RAW & best jpg appear identical on Sony DSC-R1

On 13 Apr, 21:06, Alfred Molon wrote:
In article .com,
says...



I see no reason in principle why a high quality jpg should not be as
good as a RAW file, but the conventional wisdom seems to be that RAW
files have the ultimate edge on fine detail. So having been impressed
by the image quality of the highest resolution jpg images from my Sony
DSC-R1 I was prepared to be even more impressed by the quality from
those huge RAW files.

Not being able to see any significant difference between the too, I
blamed handshake, poor focus, etc.. I ended up taking a photograph of
a brick wall with very carefully adjusted focus from a solid tripod
with delayed shutter release. And I still couldn't see any significant
difference. The jpg files by the way seem to vary between 3 and
4.9Mbytes in size.

So if I can't see any significant difference at any magnification
between these large jpgs and the much larger RAW files, is there any
advantage to be gained from RAW files, for this camera, of course?

Actually JPEG files of the R1 suck terribly. On the one hand, they do
not have as much resolution as a properly processed RAW files from the
R1. Do a test, shooting RAW+JPEG.

I did several careful tests involving a variety of subject matter. I
could detect occasional very small differences down at the pixel
level, but only by looking very carefully at very high magnifications.

Then there is the issue of the white
balance, which most of the time is off in R1 JPEG files,

I can only speak from experience of my own R1, which has impressed me
with how good its natural white balance is. I viewed the RAW files
using Sony's own RAW file processor, and that gave identical colour
balance to the jpgs from the camera. Each of my test subjects did
contain at least one small area of pure white to test the white
balance, and I could find nothing to criticise in it. However, all
tests were out of doors at midday, nothing difficult like artificial
light or low sun. I compared the images by arranging equal sized high
magnifications of the same portions of the image, so that I could
click instantly back and forth between the two images to see if
anything moved. which is a very sensitive test.

and lastly you
have a better control over exposure with RAW files.

I'm quite prepared to accept that if I want to do more than a little
post processing that it would be better to start with a RAW file. My
question concerned detail resolution.

Chris Malcolm

"Chris Malcolm" wrote in message
oups.com...
I see no reason in principle why a high quality jpg should not be as
good as a RAW file, but the conventional wisdom seems to be that RAW
files have the ultimate edge on fine detail. So having been impressed
by the image quality of the highest resolution jpg images from my Sony
DSC-R1 I was prepared to be even more impressed by the quality from
those huge RAW files.

Not being able to see any significant difference between the too, I
blamed handshake, poor focus, etc.. I ended up taking a photograph of
a brick wall with very carefully adjusted focus from a solid tripod
with delayed shutter release. And I still couldn't see any significant
difference. The jpg files by the way seem to vary between 3 and
4.9Mbytes in size.

So if I can't see any significant difference at any magnification
between these large jpgs and the much larger RAW files, is there any
advantage to be gained from RAW files, for this camera, of course?

Shhh - careful. The emperor does not take kindly to having his nudity
pointed out.

Those who want to make comments about raw will kindly confine themselves to
theoretical discussions of bit depth, posterization, histograms, data loss,
and the ability of certain gifted people to see the "obvious" advantage of
raw processing for well-exposed originals. The use of actual image examples
to show this makes any difference is strictly forbidden.
--
Mike Russell
www.curvemeister.com/forum/

On Apr 14, 5:04 am, "Mike Russell" -
MOVE wrote:
"Chris Malcolm" wrote in message

oups.com...



I see no reason in principle why a high quality jpg should not be as
good as a RAW file, but the conventional wisdom seems to be that RAW
files have the ultimate edge on fine detail. So having been impressed
by the image quality of the highest resolution jpg images from my Sony
DSC-R1 I was prepared to be even more impressed by the quality from
those huge RAW files.

Not being able to see any significant difference between the too, I
blamed handshake, poor focus, etc.. I ended up taking a photograph of
a brick wall with very carefully adjusted focus from a solid tripod
with delayed shutter release. And I still couldn't see any significant
difference. The jpg files by the way seem to vary between 3 and
4.9Mbytes in size.

