More megapixels vs sensor size

[Soocom1]

Look up lines of resolution, reproduction ratio and the physics behind image circle vs image plan size (format size)

There are other physical factors involved.

 

[RacePhoto]

And what use is any image if we do not view it? And to view it, we must enlarge it to viewing size. Since the cropped image is smaller, it must be enlarged more to be the same viewing size, and more enlargement reduces the resolution in that image. If printing both at same size, then like instead of printing at maybe 300 dpi, same size is only 200 dpi (assuming 1.5x crop). If that enlargement is carried too far, it becomes a pretty sorry image. Size is very important. Dpi is about inches.

That's where I thought you and I were speaking about... two different things. Yes, for prints, also crop room to adjust an image composition, there's a big difference.

I was only referring to the images themself, and given the same quality of the sensor, the image quality is the same, just cropped. And how in theory, using the sweet spot in the center of a lens, could make the images actually better on a crop camera.

But yes, I agree 100%, if I was making an 8x10 print of an image, the full frame would have more pixels and be physically better quality.

I don't print much. I think more in terms of digital on computers and news or functional, not art prints. I can see how two different purposes would have a different quality and size demands. I just counted and I have nine of my photos printed. Most of those were free offers for a sample print, like a canvas border-less, small poster, or 8x10. I'm not sure that anything hanging framed, is something I paid for. Oh wait, one panorama that's 36x12. (the frame was a dollar at a church sale)

I don't own a color printer, well I do, a laser, but that's no good for images. If someone wants one of my photos, as a print, I put it up on Fine Art America and let them pick everything. My other use is Editorial/News which is for news use, not for prints. Or for websites, where a 12MP image is pretty large.

Like the question, what camera should I buy and what lenses? "What are you going to shoot and do with it?"

I think we agree, I was just speaking from a different perspective.

ps I just bought a Canon RP. I had a 1Ds for a short time and I was impressed with the color and depth of the few images I took with it, before I sold it. Came as part of an equipment purchase, a "deal", too good for me, plus the buffer was too small for bursts. I'm hoping that it's the full frame part, not that the 1Ds was somehow different than a new R sensor?

The reason I decided to move to an R from all my ##-D cameras (CF cards) was 26MP and the camera is able to write to memory cards, high-end UHS-II SD card, from the buffer faster than it can take new photos. With something a little less, I'll still have an 80 something buffer. Yes, 5fps/4 with AF active but I also seldom shoot big bursts, however I do shoot one after another.

4 frames per second? Good enough for me.

Nope I don't do fine art, I don't shoot RAW, everything is about speed., plus fast editing and uploading ASAP.

 

[WayneF]

I was only referring to the images themself, and given the same quality of the sensor, the image quality is the same, just cropped. And how in theory, using the sweet spot in the center of a lens, could make the images actually better on a crop camera.

But yes, I agree 100%, if I was making an 8x10 print of an image, the full frame would have more pixels and be physically better quality.

I don't print much. I think more in terms of digital on computers and news or functional, not art prints. I can see how two different purposes would have a different quality and size demands. I just counted and I have nine of my photos printed. Most of those were free offers for a sample print, like a canvas border-less, small poster, or 8x10. I'm not sure that anything hanging framed, is something I paid for. Oh wait, one panorama that's 36x12. (the frame was a dollar at a church sale)

I don't own a color printer, well I do, a laser, but that's no good for images. If someone wants one of my photos, as a print, I put it up on Fine Art America and let them pick everything. My other use is Editorial/News which is for news use, not for prints. Or for websites, where a 12MP image is pretty large.

Like the question, what camera should I buy and what lenses? "What are you going to shoot and do with it?"

I think we agree, I was just speaking from a different perspective.

Sorry, it seems unimaginable, but we don't agree. You only want to speak of the lens image, and I do agree the same lens does do the same thing every time (at its same settings), and that the cropped sensor only crops that image smaller.

But so what? To use the image, we must capture it and view it somehow. The cropped image is smaller (significantly fewer mm of sensor size), and viewing the same framed image content (meaning showing the same field of view dimensions) then requires more enlargement (to view the smaller one at same size), be that in 10 inches of print or 10 inches of monitor screen size. That 10 inches is enlarged from a few mm of size. The detail of the smaller image is stretched more, as the viewed inches increase, so the pixels are further apart, so the dpi (which is viewed resolution) decreases inversely proportionately to the enlargement. The original pixel detail does not change, but it becomes coarsely shown, with many interpolated pixels inserted to fill the larger space, fake blank pixels detrimental to the viewing, and reducing viewed resolution. And looking awful if taken very far.

