Panasonic G1 - some AF numbers

For those watching this rather interesting camera, Imaging Resource has
some numbers from a production model.

The G1's AF lag is around 0.35-0.39s.

To put that into perspective, Canon 40D is ~0.13s, Sony A200 ~0.19s,
Nikon D60 ~0.26s, Olympus E420 ~0.14s...

Not quite up to dslr speed, but not too shabby. Compared to p&s
cameras, the Oly SP550 is 1.1-1.5s, Pan FZ18 ~0.5s, Canon G9 ~0.6s.

Be interesting to see how the rest of this camera pans out. I'm looking
forward to having a good hard look at that EVF...

 

It does look interesting. But $800??? I would think that by removing
the mirror box, prism, etc., which simplifies assembly and by going to
a smaller sensor, they could save some cost over over something like a
D60 with a similar lens. But it's like $200 more! Oh well.

And sure, it's smaller. But only slightly. It not small enough to
replace a pocket P&S so I'm not sure the slightly smaller size and
weight is worth the extra cost. We're talking about 124x84x45 vs
126x94x64 mm in size for the body. And weight is 380g vs 495g for the
body alone. 630g for the G1 with lens and battery vs. about 730g for
the D60 with an equivalent lens and battery. So even though it's
smaller, not enough so to make much difference in terms of
portability.

It'll be really interesting to see whether the new "gee whiz" nifty
factor of the camera is enough for it to compete. I'd wait a bit
since I suspect you'll see that high price drop quickly.

 

I wonder how much of that is actual cost recovery, and how much is a
pretty accurate guess on what type of buyer they will attract and what
the market will stand...

I sort of agree, but can't help thinking that this *could* be a very
interesting format if/as it develops. We all recognise that the high
iso performance of 4/3 will always be less than larger sensors, but the
key is that the noise performance of the later 4/3 sensors is actually
not that bad, and it is *significantly better* than high end p&s.

The other major limiting factors for the p&s are:
- lens range/flexibility. The G1 addresses that, even if there are some
limitations on the current choices. *If* the format takes off, there
will be few limitations in the future as Tamron/Sigma/etc jump on board.
- viewfinder quality/ability to manual focus. The reports on the G1
viewfinder are fairly promising, but I'll take a wild guess that we are
still one generation (at least) away from an evf that truly competes
with a good ovf.
- AF speed. It seems the G1 has respectable performance there, although
again I'd like to see what the next round brings.

So, to me, it appears as if the G1 is (almost) the first true 'bridge'
camera. I'll wait for the second or third, I think..

I have the feeling it will leap off the shelves, frankly. There have
been very few truly innovative camera formats of late, and this just
might be 'the one'. Of course the price will drop, and there will
probably be some issues with the first few runs of the camera that will
see its used value drop dramatically also. It's all good, and I wish
Panasonic and Olympus well in this new venture... Diversity rules!

 

I think they're going to have to do a little better though, at least
for that price and for what the competition, such as a cheap low-end
DSLR like a D60 brings to the table for $200 less. Not even counting
the lens choices available since this G1 is something fairly new
compared to the Nikon F mount, a few of the specs even in addition to
what you pointed out above (noise, Autofocus, EVF vs OVF for the G1
vs. the D60) are in need of improving if it wants to compete at that
price point:

- Flash sync speed of only 1/160 (G1) compared to 1/200 (D60)
- Flash guide number of 10.5 vs. 12 at ISO 100.

There's other things also, like the CLS system that the D60 supports
with the built in flash as a controller. And why is the Bulb shutter
limited to 8 minutes with the G1? I know that's tiny and nit-picky
and probably will never come up as a limitation in the real world, but
it doesn't make sense. Maybe they'll change that with a future
firmware release.

Yes, those things are tiny nit-picks and you can find things about the
G1 that are a bit better than a D60 also.

But I think a revolutionary new system has to be more than just tiny
nit-picks better than the establishment for it to take hold. This one
is not, and the price is higher.

So do I. Hopefully you're right and this is just the first in a
series and the next few offerings will provide something
revolutionarily better than the cheap DSLRs of today.

It does look interesting as a replacement for a bridge camera. But
even there, the price is over twice as much as a lot of decent bridge
cameras. And I'm not sure you're getting all that much more for that
much more money. Maybe the G2 will be different, or a half price G1.

Steve

 

They should have two lists of tests. In daylight and then in low-light
situations where contrast-detection cameras will always win.

