KenRockwell.com: Nikon D200 Image Quality Settings

[Jason Dunn]

http://www.kenrockwell.com/nikon/d200/quality-settings.htm

"What's the best image quality (QUAL) setting? JPG or raw, or raw + JPG? If JPG, should you use FINE, NORMAL or BASIC? Should you use Size Priority or Optimal Quality JPG Compression? How about Large, Medium or Small image dimensions? If raw, should you use compressed or uncompressed? In a fantasy world you'd use uncompressed NEF or FINE JPG, but after you shoot for a while you'll get logjammed with all the data you have to store, forward, process and archive. Ideally we want a setting that gives us spectacular quality with as small a file size as possible. This also lets you get the greatest number of shots on a card and download them the fastest."

I enjoy Ken Rockwell's writing because he often challenges the norms that we all assume to be true. If you have a big storage card, it's best to shoot your photos in the highest resolution, with the least compression, right? Well, maybe not. Ken shoots some photos and does comparisons here that may surprise you - I did some tests of my own a few months back, comparing JPEG Fine to JPEG Normal, and was unable to discern a difference with any image I shot, even at 200% zoom. So since then, when I shoot JPEG, I leave it at JPEG Normal. This article is a must-read for any digital photographer, even if you own a different digital camera. I tend to slightly tweak most images, even JPEGs, after shooting, so I don't know that I'd ever want to go down to JPEG Basic, but even after one re-save of a JPEG Normal file I have yet to see any artifacts. Sometimes you can make your life easier by second-guessing your assumptions!

I take issue with one of Ken's recommendations though: dropping the resolution of the image. Near-invisible compression artifacts are one thing, but a smaller image at a lower resolution is quite another. Twenty years from now we may be staring at 12,000 by 10,000 pixel resolution monitors, so having images with the largest "physical" pixel dimensions is important. I look back now at images I took with my first digital camera, a Kodak DC265, and the 1152 x 768 resolution images don't even fill the screen of any of my monitors.

[Jason Dunn]

http://www.kenrockwell.com/nikon/d200/quality-settings.htm

"What's the best image quality (QUAL) setting? JPG or raw, or raw + JPG? If JPG, should you use FINE, NORMAL or BASIC? Should you use Size Priority or Optimal Quality JPG Compression? How about Large, Medium or Small image dimensions? If raw, should you use compressed or uncompressed? In a fantasy world you'd use uncompressed NEF or FINE JPG, but after you shoot for a while you'll get logjammed with all the data you have to store, forward, process and archive. Ideally we want a setting that gives us spectacular quality with as small a file size as possible. This also lets you get the greatest number of shots on a card and download them the fastest."

I enjoy Ken Rockwell's writing because he often challenges the norms that we all assume to be true. If you have a big storage card, it's best to shoot your photos in the highest resolution, with the least compression, right? Well, maybe not. Ken shoots some photos and does comparisons here that may surprise you - I did some tests of my own a few months back, comparing JPEG Fine to JPEG Normal, and was unable to discern a difference with any image I shot, even at 200% zoom. So since then, when I shoot JPEG, I leave it at JPEG Normal. This article is a must-read for any digital photographer, even if you own a different digital camera. I tend to slightly tweak most images, even JPEGs, after shooting, so I don't know that I'd ever want to go down to JPEG Basic, but even after one re-save of a JPEG Normal file I have yet to see any artifacts. Sometimes you can make your life easier by second-guessing your assumptions!

I take issue with one of Ken's recommendations though: dropping the resolution of the image. Near-invisible compression artifacts are one thing, but a smaller image at a lower resolution is quite another. Twenty years from now we may be staring at 12,000 by 10,000 pixel resolution monitors, so having images with the largest "physical" pixel dimensions is important. I look back now at images I took with my first digital camera, a Kodak DC265, and the 1152 x 768 resolution images don't even fill the screen of any of my monitors.