So if I can't see any significant difference at any magnification
between these large jpgs and the much larger RAW files, is there any
advantage to be gained from RAW files, for this camera, of course?

Shhh - careful. The emperor does not take kindly to having his nudity
pointed out.

Those who want to make comments about raw will kindly confine themselves to
theoretical discussions of bit depth, posterization, histograms, data loss,
and the ability of certain gifted people to see the "obvious" advantage of
raw processing for well-exposed originals. The use of actual image examples
to show this makes any difference is strictly forbidden.
--

While I agree in general that higher bit depth and starting from raw
are both overrated (not that I'm much good at processing, but...),
there are times when it does make a difference, if for no other reason
than that it offers greater control over various conversion
parameters. For example,
http://www.pbase.com/al599/image/75823118
and
http://www.pbase.com/al599/image/75823122
are crops, one from the camera's jpeg and the other from a raw
conversion I made. While both have numerous problems (mostly related
to my "processing" experiments), I think colour smoothness is much
better in the converted example (admittedly, perhaps it's possible
that I could get the jpeg to look better, but I did not manage to).
This has more to do with the type and amount of NR done by the camera,
I suppose, but it was set to "off" (which isn't off, it still tries to
smooth the colour channels at higher ISOs), so if you don't like the
result, tough. Or this
http://www.pbase.com/al599/image/73360346
and this
http://www.pbase.com/al599/image/73422697
(these are quite large in full-size). Look at the gray book on the
right: the camera jpeg renders it as smooth gray, while in fact it is
not! Note that here NR was "low" for the jpeg, and I did some extra NR
and sharpening on the raw converted result (I could not get the jpeg
to look like that no matter what I tried; the jpeg I have there is
completely unprocessed, so can be improved, but not to the level of
the raw converted file-at least not by me). Again, it primarily has to
do with the amount of NR applied.

Also with some converters one does get slightly more resolution than
with camera jpegs, but I found the difference to be completely
inconsequential. What is useful is if the noise is rendered as higher
frequency, because high frequency noise prints much better (and I
would think that a converter that gives higher resolution would also
give a noise spectrum with higher frequencies dominating-but did I
plot spectra? No!).

But in general I think you're right, the differences are minute and
only significant in fairly unusual cases, and the price paid (in time
and effort) usually not worth it.

"acl" wrote in message
ps.com...
On Apr 14, 5:04 am, "Mike Russell" -
MOVE wrote:
"Chris Malcolm" wrote in message

oups.com...



I see no reason in principle why a high quality jpg should not be as
good as a RAW file, but the conventional wisdom seems to be that RAW
files have the ultimate edge on fine detail. So having been impressed
by the image quality of the highest resolution jpg images from my Sony
DSC-R1 I was prepared to be even more impressed by the quality from
those huge RAW files.

Not being able to see any significant difference between the too, I
blamed handshake, poor focus, etc.. I ended up taking a photograph of
a brick wall with very carefully adjusted focus from a solid tripod
with delayed shutter release. And I still couldn't see any significant
difference. The jpg files by the way seem to vary between 3 and
4.9Mbytes in size.

So if I can't see any significant difference at any magnification
between these large jpgs and the much larger RAW files, is there any
advantage to be gained from RAW files, for this camera, of course?

Shhh - careful. The emperor does not take kindly to having his nudity
pointed out.

Those who want to make comments about raw will kindly confine themselves
to
theoretical discussions of bit depth, posterization, histograms, data
loss,
and the ability of certain gifted people to see the "obvious" advantage
of
raw processing for well-exposed originals. The use of actual image
examples
to show this makes any difference is strictly forbidden.
--

