Sigma DP2 - RAW versus JPG

[Don Ellis]

It’s official… if you’re shooting JPG with your DP2, you might as well donate it to the Salvation Army and invest in a pair of mediocre cameras for the same price.

I’m not in the business of doing photo tests, especially since I shoot RAW on every camera I have that does RAW… which is all of them. But I have heard some people say they thought JPG was pretty good, so today at lunch I took a few photos.

I’ll post two images for now, but I’ve taken others that show the difference between RAW and JPG is substantial – and lead me to the conclusion that if you bought a DP2 for image quality, you need to be shooting RAW.

Click on these images to download JPGs saved at 90% quality. The best way to do a comparison is to layer the JPG on top of the RAW in Photoshop (or any program that supports layers). Then click the top layer on and off to see the difference as you move around the image.

How I processed: I converted the RAWs to TIFs using Auto in SSP 3.5. Then I brought both TIF and JPG into Photoshop CS4 and hit Auto-Contrast. In the case of the Puddle, I set the white balance for both images. I didn’t sharpen anything.

I’ve named the files so they’re pretty easy to identify once you’ve downloaded them.

Puddle, RAW



Puddle, JPG



Bottles, RAW



Bottles, JPG



At first glance, and in these sizes, these photos may look similar, so I urge you to download the full-size photos if you're wavering between RAW and JPG. But that's all I'll do, is urge... I shoot RAW, so I really don't need the competition... please, shoot JPG... do me a favour.

Other photos that I took yesterday included trees. One of the things that impressed me when I was looking at DP2 images before I bought the camera was the definition of individual leaves. They are much more muted in JPG than RAW. If I find time, I’ll post an image of those later on.

For now, you get water… in bottles and on the sidewalk. Hopefully, you’ll forgive the tripod leg in the puddle photo -- it's there to point out the big "D" (for Don) just above it. That's your Seal of Authenticity.

Don

Ok, I'm waiting for artwork so I might as well do something constructive... here are the "tree" shots...

Trees, RAW



Trees, JPG



After you download the full-size files (by clicking on these two), take a look at the trees on the left and in the centre, at the bushes along the sidewalk, and at the glass above the Chanel sign.

Again, layering is more effective than side-by-side comparison in revealing subtle differences.

 

[DjordjeJ]

Thanks for examples Don. Yes difference is visible,especially with puddle photos.
But, how come JPG version has more reflections than RAW version, like You used PL in JPG version. I assume, You didnt.

 

[Don Ellis]

Thanks for examples Don. Yes difference is visible,especially with puddle photos.
But, how come JPG version has more reflections than RAW version, like You used PL in JPG version. I assume, You didnt.

Very good question... it is like a polarizer, isn't it? I noticed that. Not even I am unscientific enough to use a polarizer in one shot and not the other. So now we have another benefit: Use RAW and get a Free Polarizer!

I also noticed that with a shot I did yesterday... the RAW image "saw" through some glass to pick up detail on the other side, which the JPG did not. (After I said this, I decided to post the images I was referring to... Trees, added to the first message.)

Don

 

[Tim]

Thanks for posting this Don.

It seems curious!? The in camera RAW to jpg conversion must more than just a set of parameters. For example the noise reduction algorithm in SPP must be somewhat more capable than the in camera algorithm? Perhaps you need more cpu performance for the Mac/Win than the in camera cpu can handle?

Thoughts?

 

[Don Ellis]

Thanks for posting this Don.

It seems curious!? The in camera RAW to jpg conversion must more than just a set of parameters. For example the noise reduction algorithm in SPP must be somewhat more capable than the in camera algorithm? Perhaps you need more cpu performance for the Mac/Win than the in camera cpu can handle?

Thoughts?

I remember an American novels exam I once had in college -- we had 50 minutes to write our essay and I spent 25 minutes trying to understand the question.

Same situation here... but after spending a few minutes re-reading, I see the one example you mention is noise reduction -- in-camera versus in SSP. I'm not sure how that's a factor at all, since these were shot at ISO50 and shouldn't need any noise reduction at all.

Whatever information is deleted or revealed in these photos is, I would guess, a product of other factors.

One thing to consider is that if you're shooting or saving JPG, you have only 8 bits of data to work with; whereas, the DP2 RAW files have 12 bits of data.

To my understanding, that's the reason you should convert to 16-bit TIFs. If I may offer an example for people who don't quite understand this:

[pamo]If you think of 8 bits as a gallon, then 16 bits is two gallons, and 12 bits is a gallon and a half.

If you shoot in RAW, you have a gallon and a half of information. If you dump that gallon and a half into a JPG bucket or an 8-bit TIF bucket that only holds a gallon, you're going to lose half a gallon of information.

But if you dump it into a two-gallon, 16-bit TIF bucket, none of the information is lost because the bucket is big enough to hold it all. [/pedantic analogy mode off]

Now, there's every chance that you're talking about more than just in-camera processing of noise, so I'll let you have a second chance at questions.

Don

 

[Tim]

Hi Don,

I'm convinced that the RAW that is processed in SSP yields a better image - your samples show that - thanks for your effort. I am not really needing the in camera jpgs, this is just a curiosity discussion for me. I am happy to process my RAW files.

Correct me if I am wrong but all digital camera's take RAW images, most then process it to a jpg only in camera. Unless you have a camera that can also save the RAW or in the case of the DP2 only saves the RAW.

It remains a mystery (to me) as to what goes on in a camera to process the RAW file to a jpg - who sets these parameters at Sigma or any other camera manufacturer for that matter? Someone (the camera designer/s?) make choices regarding the RAW-jpg in camera parameters.

At least we get great pictures! :bugeyes:

 

[Don Ellis]

I'm convinced that the RAW that is processed in SSP yields a better image – your samples show that – thanks for your effort.

You’re welcome… these things are often as much for me as anyone else. I’ve always shot RAW but these photos, and a few others I didn't post, were enough to remind me never to shoot JPG.

Correct me if I am wrong but all digital camera's take RAW images, most then process it to a jpg only in camera.

That sounds right, anyway. Once the sensor captures the Really RAW data, it would be converted to JPG (or to RAW with your settings).

Unless you have a camera that can also save the RAW or in the case of the DP2 only saves the RAW.

Actually, there’s a full-size JPG buried in the DP2’s RAW file. It’s extracted (not converted) with the “Convert to JPG file” command.

It remains a mystery (to me) as to what goes on in a camera to process the RAW file to a jpg – who sets these parameters at Sigma or any other camera manufacturer for that matter? Someone (the camera designer/s?) make choices regarding the RAW-jpg in camera parameters.

Was that a question or an answer? Or a question and answer?

Designers, engineers, and maybe even a photographic consultant all get together over cocktails and figure out what looks best to them and to the average photographer based on criteria they’ve amassed or guesses they’ve made. Then the engineers include tweaks so people can set sharpness, contrast, saturation, etc., etc. And there you go.

At least that’s how it works in my mind. There could be a dartboard involved.

At least we get great pictures! :bugeyes:

Well, some of us do... I'm still working on it. But it's not the camera's fault.

Don