Though I fail to understand how autofocus speed is ever a selling point to
someone that knows what they are doing. I never look at nor consider the specs
on autofocus in any camera that I buy, that feature is always last on my list.
Autofocus is insignificant, like buying a camera based on the kind of neckstrap
that comes with it. I'll occasionally use it to get the lens quickly into the
range needed then I take over from there, as any decent photographer should.

I can only guess that for the snapshooter it's very important. Lots of
snapshooters that hope their purchase will turn them into a professional. What a
shame that they are the ones defining the market and designing the next cameras
with their purchases. The real professionals then have to put up with the
end-result of their majority-rules inexperience and no-talent choices in life.
There's something consummately wrong in that mechanism.

 

Really promising! I didn't think Panny would be that good that fast. If the
EVF is good, this may finally signal the end of the mirror box.

 

Full review now up:

http://www.imaging-resource.com/PRODS/DMCG1/DMCG1A.HTM

Gets a pretty good wrap..

I'll bet dpreview will be a little annoyed they were beaten, and will be
not far behind.

 

The Panasonic DMC-G1 isn't a DSLR camera nor is it a Canon camera, so
I'm not so sure that depreview.com will be rushing to do a full review
any time soon.

 

Not bad, but still a few problems. Most comparisions to it are
similar to existing cameras so I concentrate on things that are
different, like sensor size, EVF vs. OVF, etc.

Obviously, the high ISO performance wasn't as good as the Rebel XSi.
That's to be expected due to the smaller sensor and higher pixel
density. Is a slightly smaller body/lens worth that when it's not
really smaller enough to change it's portability? Personally I don't
think so. But some might.

I'm kind of dissapointed with the EVF. It's only 800x600 resolution.
I'd want to see "high def" resolution 1920x1080, or at that 4:3 aspect
ratio maybe double it to 1600x1200 before it could come close to
matching the detail you might see in an OVF.

Also, from the review:
"Another difference between the EVF and the LCD is the relative
contrast ratio. When looking through the EVF at a scene with a high
contrast ratio, detail in shadow areas falls off abruptly into
relative blackness, while the same shadow scene on the LCD still holds
detail. It's just a difference in the amount of contrast the LCOS chip
inside is able to display compared to the LCD."

Is that going to lead people to use the LCD and have to hold the
camera out at arm's length? Probably not. The EVF may be good enough
for framing, but I'd like to see something at least approaching a good
OVF if this really is the wave of the future.

Also, the reviewer says that there's a noticable time lag between real
time and what you see in the EVF. To me, that's unacceptable and
absolutely must be improved before I'll use an EVF. That just adds to
the already slower shutter lag from a DSLR.

And no live view when shooting in continious mode, only the captured
shots at the frame rate? Seems like that would make tracking a fast
subject (just when you'd want to use continuous shooting) very
difficult.

If this camera has the best EVF that's available (which it seems to),
then EVF's still have a way to go.

Also, the screen blanks out before taking an exposure so the shutter
can close and then re-open during exposure. That doesn't solve the
problem of the short screen blank on a DSLR to move the mirror out of
the way. I would hope that new technology can solve old problems but
this one hasn't.

So the main conclusions I get from reading that review is that in most
respects, it's not all that much different than something like the
Rebel XSi. It's main benefit is that it's a bit smaller but not
enough so to make a major difference in portability. But it has
several detrements such as noticeably lower image quality at high ISO,
slower AF, and a good EVF but not nearly up to OVF standards in terms
of image quality and ability to track a moving subject.

I seriously hope their next effort is better. This one could be fun
for early adopters and techogeeks. But for photographers I don't
think it's ready for prime time.

Steve

 

Don't be quite so dismissive of the 480Kpixel (800x600) DMC-G1 EVF, the
Konica-Minolta A2 has 640x480 (307Kpixel) and is almost (but not quite)
good enough for manual focus without using the focus-zoom/mag feature.
It's not a great jump in rez (307 to 480Kpix), but considering that this
is a brand new EVF design I believe that it would be prudent to at least
examine it yourself before concluding that it's rubbish.

There may be some setting in a menu which can change the contrast of the
EVF, but maybe I'm hoping for too much?

That does sound like a bit of a deal-killer.

That's a bit disappointing, adding that to the reported live-view lag
and that seems a bit sad.

I was hoping that the EVF would turn out to be vastly better than it is,
as Panasonic was touting it as a "1.4Mpix full-colour EVF".
Sadly it seems to be a repeat of the KM A2 "900Kpix EVF" fiasco, where
it was required to divide it by 3 to get the actual number of
full-colour pixels in the EVF and that was a bit disappointing (to say
the least).