[Vincent Ferrari]

I think, as usual, he's off the mark. Close, but off the mark.

Here's my philosophy:

1. If your camera shoots RAW, shoot RAW. Always 100% of the time use as much data as your camera is willing to provide you. If you're not going to shoot RAW and only shoot JPG, you can get comparable results shooting with any higher end P&S. Seriously.

2. There's no magic in third party RAW decoders. The spec for RAW is given to these manufacturers so they can decode it in the first place. Adobe doesn't reverse-engineer Canon or Nikon's spec; they use their spec. There are no features in a camera's raw spec that aren't available to any program that can read it. His investigation into the topic amounts to, "I don't use them, but I know." Whatever dude. If you want to argue convenience, go for it. I agree on that point. For convenience sake, I use Canon's software exclusively and as tempting as Aperture is, I won't touch it because Canon's works so well for me.

3. Is he honestly making the point that people can't tell the difference between different levels of JPG? Sure, Ken, you can shoot lower quality JPGs if you plan on not cropping anything you shoot and using it in medium to small sized prints. Look at his samples. The differences are glaringly obvious. You could make the point that at the full size, the picture is not terribly different. That's obvious. One of the advantages to cropping a picture is that you can mess up the framing and still fix it.

4. Finally, there's one thing Ken always seems to miss when discussing RAW. Most people that aren't great professional photographers don't get every shot right on the first try. With RAW, you have the opportunity to almost take the photo over a second time. You can simulate exposure compensation in JPGs, but it's obviously better in RAW (do some sample tests. RAW has saved my ass on numerous occasions). So is White Balancing (something you cannot do effectively in JPG; I don't care who says you can), sharpening (If you sharpen a JPG, you sharpen artifacts; fact, not opinion), and even color temperature.

Is RAW slower to work with? Of course. Is it more storage hungry? Absolutely. Is it worth using for most people? Depends. Should you take advantage of it if your camera can do it? You'd be absolutely foolish not to, despite the page after page after page Ken's written that says otherwise.

[Jerry Raia]

Still new to this but my cameras will shoot both so I have been shooting RAW + JPEG. If I like the JPEG that's the end of it other than cropping etc. If I don't then I take the RAW to Adobe. Up till now I have been using the largest size and lowest compression for the JPEG side. I read what he wrote and vincenzosi's comments and am convinced I don't need to change what I am doing.

[Vincent Ferrari]

RAW+JPEG is a smart move. You get the benefit of the camera doing its magic and such, but you also get a RAW image in case it isn't quite what you wanted. That's definitely a good compromise.

I tend to like my images "flatter" than what the camera gives me in JPEG mode, so I shoot RAW and adjust accordingly. I find the Rebel XT makes reds a bit too punchy and fleshtones a bit on the warm side.

Just sayin'.

[Jerry Raia]

On the jpeg issue. If I, for example, am always resizing them to make them smaller am I not loosing information every time I re save a jpeg? If I know I'm going to need a smaller image would it make more sense to start off with one from that perspective?

[Jason Dunn]

Originally Posted by Jerry Raia

On the jpeg issue. If I, for example, am always resizing them to make them smaller am I not loosing information every time I re save a jpeg? If I know I'm going to need a smaller image would it make more sense to start off with one from that perspective?

Well, normally if you're making smaller versions, you're keeping the original. If you're resizing the original to keep as your master copy, then yeah, it's better to drop the resolution to whatever size it is you want to keep.

[Jason Dunn]

My, my, Vincent, you're at your best here! ;-)

Originally Posted by vincenzosi

1. If your camera shoots RAW, shoot RAW. Always 100% of the time use as much data as your camera is willing to provide you. If you're not going to shoot RAW and only shoot JPG, you can get comparable results shooting with any higher end P&S. Seriously.

