What is the minimum usable megapixel count ?

I still use my 1Dc regularly; it features a majestic 4.1MP on a CCD sensor. It works perfectly well for web use. When printing, the limit is 10x8 at 250dpi.

Both these rugby shots are cropped to about 2/3 original size.

IMAGE LINK: http://s17.photobucket​.com …ionship/2A5D130​7.jpg.html

IMAGE LINK: http://s17.photobucket​.com …ionship/2A5D133​7.jpg.html

I started with a 300D with the same 6.3 Mpx sensor as the 10D and only upgraded it to a 20D with 8 Mpx last year. Yes more would definitely be nice, as I almost always focal length limited even when I hire a 100-400 L. 18 Mpx would be super but I can see no way that I can acquire a7D by legal means. It would have to be a 7D as I will not use SD cards in a DSLR, I've had just too many issues with them in the past.

Alan

The best thing about these threads is how the 'optimum' sensor size changes over the years. In 2025 somebody will post the exact same thread bemoaning the obsession with GP and how 500 MP should be enough for anybody.

Bill Gates said "640K ought to be enough for anybody," and to this day he keeps smacking his head and muttering, "When will I learn to keep my mouth shut?"

cdifoto wrote in post #16240322
Fixed that for you. Most normal people stand back far enough to see the image for what it is.

No, normal people get as close as physically possible.

hollis_f wrote in post #16240524
The best thing about these threads is how the 'optimum' sensor size changes over the years. In 2025 somebody will post the exact same thread bemoaning the obsession with GP and how 500 MP should be enough for anybody.

Or vice-versa. That 2GP isn't enough...

tzalman wrote in post #16240649
Bill Gates said "640K ought to be enough for anybody," and to this day he keeps smacking his head and muttering, "When will I learn to keep my mouth shut?"

I'm not sure he's really that bothered by having made that statement, if he said it at all. Other than the flak he's caught over the years.

Tom Reichner wrote in post #16240313
Many publishers prefer that submitted images have enough pixels to allow them to print at 300 or 330 dpi without having to interpolate (upsize). For a double-page spread of 10 inches by 17 inches, at 300 dpi, this would mean an image 5100 pixels on the long side - which, for a 3:2 aspect ratio sensor, means 3400 pixels on the short side . . . a total of 17.34 megapixels.

Which publishing process actually outputs 300 dpi?

I think the actual resolution can be extremely low. 2 megapixels is enough for everything.

Dynamic range is the main determinant of quality.

DocFrankenstein wrote in post #16241540
Which publishing process actually outputs 300 dpi?

I think the actual resolution can be extremely low. 2 megapixels is enough for everything.

Dynamic range is the main determinant of quality.

300 dpi is indeed the number requested by every printer I've been involved with.

Personally its DR I need more of.

Lowner wrote in post #16241583
300 dpi is indeed the number requested by every printer I've been involved with.

Personally its DR I need more of.

My lab basically says "240 dpi or just leave it alone" because DPI doesn't mean squat in a digital file.

DocFrankenstein wrote in post #16241540
Which publishing process actually outputs 300 dpi?

I think the actual resolution can be extremely low. 2 megapixels is enough for everything.

Dynamic range is the main determinant of quality.

Doc,
I don't know much about printing processes, output, etc. What I do know is that the publications and agencies I submit to have requirements, and that if my images don't have enough pixels to comply with their requirements, my images get rejected, and I don't get paid. A lot of native pixels are required by almost every publisher I have ever worked with.

Below is something I just copied & pasted from the submission guidelines of one of the publishers I work with. Their pay rate is based on the size at which they print your image, so one really wants to be sure that their images are able to be printed as a two-page spread. Because my cameras do not have enough pixels, this publisher has never used one of my images as a lead spread. I've had a lot of full page images published with them, and a couple covers, but never a "double truck".

Digital photos must meet minimum resolution requirements for printing.
At minimum, we prefer a photo to be at least 8”x10” at 300 pixels per inch. Ideally, we prefer photos that are 11”x17” (a full 2-page spread in the magazine) at 360 pixels per inch. Since our magazine is printed stochastically, we require slightly higher resolution than standard line-screen printing. The bottom-line: the higher the resolution and size of your image, the larger we can feature it in our pages.

