Megapixels

[The_Traveler]

The content of the Myth column was as expected but a couple of the letters were much more on point

The article assumes that the only difference in digital cameras is the megapixels. This, certainly, is one of the things the media is constantly putting forward, as well as the camera companies and the suppliers to the camera companies (they make more money on higher-priced, latest technology products).
The truth, of course, is that the number of megapixels is only one difference that shows up in the result. Other differences include plastic vs. glass lens, a cheap glass lens vs. a high quality, multicoated lens, the speed of the camera, the way the aperture opens and closes, the quality and size of the CCD, the firmware inside the camera, etc.
As a result, a camera with 8 megapixels may produce poorer images than a better camera with only 6 megapixels. And that difference would be observable by anyone.

No reason to challenge your basic premise: 5 megapixels worth of data makes for a very nice image, assuming little or no cropping.
But that doesn’t mean that a 13MP camera will take the same quality picture as a 5MP one. More likely than not, the 13MP model will have less noise, more range, better quality at high ISO equivalents, etc.
Well-lit settings with cute kids who don’t squirm too quickly are important to most of us, but they’re not very challenging to cameras. You don’t have to worry about purple fringing because the soft-focus look minimizes contrasts. You can use a nice, low ISO rating & get a good exposure. Etc.
Not to mention all the other features that manufacturers feel obliged to toss in when building a premium camera.
So: partial agreement only. It’s not really clear (unless all you care about is blowups of the kids) that a 5MP camera is the functional equivalent of the 13MP monster.

And somewhere in the responses someone also mentioned that at full frame and with printers that produce the same quality print image from 200 pixels/inch through 300 pixels/inchthe printer is the limiting factor. Try cropping each frame to 1/4 of the original frame size, then print and the difference between 5 and 10 megs will show up with even these printers. And it will show in picture tones and noise.


I traded up from my D70, without qualms, for a couple of reasons, better camera build and features but most important different and improved sensor. And it wasn't the 2 M difference, it was the better response to the same input image!!

 

[ksmattfish]

The term "megapixel" will become meaningless as new technologies are introduced. As people have mentioned, not all megapixels are created equal. Sigma is about to release several cameras with the 3 layer Foveon sensor that they are advertising as 14 megapixels, but the actual resolution is 4.6 megapixels (2652x1769). If the manufacturers can't agree on a definition of megapixel, then they can redefine it to mean anything they think will lead to more sales. In the end we will have to go back to looking at real photographic prints, and make any decisions about quality with our own eyes, instead of relying on spec charts.

 

[Alpha]

I had it out with a guy at my local pro shop recently about this issue. I was talking to him about the BetterLight scan backs, the biggest of which can capture an image 14836 x 20072 pixels. What he asked me was "why on earth would you need that many pixels?" The real answer that is that it's a question of the relationship between resolution and coverage.

Let's take 4x5 for example. The native resolution of negative film is usually around 3000dpi. So a 4x5 negative is about 12,000x15,000 pixels. Now, conceivably, I could build a sensor the size of a postage stamp that could capture that much information. Would my image be as good as a 4x5 negative? Yes and no. It would have as much detail, but it wouldn't be able to capture the same image because of the crop factor/distortion, which is what people are talking about when they speak of small vs full-frame sensors. This occurs because 1) A sensor that is not full frame is just that-- it's not capturing the entire frame (image). This is precisely the same thing that would happen if you tried to shoot with a 5x7 lens on a 4x5 camera, i.e. the resolution is still the same, but the image that is being projected onto the film is actually much larger than the film area, and therefore you can only capture part of the frame. In order to capture the full frame, you'd need a bigger back/sensor. 2) You're distorting the image basically by taking something that was intended to fit in a particular area and squeezing it into something smaller. Now back to the 4x5 analogy: In order for you to capture the same image as an actual 4x5 film camera, you literally need a 4x5 inch sensor, that can capture 12,000x15,000 pixels.

A side note on the printer issue: printers are always the limiting factor in digital work, no matter what kind of printer it is. In fact, sometimes, inkjet printers can be the best out there. Consider that a LightJet prints at 300 dpi. I mean it's an impressive machine that costs more than most people's homes, but 300dpi is laughable. Or you could try printing on something like a ProPalette (for those who don't know this was a device made by polaroid that takes a roll of film and exposes a digital image via computer onto it). Even the ProPalette is 300dpi. I pay very careful attention to my resolution and image size. My motto is to scan negs 1:1 @ 3000dpi and resize as needed for printing or web applications. I know that isn't practical for everyone, but I find working with archival-quality scans to be the most reassuring part of my entire digital darkroom setup.

 

[cal_gundert05]

Using your car analogy, I believe I own a bicycle...and a crumby bicycle at that.

I've heard of people beingself-conscious walking around with telephoto zooms and other big lenses, but trying walking around with this.

 

[ksmattfish]

Now, conceivably, I could build a sensor the size of a postage stamp that could capture that much information. Would my image be as good as a 4x5 negative? Yes and no. It would have as much detail, but it wouldn't be able to capture the same image because of the crop factor/distortion, which is what people are talking about when they speak of small vs full-frame sensors. This occurs because 1) A sensor that is not full frame is just that-- it's not capturing the entire frame (image). This is precisely the same thing that would happen if you tried to shoot with a 5x7 lens on a 4x5 camera, i.e. the resolution is still the same, but the image that is being projected onto the film is actually much larger than the film area, and therefore you can only capture part of the frame. In order to capture the full frame, you'd need a bigger back/sensor. 2) You're distorting the image basically by taking something that was intended to fit in a particular area and squeezing it into something smaller. Now back to the 4x5 analogy: In order for you to capture the same image as an actual 4x5 film camera, you literally need a 4x5 inch sensor, that can capture 12,000x15,000 pixels.

