding of how noise and DR "works" is of
course gathered by reading various posts and articles. I wish more people
would give constructive criticism about the DR test, then DPR just might
change it to something more useful. It's the weak point of otherwise
mostly very good reviews.

Yes,
I remember a dude in the Oly forum who claimed he could make a test
that would match the DR of some Oly with any Canon (suppose not a FF).

I think the "dude" is Jay Turberville with whom you've had a few exchanges
in this thread! As I recall, his main point was that DPR imposes an arbitrarily
chosen clipping limit at 2% luminance (as an alternative to the undefined
noise floor), and that you could potentially change the result of the test
simply by taking the file to Photoshop and change the gamma (the middle
value of the Levels dialogue).

So there you have yet another reason not to take the numbers they list
too seriously.

And we haven't even mentioned their fixation with highlight range yet...

Diffraction never gets worse from smaller pixels, they are just
better at recording it.

That's the same thing, isn't it? Or am I missing something?
Basically the discs start to overlap => resolution doesn't improve,
despite more pixels.

Well, there are lots of people who mistakenly think a higher pixel density
on the same sensor size will cause more diffraction in the image.

The use of the concept "diffraction limit" is also misleading. Some people
think it's a sharp cutoff, but it's rather a gradual loss of contrast. Different
sources add up to lower the MTF and at the diffraction limit the airy disc
and the pixel pitch (and let's assume the AA filter blur width) are the
same, so their contribution to lowering the MTF are of a similar magnitude.
Beyond that, the diffraction will be the main limit, but a higher pixel density
will still make a difference, with diminished returns, for a while.

Just like with lens aberrations, there is big gap from the pixel pitch where
you start seeing it, to the pixel pitch where no further practical
resolution can be extracted.

Conversely, below the "diffraction limit" (with an aberration-free lens) the
sensor is the main course of lack of resolution. That the diffraction limit is
further away on a coarse pixel-pitch sensor just says that the aperture
range where we suffer from the limited sensor resolution is broader!

15mp is insane IMO. I'm really interested in a
DR comparison between the LX3 and the G10, the winner probably gets
my money.

Yes, too bad there are no useful DR tests around.

Is that the Panny original bag you bought for your FZ5? Does it have room
for the FZ18?

Just my two oere
Erik from Sweden

 

[GregGory]

Almost all DSLR lenses have to be stopped down one or two
stops the get good results,

This is simply not true with the typical Olympus ZD or
Panasonic/Leica lens. They typically are quite good wide open.

3 lenses have been tested by DPR, stopping down is almost not
required with the 14-42 kit lens, but it's slow, so the the stopping
down is almost a 'built-in' feature.

The other two are better stopped down.

Now you are shifting the issue. The point you made wasn't that lens
performance improves if you stop down. The point you made was that,
"Almost all DSLR lenses have to be stopped down one or two stops to
get good results."

OK, you do have a point, but essentially it's question of defining 'good' performance. Pany users have been spoiled with HQ glass right from the beginning, and when people spend 2-3 times more on an 'upgrade', they want significant improvement in the IQ department. That probably means the best these 'kit' lenses can produce. Some of the kit lenses out there are just a joke compared to an FZ, and yes I know, the Oly kitties are decent.

I just reviewed the three DPReview lens tests and each review said
that the lens tested delivered either "good" or in the case of the
12-60mm "very good" results wide open. So in the opinion of the
reviewer, you don't need to stop down to get good results with those
lenses.

I went through the flash 'test-bench' and except for the kit lens, there are significant improvement by stopping down. Not to mention, the 12-60 better excel for its price, and there's still much to gain from stopping down...

While I haven't formally tested my 50mm ZD macro, 17.5-45mm ZD,
14-50mm Leica, 40-150mm ZD, 14-45mm ZD, or 70-300mm ZD, my impression
from using them is that they all give "good" results wide open. I
never find myself thinking that I'd better stop down in order to get
good results.

So again, it's the definition of 'good'. Maybe my perception of 'good' glass is equivalent to 'excellent' for others... Maybe

--

 

[GregGory]

The G1 does not compare well to the E3. It is much more
appropriately compared to cameras like the DMC-L10, E420 and E520 -
perhaps also the upcoming E-A1

Concerning the VF size and the swivel LCD, yes they are comparable,

Actually, I think the EVF in the G1 and the optical finder in the E-3
are significantly different. For instance the G1 image is much
larger. In this area, we see more differences than similarities when
comparing the G1 to any 4/3rds DSLR.

