Jump to content

Megapixel quality


Recommended Posts

I've been shooting my RD-1s for a while now, and am very pleased in general

with its quality. I've been using it at EI 400 and 800, mostly, and blowing up

to 16x20 size or a bit larger. I don't fully understand all the nuances of

digital, but to my surprise, there doesn't seem to be much difference between

the small file and large file settings with respect to quality (that I can see,

at least). With the digital M coming out at 10 megapixels, I'm curious for

what applications the 10mpix vs 6mpix difference will make a real difference?

It must make a difference, because the digital SLR companies are locked in a

megapixel race it seems.

 

I read somewhere that National Geographic restricted digital images to single

page photographs while they accepted only film images for double-page spreads.

I find this quite suprising as I would have thought a double page spread was

less than 16x20 and I would have thought for certain the best digital cameras

delivering up to 10mpix would provide quality good enough for that.

Link to comment
Share on other sites

Hi Frederick!

 

I assume that you're talking about JPEG files. And if you cruise the dpreview site, you will see some instances where, for example, the top JPEG quality settings look pretty much alike (and in some cases, where the highest JPEG quality output actually looks as good as, or better than, the TIFF output!). Also, the visibility of artifacts that are created by increased JPEG compression (and other in-camera processes) depends somewhat on the image itself...such as, on whether it contains areas of closely spaced lines.

 

But as you are thinking, the megapixel war is sometimes irrelevant to a photographer's artistic goals. At exhibitions, I've sold several 12x18 prints of one of my earliest digital-infrared images for more than $120 each...and it was shot with a 0.8-megapixel camera!

 

The image has the last say, regardless of the megapixel rating!

 

Sincerely,

 

Dave

Link to comment
Share on other sites

Hi Dave. Well put. I'm still selling images made with a 3.2 mp digicam. While I now have a 10 mp DSLR I can honestly state that for smaller images there just isn't much difference. The real advantage of 10 mp is that interpolating isn't necessary for larger prints. But just going back to evaluating an image - there's no reason for a big contretemps over mp number, an image is an image is an image. It should stand alone on its own merits, or lack thereof.

Cheers,

John

Link to comment
Share on other sites

In the days of film photography there were always people who said that 35mm didn't produce good enough quality for them, and in the digital days there will be people who say that 10mp doesn't produce enough quality for them. That's why there were and will be options for those who insist that bigger is always better.
Link to comment
Share on other sites

I believe the submission standard for best quality prints with current technology is still 300 pixels per inch. I've read in a recent photo article that some more recent printers can even deliver good quality at 150 ppi for exhibition. But as a number of others have noted in the forums, the viewing distance must also be taken into consideration. Comparing the ppi requirements of a gallery print to those of a billboard is an obvious example.
Link to comment
Share on other sites

"I'm curious for what applications the 10mpix vs 6mpix difference will make a real

difference? It must make a difference, because the digital SLR companies are locked in a

megapixel race it seems."

 

As far as I understand, pixel count, expressed as megapixels, is just the multiplication

product of the number of horizontal and the vertical pixels. Your RD-1 has a 6.1

megapixel APS C-size CCD sensor with 3,008 (horizontal) x 2,000 pixels (vertical), and

should be able to give you decent 16ᄀᄄ x 20ᄀᄄ prints.

 

In fact, a 10MP camera only have 40% more linear resolution than that of a 5MP camera.

And this kind of difference is almost invisible to most, if not all, people. For any visible

improvement in resolution, you need at least a doubling of linear resolution. This means

you need to quadruple your megapixels to get any visible results.

 

I believe the ᄀᆬmegapixel warᄀᆭ is pretty much a marketing war that could quickly replace the

ᄀᆬmegahertz warᄀᆭ. I just bought my Casio EX-Z850 (8.1MB) the day EX-Z1000 (10MB) was

released, with a bargain of course.

 

So, to answer your question Frederick, the difference between a 6MB and a 10MB, when

you blow up to 16" x 20", should be minimal, if at all visible.

Link to comment
Share on other sites

Any more, higher megapixels really only relates to the professional publishing world. As

stated above, 300 pixels per inch is the gold standard in quality publishing (twice the line

screen, which in good magazines is at least 150 lines per inch). 10 megs will cover a little

more than 8x10 at 300 dpi when open in the industry standard Photoshop. This is good

for cover or full-page plus. 6.1 megs streeeeeetches to make a full page, but I've done it

(with a 5 meg Nikon Coolpix 5000, no less).

 

With art or home printing, the rule is, is if looks good to you (and your clients), it is. But if

you put your nose to the print, in most cases the higher meg files will look crisper when

printed full at 300 dpi. 300 dpi looks best to the human eye, anything higher is so fine as

to be lost on us.