While I agree in general that higher bit depth and starting from raw
are both overrated (not that I'm much good at processing, but...),
there are times when it does make a difference, if for no other reason
than that it offers greater control over various conversion
parameters. For example,
http://www.pbase.com/al599/image/75823118
and
http://www.pbase.com/al599/image/75823122
are crops, one from the camera's jpeg and the other from a raw
conversion I made. While both have numerous problems (mostly related
to my "processing" experiments), I think colour smoothness is much
better in the converted example (admittedly, perhaps it's possible
that I could get the jpeg to look better, but I did not manage to).
This has more to do with the type and amount of NR done by the camera,
I suppose, but it was set to "off" (which isn't off, it still tries to
smooth the colour channels at higher ISOs), so if you don't like the
result, tough. Or this
http://www.pbase.com/al599/image/73360346
and this
http://www.pbase.com/al599/image/73422697
(these are quite large in full-size). Look at the gray book on the
right: the camera jpeg renders it as smooth gray, while in fact it is
not! Note that here NR was "low" for the jpeg, and I did some extra NR
and sharpening on the raw converted result (I could not get the jpeg
to look like that no matter what I tried; the jpeg I have there is
completely unprocessed, so can be improved, but not to the level of
the raw converted file-at least not by me). Again, it primarily has to
do with the amount of NR applied.

Also with some converters one does get slightly more resolution than
with camera jpegs, but I found the difference to be completely
inconsequential. What is useful is if the noise is rendered as higher
frequency, because high frequency noise prints much better (and I
would think that a converter that gives higher resolution would also
give a noise spectrum with higher frequencies dominating-but did I
plot spectra? No!).

But in general I think you're right, the differences are minute and
only significant in fairly unusual cases, and the price paid (in time
and effort) usually not worth it.

I appreciate your providing examples, and also the fact that we don't differ
significantly, at least in general terms, in our assessment of the value of
raw over in-camera processing.
--
Mike Russell
www.curvemeister.com/forum/

In article .com,
says...

I did several careful tests involving a variety of subject matter. I
could detect occasional very small differences down at the pixel
level, but only by looking very carefully at very high magnifications.

I've been using the R1 for over a year now, and have been observing the
same thing again and again. The JPEGs simply do not have the resolution
you get if you process that RAW files. Phil Askey reaches the same
conclusion in his dpreview test:
http://www.dpreview.com/reviews/sonydscr1/page19.asp

"My only disappointment with the DSC-R1 was that its JPEG images out-of-
the-camera didn't have that "digital SLR look". They appear a little
over-processed and over-sharpened for my liking and didn't exhibit the
per-pixel sharpness we're used to seeing from digital SLR images (yes
you can reduce sharpening and post-process every image).

You can get this per-pixel crisp appearance from the R1 but only by
shooting RAW (at 20 MB a go) and then converting in a third party
converter (currently only Adobe Camera RAW). It's a pity that Sony
appear to have just dropped the same 'consumer grade' algorithms (for
demosaicing, sharpening, color) into the DSC-R1 when it really would
have gained considerably from something a little more sophisticated. It
would also be refreshing to see more sensor manufacturers taking a risk
on using lighter anti-alias filter which would mean images would require
less sharpening. "

I can only speak from experience of my own R1, which has impressed me
with how good its natural white balance is. I viewed the RAW files
using Sony's own RAW file processor, and that gave identical colour
balance to the jpgs from the camera. Each of my test subjects did
contain at least one small area of pure white to test the white
balance, and I could find nothing to criticise in it. However, all
tests were out of doors at midday, nothing difficult like artificial
light or low sun. I compared the images by arranging equal sized high
magnifications of the same portions of the image, so that I could
click instantly back and forth between the two images to see if
anything moved. which is a very sensitive test.

Maybe we have different standards? I get these shots by processing R1
RAWs:
http://www.molon.de/galleries/Morocco/Merzouga/Sand/

The out-of-the-camera R1 JPEGs have much different colours.
--

Alfred Molon
------------------------------
Olympus 50X0, 7070, 8080, E300, E330, E400 and E500 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site

Those who want to make comments about raw will kindly confine themselves to
theoretical discussions of bit depth, posterization, histograms, data loss,
and the ability of certain gifted people to see the "obvious" advantage of
raw processing for well-exposed originals. The use of actual image examples
to show this makes any difference is strictly forbidden.