I think the OF size and the swivel LCD are very significant attributes to any camera buyer, especially in the 'DSLR-like' market. Personally, I would never ever buy a small VF DSLR again, the minimum size for me would be the D90, 40d or the K10d. So even though I REALLY like the E520, the OVF is a dealbreaker for me, and I know for many others too. Some people are also crazy about swivel LCDs. Eg. there are hoards of superzoom buyers, who only consider the Canon Sx, FZx0, and Sony H50 cause the fixed LCD is a dealbraker for them on all the other models.

So for people who want a decent OVF and a swivel LCD, the E3 is the only option. Yes, it has tons of other pro/ semi-pro features, but it doesn't mean that every buyer wants and needs every feature...

The E-420 has the same finder as the E-520 and L10.

Oooh, I thought it at least had a smallish prism... Bugger.

And the E-420 does IS if you simply purchase a Panasonic IS lens for it.

Great, slap on the L10-kit and your E420 would be MUCH larger than
the G1 kit, with a sub-puny VF, and no swivel LCD... The E4x0 is
simply not competition for the G1, there might (probably will) be a
smaller G1 with smaller VF and no swivel LCD.

Fair enough. The E-520 is a better target for comparison. The
DMC-L10 is better still. But the E-3 is such a different kind of
camera - large, fast AF, waterproof and designed and marketed toward
a pro or semi-pro market that such a comparison really doesn't fit.

The L10 would be a closer comparison if it had a decent OVF, it doesn't so for many many people, it doesn't compare. For those who don't care about the OVF, it's close. The E520 doesn't have a decent OVF nor a swivel LCD...

So which one you want to compare it to depends entirely from your personal point of view, but my point is, comparing the G1 to the E3 is just as valid, as comparing it to the E520, L10, or E420.

Getting back to the issue of appropriate price, the G1 could
reasonably sell for a bit more than the E-520, but it really needs
stay somewhat close to that price point if they want to sell a lot of
them.

It will, save for the 6 months 'lag' in it's PLC. In the spring, 500 USD/Euro sounds realistic.

--

 

[GregGory]

Welcome. I'm not sure this is written down anywhere; it's my own
"analysis",
so it may be wrong. My understanding of how noise and DR "works" is of
course gathered by reading various posts and articles. I wish more
people
would give constructive criticism about the DR test, then DPR just might
change it to something more useful. It's the weak point of otherwise
mostly very good reviews.

Hmmm, I'm still waiting for a DR test for compacts... Even it it's flawed and not very precise, it would still give an indication, since DPR wouldn't be so dumb as some sites where they 'test' the DR in standard contrast setting...

I think the "dude" is Jay Turberville with whom you've had a few
exchanges
in this thread!

Lol, good to know that evidence is right at hand

As I recall, his main point was that DPR imposes an

arbitrarily
chosen clipping limit at 2% luminance (as an alternative to the
undefined
noise floor), and that you could potentially change the result of the
test
simply by taking the file to Photoshop and change the gamma (the middle
value of the Levels dialogue).

So there you have yet another reason not to take the numbers they list
too seriously.

And we haven't even mentioned their fixation with highlight range yet...

Well, there are lots of people who mistakenly think a higher pixel
density
on the same sensor size will cause more diffraction in the image.

The use of the concept "diffraction limit" is also misleading. Some
people
think it's a sharp cutoff, but it's rather a gradual loss of
contrast. Different
sources add up to lower the MTF and at the diffraction limit the airy
disc
and the pixel pitch (and let's assume the AA filter blur width) are the
same, so their contribution to lowering the MTF are of a similar
magnitude.
Beyond that, the diffraction will be the main limit, but a higher
pixel density
will still make a difference, with diminished returns, for a while.

Well, I can only confirm that the 'diffraction calculator' doesn't apply fully to my findings on the FZ5 and FZ8. Maybe we have weaker AA filters than the calculator assumes (suppose it's 'extrapolated' from DSLR findings)? This is purely an idea, I have no evidence what so ever. But I don't expect I can 'play' with the aperture too much with the G10 at 140mm without softening the image severely, so I cannot use the extra MP, but lose some DR and burst.

Just like with lens aberrations, there is big gap from the pixel
pitch where
you start seeing it, to the pixel pitch where no further practical
resolution can be extracted.

Conversely, below the "diffraction limit" (with an aberration-free
lens) the
sensor is the main course of lack of resolution. That the diffraction
limit is
further away on a coarse pixel-pitch sensor just says that the aperture
range where we suffer from the limited sensor resolution is broader!

