Which camera has the best dynamic range?

David J Taylor wrote:
John McWilliams wrote:
[]
RAW files can be converted to a viewable format by non-linear or
linear, gamma corrected, or not, depending on the software and
settings of the user.

Saying JPEGs have more dynamic range is simply wrong.

To clarify, John, I was talking about the JPEG files which come straight
from the camera, not those from a RAW to JPEG conversion.

Given the class of camera under consideration (Fuji F100fd and Nikon
Coolpix 610), would you like to name one which uses a linear rather than a
gamma corrected representation in the JPEG? I would be very surprised if
any did.

I am not familiar with either camera, but in any event, it's irrelevant
to my statement: JPEGs do not have inherently more dynamic range than
the RAW from which it's processed.

--
john mcwilliams

John McWilliams wrote:
David J Taylor wrote:
John McWilliams wrote:
[]
RAW files can be converted to a viewable format by non-linear or
linear, gamma corrected, or not, depending on the software and
settings of the user.

Saying JPEGs have more dynamic range is simply wrong.

To clarify, John, I was talking about the JPEG files which come
straight from the camera, not those from a RAW to JPEG conversion.

Given the class of camera under consideration (Fuji F100fd and Nikon
Coolpix 610), would you like to name one which uses a linear rather
than a gamma corrected representation in the JPEG? I would be very
surprised if any did.

I am not familiar with either camera, but in any event, it's irrelevant
to my statement: JPEGs do not have inherently more dynamic range than
the RAW from which it's processed.


It's not your misunderstanding of a camera's dynamic range that matters
as much as whether or not you can display it all in a single picture.

If you consider 3 stops to the left of centre will produce an image from
a RAW file that has detail in white and near white areas and 3 stops to
the right will produce detain in dark areas... It is *impossible* to
display a picture with both the left and right detail *IF* the central
portion is correctly displaying the dynamic range of the sensor.

Where David is (seemingly) unable to elaborate on his statement is when
you and the idiot from the frozen wastes of Canada jumped in and try to
discredit him with twisted bull**** about what you perceive the dynamic
range to be.

Well here's some information for you two, that I suspect neither of you
would prefer to read.

The Dynamic range of a camera is that point at which detail in
highlights and shadows in a correctly exposed image cease - *not* an
image tampered with in Photoshop.

When the detail stops being displayed in light areas and detail stops
being available in dark areas, the dynamic range has been exceeded.

It is entirely possible to manipulate the dynamic range during
processing. The resulting JPEG developed from a RAW file may well exceed
the dynamic range of the camera by a considerable amount - but that is
tampering with the image, not a description of the true dynamic range of
a camera.

When (most) Digital cameras save a JPEG file, it will have had some post
processing done to it by the camera's computer and it *will* exceed the
dynamic range of the camera and therefore the dynamic range of an
unprocessed RAW file. Not enough for those people who refuse to consider
ND filters to control contrast and believe they need a $1000 program to
do it after the shoot.

The fact you may be able to extend the dynamic range of a photo further
that the range of the camera with manipulation of the RAW (or JPEG)
image during development, does not change the fact a RAW file - straight
from the camera - has less dynamic range than a JPEG file - straight
from the camera.

Further to this; Instead of trying to extend the dynamic range of the
camera, it is entirely practical to *compress* the dynamic range in the
camera and avoid the need to post process entirely... Unless of course
you actually enjoy this sort of stuff. I prefer to use Photoshop to
produce art photos rather than recover disasters that could have been
avoided with knowledge of photography.

Alan Browne wrote:
[]
JPG's 8 bits/color is compressed DR, not more DR. The 'loss' is in
graduation 'tween colors. JPG cannot contain an expression of more
information than the original raw, compressed or otherwise.

The key point is that in-camera JPG leaves you with much less in terms
of options than post-processed raw.

With an 8-bit linear coding, the ratio between maximum and minimum signal
level is 255:1.

The typical RAW data is 12-bit or 14-1bit, having a ratio of max/min of
4095:1 or 16383:1.

With JPEG, taking 2.2 as the typical gamma correction, the ratio is
255^2.2:1, or about 200,000:1.

Agreed that JPEG compromises on the accuracy with which any particular
brightness can be represented, but the range of values which can be
represented is greater with JPEG than RAW.

Cheers,
David

John McWilliams wrote:
[]
I am not familiar with either camera, but in any event, it's
irrelevant to my statement: JPEGs do not have inherently more dynamic
range than the RAW from which it's processed.

Of course not, but on its own, JPEG has the greater dynamic range. See my
earlier post for the numbers.

David

Shon Kei Picture company wrote:
[]
Where David is (seemingly) unable to elaborate on his statement is
when you and the idiot from the frozen wastes of Canada jumped in and
try to discredit him with twisted bull**** about what you perceive
the dynamic range to be.

I have given the numbers which justify my statement in a posting this
morning.

Well here's some information for you two, that I suspect neither of
you would prefer to read.

