Sign in with
Sign up | Sign in
Your question

D70 raw vs jpeg

Tags:
Last response: in Digital Camera
Share
July 25, 2005 3:48:18 PM

Archived from groups: rec.photo.digital (More info?)

http://www.geocities.com/angels2000photos/raw.jpg
Some comparisons between RAW (converted with RSE) and ex-camera jpegs
from a D70.
No sharpening applied with RSE, jpgs shot in fine (best) setting. Same
lens, same time (give or take 5 seconds), no camera shake or focus
difference. Repeated several times, same result each time.
With jpeg basic and normal, resolution loss is far more pronounced.
Crops are 1:1.

Yes, you can see this difference in a (large) print.
With a sufficiently sharp lens, aliasing "jaggies" are usually visible
at the pixel level when shooting raw. I print with an Epson R1800,
which can resolve detail to (over) 600dpi. Whether it is the result of
ink droplets "blending" - or more probably anti-aliasing applied by the
printer drivers when they resample the image, I never see jaggies in
prints at all.

More about : d70 raw jpeg

July 25, 2005 3:48:19 PM

Archived from groups: rec.photo.digital (More info?)

The only difference I notice is in the color.


"frederick" <nomail@nomail.com> wrote in message
news:1122248823.556973@ftpsrv1...
> http://www.geocities.com/angels2000photos/raw.jpg
> Some comparisons between RAW (converted with RSE) and ex-camera jpegs from
> a D70.
> No sharpening applied with RSE, jpgs shot in fine (best) setting. Same
> lens, same time (give or take 5 seconds), no camera shake or focus
> difference. Repeated several times, same result each time.
> With jpeg basic and normal, resolution loss is far more pronounced.
> Crops are 1:1.
>
> Yes, you can see this difference in a (large) print.
> With a sufficiently sharp lens, aliasing "jaggies" are usually visible at
> the pixel level when shooting raw. I print with an Epson R1800, which can
> resolve detail to (over) 600dpi. Whether it is the result of ink droplets
> "blending" - or more probably anti-aliasing applied by the printer drivers
> when they resample the image, I never see jaggies in prints at all.
Anonymous
July 25, 2005 3:48:19 PM

Archived from groups: rec.photo.digital (More info?)

frederick wrote:

> http://www.geocities.com/angels2000photos/raw.jpg
> Some comparisons between RAW (converted with RSE) and ex-camera jpegs
> from a D70.
> No sharpening applied with RSE, jpgs shot in fine (best) setting.

Are you absolutely sure about that? The sharpness differences on the
samples you show should *not* be an effect of JPEG compression unless
there is some insane quality setting used on the camera.

The RSE shot looks like it has applied unsharp masking or some other
heuristic demosaicing fixup whether you asked for it or not. Look
carefully at the individual pixels in the L of De Luze..

It shows Gibbs ringing (as does the D and E). This strongly suggests
that the left hand image has been subject to unsharp mask or similar.

The image on the web is a JPEG and it can faithfully reproduce it.

Regards,
Martin Brown
Same
> lens, same time (give or take 5 seconds), no camera shake or focus
> difference. Repeated several times, same result each time.
> With jpeg basic and normal, resolution loss is far more pronounced.
> Crops are 1:1.
>
> Yes, you can see this difference in a (large) print.
> With a sufficiently sharp lens, aliasing "jaggies" are usually visible
> at the pixel level when shooting raw. I print with an Epson R1800,
> which can resolve detail to (over) 600dpi. Whether it is the result of
> ink droplets "blending" - or more probably anti-aliasing applied by the
> printer drivers when they resample the image, I never see jaggies in
> prints at all.
Related resources
Can't find your answer ? Ask !
Anonymous
July 25, 2005 3:48:19 PM

Archived from groups: rec.photo.digital (More info?)

On Mon, 25 Jul 2005 11:48:18 +1200, in rec.photo.digital frederick
<nomail@nomail.com> wrote:

>http://www.geocities.com/angels2000photos/raw.jpg
>Some comparisons between RAW (converted with RSE) and ex-camera jpegs
>from a D70.
>No sharpening applied with RSE

Be aware in RSE that settings of ZERO do NOT necessarily mean there is no
adjustment. IE sharpening set to zero does not mean no sharpening is
applied, IIRC.
----------
Ed Ruf Lifetime AMA# 344007 (Usenet@EdwardG.Ruf.com)
See images taken with my CP-990/5700 & D70 at
http://edwardgruf.com/Digital_Photography/General/index...
Anonymous
July 25, 2005 3:48:20 PM

Archived from groups: rec.photo.digital (More info?)

