DMC-GH1

The vertical dimension of the sensor is unchanged, as the vertical pixel count (3000).

if the horizontal dimension is the same with more horizontal pixel count (4352), we have a problem, the pixels are not square anymore.
So, I think it is an error...

mea00jmf said:

On the panasonic website they have the sensor dimensions listed as
17.3 x 13 mm.

Here: http://panasonic.net/avc/lumix/systemcamera/gms/gh1/specifications.html

This is the standard 4/3s dimensions but if that is the case then
they have crammed more pixels on the sensor. Bad for noise.

I suspect that some marketing person has just copied the stats from
the G1 without understanding or checking. However it may not be the
case so can any one confirm that the GH1 sensor is physically bigger
than the standard 4/3rds version in the G1? I know it seems logical
that it would be but it needs to be confirmed.
--

Click to expand...

You can with a multi aspect sensor...

stefaan said:

You can't simply increase sensor size without changing mount and/or
lenses.
Since the older lenses work on the GH1 ...
--

Click to expand...

honza_pl said:

It's clear that 16:9 is more affected by vignetting as 4:3

Click to expand...

No it' not clear! Care to explain?

To me the outer circle touches both corners (and since they are at the same distance) so vignetting is equal.

Even if you are talking about the zones outside the inner circle, I still don't see that as those zones seem to have the same area...
--
Duarte Bruno

tkbslc said:

Corkcampbell said:

make yourself a little more knowledgeable about what Panasonic has
been doing lately. You might also wait for more detailed studies and
reviews of the new camera before making inaccurate generalizations.

Click to expand...

While in general, I agree that we should wait for the hard facts,
nobody has overcome pixel density physics yet.

Click to expand...

What would 'pixel density physics' be? This is a branch of the science I have never heard of.

tkbslc said:

15MP on the Canon
50D is starting to look like the noise limit for APS-C,

Click to expand...

In what sense? The 50D has better IQ than any other APS-C Canon, according to the posts ofthe majority of those who actually use it and have other Canons to compare it with. According also to the test results, it's only DPR's skewed commentary which has given it an undeserved reputation.

tkbslc said:

so I can't
wait to see what a 50% smaller sensor with 14MP will look like.
Probably not bad up to ISO 800 like all the other 4/3 cams.

Click to expand...

It will have the same characteristics as other FT cameras with Panasonic sensors. Why would it not?

--
Bob

LLS1 said:

honza_pl said:

interesting - so size of sensor is bigger as usuall 3 ones ?

if it's so, 16:9 will use wider area as 4:3 ? nice idea, but it can
bring problem with vignetting

Click to expand...

With an oversized sensor all aspect ratios can use the same diagonal.
If it was just a normal 4:3 that was cropped, the new 16:9 picture's
edges wouldn't be as close to the image circle edges as with the 4:3.
With an oversized sensor even the 16:9 can use the whole image
circle. There shouldn't be any more vignetting since both areas fit
inside the same circle.

Click to expand...

As I remeber, the FT system defines the image diagonal, but not the aspect ratio (although it's generally been 4:3). Presumably, if lenses produce the required FT image circle, any aspect ratio with the same diagonal will work within it.
--
Bob

BJL said:

honza_pl wrote:
...

Click to expand...

Last September there was a discussion on this where I posted ( http://forums.dpreview.com/forums/read.asp?forum=1000&message=29441938 ) an image circle resolution schematic: i528.photobucket.com/albums/dd330/Ajax9000/mFT/mft_image_circle.png.

In part of the post I noted:

BJL said:

... The key point is that the Four Thirds System
Specification specifies that the diagonal is 21.63 mm. Regarding mFT,
the FT website says "The Micro Four Thirds System is designed to be
compatible with the aspect ratios specified for the Four Thirds
System, including 4:3, 3:2 and 16:9." This does not give sufficent
information (I can't find the detailed standard) to be definitive re
a wider sensor or cropping.

Click to expand...

DPReview did an LX3 schematic in November 2008 as BJL noted.

(BTW, that page has an error in the description. Where is says "The key thing is that the LX3 even uses a crop from the sensor at 4:3 ratio, ..." It should say "The key thing is that the LX3 even uses a crop from the sensor at 3:2 ratio, ...")

And honza_pl did a schematic earlier in this thread ( http://forums.dpreview.com/forums/read.asp?forum=1000&message=31207485 )

Below is my take on what Panasonic seems to have done:



… and a few hypotheticals of slightly lower resolutions sensors that may be better suited to the HD video task (i.e. they use integer downscaling):

in case of 16:9 not only corners are close to outer circle, but left/right sides too (not much, but you can see that they are out of inner circle - drawn with - - - - line)

This seems very plausible and if panasonic haven't implimented it like this there designiers have made a huge error.

However the Panasonic website continues to state the sensor size is 17.3 x 13. Does anyone have conclusive proof that it has been implimented as the diagrams state? Therefore maintaining an active sensor area of 17.3 x 13 in 4:3 mode but with a larger width in 3:2 and 16:9

http://panasonic.net/avc/lumix/systemcamera/gms/gh1/specifications.html

If you could provide the link that would be great.

honza_pl said:

in case of 16:9 not only corners are close to outer circle, (...)

Click to expand...

No they are not closer. They are all (4/3 and 16/9) touching the outer circle so they exactly at the same distance from the center, and since vignetting rises as the distance from the center rises, they should have the same exact amount of vignetting!

but

honza_pl said:

left/right sides too (not much, but you can see that they are out of
inner circle - drawn with - - - - line)

Click to expand...

Yes at first sight they are when you look at the sides, but I think your brain is selectively forgetting that 4/3 area that is vignetting out of the ----- circle, between the top horizontal blue and pink lines.

So considering the 16/9 crop and the ----- line, it vignettes more on the sides but less on top and bottom, and just by sight I'm sure that if you calculate those vignetted areas they'll both add up to the same.
--
Duarte Bruno

but vignetting effect is made by worst areas (corners by both of 4:3 and 16:9 and sides in case of 16:9), not as average amount from whole area

0lf said:

The vertical dimension of the sensor is unchanged, as the vertical
pixel count (3000).
if the horizontal dimension is the same with more horizontal pixel
count (4352), we have a problem, the pixels are not square anymore.
So, I think it is an error...

Click to expand...

yes, you are right

honza_pl said:

but vignetting effect is made by worst areas (corners by both of 4:3
and 16:9 and sides in case of 16:9), (...)

Click to expand...

And?

The worst areas are the most distant (as you said: "corners by both of 4:3") and not the sides, have the exact same vignetting, as they are AT THE SAME DISTANCE from the center.

I think you need to take one such extreme vignetted picture and do some crops to convince yourself.
I'm stopping right now!
--
Duarte Bruno