MP myth: My own experiment

[maxi]

We all know MP don´t mean a thing unless you really need them for your specific purposes (ie: big crops, huge prints, macho comparisons).
But still, when I told my mom yesterday that I was keeping a nice 3MP P&S for myself (to carry everywhere the rest of my stuff would be impractical) and that I had a bunch of them to sell really cheap (I´m kind of in the business) in case any relative/friend wanted one, she replied with: But 3MP is so little...
That comes from a woman that cannot program a VCR

This made me decide and finally do this comparison I have wanted to do for a while (at first I thought of showing it to customers, but I dunno if it´s a good idea )

Ok, so here. I´m posting 6 different pics (5 here + 1 later), taken with:

a 2MP camera
a 3MP camera
a 4MP camera
a 5MP camera
a 6MP camera
and a 35mm camera

All the cameras are of the same brand, all of them but 1 are taken by me in the last 3 years (I think).
I adjusted the aspect ratios and I set the resolution to about 1200x800 which fills my macbook screen and is too big for the web and probably the biggest your average point and shooter will ever need to show his vacations to his friends or the last thing his baby did.
I saved them as JPEGs at quality: 60

My conclusion??? I cannot tell the difference in resolutions, what I can tell though is the difference in sensor sizes and 4-5 years of development in the field. So investing in a larger sensor is much more intelligent than investing on a smaller sensor (I'm talking P&S vs. APS size here, I haven't yet tried neither full sensor or 4/3 cameras)... and between 2 equally sized sensors with a 1MP difference, I may even choose the one with lower MP.

Still, I'm very much curious about what do you guys think, can you tell which pic is which? or can you only tell them apart from image quality unrelated to resolution?

 

[maxi]

Here's the 6th pic

 

[dopey4800]

Here's the 6th pic

so you've proven that at 1800x1200, you can't tell the difference between shots. not really a groundbreaking test. not really a test at all, just rather a reason to show off 6 snapshots.

what exactly is your point with this?

you've really shown us nothing here and proven even less.

let's see 100% crops of the same scene with a Rebel Xti, 30d, D200, D80 and D50, maybe toss in a few pro cams, and some P&S junkies.

Secondly, MP is not in any way tied to resolution. that's solely in the lens and its resolving powers. MP, rather, is a misleading count that is used to describe the size of the image.

I've seen 4mp images from a pro-DSLR than crush 6-8mp counterparts IN RESOLUTION. Meaning that the combination of light, shutter, apertue, ISO and lens in the 4mp image were able to capture optimum resolution AT 4MP.

I bet if i had all of those variables at 8mp, i'd still get optimum resolution, just at a larger file size.

Nothing more to see here. move along.

 

[robbieduncan]

It also sounds like the compression used on these images is quite severe (I'm assuming that maximum quality, minimum compression is 100 on the scale you've used). This compression will smere out most of the differences in resolution anyway...

 

[Bibulous]

I´m kind of in the business

Wow

Edit:

Maybe should have taken 3mp, 5mp, 7mp cameras from the same manufacture and made 4x6, 5x7, 8x10 and 11x17 prints from one image from each camera, preferably all with common settings and subject. Then compared the image quality to see if you were only getting more low quality pixels on the higher MP cameras.

 

[billtanderson]

If you throw away all the extra information that the highter MP cameras generate then yes the pictures will look pretty similar on screen.

I'm no expert but it might even be possible to produce a worse picture by squeezing a 6MP original down to 1200x800 rather than taking it at that size, depending on how you do it.

Printing I think is where the big differences really appear - really good prints are usually 300dpi or so. 1200 x 800 gives us only a 4 x 3 inch print. In practice some things look OK at 150dpi, so you could do some 8 x 6 prints.

The higher MP cameras will allow high quality prints at correspondingly larger sizes.

 

[dopey4800]

Wow

sad isn't it?

im kind of a big deal too. haha.

