Beyond Megapixels
TheTechLounge points to this "first of a three-part series of editorial articles examining current digital photography hardware, as well as the author's views of what is to come." It boils down to the excellent point that pixel count alone is not the way to evaluate digital camera capabilities.
Most people didn't care about resolution in the analog world. The fact that many people considered APS cameras to be better than 35mm is simple proof of this.
This seems analogous to consumer computer makers moving away from advertising GHz and MB.
It's what you (can) do with it that counts.
if we are so 'green aware' why don't inkjet printers ever have green ink?
every day http://en.wikipedia.org/wiki/Special:Random
When you're dealing with digital you quite simply need pixels. You need to decide what size pictures you intend to print or whatnot and get an MP count to match. You can't get a 1.0MP camera and do large prints of any quality.
Of course you also need picture quality. But it really doesn't matter how good the colors are if you're only getting a postage stamp image.
I have a 2.0 megapixel camera which I intend to replace eventually. Not because of the pixel count, but because of the image quality. I have a few pictures where a small branch got just a bit into the frame. The camera focused on that little branch and blurred the rest of the picture. There's no manual focus so all I can do is watch what's in the view carefully.
It also doesn't react intelligently to low light. Although with a bit of modification I can turn that into a feature as I can take time lapse photos to get good pictures in very low light.
As with all things, you need to pick the versions with the features you need.
Ben
Work Safe Porn
If we can't use just one metric to identify the quality of a digicam, we'd have to do with something like a (megapixel,sensor size,optical zoom) triplet. Most of us already know to look for more than one feature while buying PC's. It would be nice if somebody well-versed in the mysteries of digital camera technology would standardize the set of features that I should be looking for as a consumer....
...electronics: Cheap ...optics: Expensive
Look at screens. Graphics cards have improved massively (electronics), screens (optics) used to be 1024x768 quite a while back, and typically aren't more than 1600x1200 now. The LCDs will hopefully change that though, since they're much more scalable (make more pixels) than a CRT (move beam faster).
Same with digital camera. The back-end is getting much cheaper, multi-MP CCDs and other electronics, but good optics in the lens is still damn expensive.
I read a piece recently about HDTV cameras. There were rumors that a certain camera would be sub-10.000$. The official comment basicly said "we can't tell you the real price yet, but you're smoking crack. the lens alone is in the 7-9.000$ range".
That being said, most digital cameras today should be just fine, if you don't try to take "impossible" pics. If the sun is saturating the CCD, it won't happen. If there's light casting ugly shadows, fill it in or you'll never get rid of them. There's a lot more bad photography than bad cameras...
Kjella
Live today, because you never know what tomorrow brings
The article seems to be making the argument that a smaller format sensor won't be as sensitive as a larger sensor, but I'm not sure I buy this.
The example he gives of buckets of water is flawed, since falling rain isn't *focused* like light is. Light entering a lens is just being focused on a smaller area. Sure the area is smaller, but it's also brighter.
A larger sensor just requires the projected image to be spread out further. Of course, maybe if you got too small, you'd run into the same limits optical microscopes do, but I don't know that it's near that point yet.
Maybe the author was thinking of regular film cameras where a larger format negative captures more detail? Still, this is because the level of detail film can capture would be about the same per sq inch (so larger format, more detail). What I'd really like to see are some actual tests, and not just some author's wild speculation.
An 8 megapixel ccd behind a cheap lens is going to look worse than a 1 megapixel ccd behind a high quality lens. Look at the pictures of mars, they were taken using a 1mp camera.
Of course, the additional detail is nice. But to be really usable to blow images up (which is probably the only reason for going higher than 4-5mp), the following problems have to be solved.
1. Noise has to be reduced. Especially in dark pics. Less of a problem now, but still an issue. Of course, if you're taking a 8mp camera and printing out an 8x10, you probably won't be complaining. Zoom in to 300-400% and you will be easily able to see it (and all the stuck sensors, but that is another story).
2. The lens is good enough to resolve that detail.
No, your made in china $5 lens will not be good enough. There is a reason professional film cameras have "big ass lenses".
