«

ok, I finally got autobracketing working on my Canon 10D - but I'm not
really pleased on how the camera actually does it.

It *mechanically fires* for each exposure. With continuous mode set,
and aperture priority set, you press the button and you get three
successive clicks in taking the picture.

Somewhat OK for tripods, NOT so OK for hand-held movement. What's the
advantage of having a f/1.x lens, so you can shoot at reasonable ISO
levels and high shutter speeds if the shutter click is going to
provide you with an inherent quarter second delay between snaps?

Since the only thing that is changing is shutter speed, couldn't one
click just expose three photographs in rapid succession, with the chip
itself handling the saving of raw data? I realize that you might not
be able to get as precise resolution (- for canon 10d, 2 megapixels
(out of 6) to be exact because each 'picture' would take up 1/3rd of
the camera's RAM - unless they had a 12mp spare buffer, or they wrote
pictures out to flash fast enough to make this possible) but it would
more than make up for it in cases where the contrast levels are high.

Now, the question: is there a camera that allows you to do this? I
would think it would be a simple firmware upgrade for any digital slr
that does auto bracketing, but I don't know.

I really want this feature though. Even with a tripod, autobracketing
is iffy for long range lenses - the 'click' of the camera turns into
vibration which makes the pictures not line up very well when you
recombine them.

jon

"jonathan" <> wrote in message
news: om...
<snip>
>
> Since the only thing that is changing is shutter speed, couldn't one
> click just expose three photographs in rapid succession, with the chip
> itself handling the saving of raw data?...<snip>
>
> jon

That is what is happening, ignore the sound of the click.

Rob

"Canopus" <> wrote in message news:<bp5h60$1l7780$>...
> "jonathan" <> wrote in message
> news: om...
> <snip>
> >
> > Since the only thing that is changing is shutter speed, couldn't one
> > click just expose three photographs in rapid succession, with the chip
> > itself handling the saving of raw data?...<snip>
> >
> > jon
>
> That is what is happening, ignore the sound of the click.
>
> Rob

I hardly think this to be the case (you can see the mechanical shutter
opening and shutting whilst taking pictures if you look directly at
the lens.)

In fact, I timed the delay by taking both one exposure, and three
separate exposures (using continuous mode), by shooting a model train
at constant speed and looking at the motion blur. Indeed, the delay is
about 1/3rd second which is where they get their 'burst of 9 in 3
seconds' statistic.

so for the canon 10D at least, this is out. what camera are you using
that shows different, and what leads you to think that you can ignore
the click?

jon

"jonathan" <> wrote in message
news: om...
> "Canopus" <> wrote in message
news:<bp5h60$1l7780$>...
> > "jonathan" <> wrote in message
> > news: om...
> > <snip>
> > >
> > > Since the only thing that is changing is shutter speed, couldn't one
> > > click just expose three photographs in rapid succession, with the chip
> > > itself handling the saving of raw data?...<snip>
> > >
> > > jon
> >
> > That is what is happening, ignore the sound of the click.
> >
> > Rob
>
> I hardly think this to be the case (you can see the mechanical shutter
> opening and shutting whilst taking pictures if you look directly at
> the lens.)
>
> In fact, I timed the delay by taking both one exposure, and three
> separate exposures (using continuous mode), by shooting a model train
> at constant speed and looking at the motion blur. Indeed, the delay is
> about 1/3rd second which is where they get their 'burst of 9 in 3
> seconds' statistic.
>
> so for the canon 10D at least, this is out. what camera are you using
> that shows different, and what leads you to think that you can ignore
> the click?
>
> jon

Well I may be wrong, but, I was experimenting using bracketing while
shooting motorway traffic from a bridge after dusk. What I noticed was that
except for the various degrees of exposure all the pictures looked the same
except the ones taken fractionally later than the first had the trails of
car lights a little longer, but, in the same place.

I hope that scans, it's 5:30 a.m. here and I've just finished work.

I'm using a Fuji S602Z by the way.

Rob

In message < >,
(jonathan) wrote:

>so for the canon 10D at least, this is out. what camera are you using
>that shows different, and what leads you to think that you can ignore
>the click?