 

[dopey4800]

Wow

Edit:

Maybe should have taken 3mp, 5mp, 7mp cameras from the same manufacture and made 4x6, 5x7, 8x10 and 11x17 prints from one image from each camera, preferably all with common settings and subject. Then compared the image quality to see if you were only getting more low quality pixels on the higher MP cameras.

but what would that prove?

he would need to get the same scene, and different style cameras and show us 100% crops of the images.

At, say, 28mm, the image center should be the same for all, though the larger MP counts will show more of the image, and we can see the difference that you get from say, a 1mp crop versus a 10mp crop.

that's a test. this is...like i said... showing off snapshots.

 

[maxi]

what exactly is your point with this?

you've really shown us nothing here and proven even less.

Secondly, MP is not in any way tied to resolution. that's solely in the lens and its resolving powers. MP, rather, is a misleading count that is used to describe the size of the image.

I've seen 4mp images from a pro-DSLR than crush 6-8mp counterparts IN RESOLUTION. Meaning that the combination of light, shutter, apertue, ISO and lens in the 4mp image were able to capture optimum resolution AT 4MP.

I bet if i had all of those variables at 8mp, i'd still get optimum resolution, just at a larger file size.

Nothing more to see here. move along.

Actually, you said what I wanted to say, only much more clearly (I'm not a native english speaker, you know?).

When I did this I was aiming at people (like my mom, or friends or whatever) that think that MORE MP IS BETTER, which I find it's not the case for the use the average person gives the camera.

Of course this information is useless for you, but I've seen loads of people ask for advice on p&s and general noob advice and I thought this would be useful, in fact, I have actually shown this to the kind of people I'm talking about and most have been clearly impressed, for they thought more MP meant automatically better images .
I'm sorry you found it sad.

 

[ChrisA]

Let me see if I understand this..... If you take a high resolution image and transform it into a low resolution image the result looks just like a low rresolution image.

What this means is that if you are shooting images for display on a computer screen there is no need for more pixels then will fit on the screen. That's correct but you have to allow some room for crops and other image processing.

 

[dopey4800]

Actually, you said what I wanted to say, only much more clearly (I'm not a native english speaker, you know?).

When I did this I was aiming at people (like my mom, or friends or whatever) that think that MORE MP IS BETTER, which I find it's not the case for the use the average person gives the camera.

Of course this information is useless for you, but I've seen loads of people ask for advice on p&s and general noob advice and I thought this would be useful, in fact, I have actually shown this to the kind of people I'm talking about and most have been clearly impressed, for they thought more MP meant automatically better images .
I'm sorry you found it sad.

I agree with you on the point that MP, by and large, does not matter.

but your methods of testing are horrifically flawed. First off, you show six completely different styles of imagery in a format that levels any quality benefits of a larger camera. and yes, there ARE benefits.

secondly, you equate resolution with MP, which is not the case. MP is basically a hollow term to describe file size. Resolution is determined by much more. And you obliterate the resolution advantages you might get from, say, an Rebel Xti by crushing the file size down to 1800x1200 (not even the native aspect ratio of many cameras) and then compress the bejusus out of them

I could easily make a great 1200x1800 web image from a 1mp cam. Consequently, I could make a horrible image out of a 10mp image downsized to 1200x1800.

As i said, if you want to see the true differences, you can illustrate that by showing 100% crops of the same scene.

Also, think about this: your major differences in the images are MOST LIKELY due to the LENSES USED. A 17-55 AFS-DX Nikkor at f11 is going to be infinitely sharper than a 5.5mm lens at f4 on a P&S cam.

This affects the DOF, the sharpness, the contrast, the color, and most importantly, the resolution. You're right about 3mp not being bad for most general uses.

That said, not all 3mp cams are created equally. If you have me the option of a D200/30d at 8-10mp with high end nikon or canon glass, i'd take it over the 3mp COOLPIX anyday.

That said as well, I'd take a Canon D30 (3mp DSLR) with that same high end lens over the 3mp P&S.

And I will say that, even at 4x6, the benefits gained from the larger sensor, the DSLR technology and control (not to mention excellent glass) and the amount of flexibility I'd have with that DSLR over a 7 or 8 MP P&S.