Nah. You're way off base. RAW is a big processing pain in the ass, and there's just no way in hell that EVERYONE should shoot RAW ALL THE TIME. That's insane. You mean to tell me that if you're doing casual pics at a birthday party or something, you shoot in RAW? And that if a first time DSLR-buyer should start out shooting RAW? And the comment about a DSLR shooting JPEG being the same as a P&S...that's just plain goofy. ;-) I'm not saying RAW is bad - it has it's purposes - but to say that it's the "only" way to go is a gross overstatement.

Originally Posted by vincenzosi

3. Is he honestly making the point that people can't tell the difference between different levels of JPG?...Look at his samples. The differences are glaringly obvious. You could make the point that at the full size, the picture is not terribly different. That's obvious.

But that IS his point: that the minor JPEG artifacts present are so minute, so hard to see, that you'd need a loupe to find them in a 4x6 print (or even 8x10 for that matter). Sure, you can see a slight difference between the RAW image and the JPG BASIC, Large image, but that's only zooming in by a HUGE amount. I estimate that the grid is 7 x 7 total squares equal to the size of the crop area he selected, so that crop area is 1/49th the size of the entire area. I suck at math, so I can't quite figure out what zoom it would be equivalent to, but if we think of that image being printed up as an 8 x 10 print, with 1/49th of it in the centre being the zoomed in area...the physical size of the print would have to be, what, six feet wide in order to get the same zoom level as the crop in the real world?

Originally Posted by vincenzosi

One of the advantages to cropping a picture is that you can mess up the framing and still fix it.

How much are you cropping though? If you're cropping more than the edge of something that might have popped into frame, then you as a photographer need to work on your framing of images.

Sounds like this might be a good topic for a podcast. ;-)

[Doug Johnson]

But with storage as cheap as it is, why sacrifice even the "little" loss in quality from Fine to Normal, let alone Fine to Basic (or whatever he calls it)? The JPEG noise in his crops is considerable, and with hard disk space as cheap as it is, why skimp that little bit?

Personally I use RAW for things that are really important -- bridal, engagement, wedding, and family pictures, mostly. Oh, and anything that I might ever blow up like landscapes. Then Fine JPEG for everything else. When I need JPEGs of any RAW pictures I use Adobe Bridge to batch convert all of them simultaneously. Pretty easy.

Just one example: There was one picture that I shot a couple years ago (JPEG) and didn't think much of it at the time, but later on I decided I want to give a copy of it to my parents for Christmas, so I had it enlarged to 20x30. If I had taken that picture with anything other than Fine, I don't think I would have liked it at that size (based on Ken Rockwell's article).

[Jerry Raia]

So really here I am reading a space debate. No one is arguing that using a huge jpeg is bad, just perhaps a waste of space. On the other side no one is arguing that there is no loss, just how much and if you can see it. Perhaps the only other issue might be the maximum burst speed. If you are just using jpegs and shoot M instead of L you can sustain higher speeds correct?

[Jason Dunn]

Originally Posted by Jerry Raia

So really here I am reading a space debate. No one is arguing that using a huge jpeg is bad, just perhaps a waste of space.

Yeah, pretty much. And while hard drive space is certainly cheap, throwing more storage at a problem isn't always the solution. It's certainly not a bad thing to have lots of storage, but in my case, since I use Foldershare to synchronize all my photos to all my computers, my laptop with the 80 GB hard drive is starting to run out of space...so if I think long term, looking at the photos I'll take in 2007, saving space by using JPEG Normal isn't such a bad plan. Again though, that's just my situation, not everyone will be constrained the way I am.

Originally Posted by Jerry Raia

Perhaps the only other issue might be the maximum burst speed. If you are just using jpegs and shoot M instead of L you can sustain higher speeds correct?

That's correct - the buffer on the camera can fit more images in if they're smaller, so using JPEG Normal instead of JPEG Fine would give you more sustained burst...while using RAW gives you much, much less. Depending on how you shoot though, it might not be a huge issue, because most cameras have sizeable buffers that are hard to fill (my D200 is that way at least).