EDIT: Doc, I am not sure I understand your statement about two megapixels being "enough for everything". If my purpose is to create a portrait of a wild bird, with nicely detailed feather filaments, I need far more than two megapixels. In fact, there have been times when the 15 megapixels on my 50D's sensor were not enough to show all of the intricate feather detail the way I wanted to show it. Shooting multiple frames, then stitching them together, does not work with birds and wildlife, because they usually do not stay still long enough for me to recompose and take additional images. A huge number of pixels is the way to show a great amount of tiny, intricate detail in one's subject when only one frame can be taken.

Attached is an image to illustrate my point. The first image is a portrait of a quail, it is the full uncropped image. It is framed and composed exactly the way I wanted it to be. The 2nd image is a 100% crop. This shows that the feathers in the image are nicely detailed. I can see each filament resolved independantly. The individual filaments do not blur into the adjacent filaments. This is the way the bird's plumage "really looks", and it is how I wanted to capture the portrait. I guarantee you that there is no way that the feather detail could be as accurately resolved if I had only had two megapixels to work with. So, two megapixels are not enough for everything.

Tom Reichner wrote in post #16241653
Doc,
I don't know much about printing processes, output, etc. What I do know is that the publications and agencies I submit to have requirements, and that if my images don't have enough pixels to comply with their requirements, my images get rejected, and I don't get paid. A lot of native pixels are required by almost every publisher I have ever worked with.

Below is something I just copied & pasted from the submission guidelines of one of the publishers I work with. Their pay rate is based on the size at which they print your image, so one really wants to be sure that their images are able to be printed as a two-page spread. Because my cameras do not have enough pixels, this publisher has never used one of my images as a lead spread. I've had a lot of full page images published with them, and a couple covers, but never a "double truck".

Digital photos must meet minimum resolution requirements for printing.
At minimum, we prefer a photo to be at least 8”x10” at 300 pixels per inch. Ideally, we prefer photos that are 11”x17” (a full 2-page spread in the magazine) at 360 pixels per inch. Since our magazine is printed stochastically, we require slightly higher resolution than standard line-screen printing. The bottom-line: the higher the resolution and size of your image, the larger we can feature it in our pages.

Actually that's talking about PPI. PPI is not DPI. It's correct in its terminology. But their printing is nowhere nearly "sharp" enough to warrant such strict requirements. I'll still wager that if they like a shot enough, they'll run it no matter what size the file is.

tzalman wrote in post #16240649
Bill Gates said "640K ought to be enough for anybody," and to this day he keeps smacking his head and muttering, "When will I learn to keep my mouth shut?"

I don't think he ever said that. But IBM did make such a claim very early in the IBM PC life.

At that time, the PC was sold with 16kB of RAM (one memory bank) and could be updated to 64kB. It later got 64kB memory banks allowing up to 256KB. And yet later the 64kB memory modules got replaced with 2x256KB + 2x64KB = 640KB. And some AT machines installed 4x256 but then needed clever mapping since the remaining memory were in memory regions either not accessible or reserved for other hardware.

When you ship a machine with 16kB of RAM, you don't really feel bad about having a hard limit of 640kB in the design - i.e. being able to increase the amount of RAM with a factor 40.

Especially since the processor itself had a limit of 1MB, so reserving 10/16 of the address space for RAM and the remainder for BIOS (and their BASIC in the early machines) and I/O is a quite sound design decision.

Some printing technologies uses a very high resolution system, but with very low quality of each pixel the printer outputs. So in some situations, the colors might be just "on" or "off". Send a drop of color or don't send a drop of color.

Some other printing technologies have quite low resolution, but every pixel can represent a great dynamic range.

So big care is needed when debating how many pixels that is needed. If you equalize camera resolution with the resolution of the ink-jet drops of a high-end ink-jet printer, then you would get a very tiny printout whatever camera your economy could afford.

tzalman wrote in post #16240242
Personally, I need high quality 20x30 prints, for which 54 Mp would be optimum, but I'm forced to acept the "close enough" of 21 Mp.

In the Beginning Photography class I teach, I exhibit several prints and ask students to guess how many megapixels are in each. One is an 8x10 taken with a 2.1 MP camera (circa 1999). One is a 24x36 made up of about 3.5 MP (8.1 MP horizontal shot, cropped to vertical).

The answers are all over the board, but no one mentions any lack of detail in either image. All they ever say is, "That's beautiful."

Go figure.

The point I try to get across to the students is, megapixels are overrated.