The Dynamic range of a camera is that point at which detail in
highlights and shadows in a correctly exposed image cease - *not* an
image tampered with in Photoshop.
[]

The part of the definition which is critical here is "cease". What do we
mean by "cease", and is there a standard definition? With a typical
digital camera, in the raw file, once the maximum value is exceeded,
that's it. No more dynamic range. With film, it's a gradual compression,
so what percent contrast or whatever is used? Again, with digital, what
contrast is used at the low end? Should we call the toe of the dynamic
range where a 20% contrast can still be perceived (i.e. relative light
values of 100 and 120 produce distinct levels in the RAW file)?

Is there anything from film photography which can be carried across
meaningfully to digital to help in a more useful dynamic range definition?

David

ASAAR wrote:
On Sun, 31 Aug 2008 19:20:34 GMT, David J Taylor wrote:

If you feel I have not backed up my opinions sufficiently, I will
try and expand or explain.

No, I don't believe that you will. You have only to examine the
quotes in your own replies to see the issues you've avoided
addressing. That goes for this latest reply of yours. You've had
many opportunities to expand or explain, but it's not even a matter
of backing up your opinions sufficiently. You have yet to make a
reasonable start. It means nothing to you to have me repeatedly say
that Fuji doesn't provide full resolution 12MP shots at its highest
ISOs, and Fuji makes this clear on their website, in their catalogs
and in their manuals. You continue to state that you wouldn't trust
a company that offers those high ISOs with 12mp shots, when even the
most clueless dummy knows by now that Fuji doesn't and has never
offered that in any of their cameras. You're quick to see insults
before they materialize, yet you don't mind tarnishing your own
reputation with what now can only be described as your own
incorrigible behavior, which continues year after year. You can
only play "rope-a-dope" for so long before everyone catches on.

ASAAR,

The statement that I made was, quite intentionally, intended to be that an
acceptable 12MP ISO 12,800 image is an unreasonable expectation from a
small sensor camera. Indeed, it would be interesting to do the sums and
see just what physical size of sensor was required to do this. I did
/not/ intend to say that Fuji (or any other company) provided this
combination of settings in their camera.

You say that 6 x 4 inch photos might be usable from the ISO 12,800 setting
offered by Fuji - so being kind that might be a 1200 x 800 pixel image,
i.e. about 1MP. This may be consistent with the other claims made for
some Fuji cameras. It would be interesting to have a link to an ISO
12,800 image - I didn't see a full review on the D P Review Web site.

I still trust engineering and physics more than marketing claims.

David

On Mon, 01 Sep 2008 06:20:27 GMT, David J Taylor weaseled:

The statement that I made was, quite intentionally, intended to be that an
acceptable 12MP ISO 12,800 image is an unreasonable expectation from a
small sensor camera.

It appears that you're now trying to make us believe that you
can't or won't read. Nobody has ever said that 12MP ISO 12,800
images were possible. You've been explicitly told several times
that at the highest ISOs, the Fuji camera only is capable of taking
reduced resolution pictures. What you've done is warned people that
not only a particular Fuji camera, but all cameras made by Fuji are
suspect, because YOU would have a problem with any company that
offers features that only you are aware of. What's unreasonable is
getting a straight answer from you.



Indeed, it would be interesting to do the sums and
see just what physical size of sensor was required to do this. I did
/not/ intend to say that Fuji (or any other company) provided this
combination of settings in their camera.

You're really adept at using what are called "weasel" words,
David. No, you didn't explicitly state anything. But you imply
much, and this is what you said :

Does the F100fd offer 12MP at ISO 12,800? If so, I would expect the
results to be completely unusable, and hence I would have considerable
reduced trust in a camera (or should it be the company?), which has
unusable settings?

When you ask that question and then repeatedly ignore the answer,
most likely because it contradicted your unreasonable assumptions,
it only leads us to have considerable reduced trust in you and your
motivations.


You say that 6 x 4 inch photos might be usable from the ISO 12,800 setting
offered by Fuji - so being kind that might be a 1200 x 800 pixel image,
i.e. about 1MP. This may be consistent with the other claims made for
some Fuji cameras. It would be interesting to have a link to an ISO
12,800 image - I didn't see a full review on the D P Review Web site.

1MP is one of the resolutions offered at ISO 12,800. Since I
replied to you that the F100fd can "produce small 4"x6" snapshots of
usable, if not good quality", you're free to assume either that I
know this because I took some pictures myself with the camera at
that setting, or that I saw some of the 12,800 ISO images in a
review. Hint: read my reply to the OP. DPReview doesn't test all
cameras. When they miss one, Google can be your friend.


I still trust engineering and physics more than marketing claims.

As do most reasonable people. But it appears that you're using it
here to defend your bogus assumption that Fuji's camera provided a
12MP 12,800 ISO option. Where's the marketing claim that you trust
less? Only in your imagination. Fuji *never* made that claim, and
stating this yet again, after being told that such a claim and such
an option never existed shows how little you care about the truth.

