Whats the Difference??

[Iron]

Between a full size sensor and a non full size sensor??
How does it affect a picture??

[random one]

full size (aka. full frame) sensors capture the whole image (relatively speaking) and non full size (aka. APS-C) sensors capture a middle portion of the whole image.

it's kind of complicated, but simple at the same time. take a look at this website and scroll down a bit until you see the diagram and the example pictures. you can read everything if you want, but the pictures and diagram should give you an idea.

http://www.minasi.com/photos/dslrmag/

this is just some random website i just found in a google search. its about 2 canon models, but still explains the full frame/ crop difference.

[TheDigitalMage]

There are many ways to look at this...

Zooming a lens and cropping an image have the same effect - i.e. to take a (the center, in this case) portion of the image and make it the full image; making it seem as if you were closer to the subject than you really were. Many foibles found in lenses are manifested around the edges. The smaller sensor simply discards what would be the edges, on a full-sized sensor, and the defects along with it.

Lets say you like to shoot birds (UFOs - whatever). You're in luck! I have a 300mm lens. With a Tamron doubler it becomes a 600mm lens. The aps-c size sensor contributes another 50%, making the rig a 900mm lens. Check into the price for a 900mm setup for full frame - lol

Some people will exaggerate the negative. I'm not sure there is one. I've seen excellent images from cams containing sensors that are much smaller than aps-c! People even say the 50D is no good, basically because it lacks a full-sized sensor on which to place it's bumper crop of pixels. I have images, that say otherwise...

[Photo-John]

The crop factor of the APS-C sensor vs the full frame sensor is the obvious difference. The tradeoffs there are the increased reach of the smaller sensor vs the wider angle potential of the full frame sensor. The wide-angle issue used to be more important but there are plenty of APS-C specific super-wide lenses now.

The two previous replies didn't mention pixel size, which I think is the most important difference between full frame and APS-C sensor DSLRs. A sensor with more surface area will always offer the potential of better image quality than a smaller sensor. Simply put - bigger pixels collect better information. That's why the image quality of the full frame Canon EOS 5D Mark II is so much better than that of the Canon EOS 50D.

Does better image quality mean you should buy a full frame camera? Not necessarily. A few years ago I made the decision to commit to the smaller sensor cameras. I realized that the extra reach of the cropped sensor and the smaller body size were more important to me than the image quality the larger sensor offers. APS-C sensor image quality is excellent now. Most people don't need and probably won't even benefit from the improved image quality a full frame camera offers. Usually, it's best to buy a less expensive camera and a better lens. That's my standard advice to photographers who are trying to decide what camera to buy. The truth is, even the cheapest current digital SLR is better than ones I was making money with 5 years ago. So analyze your real needs and budget and you'll likely find that full frame isn't something you really need.

[TheDigitalMage]

"That's why the image quality of the full frame Canon EOS 5D Mark II is so much better than that of the Canon EOS 50D."

(50D at night!)

If so it must be so much better you can't see it... so like, what good is it?

Keeping your perceptions updated is perhaps the most difficult of things. However, your post seems to contain obvious errors. Like saying larger pixels are better? Is that why manufacturers of printers have striven so hard to make higher DPI machines? Resolution is the name, of the game. And the 50D has a boatload of that!

[brmill26]

"That's why the image quality of the full frame Canon EOS 5D Mark II is so much better than that of the Canon EOS 50D." If so it must be so much better you can't see it... so like, what good is it?

Your image was taken at ISO160 on a tripod. Reshoot it handheld at ISO6400 and I will give you an absolute guarantee you'll see a massive difference. Alternatively, print it out on a 40x60 print and compare it to a 5DMKII. I bet you'll be able to tell a difference.

The point being, those are the major advantages of full frame - extremely low noise at very high ISOs and maximum detail. If your type of photography doesn't require either of those, then you're right, you don't gain much by it. Just the same as the Rebel XTi will take every bit as good of a nightscape skyline photo as the D50 at ISO100. It all depends on what you need the camera to do. If you're not pushing it in extremely low light or extremely detailed applications, then you won't see a huge gain. Likewise, on a bright sunny day outside, many good P&S cameras can take just as nice a photo as an SLR.