1q2w3e4r5t6y7u8i9o0pqawsedrftgthyjukilo;p'azsxdcf
If it was me, I would take a 3 MP DSL over a 5MP compact consumer camera....
IMHO
Even with good lenses and modern low noise sensors, digital cameras has a rather narrow exposure range as compared to classical photography. Shooting with negative film you can get something like twice the exposure range, compared to any ordinary digital camera (i.e. you will be able to see more details in both the dark and light areas of your photo)
lenses are of course top priority, but at the same time your mars rover example also supports the article's contention. quating from link:
.35 by .26 inches). The Pancam has just a million sensors spread across a chip that's 12 by 12 millimeters -- nearly a half-inch square.
A Sony DSC-F717, with a street price of around $600, has 5.2 million sensors (or 5 megapixels) on a chip that is 8.8 by 6.6 millimeters (or
People who go all digital really do need high MP. I have a 4MP, but I'm worried that if I get a really unbelievable shot, I won't be able to blow it up as big as I want. At least with 35mm, I always had the option of making it poster size if I were so inclined. I agree with what the article says, but if you're actually replacing film, you might as well get as close to the capabilities of film as possible.
Also, with the near ubiquitousness of photo editing software, almost everyone has the ability to crop and edit images. Not only would you rather have more pixels for any kind of editing, but with high MP, you can crop even a small portion of a picture and still get a decent 4x6.
First of all, no one has mentioned DYNAMIC RANGE yet. This is the range between absolute black and absolute white. Whether you are using film or digital, this range is crushed compared to the human eye. Digital dynamic range tends to be worse than film, which is one reason film isn't going to go completely away any time soon.
Greater dynamic range will give you better details in your shadows and highlights. This is very important for the serious photog, although probably not important for snaps of your kid's Bar Mitzvah.
The other thing that matters is the actual size of the CCD. Manufacturers are using various technical tricks to squeeze out more pixels from the same size CCD, and the results are sometimes pretty bad. The worst problem I've seen was purple fringing in bright red objects that were backlit. Totally ruined an otherwise beautiful photograph.
The bigger the CCD the better.
It's not offtopic, dumbass. It's orthogonal.
Some interesting points there mate. However, don't dismiss software zooming as being useless. The software zoom on most digital cameras take the picture, then software zoom, then convert the raw data into formats such as jpg. If you were to just convert to jpg and use software to zoom, you would be zooming in on the artifacts of the jpeg compression. Therefore, software zooming can give you that little ooomf. I'm ignoring the fact that some cameras don't compress the raw data though. In which case, just ignore me :p
Why not shoot in film and use a film scanner? I've got a 30 year old (Minolta X-700) camera that has been with me through a lot. The thing will not die and just keeps on going. I just have to change the battery once a year or so. I usually develop my photos at a grocery store. Ask to have it developed and cut only - no prints. It costs me 1.25 per roll and I have it in about 20 minutes. Later I scan them in myself, get 11 Megapixel images with 48 bit color, scanned 8 times to minimize noise. (They're about 62 Meg TIFF images) that I can print with up to 13x19 on my Epson 2000P printer. The best part is, in 5 years I'll buy the newest and greatest film scanner and I have the option to re-scan the images at 20 Megapixels or whatever. That's my solution at least. By the way, the scanner was only 500CAD ;-)
Foveon cameras have one three-color sensor per pixel, but for PR purposes, they, too, count R, G, and B as separate pixels. For example, the Sigma SD-10 mentioned in the article has an imager 2268 x 1512 pixels, but is listed as a "10.8 megapixel" camera. For Foveon units, divide by 3.
Foveon cameras, since the R, G, and B sensors are at the same place, don't generate color artifacts at black/white boundaries. This eliminates one of the main effects that makes "digital" look worse than film. Of course, if you compress to JPEG, you get color artifacts anyway, but that's a JPEG problem, not an imager problem.
At the end of each the articles in this series, I will comment on what I think camera developers have done right and wrong, and what I think is important to the photographer who wants to produce better photographs.