When you put a 150mm lens designed for 5x7 coverage on a 4x5 camera, it takes exactly the same photo as a 150mm lens designed to cover 4x5. Except that since you are using more of the center of the image circle you would be less likely to get edge and corner distortions. It's not crammed or distorted to fit the smaller format. Most of my 4x5 lenses cover 5x7, and some even cover 8x10. If the lens only covered 4x5 I couldn't use the view camera movements. The only 4x5 lenses that actually only have a circle of coverage of 4x5 are the cheaper press lenses that weren't designed for use on a view camera.

Any given focal length will be a wider angle of view on a larger format, so you might have to change where you take the photo from if you were trying to match angles of view between different formats, but it doesn't affect resolution or image quality in the camera. Except as I mentioned above, if you are using a lens with greater coverage than you need, you are using more of the best part of the lens.

If you somehow loaded a view camera with 35mm film, and took photos with a 90mm lens designed for 35mm film, and with a 90mm lens designed for 4x5, you will get exactly the same composition and angle of view, except the lens designed for 35mm would probably be sharper. Small format lenses are designed with much more precision, because the smaller format, and increased enlargement ratios, are more demanding.

If I had a APS-C sized sensor with 12000x12500 resolution I'd be using the lenses that were designed to be used with it, or my current small format lenses. Large format and medium format lenses don't perform as well regarding resolution and sharpness as small format lenses, because they don't need to. The huge chunk of film makes up for it.

When they introduce the small format digital sensor that has the resolution of 4x5 film they will also introduce lenses designed to be used with it. It's my opinion that we'll see a small format digital camera (DSLR sized) that can beat the resolution and dynamic range of 4x5 film within 10 years. It may cost more than anyone here can afford, but it will exist.

EDIT: Oh, and the native ISO of that super DSLR will be ISO 6400 or better....

 

[ksmattfish]

I've heard of people beingself-conscious walking around with telephoto zooms and other big lenses, but trying walking around with this.

The flip side of that tiny camera is this...

http://www.engadget.com/2006/09/13/carl-zeiss-creates-over-five-foot-long-telephoto-lens/

 

[boclcown]

"Personally I wouldn't automatically assume more megapixels is better. There are a couple of reasons for this; firstly the larger file sizes mean more storage and more money spent on it."

It is true that more megapixels produce larger files, but in my camera, I can turn down the Megapixels to 5.5, and even 2.5. Someone who has a 6 megapixel camera never has the option of 10 megapixels, even if a situation calls for them. Anyways, storage is becoming less and less of an issue due to dropping prices in CF and SD cards (and a rise in capacity).

 

[Alpha]

Matt, my point was simply that I think to a great degree, this is a question of field of view, and not simply megapixels. People who take architectural photos don't shoot LF because the quality is better.

 

[The_Traveler]

Matt, my point was simply that I think to a great degree, this is a question of field of view, and not simply megapixels. People who take architectural photos don't shoot LF because the quality is better.

They shoot LF because of the swings and tilts of LF cameras.

 

[Mithlondor]

All hail film.

That is all.

 

[castrol]

Here are two pictures I took with two different cameras. They were both rated
at 6.1 megapixels and were taken as such. I shrunk them down a bit
just to make it easier. I can post the full res versions if anyone is really interested.
It was just a quick comparison just to show some folks that all 6.1
megapixels are not created equally and why the word "mega-pixel"
actually means squat to a photograph.

 

[Alex_B]

good example of what i keep telling people

megapixels beyond 7 or 8 for p&s cameras is an advertising gag.

megapixels beyond 10 in DSLRs makes you buy extremely expensive lenses, because else you see nothing of the anticipated resolution

megapixels on a high-res scan from small-grain film in MF or LF cameras can be very real and convincing though (if you go to ultra-large high-res printing or tiny crops)

 

[Jeremy Z]

The camera technology I am most looking forward to is when full frame 35mm sized sensors are made to be affordable. It will happen, eventually, and we'll look back on these days of magnification factors and laugh.

Having a 5D and being able to use 35mm EOS optics is a nice dream for many of us. I guess it is a matter of priorities. If I didn't have a super-nice sport-touring motorcycle, I could have a super-nice camera outfit instead.

 

[THORHAMMER]

Since were dealing with point and shoots, and dslrs, sensor size is one of the largest factors. MP count doesnt mean too much nowadays, just the way the P&S cameras are marketed you need to memorize the MP's to know which year it came out... lol

the sensor size, and then the lens makes the largest difference, but
in the canon world a proper photographer can use a 6.3 - 12mp and you shouldnt really be able to tell at 8x10 sizes.

sure you will get less grain in larger prints with the larger sensor/mp combo

I would be curious to know if you used the same lens with the 5D as you did with the 20d, and if your processing was the same. That aside, i will concurr the 5D is a lot better in larger printing due to the sensor and almost no image degradation due to the sensor size.

people should still be able to get decent 16x20 and 18x24 prints from the 20D maybye larger if your not super picky...

 

[binglemybongle]

Right........

Best results - big CCD, high megapixel
Worst results - small CCD, low megapixel
Average - small CCD, big megapixel OR big CCD, small megapixel.


What i want to know is when you say a 10MP sensor compared to a 6MP sensor will create more noise (same sensor size), will the 10MP image be worse than the 6MP at the smae print size or just when you print it larger?

Sorry i've confused myself there!

Will the 10MP and 6MP be the same if both printed say 10" by 10" or will one be "worse" than the other?