I think that the biggest selling point a 3mp P&S has over ANY DSLR is its compact size. nad that DOES NOT affect image quality in a scientifically measurable way. The only way it affects IQ is that I'm more likelty to carry my P&S everywhere I go.

BUT even then, I'm certain that the image i get with my 3mp cam isn't going to stand up to my D200. or even an old 1dmk II (4mp) or D1h (2.74mp).

I agree with your conclusions, BUT i think you method of testing is sad and doesn't lend any credibility to your claims in the slightest. I just agree with your conlusions because i've seen this market evolve and I have a lot of knowledge about it.

The general MR user might go out and buy a $50 3mp P&S from you with the thought that it will outperform his D2x that he paid $5000 for. If he does, he's probably a stupid lawyer who bought the D2x as neck jewelry.

But he's certainly not gonna buy his next cam from you after hearing about this MP theory and then not seeing it prove true in practice.

edit: the thing you're not realizing is that when you show this stuff on a screen it's not going to follow through to printing. and then you'll have joe dad who takes pics of his kid's hockey game with his 3mp P&S and is pissed because his crops look terrible. That guy would benefit from the flexibility of a SLR, even if it were 3mp, if only for the flexibility of lenses he can use, thusly maximizing his image.

I think you're basically doing this to sell 3mp cams, which is fine, if not a bit misleading. Well. Not a bit. VERY misleading

 

[Bibulous]

I agree with your above post completely. This is how camera shops (assuming the OP is in sales) get a bad name.

but what would that prove?

he would need to get the same scene, and different style cameras and show us 100% crops of the images.

I was thinking more about print size and image quality instead of working on screen with crops, two sides of the same coin?

 

[maxi]

@Dopey:

First of all, I know the difference between resolution and MP. Why did I say resolution? probably because we are used to say "resolution" in spanish and even most of the people in the interweb use both interchangeably (not that that makes it right though )

I didn't intend this to be a rigorous test, I just grabbed 6 pics I had in my library and compared them to show my mom that MP numbers are not the whole story. I just posted them here as an afterthought.

If I wanted to show you a "real" test, I would refer you to the resolution charts in dpreview or some technical site, but that doesn't make sense, why would I compare the resolution of a >$1000 DSLR with that of a $100 P&S ?? there's absolutely no point there.

In this I agree with ken rockwell (ohhh controversial!! ): technical numbers mean squat, the pictures are what matter.

You can obviously recognize which of the pictures above was taken with a 105mm DC f2, but many might not and perhaps realize that they are better off with a smaller P&S or maybe even decide between 2 different P&S because they may not even realize that what seems so obvious here (that a downsized high resolution file is the same as a low resolution one).

Me? I'll pick a DSLR or even medium format any day, but still: If I'm not going to have my dslr with me and won't be able to take the picture, then that is obviously worse than having taken the pic no matter how bad the pixel count is.

And one last thing, I wasn't trying to make P&S look like a good choice or the other way round, nor am I trying to sell no one anything. I don't even sell cameras (i fix them) so I have nothing to gain from this (and who would travel all the way to south america to buy a 3MP camera? )

I enjoyed reading you point of view and respect it, but I think you are more focused on trying to prove me wrong on my "methods" (when I never said this was a scientific thing) than anything.

 

[andiwm2003]

take your 2MP picture and your 7MP picture and make a 20inchx30inch print from them. you will see that the 2MP picture shows visible pixels and less detail than the 7MP picture. therefore the 7MP camera has higher image quality (assuming the pictures are both exposed and focussed correctly).

the test you did is useless.

and yes the differnece between 6MP and 7MP will be very small or even undetectable. but the difference between 2MP and 7MP is very significant.
and also yes, if you only want to see your picture on a monitor or on small printouts 3MP is good enough.
and most importantly the only really bad pictures are the ones you didn't take. so your argument to have better a 3MP than nothing is correct.

 

[maxi]

I agree with your above post completely. This is how camera shops (assuming the OP is in sales) get a bad name.

Actually, my intentions were completely the opposite. I was trying to show that there's more to digicams than pixel count, you obviously know that, but many don't and those are the ones screwed by camera shops saying that they need the latest camera in order to take pics (and don't get me started on those who try to sell you the lenses and accesories that give them more margin ).

