February 6, 2007 4:00 AM PST
More megapixels, better photos: Fact or fiction?
- Related Stories
-
Vista to give HD Photo format more exposure
January 26, 2007 -
Adobe taps the power of negative thinking
November 20, 2006
Digital camera makers continue to squeeze ever more megapixels into their products. But does that make for better pictures?
Some experts say no. Image quality isn't improving, they say, and some fear it may actually be degrading as the megapixel race escalates.
"There is definitely a decrease in image quality," said Dave Etchells, editor of a camera reviews Web site, the Imaging Resource, which performs extensive camera tests. "There have been some improvements in semiconductor process technology for sensors, so it's mitigated the problem a bit, but there overall has been an increase in image noise."
The basic concern is that smaller pixels on camera sensors means less sensitivity to light, leading to image noise such as off-color speckles or rough edges, worse performance in dim conditions, and the loss of finer tonal gradations such as the subtle shadows of a white wedding dress. Point-and-shoot cameras, with their small sensors, are the chief culprits.
Camera makers disagree, saying consumers have an appetite for higher-resolution images--for making larger prints or cropping to focus on specific details--and that image quality has indeed improved overall. But even if they're correct, they have a growing perception problem among influential camera experts and enthusiasts.
Some of the blame can be laid at the feet of consumers, who fixate on megapixels as a measure of quality. It's the same thing that happened with PC processor megahertz and flat-panel TV dimensions. "The word 'megapixel' is a marketer's dream. Every consumer believes more is better," said Chris MacAskill, chief executive of SmugMug, a Web site that hosts photos and lets users print them.
"The trouble is megapixels stopped mattering once we passed 6 of them," MacAskill said. "One in a million shots would benefit from more than 6 megapixels, while every indoor shot would benefit from less noise."
Not so fast, experts
Canon, which market analyst firm iSuppli estimates sold 20 percent of the 96.4 million cameras shipped in 2006, sharply denies there has been any degradation of image quality with its PowerShot point-and-shoot cameras.
"When all else is equal, our 10-megapixel models tend to produce better detail than lower-resolution models at print sizes of 8x10 (inches) and larger," said Chuck Westfall, director of media and customer relationships at Canon. That holds true even while holding other factors constant, he added. "For example, it's fair to compare the PowerShot SD900 Digital Elph at 10 megapixels vs. the PowerShot SD550 at 7.1 megapixels because both cameras have the same lens and sensor dimensions," he said.
Sally Smith Clemens, a product manager for Olympus Imaging America, added that reputable camera makers are careful to improve image processing to ensure that the overall picture is good even if noise levels increase in a sensor. "It's not just the resolution of the sensor that determines the final image quality. It's optics, the color management, the technology in image-processing engines," she said.
Why increase pixel counts?
There are advantages to increasing the number of megapixels. Larger prints that require a minimum pixel count can be easier to make, and consumers can crop images to focus on just the subject matter they want.
But there are costs, too. Among the more obvious burdens: Camera image-processing chips have more data to digest; memory cards and hard drives fill up faster; and photo editing puts greater space, memory and time demands on computers.
More subtle problems also are possible. Camera image sensors rarely get larger from one generation to the next, so squeezing more megapixels out of a sensor means each pixel on the sensor is smaller. In most of the chip business, smaller electronics are dandy, but with cameras, they translate to less light per pixel.
That light difference means it's harder to distinguish the signals produced by light from the electronic noise in the sensor. The idea of making the signal-to-noise ratio worse may sound pretty technical, but possible consequences are easily understood: Images suffer from color speckles, and cameras work poorly in dimmer conditions such as indoors.
"If you try to cram more pixels into the same amount of space, you risk getting signal degradation because you're not getting as much light into the same pixel," said Chris Crotty, an analyst with iSuppli.
It can be tough for consumers to understand why they might not want to snap up the most megapixels possible. "People can understand the idea of more numbers is better," Crotty said. "But signal-to-noise, fill factors, dynamic range, blooming--these are concepts most people aren't going to understand."
--Stephen Shankland
Panasonic, a newer entrant to the digital-camera market, also emphasizes noise reduction. "Historically, you really had to make a choice between reducing image noise or preserving image detail," said Richard Campbell, Panasonic vice president of imaging for Panasonic Consumer Electronics. "However, Panasonic has made drastic improvements in reducing overall picture noise without sacrificing detail, specifically in its recent improvements in the Venus Engine 3 processor for 2007 models."