"RON" <joycegolf@adelphia.net> wrote in message
news:Jf-dnXPZtfe5rHnfRVn-2A@adelphia.com...
> The only difference I notice is in the color.
>


Seems to me that the RAW images are crisper then the JPG images. However
the color saturations is higher in the JPG images.

--

Rob
July 25, 2005 7:19:40 PM

Archived from groups: rec.photo.digital (More info?)

Robert R Kircher, Jr. wrote:
> "RON" <joycegolf@adelphia.net> wrote in message
> news:Jf-dnXPZtfe5rHnfRVn-2A@adelphia.com...
>
>>The only difference I notice is in the color.
>>
>
>
>
> Seems to me that the RAW images are crisper then the JPG images. However
> the color saturations is higher in the JPG images.
>
You can make the raw images look almost exactly like the jpgs by...
Gaussian blur: 1 pixel rle in Gimp / 0.5 pixel in PS.
The ex-camera jpgs are just not as sharp. It is not disputable - and is
not just my opinion. It's also not the only reason I almost always
shoot raw.
Yes, there is a difference in saturation. Neither is better, just
different at 100% default.
Anonymous
July 26, 2005 2:00:52 AM

Archived from groups: rec.photo.digital (More info?)

"frederick" <nomail@nomail.com> wrote in message
news:1122248823.556973@ftpsrv1...
> http://www.geocities.com/angels2000photos/raw.jpg
> Some comparisons between RAW (converted with RSE) and ex-camera jpegs

Sorry I do not have a Nikon - I have the Canon 1D and 300D. But maybe this
is still relevant:

My experience has been my JPEG images can look the exactly same as my RAW
files, if I get the exposure just perfect. I mean just, just, exactly,
perfect. But if I shoot with RAW, I later dial in perfect results every
time, even if I was off plus or minus one stop when captured. Journalists
and others you speak to might say "Don't use RAW as a crutch, and I agree,
somewhat. I'm getting better and better at hitting the mark the first time,
but I often don't and so why take a chance? Memory is cheap. People tell
newbies to stick with JPEG because it is easy. Well, it may be easier to
process but if the end result could be better, which would you choose?
Anonymous
July 26, 2005 7:39:47 AM

Archived from groups: rec.photo.digital (More info?)

I think the contrast of the RAW images is higher than the one of the JPEG
images. The resolution may be different, but it is very difficult to tell by
how much.

Can you repeat this using just B&W lines or a B&W test chart? This way, you
can rid this test from color differences that highly impact they way we
perceive the images.

Gregor

"frederick" <nomail@nomail.com> wrote in message
news:1122248823.556973@ftpsrv1...
> http://www.geocities.com/angels2000photos/raw.jpg
> Some comparisons between RAW (converted with RSE) and ex-camera jpegs from
> a D70.
> No sharpening applied with RSE, jpgs shot in fine (best) setting. Same
> lens, same time (give or take 5 seconds), no camera shake or focus
> difference. Repeated several times, same result each time.
> With jpeg basic and normal, resolution loss is far more pronounced.
> Crops are 1:1.
>
> Yes, you can see this difference in a (large) print.
> With a sufficiently sharp lens, aliasing "jaggies" are usually visible at
> the pixel level when shooting raw. I print with an Epson R1800, which can
> resolve detail to (over) 600dpi. Whether it is the result of ink droplets
> "blending" - or more probably anti-aliasing applied by the printer drivers
> when they resample the image, I never see jaggies in prints at all.
July 26, 2005 1:53:30 PM

Archived from groups: rec.photo.digital (More info?)