By the generality of this statement, the author doesn't seem to have much resource on reviewing digital cameras case by case, which is necessary to make any useful assessment at all. I recommend this site for getting camera reviews.
They provide full review of some cameras (mostly prosumer kinds), which would include ISO sensitivity comparison against similar cameras, color tone test, auto focus test, lens distortion/shading, and tons of others. My personal favorite is the resolution chart.
I once had a signature.
Now the same thing has happened with cameras. It's all about megapixels. Your average consumer won't do enough research to learn about how the camera works, all they know is megapixels.
But what can be done? Instead of producing higher quality optics such as that on the mars rovers(1MP mind you), we get more megapixels with crappy everything else.
Lenses are very good now. Anything that produces a repeatable distortion can easily be corrected for at the factory - digital cameras have large DSPs in them to handle the image compression work, those same DSPs can very easily apply a distortion correction to the camera to correct for minor lens flaws.
Good lenses are much more important in the analog world, where literally, what you see is what you get.
..don't panic
The reason Spirit's 1MP sensor is able to produce such great images is not just because of the lens, but also that the size of each pixel is so much larger than prosumer-grade pixels. "Perhaps most important, the sensors on Spirit's CCDs are bigger" and "Each tiny Pancam sensor, measured in microns, is nearly four times as big as those on the Sony.
In the consumer market, which Dalsa does not target, 5-megapixel cameras often use the same size CCD as a 3-megapixel camera. More pixels are simply crammed onto the same-size chip.
"The pixels themselves get smaller," Myles said. "This has an impact on image quality."
Why? For one thing, smaller pixels are less light-sensitive.
Right as far as dynamic range and noise are concerned. Wrong as far as "detail resolved" is concerned. A small 8 Mpixel sensor, given sufficient light, will resolve more detail than even the largest 4 Mpixel sensor. Furthermore, in particular for digital SLR sensors, you are better off taking the higher resolution and smaller pixels and removing noise in software than to limit yourself by an otherwise equivalent lower resolution sensor.
Foveon's images have not lived up to the hype in tests, and there is no reason to believe that they would. The Foveon sensor really does have 1/4 the spatial resolution of a regular CCD sensor. In return, it avoids some color artifacts and requires a bit less post-processing. But that turns out not to be a very good tradeoff.
I definitely disagree. Check out DP Review's review of the Sigma SD10 which uses the Foveon sensor. You'll see images from the Foveon sensor that have been upsampled to match those of a Canon dSLR. The Canon does appear to resolve a bit more detail, but remember that the Sigma's images have been "digitally zoomed" from 3.4 MP to 6.3MP.
As for Fujifilm's new sensors that are designed to improve dynamic range, compare one of the pictures here (try the one with a lot of window reflection) with another picture of the same subject. You'll see that in the shadowed areas you can resolve more light detail by using the Fuji. It's not a huge difference but it is one that some people will appreciate.
Agree that ergonomics should be prioritized. It's amazing how many useless "features" the camera makers are adding to jack up their marketing feature list at the expense of usability. There are just WAY too many options. They could get rid of almost all the buttons on a Digital camera for even the pros. I really wish they would simply cut out switches and menu options and make it so that you DON'T need a manual to operate it. My favorite camera is still a fully-manual 35mm Nikon FM2. Either that or an 8x10.
Things I wish manufacturers did:
1. Store data in RAW format. (Thanks to Sigma for pushing this.) This get rid of the useless "low/medium/high quality" switch on the camera. There goes one pointless switch.
2. Store all data at the highest resolution. Get rid of the "small/medium/large" switch. If I needed to store more pictures on my card, I would have bought a higher-capacity CF drive. I can get 4GB models now. That should be enough to store hundreds of pics. Another pointless switch, gone...
3. Get rid of in-camera white-balance setting, and do this on the computer or laptop or even palmtop to simplify the camera and force the complexity outside. (Again, thanks to Sigma) This can be done on the computer if needed with the RAW file. Most amateur users have NO idea what the hell white-balance means anyways. A third pointless switch gone..