Image-processing chips, while more powerful and sophisticated in current cameras, can cause problems, though, as they try to remove noise.
"What you often get along with the additional pixels is more noise, which ends up getting smudged away oftentimes by in-camera noise reduction software," said Jeff Keller, editor of the Digital Camera Resource Page.
Keller recently advised most readers to pick the 8-megapixel Canon PowerShot A630 over the 10-megapixel A640. "Most people don't need the extra resolution. Plus, the file sizes are larger, and the camera is slower," he said.
Image sensor makers are also changing to try to deal with the noise issue. In particular, they're working on reducing the size of supporting electronics so more sensor area is devoted to gathering light. Compensation strategies include the following:• Olympus began using "NMOS" sensor chips that decrease the area taken up by electronics, Smith Clemens said. And in low light, results from multiple pixels are ganged together, a move that averages out noise but reduces the ultimate image's resolution.
• Sensor maker Micron has technology to let multiple pixels share the same circuitry so less is needed, said Suresh Venkatraman, director of digital camera work at Micron.
• Canon said its image sensor pixel sizes shrank only "fractionally" when it moved from its 8-megapixel EOS Digital Rebel XT to its 10-megapixel XTi sequel, a move made possible by reducing the amount of space between the sensor pixels. (The company also said it improved the sensor sensitivity.)
• And Fujifilm's SuperCCD sensor grid is rotated 45 degrees compared with ordinary sensors, to devote more area to light-gathering, said David Troy a senior product manager at Fujifilm USA.
Do consumers care?
Having better technology and getting consumers to buy aren't the same thing. Fujifilm, for example, has technology that uses two sensors for each pixel--one for low-light sensitivity and one for brighter conditions.
"Those...cameras did a fantastic job with highlight detail," Chris Crotty, an analyst with iSuppli. "It was an admirable effort, but it really didn't get any traction in the marketplace."
But there's evidence of a more sophisticated understanding of the situation--in part the result of the fact that most digital-camera buyers aren't buying their first digital cameras. Among features camera makers are beginning to add to point-and-shoot models are face detection to set focus and exposure better, lenses that gather more light, and zoom lenses that work better with wide-angle views.
"Megapixels will always be important, but as consumers become more educated, it will be the addition of other features that make the purchase of their new camera more worthwhile," said Richard Campbell, vice president of imaging for Panasonic Consumer Electronics. "As the digital-camera market matures, consumers are becoming aware that lens quality, processor quality and image stabilization technologies are at least as important as pixel counts when determining image quality."
Even better, those features aren't mere marketing hype. "Face detection is a technology that has real tangible benefits to consumers," Etchells said. "You will get more in-focus photos and get the exposure right."
Don't expect an end to the megapixel race, though, Crotty said. The firm estimates that digital cameras had an average of 5.7 megapixels in 2006 and forecasts it will increase to 6.5 megapixels this year and to 9.2 megapixels in 2010.
SmugMug's MacAskill thinks that's a shame.
"We went past the point where more megapixels made a difference years ago," MacAskill said. "In the last 3 million prints we've made for very discriminating eyes, none were returned for lack of pixels."
See more CNET content tagged:
sensor,
camera maker,
image quality,
camera,
noise
As for what's **really** needed to improve digital cameras: Better dynamic range (ability to expose for darker and lighter ares of the image at the same time). The loss of 1-stop by going digital hurts. I'd like to see digital give that back, and more. That would be a huge improvement for snapshooters, serious amateurs, and pro's alike.
--mark d.
http://www.summitpost.org/user_page.php?user_id=26307
However, just about every printing facility (whether it be Kodak, Fuji or your own photo printer) allow you to crop pictures before printing, so you do not need to ever change the original on the camera's memory card, just the version that you are printing.
Trust me my friend, almost everyone does a little bit of cropping on the mall self service photo printer terminal before printing.
just increasing he size of a picture but the sensetivity panel as well
and that could make a gr8 difrence in the price of the
product,unfortunately most of the cameras have been squeezed
into alot of mega pixels so it would atract the custumer. but what
is the difference?? only the size of the picture and not the clarity
and how much zoming u can do withough losing details. try it ur
self buy 3 xmebapixel for example of 3 diff. brands and u ll see
that non is like the same.the fact is also applyed on mobile phones
And that includes low enough noise rations and light to dark ratios.