Keeping your perceptions updated is perhaps the most difficult of things. However, your post seems to contain obvious errors. Like saying larger pixels are better? Is that why manufacturers of printers have striven so hard to make higher DPI machines? Resolution is the name, of the game. And the 50D has a boatload of that!

You realize John owns this site, does most of the reviews, attends most of the major trade shows, and is a professional photographer, right? His job, by definition, is to keep himself updated on all things photography and camera related. He knows the technical stuff in and out, as well as having shot and tested nearly every camera on the market.

Pixels on camera sensors are not the same as dots on printers. In fact they're entirely unrelated. Camera pixels are like tiny lenses. Bigger lenses take in more light which, as John said, is equivalent to more information. When the digital circuits amplify this information, you get a much more detailed, cleaner result from better information inputs. That results from larger pixels (aka photosites).

Resolution, on the other hand, is almost entirely driven by marketing now. They want big numbers on the side of the box. But resolution is beginning to reach its technical limits. Squeezing more pixels onto a fixed area sensor means the pixels have to be more and more compact. This limits the amount of light they can take in, which limits the quality of information. Thus, when it's heavily amplified (ISO is simply amplification; high ISO = high amplification), you get tons of noise b/c the information taken in isn't as good. That makes the noise reduction have to work harder, which comes at the expense of detail. Thus, resolution is a false promise b/c you can have an image with a higher technical resolution (ie, it is larger), but actually less detailed b/c of the required noise reduction. Even if you turn NR off, you simply get a larger image with more noise, not more detail.

The 50D is landmark because it marks the first time that even though sensor technical resolution increased from the 40D, actual detail captured did not improve very much if at all. That's b/c the smaller pixels are generating enough noise that NR has to run at a higher level to maintain a clean image, thus wiping out any detail gains. On top of that, since crop sensors use only a small portion of the glass, it means that portion of glass must transmit all the detail that would normally be transmitted by a larger area of glass. At some point, physics dictate that the glass simply cannot transmit more information per molecule. Some people argue we're beginning to reach that point for all but the very best lenses - ie, consumer lenses cannot physically transmit enough information for the 50D's sensor. Whether that's actually true or not is very debatable. But the overall point is that most people agree crop sensors are reaching the point of diminishing returns on resolution. Certainly improvements in noise processing will allow things to progress more, but at some point glass can only transmit so perfect a ray of light per area, and that practical limit is on the horizon.

The old saying applies - it's quality, not quantity. Larger formats have always produced more detailed images (35mm vs. 120 medium format, MF vs. large format 8x10 sheet film). It's simple physics. But the *reality* is that 99.9% of us will never print large enough to need the resolution of an 8x10 sheet film. For most people, 35mm is (was) plenty enough quality for their needs, and it was much cheaper and much easier to handle. The same thing applies here. Full frame has its advantages, but it also has costs. How that costs / benefits analysis shakes out depends on what you plan to do with it.

[TheDigitalMage]

"Your image was taken at ISO160 on a tripod. Reshoot it handheld at ISO6400 and I will give you an absolute guarantee you'll see a massive difference. Alternatively, print it out on a 40x60 print and compare it to a 5DMKII. I bet you'll be able to tell a difference."

I'm sure! Just as shooting any image WRONG would produce a visible difference ... (wtf?)

I'm almost 60 and I've been exposed to techno-babble far more than most. Your post represents what has been said. Mine is the product of having the camera.

It all boils down to this: "According to the numbers, it should... but it DOESN'T." Numbers say pretty well what you want, once you learn to manipulate them...

Of course a more densely packed sensor, OF ANY SIZE, will cough up more noise. That, however, certainly does not support your argument, which is that the larger sensor inherently has more IQ. If you put the same density on the larger sensor, you get the same result.

Yes, you can show an image shot in conditions that favor your cam and it will be better (less noise). I can (and have - lol) show another image, shot in conditions that favor my cam, and it has better IQ. Plainly, nither is better all around.

Your last line, with which I fully agree, tells the tale. Most of us want to take pictures in the light. In that situation, the cheaper cam is the definite winner, because it offers more resolution.

[spiraleyes]

TDI, I've seen the images that you post and repost all over this site. I'm sorry, but I haven't been impressed with any of them. Your examples don't really make your case.

