Disadvantage of RAW

[Glenn W.]

Well... you could buy a 40D, get 17 shot RAW buffer... buy more memory cards and hard drive space...

The 14bit RAW files on the 40D really eat ups space... like 13MB/image... but hey, memory is cheap. (not exactly free, but not that bad)... and like I say the 17 frame RAW buffer of the 40D does seem to be of reasonable size that it's not much of a issue for me...

You never give anything up by shooting in RAW.

Not true. I shoot raw 99% of the time, but I give up a deeper shot
buffer. Instead of 20 shots continuous on my 20D, I get 6. Not a
problem most of the time, but something that has to be factored in
occasionally.

 

[John Sheehy]

Maybe canon could have add Custom function to apply dark frame
substraction in RAW and bring it in firmware upgrade. That would not
be pure RAW, since it will not contain data as they are read from
sensor. But i think it will be aceptable, since it will bring lower
RAW noise with no loss of detail. Every pixel will have value from
open shutter substracted by value from closed shutter.

Dark frame subtraction always applies to RAW files in Canon DSLRs.

--
John

 

[Mark Shaxted]

--
Regards
Mark

 

[GaborSch]

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

 

[Mark Shaxted]

Let me quantify that...

DFS requires an equal exposure time at the same ambient temperature. This DOES NOT happen in current Canon hardaware (without CF-N #)!

--
Regards
Mark

 

[John Sheehy]

No it doesn't...

Read my post again ... s l o w l y

--
John

 

[John Sheehy]

The 14bit RAW files on the 40D really eat ups space...

And is a marketing gimmick for the most part; only useful for stacking huge numbers of images together. 14-bit RAW data should be a non-defaiult option, only. Thanks, Canon, for wasting time and CF and hard disk space.

--
John

 

[dwharrison]

Or alternatively, I could live with the equipment I have now and not spend additional money given that my wife was just laid off yesterday while on maternity leave, and therefore we're going to need to be cautious for a while.

At any rate, even on a 40D, the capacity of the buffer for raw is less than for jpeg. Not a problem for most uses, but it is true that buffer capacity and CF card capacity are reduced by shooting now. I think it's worth it for 99% of my usage, but that doesn't mean that raw doesn't have a price.

 

[MitchAlsup]

Since the instruction sets and capabilities are fundamentally
different, why would you assume that one would produce the same
results as the other

There is no such connection. The result does not depend on the actual
computer but on the algorythm.

But the result IS dependent upon the number of bits the processor uses to perform the computations! or store those results in memory.

Similar problems plagued floating point computations for decades before IEEE 754 (partially) standardized floating point arithmetic. Algorythms that ran just fine on a DEC VAX would crash and burn on a CDC 7600 or CRAY 1 due to sublte differeneces inthe way FL operations were performed and the precision of each result.
--
Mitch

 

[GaborSch]

But the result IS dependent upon the number of bits the processor
uses to perform the computations! or store those results in memory

No, it is not.

Similar problems plagued floating point computations for decades
before IEEE 754 (partially) standardized floating point arithmetic.
Algorythms that ran just fine on a DEC VAX would crash and burn on a
CDC 7600 or CRAY 1 due to sublte differeneces inthe way FL operations
were performed and the precision of each result

1. We are talking about operations on max five digit integers with the result in the same range. The intermediate results need to be of higher precision, but nothing of really high precision .

2. An algorythm specifying a model for such calculation does not contan how to achieve that on the bit level, but what is to be achieved. Except, of course, if those creating the specification, and/or those, who coded that, were suckers .

An interesting example for this is the JPEG encoding/decoding specification. It comes with the depreciated (but still the best) old style flow charts as explanation . The specification states, that you don't need to code according to that flowchart, but the result is supposed to be the same.

--
Gabor

http://www.panopeeper.com/panorama/pano.htm

 

[wiggims]

I found that when first using DPP I was getting unacceptably noisy images as well. What I soon realized is that by default it boosts the sharpening. Make sure that you do no sharpening in DPP. Do it with another program after the shot is converted. Process the same image with and without sharpening from DPP and you will easily see the difference in ugly noise that the sharpening creates.

--
http://wiggims.zenfolio.com/

http://www.pbase.com/wiggims/galleries

 

[wiggims]

In other words set the sharpening sliders to zero and then process the image.

--
http://wiggims.zenfolio.com/

http://www.pbase.com/wiggims/galleries

 

[IA123]

Thanks wiggims, I will give it a try.

IA

 

[thw]

-------------------------------------------
See the colors of my world in:
thw.smugmug.com