Some experience printing larger images should help understanding.

 

[greybeard]

Sorry, it seems unimaginable, but we don't agree. You only want to speak of the lens image, and I do agree the same lens does do the same thing every time (at its same settings), and that the cropped sensor only crops that image smaller.

But so what? To use the image, we must capture it and view it somehow. The cropped image is smaller (significantly fewer mm of sensor size), and viewing the same framed image content (meaning showing the same field of view dimensions) then requires more enlargement (to view the smaller one at same size), be that in 10 inches of print or 10 inches of monitor screen size. That 10 inches is enlarged from a few mm of size. The detail of the smaller image is stretched more, as the viewed inches increase, so the pixels are further apart, so the dpi (which is viewed resolution) decreases inversely proportionately to the enlargement. The original pixel detail does not change, but it becomes coarsely shown, with many interpolated pixels inserted to fill the larger space, fake blank pixels detrimental to the viewing, and reducing viewed resolution. And looking awful if taken very far.

Some experience printing larger images should help understanding.

When you enlarge a digital image you are enlarging a digital description of the image that was projected on the sensor. The more pixels the sensor has, the more information the computer has to reconstruct the image. When comparing CROP FRAME vs FULL FRAME sensors with comparable mp's like a 20mp CF vs a 24mp FF (and I have both), the FF will have around 1 f/stop better low light and shadow detail capabilities because of the larger pixels of the FF and, because of the larger sensor, somewhat better detail due to the added lines of resolution. A few years ago I had a Nikon D850 (FF 46mp) and a Nikon D7500 (CF 21mp) When I shot the D850 with my 105 f/2.8 micro nikkor I shot it in crop mode to give me more distance between me and my bug subject. I compared those images with those shot with the same lens on the D7500 and they looked exactly the same because the pixels were essentially the same size through similar software. However, if I compared images shot with the d7500 with a 35mm f/1.8 with images shot with the D850 with my 50mm f/1.8, there was no comparison. Why the D850 images had more than twice the information for the computer to reconstruct an image. But, if I do the same comparing the D7500 and a D750 (24mp) at the time, the difference was not night and day. At iso 100 you had to do some serious pixel peeping to see any difference. Both images had basically the same information for the computer to reconstruct an image. However, when the ISO reaches 800 and above the difference becomes much more pronounced. ISO 1600 on the D7500 looks like ISO 3200 on the D750. So, (IMHO) If I'm shooting bugs with a macro lens or long tele stuff, both of which I will end up cropping, I'm better off shooting a crop frame 20mp than I am shooting a FF 24mp because the final image from the CF will have more information for the computer to reconstruct an image. However, if I'm shooting landscapes, portraits, or anything in low light, I'm much better off with a FF. This has been MY EXPERIENCE and is not based on what I have read on some forum.

 

[AlanKlein]

Doesn't DOF enter the picture as well.

 

[greybeard]

Doesn't DOF enter the picture as well.

It does when using lenses that have a similar field of view. When comparing FF with a 50mm with a CF with a 35mm the difference is again 1 stop. The CF shooting at f/2 will have a similar DOF as the FF shooting at f/2.8. Again 1 stop. The difference when comparing FF with 4/3 would be greater though I have no experience with 4/3rds

 

[Don Fischer]

OK, so let's say I have a 300mm lens on a Micro43rds body (so 600mm FF equiv) - and the camera shoots at 20megapixels.

Now let's say I have a 500mm lens on a FF body and the camera shoots at 45megapixels.

Would the FF get a better crop into the image vs the M43 in terms of quality?

Seriously debating switching to FF from my M43 - as in my head, the image would be better cropped in FF to get that extra 100mm than the actual longer focal length on M43. So seeing if this is also an advantage vs just more light capturing and less noise ability

Something I keep reminding myself of is to many people get all wrapped up in function and don't get form, the ability to shoot a good picture. My newest camera is a Panosonic P&S and I read it has a 1" sensor. That simply doesn't mean squat to me. What does is getting photo's I like! I blew up a photo from it to 12x24" and thask tripped my trigger. If the camera you have let's you make photo's you really like and it doesn't matter what size the sensor is!