As for the picture you posted in this thread - Sure, the colors are pretty, but there is a distinct lack of detail in the entire picture. Maybe it's just the size, maybe it's my monitor.

I challenge you to post any and/or all of your "prized pictures" in the critique forum.

[Photo-John]

However, your post seems to contain obvious errors.

Keyword - "seems"

There's nothing wrong with the 50D. It's a fine camera. But as far as low light image quality, it can't even come close to the 5D Mk II. Sure, you can take nice photos in low light with the 50D if you do it right. But with the 5D Mk II, I can do it wrong and still get good photos. With the 5D Mk II I can shoot handheld at ISO 1600 or 3200 in a dimly room and get absolutely usable images. That's partly because of the 21MP resolution, which minimizes noise. And it's partly because the larger pixels collect cleaner data.

You can't compare printer resolution to camera resolution. Both use the terms "dpi" and "resolution," but they aren't referring to the same things. A useful way to look at the larger pixel issue is to think about collecting water in the rain. Which will work better - a cup or a bucket? The bigger pixel catches more light and can deliver cleaner data with less noise.

And since you think it matters - I have both cameras right now. I don't have anything against the 50D - I bought it. In fact, I prefer it over the 5D Mk II for most purposes. But for low light photography, the 5D Mk II just flat blows the 50D out of the water. That doesn't mean you can't make a good low light photo with the 50D - just that the 5D Mk II is better. This isn't techno-babble. I'm saying this because I've used both cameras and I can even show you examples if you want.

[TheDigitalMage]

No, John, I don't doubt that you get something for the extra money.

However, we will discuss the resolution issue, if you like? Everyone is paying >$1,000 (in bad times) to upgrade to HD tv, and for the simple reason that HD has more resolution (and, therefore, works better with screen sizes never thought possible when tv was originally designed).

All you need to do, to experience the effects of resolution in a personal way, is right click on your desktop and change the resolution of your screen. This has some undesirable side-effects (like making things smaller) but in general everything looks much more defined at the higher resolutions... which is why they have come along.

I, too, have tried a 5d MK II. I didn't see anything to write home about but I did not try anything fancy in the camera store. I do not like being put in a position of knocking that, or any other particular camera. That's not my intent.

Your post was not the one that caused me to speak of techno-babble. Techno-speak becomes techno-babble only when it is parroted, and not understood by the person that uses it. There's a whole lot of that going on! A whole lot...

Mix that with momentum; with stuff that originated in the days of film (as one in this thread actually did!) and you get somewhere, but not to anyplace real. In the real place, which can be described as a nanosecond in technology, the explanations lag the facts, not the other way around! Using yesterday's axioms to describe/predict today's... let alone tomorrow's achievements is like chewing, on a brick.

Thanks to dpreview I was well aware of the noise issue before I bought the 50D. My eyes told me that cam is better - a LOT better - than the 40D I had. I had both, at the same time. I bought the 50D right (and then some!) as circuit city closed down. I could have dumped it on ebay and actually made a little money. It was the 40D that left.

Because of that one, in my perception extremely flawed review, every self-appointed "chat room engineer" on the planet is bad-mouthing the 50D. That suits me. It causes fewer people to buy the cam which, in turn, gives me the edge. But it's just not right...

[TheDigitalMage]

TDI, I've seen the images that you post and repost all over this site. I'm sorry, but I haven't been impressed with any of them. Your examples don't really make your case.

As for the picture you posted in this thread - Sure, the colors are pretty, but there is a distinct lack of detail in the entire picture. Maybe it's just the size, maybe it's my monitor.

I challenge you to post any and/or all of your "prized pictures" in the critique forum.

Ah, where to begin...

I suppose I could be referred to as TDM, but TDI? If this is an example of your powers of observation...

It so happens that the original image is sharp and the reduced image is not because I did not apply sharpening as I quickly prepped it for use, on the web. But, let's assume otherwise.

Would the image be "bad" because, in the vision of the artist (me!) the scene was soft, and dreamy, and so the piece was purposely rendered in that manner? Your post so implies.

You, nor your friends, will critique my images. For the main reason that the "quality" of an image is SUBJECTIVE, adhering to no set of rules and changing, even reversing, completely, depending on who you ask.

