Why using RAW with 300D doesn't make sense to me

a_hamann said:

I saw a couple of threads and on-line essays raising the question
"is using RAW worth it?", and giving an enthusiastic "yes" as an
answer unless you know nothing about post-processing and only print
4x6. But the question itself implies some serious downsides such as
storage requirements, transfer and processing speed, future
compatibility when used for long-term archiving, etc.

After this discussion
http://www.digitalsecrets.net/Sony/AdvancedKnow4-Q.html and some
own experimentation with various quality settings I now use three
megapixel basic quality setting of the 300D for virtually every
shooting situation. Most of the time, I see no difference in image
quality, especially if there is the tiniest bit of camera shake or
lens softness/diffraction due to the choice of aperture.

As Phil noted in his rewiew of the 300D, the in-camera
interpolation algorithms used for downsizing are quite good, and I
find the 3MP setting produces pretty much pixel perfect "Foveon
quality" images. There's just not very much real loss in terms of
noise, detail, and visible JPG artifacts compared to 6MP high
quality. And, indeed, even RAW.

Don't get me wrong, I like great detail and image quality as much
as anyone. I would always use 6 megapixel or even 12 megapixel if
that translates to a real gain in quality. You can never have
enough for cropping. Also, I frequently run out of dynamic range,
and shooting RAW does increase dynamic range a little bit. I would
guess maybe one stop from what I've seen.

But is it worth the hassle? The Capture One tools for RAW are nice
but if you know just a few tricks in Photoshop CS you can do
(almost) the same adjustments on JPGs. And archiving 1000 images at
700kb per CD (my average observed size for 3MP Normal) versus 100
RAW shots at 7MB should make a difference in terms of organizing
and accessing their photos to anybody.

I like to view it this way: By using RAW versus 3MP basic with the
300D, you place yourself maybe 6 month (or less) ahead in the
technology progress game of getting the best possible quality for
the buck. If a professional photographers (expense not an issue)
claim that RAW is the ONLY way to go that basically means that the
work they did last year is worthless now :^)

For the "run of the mill" shooter RAW seems just a bit "geeky" to
me. If you really need this you also have the rational (if not the
cash) to trade in your camera for a new one every half year or so
:^)

Just a though, Andy.

Click to expand...

malc said:

Raw is simply the best feature of this camera. Take jpeg out of
your mind for a week, and just use raw. Download C1Rebel trial
version so that you can handle them with ease.

Try this experiment in both raw and jpeg

1) Shoot a too dark picture, and a blown highlights picture.
2) Adjust the brightness with EC in raw, and levels in JPEG.
3) Compare the results.

This camera is really not all of the hassles if you only use jpeg
mode. When you shoot in raw you will realize why it is all so
worth it.

Click to expand...

Kai Hui said:

There are plenty of 3MP cameras available, and the price now is
$200. Why do you want to spend $1000 (not tomention all the lenses
and other accessories) and use it as a 3mp camera. If you can not
tell the diff. between 3mp and 6mp, why bother to buy a 6mp camera?

Kai

Click to expand...

a_hamann said:

I like to view it this way: By using RAW versus 3MP basic with the
300D, you place yourself maybe 6 month (or less) ahead in the
technology progress game of getting the best possible quality for
the buck. If a professional photographers (expense not an issue)
claim that RAW is the ONLY way to go that basically means that the
work they did last year is worthless now :^)

Click to expand...

a_hamann said:

For the "run of the mill" shooter RAW seems just a bit "geeky" to
me. If you really need this you also have the rational (if not the
cash) to trade in your camera for a new one every half year or so
:^)

Click to expand...

a_hamann said:

Fuji's
dynamic range extension really appears to be the way to go. If you
search the Fuji forum you'll find that somebody hacked the firmware
got the two seperate images from the sensor, converted the data and
merged the two exposures in Photoshop. AMAZING results. The funny
thing is that Fuji's in-camera software is not able to do the same
thing.

Click to expand...

DRG said:

a_hamann said:

Fuji's
dynamic range extension really appears to be the way to go. If you
search the Fuji forum you'll find that somebody hacked the firmware
got the two seperate images from the sensor, converted the data and
merged the two exposures in Photoshop. AMAZING results. The funny
thing is that Fuji's in-camera software is not able to do the same
thing.

Click to expand...

Andy,

In my prior posting, I described the tradeoff between dynamic range
and resolution. One way in which this tradeoff is very explicit is
in Fuji's approach. They basically do exactly what you mention -
take two separate photos and merge the results. The Fuji software
may not make this obvious in that it doesn't allow you to see the
separate images. However, this is how their increased dynamic range
is achieved. They basically have two sensors in one, with
interleaved pixels. This allows two exposures to be taken at the
same time instead of one after the other in the more traditional
multiple-exposure approach for increasing DR. However, for a given
chip real estate, you need to decrease pixel count or pixel area to
fit two sensor in one. Now, since one set of pixels is being
exposed for highlights and the other for shadow, you can devote the
larger pixel area to the latter, for which noise would be more
apparent. That is, the two interleaved pixel sets need not be equal
in size. This optimizes signal to noise ratio better, and seems to
be Fuji's approach.

Just remember that no improvement comes for free. A sensor has a
given total area, which means it can only intercept and detect a
given number of photons for a given exposure length and lens
aperture. Thus, the total noise summed over the chip is going to be
difficult to reduce.

David

Click to expand...

a_hamann said:

I would certainly trade a small increase in noise for alot of
dynamic range. After all, the highlight sensor only needs a tiny
bit of real estate compared to the low light sensor, which would be
mainly responsible for noise performance.

BTW, The "hack" I mentioned just shows that Fuji didn't take full
advatage of their great hardware because of lousy software.

Click to expand...