 

35mm film can be drummed scanned to cover a double-page spread. Grain looks "cool,"

jaggies don't, to art directors. Nat Geo will publish half a 35mm frame on their cover (and

you can tell), I wonder if they'd ever consider a 3 meg file...

Link to comment
Share on other sites

What is interesting to me is that much of the early digi work by Joel Meyerowitz was at the 1.5MP level! Digi sure has progressed in the last few years!

 

I wondered why someone (JM) who shot 8x10 and Leica early on would go for a 1.5 digi camera? I guess I'll have to ask next time I see him.

Link to comment
Share on other sites

Very interesting observation, Frederick.

 

The RD-1 has either 6mp or 3.3mp JPEG settings. Correct? That works out to be 10x6" and 7.5x5" at nominal 300ppi. To print larger, the image has to be interpolated (up-scaled).

 

I wonder if a different upscaling algorythm might make a difference. Can you try various upsampling methods and send the printer what it needs rather than letting it to the job?

 

Maybe the difference will not be truly significant, but one can only test and find out.

 

Have you tried RAW yet?

Link to comment
Share on other sites

Not all pixels are created equal. The 6mp of your RD-1 are pretty good, and you are

putting some of the best available lenses in front of them. That said, all else being equal,

more is better just like with film. Some people can't tell the difference between 35mm and

6x7 on an 8x10 print. Some people can. Obviously as the picture gets bigger, the

difference is easier to spot, but in general, larger sensor or film size and higher resolution

(more MP) means that finer details are visible and tonality (the number of shades of color

or grayscale) increases. Just like with silver grains, the more you have, the smoother the

image. Not just in resolution, but in the shades of color/grey. This usually means more

detail in the highlights and shadows as well. In any case, here is an example. One of these

is a 10mp digital, and one is a 6x7 3200dpi scan. Can you tell a difference between the

two. Which do you like better? I think the difference is apparent even on a small jpeg,

though the absolute level of detail is very similar. Both shots are good enough to print

large, but there is an obvious difference in tonality (to me anyway. Perhaps I am crazy).

<P><img src="http://www.stuartrichardson.com/drying-fish-m7ii.jpg"><P><img

src="http://www.stuartrichardson.com/drying-fish-bw-dmr.jpg">

Link to comment
Share on other sites

Stuart, the first image has more detail in both the shadows and the highlights; I assume it's the film image. The second, presumably digital, has snappier contrast in the midtones.

 

I know it's been done elsewhere before, but a more interesting comparison, at least for me, would be a digital image versus a 35mm image, either an optical print or from a negative drum-scanned at the highest possible resolution.

 

Erwin posted a new article yesterday comparing images from Adox high resolution film with Canon digital. The film provided better resolution than the sensor, though at much lower speed. The challenge is getting all that "information" on paper.

Link to comment
Share on other sites

Stuart - I like the first picture better. It seems to have far greater tonal range and more shadow detail. The other is just too contrasty, and the shadows have blocked up. It isn't a question of resolution, but gradation in this case.

 

Pico - I have not tried RAW yet, because I have only just managed to download the RD-1s manual (they did not include one on the software disk that shipped with the machine). Still have to figure out how the software works. I have heard that a) RAW provides all the true, pure information you can store and b) pictures in RAW actually look underexposed relative to pictures in JPEG. The latter observation is interesting and I will have to check it out, as I am getting quite interested in histograms, characteristic curves, and the sensitivity characteristics of sensors. The two JPEG settings available on the RD-1s are for "normal" and "large" file sizes. The large file size is 3x that of the normal and you can store 1/3 again as many shots on the card.

 

Dave, John and Paul - I can't imagine an image made with a 0.8 megapixel camera. What is the pixel count on a cell phone camera anyway? Anyway, it is heartening to know that great pictures are being made (and sold!!) with less than the most expensive "keeping up with the Jones's" technology! I am the latest of late adopters, but it seems even 5 years ago, there were 2 or 3 megapixel cameras available. A 0.8 mpix machine must go way back!

Link to comment
Share on other sites

I also like the first picture better. It is the 6x7 film image. I could probably get more detail in

the shadows and highlights on the second image, but it would result in a flatter image. I was

trying to more or less match the overall look of the too pictures, but I only used levels...the

characteristic curves of the DMR and of 6x7 Acros in Rodinal are obviously quite different.

Link to comment
Share on other sites

I would disagree that the difference between 5mp and 10 is negligible. I've been 100% digital for a couple of years now. Most of my photos are archive work for my business. The two images in the link below were shot under exactly the same conditions.