One man's garbage is another mans gold... Were it up to me, for example, there would be no country "music" and certainly no rap!

But if you continue trying to start a fight with me, sooner or later you will succeed...

[Photo-John]

I read the dpreview, too. And I actually had both the 50D and the 40D at the time (my dad has the 40D now). After reading their review I did careful comparisons of both the 40D and the 50D by making controlled test prints from both. I found that the 50D wasn't really anny better and may even have more noise at low ISO settings. That was disappointing to me because I was excited about the new sensor strategy that Canon introduced with the 50D. Alas, it didn't work as well as I'd hoped. But I have no regrets about my 50D purchase. And when I'm not comparing prints from both cameras with a magnifying glass, the image quality differences don't matter one bit. And yes, I do like the extra detail.

None of that changes the fact that the 5D Mark II is way better in low light and that bigger pixels are better. When I write my 5D Mk II review that's what I'll be listing as the main strength of the camera. Because I can pump the ISO up to 1600 or 3200 and shoot handheld without worrying about the images getting all chunky or blurry, the EOS 5D Mk II is a low light shooters dream. If I did a lot of street photography, concert photos, weddings, or parties the 5D Mk II would be a no-brainer. But I shoot mostly action sports or in the studio where low light isn't an issue. So the buying the 50D means I save some money, get more speed, and have a smaller camera body.

[cm_mtb]

This is starting to sound a lot like that "L vs. non-L lens" thread that TDM started.

[spiraleyes]

It so happens that the original image is sharp and the reduced image is not because I did not apply sharpening as I quickly prepped it for use, on the web.

If you are trying to prove a point with a subpar example, don't expect a good response. If you want to get your point across, share a quality picture. I don't doubt that you may have a better copy lurking somewhere, but how can I assume such when you don't share such?

It's apparent that you think there is enough of a difference between lenses and bodies of different price ranges to spend your money that way. The reality is that there are enough people that do think there is enough difference to pony up the cash and step up to the better products, otherwise Canon (or any other manufacturer) wouldn't have a pro lineup available. Your opinion may be that it's an ego thing, but the majority disagree.

/end

[Photo-John]

Your opinion may be that it's an ego thing, but the majority disagree.

To be fair, for many it is an ego thing. That's why I always ask people who want full frame cameras, "why?" Often, they don't have a good answer. They only have the impression that full frame (or whatever else) is better. But they don't know why. And it may not be better for them. Like I said, a few years ago I decided that full frame was not for me. And I haven't regretted it one bit.

[TheDigitalMage]

To be fair, John, the process of making prints brings so very many things into the comparison... Who knows WHAT happens in the algorithms that process the image in the name of matching the color to the printer. Obviously, the image is downsized in that process.

In my mind the only fair comparison is what has come to be known as pixel-peeping, on the same computer and monitor. Even that may be flawed if you shoot raw, as I do I just realized that, recently.

Soooo..... just because you are not seeing a difference in your prints, don't believe the difference your eyes see is not there, because it is.

There is no point in providing images. If they are resized to fit this message they are ruined, in terms of seeing their quality. If I upload a full-sized image (as I did before) my hosting company threatens to shut my site down, because of the bandwidth usage...

So if you want to see one...

http://www.mageProductions.com/IMG_0620.jpg

This is an ordinary image - not a masterpiece - but you will see the amazing detail (RESOLUTION) that no other cam can provide!

[Frog]

I think you need to do some more research first, photograhm.
Its a nice enough pic but it isn't because of the camera used.

[Photo-John]

In my mind the only fair comparison is what has come to be known as pixel-peeping, on the same computer and monitor. Even that may be flawed if you shoot raw, as I do I just realized that, recently.

For what it's worth, I worked in high end photo labs for years before I started this site. So I know a little about printing, too - optical and digital. When I make test prints, they're from controlled studio test photos of the same subject at more or less the same framing. I make no adjustments to the images and push them straight through the printer. Whatever the printer does, it does to both images. So it is a fair way to compare.