4. Get rid of the Priority switches- Aperture, Shutter, Etc.. Instead, allow the user to adjust the Aperture & Shutter on a lens ring. The ring can also have a setting for Auto. This can also be done for focusing with a Focusing ring. There- 3 buttons eliminated just like that.
5. Get rid of on-camera flashes settings (Keep the wimpy on-camera flash if you must, but leave it on Auto always, and auto-disable when external flash is connected) Pro photographers would have an external flash anyways, and any flash settings can be made on that. Another switch, gone...
There's so many useless switches on a modern Digital SLR that can be completely thrown away and still provide all the functionality anyone would want.
Some people may want all these useless features.. for them the camera vendors can have their own special overfeatured model. I would rather have one that's simple and obvious... The first Digital SLR vendor that comes out with a Camera that DOESN'T include an INSTRUCTION MANUAL, I'm buying.
1. Storing data in RAW format is always a good thing for non-casual users. But for many jpeg files are all they will ever use.
I think all high-quality cameras now can store RAW image formats.
2. Store at the highest resolution. Well, maybe, although it's a great way to save memory which ain't cheap or as large as I'd like to have it yet.
3. If you can shoot in RAW mode then you don't need the camera to do white-balance and you can do it in the computer where you have the horsepower and GUI to do it right.
4. I disagree. It's really nice to have a full manual mode but even Leica came to realize the joy of having Aperture Priority which many Leica shoots live in on the M7.
5. An on-camera flash is useful for fill and is a keeper
If you haven't held and played with a Leica Digilux 2 you should. It is a wonderful camera that works exactly like a film camera. Unfortunately they used an electronic view finder instead of a real range finder. Sigh.
As for manuals, even the M6 has a manual although it's only 20 pages long and in 4 or 5 languages.
the reason why i still prefer film over digital (aside from pure aesthetic reasons that are not worth discussing because it's a very personal thing) is the color/tonal resolution. hell, my 2700dpi 35mm film scanner can pick up the grain for some of the films i use, so 6MP cameras already have better "resolution" then what i get with film. however, the scanner struggles in distinguishing between subtle gradations especially toward the shadow end of the spectrum, and the same is even more the case in digital cameras.
it's not just the number of colors, btw; the average human eye, while amazing, is not going to notice the difference between two shades in a 16bit per channel image (my scanner is capable of 16bit RGB, i don't know of any non-scanning back digital cameras that can do the same) but can the CCD actually resolve those shade gradations to take advantage of all the bits? definitely not the case yet.
in this age of communication i'm just not getting through
Most people don't care about pixels. Kodak still advertises their "Max" film as a "general purpose" film, for example, and it is an ISO 800 speed film, with horrible grain and sharpness. Since most people don't enlarge above 4x6" prints anyway, though, they don't care.
... With a camera like that, you'll actually learn something about photography instead of keeping your camera in "auto" all the time, or relying on photoshop/gimp to do corrections later. And since you have to compose on ground glass, and each exposure "counts," you'll be more careful with composition.
Every few years or so, Kodak and a few other companies get together and decide that consumers don't care about resolution, as long as a 4x6 looks ok. The people fueling the "digital megapixel rush" are gadget heads who just want the latests and the greatest, and have a lot of disposable income on their hands.
Personally, if you really care about resolution, get a field camera or view camera. I used to shoot with one... 4x5" negatives/positives enlarge very nicely... albeit most of these cameras are huge and can weight 10lbs or more. A good compromise on size/quality is a decent medium format system. People are going crazy for the $1000-range, 6 megapixel digital cameras with interchangable lenses now. You can get a new Mamiya 645e medium format setup for that, and have tons more resolution... resolution that I don't think consumer digital cameras are going to reach in the near future (they are still chasing "35mm quality," IMHO).
Just my 2 cents.
This is what always happens when marketing starts to determine the specs rather then sound engineering. Those who don't do research buy based on the megapixel count and price. This causes a situation where the camera with the highest megapixel sensor crammed into the cheapest possible camera is the most succesful. The same thing happens with everything from printers to processors to cell-phones. The only positive aspect is the informed buyer can sometimes get good deals as a result, as the best camera for the price may not be the most popular one, and stores have to sell it for less of a markup.