Martin Brown wrote:
> frederick wrote:
>
>> http://www.geocities.com/angels2000photos/raw.jpg
>> Some comparisons between RAW (converted with RSE) and ex-camera jpegs
>> from a D70.
>> No sharpening applied with RSE, jpgs shot in fine (best) setting.
>
>
> Are you absolutely sure about that? The sharpness differences on the
> samples you show should *not* be an effect of JPEG compression unless
> there is some insane quality setting used on the camera.
>
> The RSE shot looks like it has applied unsharp masking or some other
> heuristic demosaicing fixup whether you asked for it or not. Look
> carefully at the individual pixels in the L of De Luze..
>
> It shows Gibbs ringing (as does the D and E). This strongly suggests
> that the left hand image has been subject to unsharp mask or similar.
>
Don't look too close - it's a bottle of De Luze, St Émilion 1970. The
label is more than a bit shabby.
>
> The image on the web is a JPEG and it can faithfully reproduce it.
>
I think Ed Ruf answers that - a zero setting in rse does not mean no
sharpening is applied, even though there is zero set on the slider, and
the "apply sharpening" check box is unchecked. Likewise, with "normal"
sharpness in ex-camera jpegs, I don't know what the camera is doing in
terms of applying sharpening.
I don't think the (relative amount of) blur in ex-camera jpegs is the
result of jpeg compression, more likely the way that the camera deals
with raw conversion.
I haven't compared rse to Nikon Capture. I was using DCRaw, but was
encountering problems with visible coloured artifacts around high
contrast fine detail.
In any case, detail - or even an appearance of sharpness to match the
raw image cannot be recovered in the ex-camera jpgs by applying USM
judiciously. The nearest I can get to achieving a similar (illusion of)
sharpness is applying USM, radius 0.5, amount about 100, but this
results in significant haloes, and looks terrible.
It just is the way it is - you get sharper images from Nikon dslrs
shooting raw. Not news on a d70 or D2x, I expect that a D50 is the
same. It is quite significant - but probably only if you use sharp lenses.
Anonymous
July 26, 2005 1:53:31 PM

Archived from groups: rec.photo.digital (More info?)

On Tue, 26 Jul 2005 09:53:30 +1200, in rec.photo.digital frederick
<nomail@nomail.com> wrote:


>I think Ed Ruf answers that - a zero setting in rse does not mean no
>sharpening is applied, even though there is zero set on the slider, and
>the "apply sharpening" check box is unchecked. Likewise, with "normal"
>sharpness in ex-camera jpegs, I don't know what the camera is doing in
>terms of applying sharpening.

I just checked the user manual and unchecked this should stop sharpening
from being applied by that algorithm. However, the detail extraction would
appear to be an edge sharpening algorithm.

----------
Ed Ruf Lifetime AMA# 344007 (Usenet@EdwardG.Ruf.com)
See images taken with my CP-990/5700 & D70 at
http://edwardgruf.com/Digital_Photography/General/index...
July 26, 2005 3:00:44 PM

Archived from groups: rec.photo.digital (More info?)

Dave R knows who wrote:

> "frederick" <nomail@nomail.com> wrote in message
> news:1122248823.556973@ftpsrv1...
>
>>http://www.geocities.com/angels2000photos/raw.jpg
>>Some comparisons between RAW (converted with RSE) and ex-camera jpegs
>
>
> Sorry I do not have a Nikon - I have the Canon 1D and 300D. But maybe this
> is still relevant:
>
> My experience has been my JPEG images can look the exactly same as my RAW
> files, if I get the exposure just perfect. I mean just, just, exactly,
> perfect. But if I shoot with RAW, I later dial in perfect results every
> time, even if I was off plus or minus one stop when captured. Journalists
> and others you speak to might say "Don't use RAW as a crutch, and I agree,
> somewhat. I'm getting better and better at hitting the mark the first time,
> but I often don't and so why take a chance? Memory is cheap. People tell
> newbies to stick with JPEG because it is easy. Well, it may be easier to
> process but if the end result could be better, which would you choose?
>
>
I suspect that the in-camera processing in Canon cameras is better than
Nikon, and the difference in sharpness between raw and jpeg may be far
less apparent.
When I first got my D70, I took it away for a week and shot about 300
images, all jpeg. It is something I regret - I did not know how to use
the camera well, and there are too many images that are just okay, when
if shot raw, they could be great. Maybe I could do better now that I
know the camera better. I will never find out, because I will never use
ex-camera jpg again. When I print, I only print large sizes, 8 x 5
minimum (2 images per A4). I'd never just throw an image at the printer
without spending some time adjusting it, so dealing with the RAW
conversion is just one little added step to begin with. This may not be
what anyone else chooses to do - and if I was using a camera to shoot
500 images a day, then 500 times even 30 seconds adjusting each raw
image is most of a day gone in front of a computer, no time for
photography, and would drive you a little crazy I think. I suspect that
many photo journalists may find raw shooting way too tedious - and find
it much easier to just bracket / shoot in burst mode.
July 26, 2005 3:05:37 PM