Understood, but the point is there is a price of having more pixels than the 'optimum' given the present sensor technology (DR, burst, file size etc.). I believe the optimum is closer to the 24 mp/cm2 than the 34 mp/cm2 at the moment. Yes, I've read and understand what John S. has proven, but there is one question that remains to be answered, Do his findings also apply to incremental differences among small sensors? Reading and seeing his evidence makes it clear the FZ50 sensor is much more efficient than the 40d per area. But this also applies to eg. the FZ50's glass and the most expensive L glass out there, the FZ50 wins hands down pr. AREA. But it's not really relevant, since making an L glass to the FZ50 quality would cost more than an arm (and it would be an overkill for the pixel density). The same for the DSLR sensor, what would an APS-c 1.75 micron sensor cost? So the final evidence which is missing is to subtract the 'quantity vs. quality' difference, since John compared apples and oranges (yet managed to point out a very important aspect). The 'solution' would be to compare the LX3 RAW files and the FX150/ FZ28 raw files, which are from the same generation and manufacturer, but 'conservative' vs. 'pop' MP count. The electronics might be better on the LX3, but not necessarily since the price difference might be down to the more expensive glass and LCD. The noise and DR performance difference would be very interesting indeed. Eg. I don't think the DR is independent of the pixel density. The improved DR of the D3 compared to the D300 does suggest something. Yes, electronics could be a factor, but still.

So, since I don't print large (A4 max) and my monitor is 'only' 20", I really don't benefit too much from the extra pixels. Even taking Bayer interpolation into account + some cropping 5-7 mp is all I need, except for when I want 'extended zoom', which is virtually non-existent on the FZ line. OTOH, I am pushing the DR as much as I can more than 1 out of two pics. So for me, DR is more important then resolution. Since even 'low' NR can take its toll on a pic, the 'optimum' sensor density would be the one that doesn't need NR at base iso at all (maybe a slight chroma NR). I'm interested to view some non-NR LX3 raw files and see what the true story is...

Yes, too bad there are no useful DR tests around.

Is that the Panny original bag you bought for your FZ5? Does it have
room
for the FZ18?

Well, the first one was for the FZ5, then it was stolen (yes, the bag only ) and the replacement I bought, event though it had the exact same prod. no. it was actually a bit larger, clearly made for the FZ7/8. I think a few are still around on the net. I haven't tested it with the FZ18/28, but I will go to a store before I put down my money. But I'm really in trouble choosing. The FZ28 has just hit the €300 mark, the LX3 the €400...

--

 

[linuxworks]

Some wanted in body Image Stabization without realizing that the E210
already was reportedly too small to incorporate the same and also the
fact that in lens IS is essential for video unless you like to watch
the jiggles at high zoom ranges. I don't know of any camcorder that
doesn't use in lens optical IS system as a result.

bah! there is no good reason for why 'shaking the mass' is better in the lens vs at the sensor. as long as you have full info about all the facts (and the new lenses have 2 more wires so you'd THINK they'd have a good comms path between lens and body) - then the body should have all the same info to 'go on' that the lens (and its cpu) would have. so tell me why 'shaking an element' (sensor or lens) is any different and why one is 'better' than the other? if both are done right, BOTH are effective.

zooming? tell me WHY a sensor can't adapt to the changes in its 'shaking pattern' when a lens is zoomed. I see no reason why a sensor can't adjust its distance or speed or offset just like the moving lens element can. same basic idea of shifting an element based on feedback. there is nothing inherent in one that can't be done in the other.

this 'only in the lens' argument from pany is what mostly kills interest in THEIR cameras and lenses, to me. after seeing the wonders of 'do it once; benefit many times' IBIS, I can't see going back to the 'pay over and over again' model of OIS. either way, you are shaking some element but I can't see paying for that shaker again and again! that does NOT make sense!!

most of the rest, I like. but this design and marketing principle is just a show-stopper to many of us. when there was a fixed lens, it was OK for pany to do OIS. now that they are going the 'many lens, one body' route, its NOT OK!

--
Bryan
(pic stream: http://www.flickr.com/photos/linux-works ) ~

 

[John Carson]

bah! there is no good reason for why 'shaking the mass' is better in
the lens vs at the sensor. as long as you have full info about all
the facts (and the new lenses have 2 more wires so you'd THINK they'd
have a good comms path between lens and body) - then the body should
have all the same info to 'go on' that the lens (and its cpu) would
have. so tell me why 'shaking an element' (sensor or lens) is any
different and why one is 'better' than the other? if both are done
right, BOTH are effective.