 

[AlanKlein]

Something I keep reminding myself of is to many people get all wrapped up in function and don't get form, the ability to shoot a good picture. My newest camera is a Panosonic P&S and I read it has a 1" sensor. That simply doesn't mean squat to me. What does is getting photo's I like! I blew up a photo from it to 12x24" and thask tripped my trigger. If the camera you have let's you make photo's you really like and it doesn't matter what size the sensor is!

1" is great. I use a Sony RX100iv and it gives me blowups that look great on a slide show on a 75" HDR 4K TV. Great for travel.

Flickr Search

www.flickr.com

 

[Don Fischer]

1" is great. I use a Sony RX100iv and it gives me blowups that look great on a slide show on a 75" HDR 4K TV. Great for travel.

Flickr Search

www.flickr.com

Nice secnic's you have there! Wish I was better at them. Need a lot more time trying them. Do you blow some up pretty big?

 

[AlanKlein]

Nice secnic's you have there! Wish I was better at them. Need a lot more time trying them. Do you blow some up pretty big?

Not as prints. I do make slide shows that I show on my 75" 4K TV. The 1" sensor of this camera provides great shots.

Here's an example taken with the Sony RX100iv 1" sensor. I uploaded it to YouTube at 4K (3840x2160). Set the bandwidth setting for 4K for your 4K TV or 2K for your 2K TV (1920x1080). You can also watch on your monitor but try on your TV to get an idea about how it looks big.

 

[Don Fischer]

Not as prints. I do make slide shows that I show on my 75" 4K TV. The 1" sensor of this camera provides great shots.

Here's an example taken with the Sony RX100iv 1" sensor. I uploaded it to YouTube at 4K (3840x2160). Set the bandwidth setting for 4K for your 4K TV or 2K for your 2K TV (1920x1080). You can also watch on your monitor but try on your TV to get an idea about how it looks big.

Looks like aa 58 T-Bird! Love it. Don't know how to get stuff from my computer to the TV!

 

[snowbear]

Looks like aa 58 T-Bird! Love it. Don't know how to get stuff from my computer to the TV!

One way is set up the slideshow on a computer, then feed that to the TV. I've done like this with Powerpoint (or the Mac equivalent) and sent it to a projector at a class reunion.

 

[greybeard]

One of the problems with this kind of question is that it assumes that everything else is equal, like the 300mm lens and 500mm lens both have equal performance at all apertures (not very likely in real life), and for that matter, equal equivalent apertures (remember the crop factor scales aperture too). And that the combination of the image sensor and camera software results in identical noise figures (also, not likely), and that the resolution of the 300mm lens is not the limiting factor to final sharpness when the image is cropped, and that the 500mm lens has the ability to resolve an image as well as the 30mm lens. Life isn't that perfect.

The real practical questions are like, is the maximum equivalent aperture with the 300mm on the cropped sensor the same as the 500mm on the full frame? What's the difference in cost and weight? Big glass is expensive, if it's going to be good too. And of course, if you're going to compare sensors, shouldn't they produce images with the same pixel count? Or you've thrown another variable into the mix.

And speaking of pixel count, as has been mentioned, you can't really have that discussion without including at least some reference to the image viewing angle (a combination of image size and distance from viewer). Because once you have pixels smaller than can be seen with the unaided eye with 20/20 vision, you're kind of done.

While we're on that, also don't dismiss a lower pixel count sensor out of hand just because it has fewer pixels. The two factors that don't get menioned nearly enough are, fewer pixels often mean larger pixels, which pick up more light and result in less noise. And some of the new image scaling tools are, frankly scary good. I'm playing with the Topaz apps, for example. Scaling is something they do really well.

While I personally would love to have a set of f2.8 or larger lenses on a full frame camera, I can't justify the expense and my shoulders won't carry the weight anymore. Not that FF cameras cost more, probably not, but the lenses very well might.

I'll add this video, which doesn't answer the question, but does highlight many of the trade-offs.

This video confirms everything I have written on the subject of FF vs CF

 

[RacePhoto]

Pixel Pitch.

 

[greybeard]

Comparing the 2 on a computer monitor may not show you all that much difference as would enlarging the images to say 16 x 20. You would have to go much larger to see a huge difference. However, schlepping around a FF body with a 500mm lens compared to your 4/3rds with a 300mm is a huge difference physically. If you are happy with what your are getting from you 4/3rds then, it ain't broke, don't fix it. (just my old fart humble opinion)