Can you trust your RAW converter?

BigAl-sa

Executive Member
Joined
Dec 26, 2006
Messages
6,757
Reaction score
317
Location
Pretoria
As most may know, I've been fiddling with Linux, and part of the fiddling would be trying to get my photography stuffies working. So I happened to be looking at UFRaw with Gimp, and found really terrible results. This meant as a good scientist I had to compare results with other converters that I could find.

These pics should be viewed at 100%, so click on them to get the bigger pic.

From l to r: Raw Therapee 2.4 beta 4; UFRaw 0.13; DPP 3.2 (converters all with their default settings).


From t to b: Raw Therapee 2.4 beta 4; UFRaw 0.13; DPP 3.2


Note that both RT and UFR are applying saturation and sharpening to the pics. Look at the top of the tree in the UFR image and you will see awful haloes on the edge of the tree and sky - the mast on the left is also almost invisible. I'm not sure how much is due to DCRaw, which both RT and UFR use as the conversion engine.

Camera 40d, neutral picture style, 0 sharpening and 0 saturation, so the output from DPP is more what I'd expect.
 
Did you try it with a real OS or just Linux? :P

For a canon surely the best yard stick would be DPP? If you were using a real OS
 
why are you using Linux for image editing
Windows and Mac are mature

do the jpegs work?
 
why are you using Linux for image editing
Windows and Mac are mature
Got nothing to do with editing, it's about the colour. [The Linux colour calibration software I'm using for my colorimeter is far more sophisticated than the Windows/Mac version that comes with it :p]

do the jpegs work?
Not sure what you're getting at here??
 
Here another comparison, this time with a RAW file from the Canon 350d. These pics are much closer in their conversion, showing that dcraw's algorithms for converting 350d RAW images are *much* better.

Again, t to b: Raw Therapee 2.4 beta 4; UFRaw 0.13; DPP 3.2 (DPP again in a virtualbox, the others in Linux)



Methinks it would be interesting to see how well Adobe managed to crack the 40d's RAW files by comparing ACR 4.1 with DPP. Anyone want to have a go?
 
Last edited:
Here another comparison, this time with a RAW file from the Canon 350d. These pics are much closer in their conversion, showing that dcraw's algorithms for converting 350d RAW images are *much* better.
But the top two almost identical ones differ greatly from the DPP which you would expect, coming from canon, would have the most accurate colours.

Are you absolutely certain that the virtualization is accurately displaying colour? Are you able to compare a virtualized DPP with one running in windows (on the same hardware)
 
But the top two almost identical ones differ greatly from the DPP which you would expect, coming from canon, would have the most accurate colours.

Are you absolutely certain that the virtualization is accurately displaying colour? Are you able to compare a virtualized DPP with one running in windows (on the same hardware)
Will do. Need to boot into the win disk sometime soon, so will put the VBox DPP images somewhere where I can access them.

I'm really sure it's nothing to do with Vbox though. I have seen this kind of difference between DPP and RT before with earlier versions in win.

I think what folks need to realise is that the camera manufacturers do not make the internals of their RAW file structures available to the public. Even Adobe has to hack the RAW files from the new cameras to get ACR to work. This is why they are pushing so hard for the manufacturers to accept their DNG file format instead of independent RAW formats.
 
But the top two almost identical ones differ greatly from the DPP which you would expect, coming from canon, would have the most accurate colours.

Are you absolutely certain that the virtualization is accurately displaying colour? Are you able to compare a virtualized DPP with one running in windows (on the same hardware)
Ok, I did the DPP conversion in XP proper, same PC, same monitor, same calibration curve and I can see very little difference (same spot above eye: (222, 170, 255) vs (221, 175, 255)).



So, I think I can safely say that the virtualbox is not influencing the colours.

There's a newer version of UFRaw out, so I'll have a look at that too.

So, I ask again, can someone compare DPP to ACR, especially with a 40d? It would be nice to see how DPP compares with the other RAW converters too.
 
So, I ask again, can someone compare DPP to ACR, especially with a 40d? It would be nice to see how DPP compares with the other RAW converters too.
Sorry - I dont have a 40D. I can tape over the 2nd '0' on my 400D if that helps . . . :p
 
Sorry - I dont have a 40D. I can tape over the 2nd '0' on my 400D if that helps . . . :p
:D

I'm sure a test of your Mac software vs DPP would be interesting too. They would also have had to hack the RAW file to build a converter.
 
:D

I'm sure a test of your Mac software vs DPP would be interesting too. They would also have had to hack the RAW file to build a converter.
I've got ACR, Aperture, and DPP.

If you want to upload a sample cr2 I'll push them through - from there I'm guessing you'd want a straight to jpg with no other corrections apart from the ones the software applies by default?
 
Last edited:
I've got ACR, Aperture, and DPP.

If you want to upload a sample cr2 I'll push them through - from there I'm guessing you'd want a straight to jpg with no other corrections apart from the ones the software applies by default?
Missed this one somewhere along the line :( Why don't you try it on one of your own RAW files? (Don't have anywhere to upload anything other than non-RAW pics at the mo).

Found this comment where the guy is thinking along the same lines as me.
 
Missed this one somewhere along the line :( Why don't you try it on one of your own RAW files? (Don't have anywhere to upload anything other than non-RAW pics at the mo).

Found this comment where the guy is thinking along the same lines as me.
Not sure how relevant it would be - the 400D is about a year older than the 40D - much more time for everyone to tweak their settings.
 
Not sure how relevant it would be - the 400D is about a year older than the 40D - much more time for everyone to tweak their settings.
You're right of course. Only have to look at my 350d pics in this thread.
 
Top
Sign up to the MyBroadband newsletter
X