The promise it held out (for me, at least) was a cheap digital body to
use Leica lenses on via a simple metal ring adapter, even with the 2X
crop factor of the 4/3 sensor.
If the EVF isn't really "up to snuff" then I think I'll pass.

 

There is still some lag even with high shutter speeds. Also, I don't
*want* the EVF to lag even when I'm using a slow shutter speed. I
don't *want* the EVF to show an artistically blurred waterfall. Let
the LCD display a "live view" that lags, shows the effects of changing
the shutter, whatever. Then you can see what your "custom" shutter
speed is doing to the image. But I want my viewfinder to show me in
real time exactly what the camera is pointing at no matter what I've
done to the settings. If a viewfinder can't do that, it's not of much
value.

Steve

 

You're absolutely wrong about this regarding the G1. The G1's EVF lag
is in no way related to shutter speed. It has a mechanical focal
plane shutter that is not doing anything when a picture is not being
taken.

If the G1 actually did let you have it both ways that would be fine.
But it doesn't. The EVF lags real time. The EVF only shows the
captured photos when continuous shooting, not the continuous view that
you need to help track a moving subject.

By taking a fault and turning it into a "feature", you're no better
than if microsoft said that the blue screen of death was a feature
intended to calibrate your monitor's blue color response.

This is why both of the P&S cameras I own also have optical
viewfinders. I use them in some situations, and use the LCD in others
and I decide when it's appropriate to use one or the other. I don't
want the camera manufacturer deciding that I don't need to see what
the camera is pointing at. That's what they've done with the G1.

Steve

 

But DSLRs give you a choice of using the low resolution LCD (and
800x600 is low resolution) or an optical viewfinder. You don't get
that choice with the G1.

Nonsense. If you can see the detail in 1920x1080 on a 60" screen from
5' away, then you can see the detail in 1920x1080 on a 1" screen from
from 1" away as long as there's a lens that lets your eye focus that
close.

From the article:
"An SLR gives you the view at the speed of light, but electrical live
view systems introduce some lag as the image is captured, processed,
and written to the LCD. Extra lag means that you're more likely to
miss the moment you see on the screen, adding to the overall shutter
lag."

Wrong yet again. From the article:
"I found a leaf bug on the screen outside our office, and tried to get
a decent shot of him. The G1 didn't focus on it well until I popped up
the flash, which helped increase the depth of field. Even then it
seems to have front-focused on my subject just a tad, which I could
perceive on the LCD, but didn't manage to fix it before the bug got
tired of me and flew away."

Wrong yet again. The G1 has a mechanical shutter which is not silent.
Check the article again for a video of the shutter in action.

Steve

 

I'm only going by what the reviewer said. He said there is EVF lag.
It can't be due to the shutter because the shutter is not actively
doing anything except when you take a picture and maybe when it does
the depth of time preview.

I'll bet there is visible lag in the EVF. It may be small with fast
shutter time, but it's there. It's there on every single EVF and live
view LCD I've ever seen and yes, it's worse with a slower shutter.

If you actually stopped to think for a minute you'd realize there has
to be lag. All the things needed to read the photons off the sensor,
process them into an image, rescale that image for the LCD or EVF and
display the image on the LCD or EVF takes time. Slowing down the
shutter speed only makes it take more time.

Steve

 

But lets not think about the time it takes to engage a solenoid to pull a heavy
mirror out of the way, wait for that to slap loudly out of the way, shaking the
camera while it does so and destroys some image resolution in the process, then
trigger another solenoid to trigger a focal-plane shutter curtain.

Heaven forbid that you should have to consider THAT agonizingly loud and
obnoxious and image destroying DELAY.

LOL

You fuckin' moron.

 

It's exactly about the angular size of the details. That's why I gave
the example I gave. The angular size is the same. I.e., the angular
size for each pixel of a 60" 1920x1080 at 5' away is the exact same as
the angular size of each pixel of a 1" 1920x1080 at 1" away. It's
also the exact same as a 20" display 20" away. And I can *absolutely*
see the difference between 800x600 and 1920x1080 on a 20" screen from
20" away. Therefore, I can also see the difference on a 1" screen
from 1" away. And 800x600 is pathetic these days. 20 years ago it
wasn't bad though.

Again, if you took the time to think before you type, you won't look
like such a fool with every post.

If you can't tell the difference in detail between 640x480 pixels and
1920x1080 pixels on a 20" display from 20" away, then you need to get
your eyes checked. But then again, that would certainly explain why
you think the quality of images from a P&S is as good as a DSLR.