 

The full image, in both cases, is a full top-to-bottom violin (two different violins, obviously), and the examples have been extracted from that. The one on the left was shot with my Olympus E-10, at 4mp. On the right is from a Canon Powershot Pro1, with 8mp. The 8mp is 100% on the screen, and the other is magnified to match its size, which would be about the equivalent, and I'm just guessing here, of a 30x40" print or maybe bigger (if you want to scale it out exactly, the diagonal height of a violin f-hole is around 76mm, and the length of a violin is 24").

 

The difference is striking, and is completely consistant with the results I've had from both cameras. Having done the same job in the past with 35mm and Plus-X, 6x7 with Tri-X, and 4x5 in color, I have a bit of perspective on this, and my estimation is that 4mp Oly was worse than 35mm, but 8mp Canon is definitely better. I'm thinking that my next move will be to 12 or 13mp, that this will give results very similar to what I was getting with 6x7, and that will be a terminal camera, for my use. I already am making prints which easily rival what I was doing with my Leicas and film, with a lot less work, which is why after 30 years of Leica use I put the film stuff away permanently. I'll mention in passing that my technique with film was not the problem here, nor was the equipment--I think the limitation has always been the film itself. Nor do I think that digital is necessarily sharper in all the examples I've given, but it definitely can be made to *appear* sharper, and appearance not measurement, is, after all, the name of the game.

 

I hope some of this is useful. Here's the link--sorry, I don't remember how to make it live:

 

http://darntonviolins.com/images2/test.jpg

Link to comment
Share on other sites

"They come in handy for Post-Processing and obviously for cropping..."

 

I'm with John on this one. IMPO, the biggest draw of more pixels is cropping. Couple of years ago, I used to tote along a D2H for birding, but sometimes, even with 400mm, things can get really small and the extra pixels will come in really handy...

Link to comment
Share on other sites

trouble is the big two agencies will probably in a year or two accept images from this camera

(M8) for editorial use only, for most people this won't be a problem but some snappers I

suspect will hold off buying the m8. The new canon 1ds upgrade I fear will raise the bar

again leaving the 10mp count way behind.

Link to comment
Share on other sites

Michael. I am not going to say that you falsify your archival pictures by making them "*appear*" sharper than they would with film but is there not a danger of missing/not recording a hair-line crack in an instrument, that neatly follows a line of pixels - or doesn't. Or even of exaggerating it.

 

What I'm getting at is: is appearance the name of the game ? - Especially for archival work.

Link to comment
Share on other sites

<i><blockquote> The new canon 1ds upgrade I fear will raise the bar again leaving the 10mp count way behind.

</blockquote> </i><p>

 

Look at Nikon, Sony, Canon (and soon Pentax) -- 10MP is quickly becoming the new low-end for DSLRs. A $5,000 digi-rangefinder is sure to leave many people unimpressed, no matter how well it works.

Link to comment
Share on other sites

Hi Again John, Paul, Fred and everyone,

 

Maybe I should explain a bit more about the 0.8 MP image and its sales! It was one of the first digital-IR shots I captured, on a visit to a local botanical garden. I had just found a Tiffen 89 blocking filter for a buck in the "junk" bin at a camera store, and was going to test it out. But as we were about to drive to the garden, I realized that I had forgotten the camera, and told my wife to just head on out. She said "No...If we don't bring it, you'll regret it." And of course, she was right.

 

Anyway, the 0.8 MP camera has no hot-mirror, so it captures nicely contrasty IR images hand-held. And the image posted below is the only one from the camera that didn't break up during very careful step-upscaling in 16-bit mode to a 12x18 inch size. Many others will go to 8x10, but not to 12x18.

 

Oh, I just remembered this morning that the $120 prints of this that I have sold were all unmatted and unframed. Many buyers of late seem to prefer to choose their own mats and frames...and to receive their prints in big sturdy envelopes!

 

Anyway, here's that image. (Don't know how it will look on your screens, since my office computer makes images look darker than they are.)

 

A lot of people have asked whether the white area behind the upper-right pavilion is all cloud. There is a cloud above it, but the "lighting" behind the structure is a large bush. In IR, the bush actually backlights the structure. One buyer actually said, "it's VERY Hollywood"!

 

Sincerely,

 

Dave<div>00HmUH-31929784.jpg.84eaaf0fad733b4c8ebe67d51910c6d5.jpg</div>

Link to comment
Share on other sites

Oh...Fred, Forgot to add that the 0.8MP camera is a Polaroid PDC-700, which dates from around the year 1999 (when I was writing their digital camera user manuals). This was at a time when they had an internal product-development group that worked hard to optimize the performance of their cameras...Macbeth charts, focusing, white balance tests, etc.

 

Dave

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...