This is out for any DSLR. All DSLRs completely open and close the
shutter, before reading data, and do housecleaning tasks between shots.
What you want is only physically possible with consumer digital cameras,
which can read and zero data while the shutter is open. Perhaps it will
be possible with professional digitals of the future; I certainly hope
so, because I can imagine a lot of functionality that would be
available.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><

wrote in message news:<>. ..
> In message < >,
> (jonathan) wrote:
>
> >so for the canon 10D at least, this is out. what camera are you using
> >that shows different, and what leads you to think that you can ignore
> >the click?
>
> This is out for any DSLR. All DSLRs completely open and close the
> shutter, before reading data, and do housecleaning tasks between shots.
> What you want is only physically possible with consumer digital cameras,
> which can read and zero data while the shutter is open. Perhaps it will
> be possible with professional digitals of the future; I certainly hope
> so, because I can imagine a lot of functionality that would be
> available.

yeah I hope so too.. I have no clue why an electronic shutter isn't
possible with a CMOS chip (anyone?)

Anyways, I did an experiment with photographing light sources at
night, and - at ISO 100 f/11 - traffic lights blow out at approx
1/45th of a second, taillights at 1/20th, and background fixtures like
signs come into acceptable capture is approximately 3 seconds (after
fiddling around with the raw photos)

Now I could up this to ISO 800/1600 and probably get the 3 second
interval down to a decent amount, but this is still a factor of 90 or
so. To get things so that reasonable, nighttime photos are decent with
one snap of the shutter will require a lot more than AEB at -2..0..+2.
And a lot of post-editing in photoshop.

jon

(
ps - all of this is giving me a very good appreciation of how
sophisticated our eyes must be in processing. when you look at
headlights, taillights, etc, you see every detail of both the signs
and the lights, which indicates to me that the preprocessor in my head
is handling the 100 fold exposure factor with ease and plenty of
processing room to spare.
)

(jonathan) wrote:
> > This is out for any DSLR. All DSLRs completely open and close the
> > shutter, before reading data, and do housecleaning tasks between shots.
> > What you want is only physically possible with consumer digital cameras,
> > which can read and zero data while the shutter is open. Perhaps it will
> > be possible with professional digitals of the future; I certainly hope
> > so, because I can imagine a lot of functionality that would be
> > available.
>
> yeah I hope so too.. I have no clue why an electronic shutter isn't
> possible with a CMOS chip (anyone?)

Has nothing to do with CMOS. It totally has to do with the architecture
of the chip.

Here is a reasonably accessible explanation:
<http://www.dpreview.com/learn/Glossary/Camera_System/Sensor_01.htm>

> (
> ps - all of this is giving me a very good appreciation of how
> sophisticated our eyes must be in processing. when you look at
> headlights, taillights, etc, you see every detail of both the signs
> and the lights, which indicates to me that the preprocessor in my head
> is handling the 100 fold exposure factor with ease and plenty of
> processing room to spare.
> )

There have been CMOS active pixel sensors with logarithmic response.
One cool demo image I recall seeing had a woman next to an incandescent
light bulb in a dark room. You could see details on the side of her
face that was shaded from the bulb as well as the glowing filiment in
the bulb.

> > yeah I hope so too.. I have no clue why an electronic shutter isn't
> > possible with a CMOS chip (anyone?)
>
> Has nothing to do with CMOS. It totally has to do with the architecture
> of the chip.
>
> Here is a reasonably accessible explanation:
> <http://www.dpreview.com/learn/Glossary/Camera_System/Sensor_01.htm>
>
> > (
> > ps - all of this is giving me a very good appreciation of how
> > sophisticated our eyes must be in processing. when you look at
> > headlights, taillights, etc, you see every detail of both the signs
> > and the lights, which indicates to me that the preprocessor in my head
> > is handling the 100 fold exposure factor with ease and plenty of
> > processing room to spare.
> > )
>
> There have been CMOS active pixel sensors with logarithmic response.
> One cool demo image I recall seeing had a woman next to an incandescent
> light bulb in a dark room. You could see details on the side of her
> face that was shaded from the bulb as well as the glowing filiment in
> the bulb.

So I take it that the CMOS in a Canon 10D is a linear response? So
where do you get the logarithmic variety? And where do you get the
camera that has a small switch on the side saying 'logarithmic or
linear'?

But seriously, is 'logarithmic or linear' the only choice? What about
CMOS chips that 'sense' their fill amount and then add slower (ie: go
from logarithmic to linear based on fill rate)?

Argh. You've now made me want this experimental camera..

jon