ASAAR wrote:
On Mon, 01 Sep 2008 06:20:27 GMT, David J Taylor weaseled:

The statement that I made was, quite intentionally, intended to be
that an acceptable 12MP ISO 12,800 image is an unreasonable
expectation from a small sensor camera.

It appears that you're now trying to make us believe that you
can't or won't read. Nobody has ever said that 12MP ISO 12,800
images were possible. You've been explicitly told several times
that at the highest ISOs, the Fuji camera only is capable of taking
reduced resolution pictures. What you've done is warned people that
not only a particular Fuji camera, but all cameras made by Fuji are
suspect, because YOU would have a problem with any company that
offers features that only you are aware of. What's unreasonable is
getting a straight answer from you.



Indeed, it would be interesting to do the sums and
see just what physical size of sensor was required to do this. I did
/not/ intend to say that Fuji (or any other company) provided this
combination of settings in their camera.

You're really adept at using what are called "weasel" words,
David. No, you didn't explicitly state anything. But you imply
much, and this is what you said :

Does the F100fd offer 12MP at ISO 12,800? If so, I would expect the
results to be completely unusable, and hence I would have
considerable reduced trust in a camera (or should it be the
company?), which has unusable settings?

When you ask that question and then repeatedly ignore the answer,
most likely because it contradicted your unreasonable assumptions,
it only leads us to have considerable reduced trust in you and your
motivations.


You say that 6 x 4 inch photos might be usable from the ISO 12,800
setting offered by Fuji - so being kind that might be a 1200 x 800
pixel image, i.e. about 1MP. This may be consistent with the other
claims made for some Fuji cameras. It would be interesting to have
a link to an ISO 12,800 image - I didn't see a full review on the D
P Review Web site.

1MP is one of the resolutions offered at ISO 12,800. Since I
replied to you that the F100fd can "produce small 4"x6" snapshots of
usable, if not good quality", you're free to assume either that I
know this because I took some pictures myself with the camera at
that setting, or that I saw some of the 12,800 ISO images in a
review. Hint: read my reply to the OP. DPReview doesn't test all
cameras. When they miss one, Google can be your friend.


I still trust engineering and physics more than marketing claims.

As do most reasonable people. But it appears that you're using it
here to defend your bogus assumption that Fuji's camera provided a
12MP 12,800 ISO option. Where's the marketing claim that you trust
less? Only in your imagination. Fuji *never* made that claim, and
stating this yet again, after being told that such a claim and such
an option never existed shows how little you care about the truth.

ASAAR,

I made no assumption about the brand of camera in my statement about
small-sensor cameras and very high ISO speeds. If you want to read in an
assumption - that's up to you. I have never claimed that any camera
offered 12MP at ISO 12,800, and therefore stick by my point that I would
highly mistrust any that did (with the present state of sensor
development). You would as well.

I have advised people that:

- if they want to use such high ISOs check the results. Perhaps it would
be helpful if you could provide a link to such an image.

- Fuji have made high dynamic range sensors, with dual photosites.

- To check whether any particular Fuji camera does, or does not, include
such a sensor.

You are correct that I have less trust in a company which makes claims for
its cameras which are not backed by either practical results or the
relevant physics. So I would place less trust in a company which claimed
4 stops of image stabilisation when independent tests showed only 1-2
stops, or claims that simply setting a high shutter speed was "image
stabilisation". The latter claim seems, to me at least, to be
deliberately misleading.

David

On Mon, 01 Sep 2008 05:42:50 GMT, "David J Taylor"
wrote:
: John McWilliams wrote:
: []
: I am not familiar with either camera, but in any event, it's
: irrelevant to my statement: JPEGs do not have inherently more dynamic
: range than the RAW from which it's processed.
:
: Of course not, but on its own, JPEG has the greater dynamic range. See my
: earlier post for the numbers.

But no camera currently in production captures more information for its JPEGs
than for its RAW images, right? So what practical difference does JPEG's
theoretically greater dynamic range make?

Bob

On Mon, 01 Sep 2008 05:40:18 +0000, David J Taylor wrote:

Alan Browne wrote:
[]
JPG's 8 bits/color is compressed DR, not more DR. The 'loss' is in
graduation 'tween colors. JPG cannot contain an expression of more
information than the original raw, compressed or otherwise.

The key point is that in-camera JPG leaves you with much less in terms
of options than post-processed raw.

With an 8-bit linear coding, the ratio between maximum and minimum
signal level is 255:1.

The typical RAW data is 12-bit or 14-1bit, having a ratio of max/min of
4095:1 or 16383:1.

With JPEG, taking 2.2 as the typical gamma correction, the ratio is
255^2.2:1, or about 200,000:1.

Agreed that JPEG compromises on the accuracy with which any particular
brightness can be represented, but the range of values which can be
represented is greater with JPEG than RAW.

Cheers,
David

Since, as I understand it, a RAW file is basically a dump of the sensor
data, you can't do any better than that. Theoretical limitations are one
thing - practical applications are another.