So all you manufactuurers, hear this: I am still using film, and will continue to do so until you shape up.
So, here I stand with my Alpas for above water and my Nikonos for underwater.
More MP are only good if you're going to enlarge significantly and can deal with the noise, or use photo editing software to smooth it out, thus losing the advantage of more pixels.
Just like with LCD panels, bigger is not always better. Some panels have horrible viewing angles, changing colors drastically or even going negative images if you even move a bit/lean back in your chair. Even the formerly "good" ones like Dell have gone downhill significantly.
Don't buy a camera based on MP alone. Buy it on overall image quality.
Can't understand the fud from CNet now a days, you look at their reviews, and they lack depth and accuracy, and RARELY if ever correspond to consumers ratings... wake up folks, its not the advert $$s that get customers, but real reviews.
Thanks
Another advantage is that you get a more film like depth of field. With a small sensor digital camera, everything is always in focus, even when you tweak up (Or down, I forget) the F-stop.
Sometimes it is nice to be able to take a sharp picture of the subject, but have the ugly background a bit fuzzy.
Some of us need the extra pixels because we like to take poster sized shots, other times we take pictures for work related reasons, and like to see details that we did not see while we were there.
Well depth is where manufacturers will run into problems when increasing pixel count on the sensors. They have to shrink the pixels which causes the wells to shrink. This in turn increases the effect of light-related noise (photon shot noise) in the images and decrease signal to noise ratio and dynamic range. This higher noise means that manufacturers have to be a bit heavy on the noise reduction post processing which leads to washed out images.
I agree with those that said 5-6MP is the "sweet spot" for resolution - image quality tradeoff.
A quick correction to your post:
Full frame sensors do not provide a better depth of field. DoF is the function of the lens. Lenses are better quality in SLR cameras though (and you can only find full frame sensors in SLR cameras) so I see why you would assume that a better depth of field can be attributed to the sensor.
To achieve better depth of field you have to close the iris. I too, can't remember if that's going down or up in the F-number (I think up because lower f numbers mean "faster" lenses - e.g. they let in more light with the same exposure time and that corresponds to an open iris).
Cheers,
IB
So what did manufacturers do next? The almighty 'image stabilisation' hype to right something that is obviously the camera manufacturers fault in some sense.
I am no disparaging certain manufacturers of good cameras that have REAL image stabilisation via the lens mechanisms. What I am really pissed off is the use of higher sensitivity + flash and called that 'image stabilisation'. It is just taking advantage of consumers' lack of photography knowledge. Hideous is the word I would use.
Time to stop the Mega Pixel nonsense and really focus on doing a good job to mix and match a sensor that better matches everyday use (5MP is my optimum) with good lens and good image processing power. All other features such as face recognition (duh~~ what happened to good ole practice of focusing first and then press the shutter?), Wi-Fi, Frames and thin cameras are just secondary to good photographs.
It really pains me to see consumers get duped everyday.
Image stabilization reduces the effect of camera shake which is more noticeable with higher zoom lenses and indoor/night time photography. It is the 10x + optical zoom cameras and low-light-high-exposure-time photography that necessitated the development of image stabilization technologies.
It is a very useful feature and in no way dupes consumers. I consider myself knowledgeable with digital cameras, optics and sensor technologies and I would never buy a super zoom point and shoot camera without image stabilization (or a tripod :o) ).
IB
megapixels into a sensor beyond a certain point does not
improve image quality. The quality of the final picture is
determined by a combination of the lens quality, the accuracy
and quickness of the camera's autofocus, the camera's ability to
properly meter the light of the scene and adjust exposure, the
size of the image sensor (large sensors are good because they
have less noise), and, finally, the megapixel count. Unless you
are making giant poster prints, you really don't need anything
more than 6 megapixels. It's ridiculous to get a 10 megapixel
point-and-shoot camera and expect to get outstanding pictures
when the 10 megapixel sensor is paired with a relatively cheap
lens. If you want better images, then invest in an SLR digital
camera since you can use high-quality lenses with it.
I want a smaller manual. I want everyone to use the same icon for the same function. How come the viewfinder isn't a oled/lcd/whatever screen? Do we really need a SLR camera for a viewfinder to work?
Yes, more megapixels will make a denser picture, but it would be moot if you can?t get a clean image to the sensors.
They also need to allow the sensors to get little bit larger, as previously indicated, packing in more and more elements in to a small space will introduce noise. Sensitivity and speed also need to be improved.
