KenRockwell.com: Nikon D200 Image Quality Settings

[Jason Dunn]

<div class='os_post_top_link'><a href='http://www.kenrockwell.com/nikon/d200/quality-settings.htm' target='_blank'>http://www.kenrockwell.com/nikon/d200/quality-settings.htm</a><br /><br /></div><i>"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><br /><br />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!<br /><br />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&amp;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]

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! ;-)

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&amp;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&amp;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.

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?

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]

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 (http://www.foldershare.com) 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.

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).

[Jerry Raia]

Good information at these Thoughts Sites! I should come more often! :lol:

[bmhome1]

It makes no sense to invest in 10MP DSLR"s, expensive lenses and cart around shooting images at 20% of their potential simply because its hard to imagine photos needing to be any bigger than file sizes based on consumer hardware limitation compromises dating five years past.

Guess what, photos consume huge amounts of megabytes capturing all the wide dynamic range, tonality and detail the latest DSLR's finally have achieved to challange film's ability to record all subtle nuance in images.

An 8MB RAW fine-tuned and saved as gigantic 28MB Tiff produces an incredibly sharp and luminous 4x6 print with a quality 6 color inkjet, blown up to 20x24 retaining subtle tonality still.

It's like deciding there's no point in anything better than a 2MB MP3 based on quick listen or buying into the idea because someone else can't tell any loss from full file original recording, therefore archiving those pointless...being forever stuck with just the MP3 to listen to.

Just because it SEEMS there's no difference gained beyond hobbling your 10MP DSLR capturing heavily compressed jpeg's (without ever possibly regaining back the 80% pixel data discarded), choosing makes as little sense as your audio collection all MP3, rationalized from indifference regarding less obvious, yet significant quality sacrifice tiny file sizes impose.

The future is one direction towards increasing and better output and display resolution where only the very best will be acceptable and half-decade old concepts towards digital photography file sizes will be as relevent as 8-track tapes and as future-use limiting.

2MB jpeg originals made sense in 2001 and none today, hauling around a 10MP DSLR to make simply illogical and just lousy concept to promote. It's just really settling for far below a level of standards set towards all other media discussion here, an uncharacteristic low-expectations direction.

Shoot your kids and family at the HIGHEST resolution possible, no one will wish you had not decades later. Cost and time issues have become irrelevent. CF/SD 2GB cards are nearly free at under $25 each, 300GB hard drives now well below $100, which can archive nearly 15,000 RAW+JPG files, 2-4GB of intensive photo shooting sessions take all of 10 minutes to offload and 60 seconds to archive to external drives.

Where is the logic NOT shooting 10MB RAW+JPG for everything, that dream shot out of hundreds can be greatly tweaked using RAW's unique ability to capture every pixel you invested in and hoped maybe even equal to film quality. Today it now can, given a chance always using fullest data captures and not obsessing about file size exponential growth.

No one has yet figured out when priceless photos will present, the only sure way to not later regret anticipating one incorrectly is having the headroom and safety net RAW only uniquely can provide. Try a shot using the worst color balance setting possible and 2 stops overexposed and compare the jpeg attempt at saving to the easily corrected result from same RAW image file.

Now, apply that approach to all the over/under exposed flash snapshots otherwise unusable. Forget saving RAW for just pretty pictures with little real future value compared to truly later treasured family snaps as sharp and color correct as 2006 technology offers. Those images deserve all pixels captured far more, and which one's later become priceless impossible to judge today.

I sure wish I had RAW's today to tweak some of my family's 40 year old snaps gems now only existing as fading, blurry little prints, I'd kill for to have as high resolution originals. At the time they were just more silly snapshots tossed into shoeboxes, value only realized much later.

A really good discussion would be about saving all the digital images from certain aging media storage failures or hardware obsolesence.

[Jerry Raia]

Well said. :)

[Lee Yuan Sheng]

Well, I might as well comment a little...

Most modern cameras come with fairly low compression methods. Even a "Basic" file is probably around a compresison level 80-90. Still, if you are outputting to JPEGs, it does help to save your files to the lowest compression mode available because you might need to do some post processing. While limited, it certainly can be done, and the less artifacts are there, the greater latitude you'll have for editing.

Halving the camera's output and getting a similar (if slightly lower) resolution on the photos is not surprising. In a GRGB sensor (which is just about 99% of cameras out there) luminance information is mostly collected by the green photosites (since they're the most senstive to light), which form 50% of the sensor. The red and blue photosites collect less light, but help to make up for the colour information. The final image is then interpolated from the individual photosites (else you'll get a mottled mess of green, blue and red dots) to save in whatever number of pixels the camera's sensor has.

This is also why Foveon sensors (that last 1% of cameras out there) display such fine image acuity even in the JPEGs. Without needing to interpolate the individual photosites, each pixel captures the full luminance and colour information.

Since I shoot RAW most of the time nowadays, what Ken Rockwell says doesn't really matter to me. :P If I shoot in JPEGs they'll be snapshots that won't be worth poring that much effort over. :P

[Leon]

I'll keep it short: no serious photographer should take Ken seriously.

[Jason Dunn]

In general I agree with you, but a few places I don't...

It's like deciding there's no point in anything better than a 2MB MP3 based on quick listen or buying into the idea because someone else can't tell any loss from full file original recording, therefore archiving those pointless...being forever stuck with just the MP3 to listen to.