Flower shots from my folks Garden
All of these pictures were taken with my Canon-EOS10D, 420EX flash (used mainly for shadow fill), and Sigma 20mm 1:1.8 EX DG prime lens. The shots were taken hand-held in AP mode using F4.0-F16 depending on the conditions. This particular lens produces ultra sharp results at F4.0-F13 or so. The 10D (and 300D) use a 6 MPix low-noise CMOS sensor and you can see it in the above shots.
Insofar as all the discussion goes, from my point of view it all comes down to three things: Lens Quality, Sensor Quality, and Dynamic Range (of the exposure). SLR's like the 10D have gotten good enough that I don't use film any more. The lens quality is there (being an SLR and taking the same lens as the film EOS's), sensor quality is there, and while dynamic range still needs another 2-4 bits of resolution for my comfort it's still good enough for 99% of the shots I take. Film is dead, digital rendition at 11!
And I tend to agree with the few other obviously experienced comments (verses the bozo comments from people that don't know jack about taking photographs). You first need to know how to take a picture before you can take a good one. Then comes lens and sensor noise. A lens hood is important, and a good flash (articulated for bounce shots and also be sure to have a diffusor handy) is very important (even when you don't think you need it). For example, most of those flower shots I took were with flash+diffusor, even though it was a bright sunny day outside. The flash was used primarily to fill in some of the shadow (one way to correct for limited dynamic range but it also makes the shots look a lot better).
-Matt
This is exactly what I see in the photo magazines. The camera is a tool. And many pros carry compact cameras for shooting snapshots.
One magazine recently reviewed the Canon 1Ds and compared it against a Canon film camera w/ the same lense and iso 100 film. They blew up a section showing a sign w/ lettering. You could read it on digital, not on the film version.
I'm learning alot w/ my D100 that I'd never do with my wife's N80. I'm never afraid to take a bad photo or too many because it's not going to cost me anything. I get more good pictures when I do candids, bird photos, etc because of it.
Not to get too picky here, but the latest Noritsus and Fuju Frontiers are laser printers. Instead of exposing a selenium drum with an IR laser, they expose the photographic emulsion with red, green, and blue lasers. Some printers (e.g. ZBE Chromiras -- the best in the business) use LEDs, others (crappy ones) use CRTs, but most use lasers.
And not counting the cost of equipment, expect to pay $0.25 per sq. ft. for wet prints. It's going to be a lot more than 5 years before somebody has an inkjet process that can spit out 2000 4x6 archival prints in an hour for less than a nickel a piece.
aQazaQa
One reason storing RAW data at the highest resolution is not always the best idea: the bandwidth to flash needed to keep up with a stream of images being taken is greatly increased! This is because modern cameras have fast buffer memory, and slow (think hard drive) flash memory. The more images can be shuffled through the buffer before requiring a write to flash, the faster a sequence of images you can take! The consumer digital SLRs are up to ~3fps for 10's or even 100's of frames!
Megapixels mattered when you couldn't even get a good 5x7 print. Then it still mattered when you couldn't even get a good 8x10 print. At that point they stopped mattering for everyone except professional photographers who need to shoot for ads and posters and so on.
And of course realize that if you take printing out of the picture and just keep everything digital, then 1 megapixel is fine for 80% of all uses. 2 megapixels covers the rest.
The huge downside is more megapixels is that, well, the images are huge, so you spend more time tranferring them and backing them up, you get fewer images on a CD, you need larger and more expensive memory cards, etc.
Great examples BTW - yea, you are right, the marketing folks would be putting this on the packaging right now ... along with even more "digital zoom" which obviousely is a loada crap - this is another thing that I just turn OFF.
I especially got a chuckle out of the wildlife pictures from the "next state over" - yep, you'd just end up with a buncha pixels all the same color, with some noise super-imposed over 'em.
Having said that, I've seen satellite images of my house that are pretty darn impressive where issue such as atmospheric distortion become significant. But obviousely these guys are spending just a little bit more on their camera equipment than we are! ;-)