Plainly sensor shift image stabilization is possible. Whether there is any advantage in principle to either sensor shift stabilization or lens based stabilization could only be answered by someone with expertise in the area. I don't have it and it is pretty obvious that you don't either.

Nikon and Canon both use lens based stabilization in their DSLRs. Maybe that is just a matter of those companies having headed down a particular path and not wanting to change. Or maybe there are good technical reasons. I don't think either of us know.

--
john carson

 

[GregGory]

Nikon and Canon both use lens based stabilization in their DSLRs.
Maybe that is just a matter of those companies having headed down a
particular path and not wanting to change. Or maybe there are good
technical reasons. I don't think either of us know.

Canon started 'shaking' the lenses in the film days, yes the ONLY way back then for obvious reasons...

 

[xavozin]

In general, I would hesitate to buy a G1 for birding... With today's
EVF technology, the m43 system excels in most respect, but the
extreme long end for max DR and incredibly fast moving subjects not
being one of them. It would still be a lot better than any FZ, but at
the foreseeable pricepoint it better be right up there with the DSLR
counterparts.

I am interested in the G1 specifically for birding. No vibration
from the shutter is a big plus when mounted on a tripod and shooting
low shutter speeds. It may not be such a great option for birds in
flight though.

about birds in flight G1 actually might be v interesting, many reports of good tracking capabilities (has to do with the focus points' ability to move accross the frame), read here what Andy had to say about these capabilities: "it can freely track a subject across the frame in the way no SLR can match" http://forums.dpreview.com/forums/read.asp?forum=1000&message=29306625

and here a user on this forum reporting first hand experience: "Once the cam cought the subject it didnt let it go again when you moved the camera " http://forums.dpreview.com/forums/read.asp?forum=1033&message=29457273

--
Xavo

image is the only validation and most of their life isn't real.
modified from Sam Sparro - Black and Gold.

 

[linuxworks]

Nikon and Canon both use lens based stabilization in their DSLRs.
Maybe that is just a matter of those companies having headed down a
particular path and not wanting to change. Or maybe there are good
technical reasons. I don't think either of us know.

Canon started 'shaking' the lenses in the film days, yes the ONLY way
back then for obvious reasons...

this is a good explanation of WHY, historically, C and N have used opto based IS instead of 'sensor based'.

but since film is not relevant anymore (to us, here) shaking the sensor is as good as shaking the lens.

N and C are too much stuck in the past. pany has no such excuse, though. pany's reason seems to be MONEY more than history!

--
Bryan
(pic stream: http://www.flickr.com/photos/linux-works ) ~

 

[linuxworks]

bah! there is no good reason for why 'shaking the mass' is better in
the lens vs at the sensor. as long as you have full info about all
the facts (and the new lenses have 2 more wires so you'd THINK they'd
have a good comms path between lens and body) - then the body should
have all the same info to 'go on' that the lens (and its cpu) would
have. so tell me why 'shaking an element' (sensor or lens) is any
different and why one is 'better' than the other? if both are done
right, BOTH are effective.

Plainly sensor shift image stabilization is possible. Whether there
is any advantage in principle to either sensor shift stabilization or
lens based stabilization could only be answered by someone with
expertise in the area. I don't have it and it is pretty obvious that
you don't either.

is that a slam or something?

I've used both. I have experience with both. both are successful at what they do. in some cases, the IBIS has beaten the equivalent OIS (I did a test on my e3 with the l1 kit lens and the IBIS was better but a noticeable margin).

it does not take design experience to REALIZE that either strategy can be made to work and work well.

Nikon and Canon both use lens based stabilization in their DSLRs.
Maybe that is just a matter of those companies having headed down a
particular path and not wanting to change. Or maybe there are good
technical reasons. I don't think either of us know.

again, you are trying to insult me or something?

we can easily see some of the money reasons why one strategy is chosen by a vendor. we can also see historical reasons. we can see that some vendors have chosen a path and already have products out there with OIS in them (N and C) and they are not about to do an about-face and introduce alternate products that make their older ones look overpriced. yet if they did an IBIS camera, very few would continue to buy the overpriced OIS ones.

this is not at all a technical issue even though you think that only designers are able to discuss this.......

--
Bryan
(pic stream: http://www.flickr.com/photos/linux-works ) ~