But using your logic, it's foolish then to rip a CD to anything but uncompressed WAV files. Or maybe you'd "slum" so far as lossless WMA or FLAC. For me, that's too much hassle. Even with $250 headphones and good hearing, a 256 kbps MP3 sounds perfect to my ears...anything more is a waste and (in formats other than MP3) a hassle.

Where is the logic NOT shooting 10MB RAW+JPG for everything, that dream shot out of hundreds can be greatly tweaked using RAW's unique ability to capture every pixel you invested in and hoped maybe even equal to film quality. Today it now can, given a chance always using fullest data captures and not obsessing about file size exponential growth.

I shot a whole session once in RAW, and jeese, the amount of time it takes to process each and every one is PAINFUL. I might not loathe RAW so much if I had a really fantastic and fast tool, but doing each one via the Photoshop plugin really sucks. I shoot RAW when it's super important, but otherwise JPEGs are SO much easier and faster to work with. My time is important to me, and RAW is a big time suck.

[Jason Dunn]

I'll keep it short: no serious photographer should take Ken seriously.

Now, now, I've learned quite a bit from him about my D200, so even if I don't always agree with everything he says, he's helped me learn more about my equipment. ;-)

[Jerry Raia]

And I'm learning a lot from this thread. :)

[Leon]

I shot a whole session once in RAW, and jeese, the amount of time it takes to process each and every one is PAINFUL. I might not loathe RAW so much if I had a really fantastic and fast tool, but doing each one via the Photoshop plugin really sucks. I shoot RAW when it's super important, but otherwise JPEGs are SO much easier and faster to work with. My time is important to me, and RAW is a big time suck.
My time is also important but I see the RAW conversion as part of photography as much as taking the pictures themselves. But then again, I only edit my best shots and don't waste time on photos that won't make it to site or print.

A good workflow is important and will save a lot of time. Since you're using Photoshop, a lot can be automated. Someone interested in RAW should definitely read "Real World Camera Raw with Adobe Photoshop (CS)" from Bruce Fraser.

That being said, I don't want to convince anyone to use RAW. Just use what you feel most confortable with.

[Jerry Raia]

Someone interested in RAW should definitely read "Real World Camera Raw with Adobe Photoshop (CS)" from Bruce Fraser

I'll have to pick that one up. The one I have now is Camera Raw for Digital Photographers by Shepard. It is also an Adobe book.

[bmhome1]

I'm only been promoting fully utilizing the new digital output and mass archiving for future-proofing against later regret choosing lesser today.

I still can't understand rationalizing except in higher than 3fps sports situations. Tech geek's photos especially. Your 80 year old aunt might be shooting at higher quality than you. Embarrasing.

It's unfortunate that Ken Rockell has his luddite philosophy towards image quality file sizes. I find his frank, common sense camera reviews refreshing and he's brave enough to state strong opinions. The other digital reviewer much more popular, even revered by fans, had an edge too, but greatly softened now. He also became camera brand insider getting prototypes and leaks at same time. I hope Rockell uses RAW to evaluate optics, now becoming limiting factors as chips no longer will mask poor performance. The compression from strong jpegs can hide even severe chromatic abberation now obvious in high res captures.

The reversal of optics becoming weak link from chips barely producing recognizable images in less than ten years. 2008 promises 20MP chips filling up 16GB cards getting stored away on terrabyte drives. I once dreaded digital overtaking film, now that it's actually arrived the quality and limitless potential for ever more very exciting.

Personally, I get great pleasure knocking out 50 or 60 RAW fine-tuned images in an hour when in a groove or even tearing hair out spending all night on a high-dynamic-range image crafted from layering multiple adjusted exposures from one single RAW original. A dynamic range far exceeding 4x5 film. Geometrically increased potential on another level and never looking back.

Save the RAW books for 2nd or 3rd stage reference, just find software comfortable and dig in. Sheer numbers are the best teacher and more fun than confusing books targeted beyond beginners comprehension.
Bruce Fraser is one of the world's leading image graphics pros and can bury you in the first paragraph, regardless of the book title. It's REALLY subject of infinite depth and life-long learning curve. Even a side subject such as Color Management have seminars targeting attendee's already armed well read.

Again, to be perfectly clear, shoot all high-res now, archive all the GB's onto hard drives (I don't even bother with housings anymore, just bare drives filled and stowed. A 500GB 3.5" drive holding 20,000 RAW's is almost exactly the size of one VHS tape). When the mood strikes, pull out some digital negatives and experience the difference, even from one or two. It's fun, not drudgery that way.

BTW, XP freeware XSetup-Pro makes really BIG 1x2" thumbnail icons inside the folder needing no viewer software to scan images. It makes several GB folder sizes lightning fast. Vista now does the same natively although much slower, either indespensible quickly finding individual images among hundreds or even thousands directly from folder.

[Lee Yuan Sheng]

I'm curious, do you have any examples of JPEG artifacts hiding chromatic aberrations?

[Perfect Shot]

...well new to this forum and somewhat to digital image capturing... but used to shoot tons of 35 mm about 20 yrs ago. I found this forum by accident and glad that I did! I am looking forward to learning a great &amp; hope to be able to add something. I am enjoying my new d200. :D .. but have tons of questions. I read as much as I can find( to learn) , but some things need to be asked......

So, if you dont mind..

As you know, a "raw" s/w program must be purchased. I like what I read about "BIBBLE" http://www.bibblelabs.com/. Please compare and contrast to Nikon's NX ( i think that is the newest) . CPU load, ease of getting around etc...and any recommendations



THANKS &amp; Happy Thanksgiving to all,

Sincerely,
David