Archived from groups: rec.photo.digital (More info?)

Ed Ruf wrote:

> On Tue, 26 Jul 2005 09:53:30 +1200, in rec.photo.digital frederick
> <nomail@nomail.com> wrote:
>
>
>
>>I think Ed Ruf answers that - a zero setting in rse does not mean no
>>sharpening is applied, even though there is zero set on the slider, and
>>the "apply sharpening" check box is unchecked. Likewise, with "normal"
>>sharpness in ex-camera jpegs, I don't know what the camera is doing in
>>terms of applying sharpening.
>
>
> I just checked the user manual and unchecked this should stop sharpening
> from being applied by that algorithm. However, the detail extraction would
> appear to be an edge sharpening algorithm.
>
Is the detail extraction included in the "apply sharpening" checkbox
option? Yeah - I know I should check it myself, but I hate reading
manuals. When I hit "F1" using RSE, a 72 page PDF appears, and my
eyelids start to feel very heavy. :-)
Anonymous
July 26, 2005 3:05:38 PM

Archived from groups: rec.photo.digital (More info?)

On Tue, 26 Jul 2005 11:05:37 +1200, frederick wrote:

>> I just checked the user manual and unchecked this should stop sharpening
>> from being applied by that algorithm. However, the detail extraction would
>> appear to be an edge sharpening algorithm.
>>
> Is the detail extraction included in the "apply sharpening" checkbox
> option? Yeah - I know I should check it myself, but I hate reading
> manuals. When I hit "F1" using RSE, a 72 page PDF appears, and my
> eyelids start to feel very heavy. :-)

Very heavy. So heavy that they have closed and your head is
starting to sag. When you wake up you will not remember this
conversation. But you WILL remember to use the Search function in
Acrobat Reader. Fast and accurate, it will leave you feeling light
(not lightheaded) and much more refreshed than if you had to search
through a paper manual instead.
July 26, 2005 8:02:00 PM

Archived from groups: rec.photo.digital (More info?)

GTO wrote:

> I think the contrast of the RAW images is higher than the one of the JPEG
> images. The resolution may be different, but it is very difficult to tell by
> how much.
>
> Can you repeat this using just B&W lines or a B&W test chart? This way, you
> can rid this test from color differences that highly impact they way we
> perceive the images.
>

The problem is that you will for sure get moire from such a test chart
pattern. Then the impression given is that moire is a big problem,
which for me it's never been. I actually think you can see quite a
difference in the existing images, regardless of colour difference.
Maybe the answer is to do the old newspaper resolution test shot. Non
repeating patterns should be less of a problem for moire than a test
chart, and are a closer representation of reality - if you aren't in the
textile business. If you want to do the test yourself, then the 18-70
should be sharp enough.
July 26, 2005 11:33:11 PM

Archived from groups: rec.photo.digital (More info?)

GTO wrote:
> I think the contrast of the RAW images is higher than the one of the JPEG
> images. The resolution may be different, but it is very difficult to tell by
> how much.
http://www.geocities.com/angels2000photos/test2.jpg.
The resolution of a real life subject is different by quite a bit....
It doesn't matter what I do, the difference is always there. I can do
what I like to the jpeg, but I can't recover what simply isn't there.
>
> Can you repeat this using just B&W lines or a B&W test chart? This way, you
> can rid this test from color differences that highly impact they way we
> perceive the images.
>
No - if I do that, then it will show that to shoot images of test charts
(as done in camera reviews using jpeg _or_ raw), then a camera that
produces noticeably _softer_ images of real life subjects can produce
_better_ test chart shots.
!