When I looked into this I concluded the differences *were* obvious. Not
for JPEG artifacts - that's complete red herring, as the OP has spotted
- but for detail in the shadows and highlights the benefits of RAW are
plain to see.

Do my examples not count?

http://www.derekfountain.org/raw_vs_jpeg.php

--
Derek Fountain on the web at http://www.derekfountain.org/

"Derek Fountain" wrote in message
...
Those who want to make comments about raw will kindly confine themselves
to
theoretical discussions of bit depth, posterization, histograms, data
loss,
and the ability of certain gifted people to see the "obvious" advantage
of
raw processing for well-exposed originals. The use of actual image
examples
to show this makes any difference is strictly forbidden.

When I looked into this I concluded the differences *were* obvious. Not
for JPEG artifacts - that's complete red herring, as the OP has spotted
- but for detail in the shadows and highlights the benefits of RAW are
plain to see.

Do my examples not count?

http://www.derekfountain.org/raw_vs_jpeg.php

Your examples certainly do count. My statement was tongue in cheek, and
aimed at people who believe that "raw is mo better" in all situations, and
do not take advantage of the ability to individually adjust each raw image,
or even groups of raw images taken under similar circumstances.

An extreme example of this would be someone who posted (to another forum)
that batch processing of thousands of raw images from each day's shoot was
taking many hours each day, and what system should he upgrade to to get the
speed up. This, IMHO, is mere genuflection to the great god of raw, and is
unlikely to result in an increase in quality.
--
Mike Russell
www.curvemeister.com/forum/

Mike Russell wrote:
Your examples certainly do count. My statement was tongue in cheek...

Oh right. I completely missed that. It was late... }

--
Derek Fountain on the web at http://www.derekfountain.org/

"Mike Russell" -MOVE wrote in
t:

Shhh - careful. The emperor does not take kindly to having his nudity
pointed out.

Those who want to make comments about raw will kindly confine
themselves to theoretical discussions of bit depth, posterization,
histograms, data loss, and the ability of certain gifted people to see
the "obvious" advantage of raw processing for well-exposed originals.

That's because most such comparisons don't take advantage of the real
differences. RAW is a different medium, with more potential, and if you
don't *SHOOT* to use it, it goes to waste for the most part, and is only
a safety net for exposure and WB errors.

I'll shoot any DSLR with more positive EC in RAW mode than I would in
JPEG mode, because the RAW catpure is actually a lesser ISO (if you gauge
ISO by the exposure index required for a fixed highlight headroom), with
a higher signal-to-noise ratio, which you will not benefit from if you
expose for the JPEG, which is what 99.99% of the people "trying RAW" do.

The use of actual image examples to show this makes any difference is
strictly forbidden.

Same shutter actuation:

http://www.pbase.com/jps_photo/image/25925162

http://www.pbase.com/jps_photo/image/25925164



--


John P Sheehy

On Apr 13, 10:16 am, "Chris Malcolm" wrote:
I see no reason in principle why a high quality jpg should not be as
good as a RAW file, but the conventional wisdom seems to be that RAW
files have the ultimate edge on fine detail. So having been impressed
by the image quality of the highest resolution jpg images from my Sony
DSC-R1 I was prepared to be even more impressed by the quality from
those huge RAW files.

Not being able to see any significant difference between the too, I
blamed handshake, poor focus, etc.. I ended up taking a photograph of
a brick wall with very carefully adjusted focus from a solid tripod
with delayed shutter release. And I still couldn't see any significant
difference. The jpg files by the way seem to vary between 3 and
4.9Mbytes in size.

So if I can't see any significant difference at any magnification
between these large jpgs and the much larger RAW files, is there any
advantage to be gained from RAW files, for this camera, of course?

Chris Malcolm

Although really only useful for static scenes like landscapes and
still life, also consider RAW for creating High Dynamic Range (HDR)
images. You do need Adobe PhotoShop, and some more info about HDR can
be found he http://www.luminous-landscape.com/tutorials/hdr.shtml


Craig --.
www.craigwasselphotoart.com