 

It is dissapointing and not just because the 800x600 resolution is
noticably poor compared to a good OVF. It may be good enough for an
EVF, but that's the point. Other reasons besides 1) resolution it's
disappointing include 2) very poor quality in high contrast scenes,
with all shadow details completely lost. 3) temporal lag in the
displayed image from real time 4) no EVF view during continuous
shooting. Only single captured shots.

IMHO, that last one is a deal breaker. Most of the time you want to
use continuous shooting, you're also trying to track a fast moving
subject. Showing single captured images a few times a second and time
late from real time would make properly tracking a fast moving subject
almost impossible.

So yes, I wrote that it's disappointing. And it is.

The usual reason a DSLR has back or front focus issues doesn't have to
do with the sensor being misaligned. That's why you see front or back
focus with some lenses and not with others. If the sensor is
misaligned and *all* images have the same front or back focus, you can
just get that fixed. That would be great. Unfortunately it's not
that simple.

Most non-DSLR's have a mechanical shutter? Don't let our resident P&S
troll see you say that. The assertion was that the G1 was silent when
shooting. It's not. It may be quieter than a DSLR, but even with
DSLRs, some are much quieter than others. My D200 isn't all that
loud.

Steve

 

Wow, you morons are more moronic than once thought.

Get this: the 800x600 resolution is accomplished by 800x600 RGB pixels. EACH
pixel conveys ALL THREE COLORS sequentially. This is not unlike the Foveon
sensor where one pixel captures all 3 colors. THREE TIMES THE RESOLUTION FROM
THE SAME PIXELS.

Blurb from dpreview's overview: "Large, bright electronic viewfinder (1.4x
magnification, 0.7x FF equiv). The viewfinder uses a field sequential system
that mixes red, green and blue images shown sequentially at 60fps, giving an
effective viewfinder resolution of 1.4 million dots at 180 fps."

And:

"According Panasonic LCOS technology can produce much higher resolution images
than liquid crystal display or plasma display technologies. Compared to
conventional LCDs in which the back light is projected through RGB filters and
into the eye, in LCOS, liquid crystals are applied directly to the surface of a
silicon chip coated with a highly reflective aluminized layer. RGB light is then
reflected off this surface and into the eye, therefore minimizing overall color
loss often associated with the low quality of color filters in conventional
LCDs. This allows the G1's Live Viewfinder to achieve over 90% of the NTSC color
gamut (this, apparently, is unusually high).

The G1's single panel display chip shows the red, green and blue components in
succession (field sequential display) - your brain does the combining to produce
the full color image. This means you can't see the individual pixels - there's
no gap between them (there's no mesh surrounding the color filters such as in
conventional LCD displays). The viewfinder is able to refresh the three RGB
colors at a rate of 60Hz, therefore achieving the 60fps Full-Time Live View."


Please! Go freakin' educate yourselves! Would you? You get more moronic and
idiotic with every keyboard key that you press.

If I could, I'd reach through your monitor to slap the shit and the stupid out
of each and every one of you.

HOLYFUCK! I can't believe people on this planet are this amazingly stupid!

 

And yet even with all of that, the image I see in the viewfinder of my
DSLR is in real time. And when I press the shutter button, it takes a
picture much faster than a P&S can. And the focus tracks a moving
object during the time it takes to move the mirror out of the way so
that everything is still nice and sharp.

Try doing this with a P&S: Take a series of pictures of a bird flying
towards you and keep the bird in focus but have the tree it just flew
out of artistically blurred. Those are the kind of shots where a DSLR
lets you express your creativity while a P&S just lets you take a
snapshot.

It's obvious you only care about snapshots. So a P&S is fine for you.

Steve

 

Now THAT is all total BS. When people claim 3x the real resolution
simply because there's an R, G and B subpixel for every full color
pixel, they are being completely disengenuous.

It's nice not to see gaps between pixels. It certainly makes the
display look nicer than an equivalent resolution display that has gaps
between pixels. However, you're a complete idiot if you think that
matters in terms of resolution.

But like I said, the resolution may be good enough, but that isn't the
main reason this EVF is disappointing. There's also very poor quality
in high contrast scenes, with all shadow details completely lost.
Temporal lag in the displayed image from real time. And most
importantly, no EVF view during continuous shooting. Only single
captured shots.

that last one is a deal breaker. Most of the time you want to
use continuous shooting, you're also trying to track a fast moving
subject. Showing single captured images a few times a second and time
late from real time would make properly tracking a fast moving subject
almost impossible.