When you compare on the computer at 100% you only see how images compare pixel to pixel. But that doesn't really work if you're comparing a 20MP camera to a 10MP camera. If they both have the same noise level on the computer monitor, the higher resolution camera will effectively eliminate that noise in two prints of the same size. So comparing two cameras of the same resolution on the computer is fine. But comparing cameras with different resolution must be done at output size - that means prints. Your 15MP 50D might look the same as the 5D Mk II on the computer. But the 5D has 33% ore pixels - as well as larger pixels. So the larger pixels collect better data and then minimize the noise compared to the 50D. I did close comparisons of the 40D, 50D and the original 5D and the original 5D surprised me by being better than both the 40D and the 50D.

You can make all these same comparisons yourself, if you like. All of my sample images are available for you to download at full resolution. The only exception is, unfortunately, the new 5D Mk II. I have the files but they're too big for our gallery. I need to make an HTML page to share them but I haven't had time to do it yet. But they will be available. Here's a link to the sample photos page: http://www.photographyreview.com/samplephotoscrx.aspx

[Anbesol]

Wow much ado about nuthin... anyway just wanted to comment on one thing.

Lets say you like to shoot birds (UFOs - whatever). You're in luck! I have a 300mm lens. With a Tamron doubler it becomes a 600mm lens. The aps-c size sensor contributes another 50%, making the rig a 900mm lens. Check into the price for a 900mm setup for full frame - lol

People frequently say this about APS, as if the crop factor was a benefit of getting more reach, but I think its more the drawback of not using the entire lens. Throw the same lens on a full frame and the entire lens is in use, throw it on APS and only the center. Of course, framing right out of the camera is a lot different, but with resolution and IQ difference, its quite simple - crop something at 300mm to get a 450mm equivalent, essentially the same thing as using an APS sensor. Of course, aps also only uses the sweet spot of the lens (assuming it gives full coverage), which is a major benefit.

As of now, the 50D and A350 have the smallest pixel density on the DSLR market - the A900/D3x actually have a pixel density equivalent to a 10mpix aps, and they're 24. nuff said.

I'd also comment an additional benefit full frame has for portraiture - the 24-40mm (equivalency) range becomes much more usable with lower distortion, in studio settings in smaller environments and for larger groups - this is a very big deal. If you photographed a person with a 27mm equivalent on APS - watch their face and bodies turn round, on full frame you are by and large alright (varies with lens).

[TheDigitalMage]

Myself, I would select what my eyes see over "research," any 'ol day! What my eyes see is (most certainly!) what photography is about.

Soooo... like, can a totally blind man select a camera, via "research?"

[TheDigitalMage]

.... Whatever the printer does, it does to both images. So it is a fair way to compare....

Uh, John? If the software that feeds the printer downsamples both images to, for example, 300 d.p.i. x the width of the print, then the software downsamples the larger image to a greater extent than the smaller image... therefore the comparison is equal, nor fair... as the printing process certainly does not do the same thing, to both images.

In one case the software will take 5 adjacent pixels and average them, creating one pixel on the print, then seven pixels from the larger image for one, on the print. It seldom comes out even, of course... it might actually be 5.2 and 7.6.

Anyhow, the more of that, the less the final quality - FOR SURE!

You seem to be a knowledgeable guy but... somehow you are not putting what you know together... following along in the bitstream, so to speak... I don't get it.

But I say again, the only fair way to compare is by pixel-peeping images that have not been resized... or processed, in any way. There is no perfect way. That, my friend, is as good as you can do.

[Frog]

I wonder if the OP,(iron), is getting anything out of all this.

[Anbesol]

Well I sure am getting a good kick out of it ;).

[Photo-John]

As of now, the 50D and A350 have the smallest pixel density on the DSLR market - the A900/D3x actually have a pixel density equivalent to a 10mpix aps, and they're 24. nuff said.

As far as image quality goes, this is the pivotal point of the whole discussion. Bigger pixels = better image quality.

[Ultra Magnus]

But I say again, the only fair way to compare is by pixel-peeping images that have not been resized... or processed, in any way. There is no perfect way. That, my friend, is as good as you can do.

Maybe the only pointless way.

The ONLY thing, the ONLY ONE thing that ever matters, EVER, is how the picture looks in it's final intended form, whether printed or on the computer. If you are a fine art photographer printing wicked big, then shoot the same scene w/ each camera, and print them wicked big, and compare the final products. If you are printing no bigger than 8x10, then print pics from the different cameras in 8x10 and compare. Pixel peeping is a total waste of time.

BM