If you zoom in on many pictures you will see artifacts that are directly caused by the lenses. Great lenses can pack a huge difference.
If the camera had lots of processing power, you might be able to map the defects in the optics and compensate for them.
I want a camera that can do this, but it will not happen this decade.
Full frame scientific cameras/sensors top out at around 160MP these days.
IB
No quantity discount.
If a commercial product this size exists, one might imagine how large a count might be in labs or in space.
What performance means and what matters most varies significantly depending on the personand his/her needs. Someone doing a web blog doesn't need high resolution photos, while someone doing real estate cares deeply about high resolution photos (for posters, etc).
For me, 4 megapixels gets the job done---8x10 prints look great. But it depends on the individual.
Just as I said when I bought my first Pentium computer: "Windows crasher faster". I can now paraphrase about a higher pixel count: "A person will still take horrible pictures, just at a higher resolution"!
No matter how good a sensor gets inside a P&S camera, it will still come down to the quality of the optics and the speed at which the camera can take a picture. If you are after high quality shots and want to work in a variety of lighting conditions, then purchase a digital SLR.
BTW: I have a digital SLR (x2 for professional work) and a P&S (for when I travel and see something interesting).
But it's also true that if you don't have the pixels, you can't make the print in large sizes, so if you need to print 8x10, and it has to look professionally good after manipulation in Photoshop (including crop), not just "I got lucky and it worked", then you need 10+ MP. And if your sensor is smaller than "Full Format" then, I guess you deal with the noise. I'm not sure those points are clearly made in the article, they're only partially made. Only someone who's already savvy would get them.
Of course, Canons have built-in noise reduction that you generally can't turn off. Some people say, too much...
However, I don't appreciate the article when it lays the blame on the consumer for "Megapixelitis." What statistics are to be found on websites and data sheets for cameras in shopping situations? What is the primary statistic used to set prices for cameras in the industry? MP and nothing else, when it comes to consumer-level cameras. So that statement early in the article is spurious.
It's the marketers and media people who made MP the deified statistic it is, because, let's face it, the consumers needed some simple ways to compare camera technology without a lot of hassle at buying time. If the consumer believed it, it's because a lot of marketing and media effort went into creating that belief. And it's a convenient belief created by the industry itself.
Nothing wrong with that statistic. It's an important statistic. It controls the ultimate useability of the file. I can never ADD pixels to a small file, so a 1MP camera will never give me a nice 8x10 print (or even a nice 4x6).
So I object to the article implying that:
1. MP is a bad statistic that should not be used anymore
2. Consumers fixated on it of their own accord for no good purpose
The rest of the article has some excellent info about sensors in an easy to understand format, and isn't afraid to tell consumers to save money by buying fewer MP's. Since cameras are still priced by MP's, that will save $.
I don't know how much impact it will have on the industry as far as encouraging them to make better optics and find ways to reduce electronic noise in pictures (perhaps, offer Full Format cameras at lower prices?). Sadly things like that are often as popular as generosity on days other than Christmas day.
Here's a forum I found for Medium Format Digital Cameras:
http://www.photo.net/bboard/q-and-a?topic_id=35&category=Digital+Photography
http://www.philoking.com/2006/11/22/the-real-truth-about-digital-camera-resolution/
It explains the physics and reasons why this is true.
luminous-landscape.com/essays/Equivalent- Lenses.shtml
But for Large panoramic shots, the 5.5 megapixels works wonderfully, but I wouldn't say it's better. It's more a case of different tools for different jobs.
What I don't understand, though, is why any relationship between an optical zoom and the number of MPs in a particular camera? Or did I just misread Anysia's post?
--mark d.
http://www.summitpost.org/user_page.php?user_id=26307
- Megapixels=profit
-
by ysunwoo
February 7, 2007 3:43 PM PST
- The only reason to increase resolution is to gain more profits. After 3.2MP I definitely seen a loss of color accuraccy and graininess of my photos and on screen. All my cameras and my relatives cameras, no matter how top notch they are, more MP doesn't mean great night shots. I thought the digital camera supposed to let the photographer break away from the limitations of film's graininess. And that was a complaint set by photographers for years since digital cameras arrived, but they didn't listen. So, I have bought refurbs and used, believe it or not they perform better. ie: Canon S20 for $200 and A85 for $160. They performed better than my new Canons of the same price.
-
Reply to this comment
-
See all 52 Comments >>