I repeat I'm not trying to sell anyone anything, too bad I was misinterpreted.

 

[Bibulous]

Please accept my apologies, after rereading my post it sounds like a personal comment on you, which was not deserved. My local shop on the other hand….

It would be nice if everyone understood all the terms before buying, but I think back to my first digital purchase and how little I understood of the technology at that time. I could not be bothered with doing much research and just wanted a camera. This may always be the case, but every new P/S camera released lately seems to be an amazing piece of technology at incredible prices, hard to go wrong with any of them from a big picture point of view.

 

[dopey4800]

I agree with your above post completely. This is how camera shops (assuming the OP is in sales) get a bad name.



I was thinking more about print size and image quality instead of working on screen with crops, two sides of the same coin?

essentially. 100% crop will show you pretty much what something will look like printed. Though not 100% accurate it give syou the best indicator of what the image will look like, and the detail your sensor can resolve.

as to the OP's thoughts about resolution, it's time to switch your terminology. You talk about "lenses and high margin equipment" but to be honest, Nikon can take 40% markup on my 17-55 because it's THAT GOOD.

And it resolves more detail, be it on a D1h, D2h, D200 or D50 than many other lenses.

And in most cases, the consumer would be MORE SCREWED by buying based on your recommendations than if they were to got to Best Buy and walk about with a $2000 D200 or 30d package.

They'll get a nice flower shot, and they'll want an enlargement. They'll take some sports shots and want to crop in to isolate their son scroing the winning goal. They'll take some HIGH-ISO concert shots, and in your example, and with your info, and maybe even their purchase from you, they will not be able to do it.

They might not even be able to make a nice enlargement of the flower to 8x10.

I agree that MP is an overstated, overused term, but in the end, it DOES give you a sense of what you can do with an image, and though there is more to that, a newer, higher MP camera is likely to be more future proof, more capable (i.e. chances are it's image stabilised, has better jpg processing and possible, a good RAW format mode), and have better IQ overall than a smaller MP cam, be it a new generation 6mp SONY or an older 3mp or lower coolpix.

There's a reason why you have "low end" and most of the lower MP cameras have less features and bells and whistles than their newer, higher MP counterparts (such as decent frame rate, lens quality/interchangeability and IS, not to mention wireless, etc)

Add in the benefits of extra cropping space, less noise (esp when downsized for smaller prints) and better more advanced lenses, and you get a combination thats tough to beat.

I agree, again, that MP is overrated. BUT it IS an indicator of how technologically advanced a camera is. And to say that everyone could get by with less MP is in essence to say that peopple don't need things like decent frame rate, IS, etc. is a danger and is something inherently tied to many of the cameras that are 6mp and under.

Not to mention that in the newer tech case, you get an image that is overall more flexible and able to be used in a variety of situations. (i.e. poster print, screensaver, mug print, et. al)

Photography is about busting through limitations and making a great image. Why limit oneself just because MP is a marketing blah de blah that hold alot of weight?

 

[maxi]

Please accept my apologies, after rereading my post it sounds like a personal comment on you, which was not deserved. My local shop on the other hand….

Dude, no need to apologize . I know where you are coming from and haven't taken any of it personally.

As for Dopey:

Did you actually read what I posted? I have a feeling you just love to hear yourself type. Either that or my ability to make myself understood in english is far worse than what I thought it was.

You lost me when you said that MP is a good indicator of how technologically advanced a camera is. That's as much BS as the first interpretation you made of my OP. If you are one of those guys who just loves to compare specs (my pipi is bigger than yours ) then be my guest, but really... I was just trying to make a point saying that things like color rendition, lens and sensor quality, etc are very important too and you go on babbling about god knows what.

Let me repeat this in case you missed it before (it goes in bold now)
I understand what you are saying
it would be nice if you could actually understand what I was saying.

 

[FritzTheWonderM]

I got your megapixels right here folks. Take an image from an unknown original source, squeeze it to 1200x800 then compress the snot out of it. Take that, run it through Genuine Fractals to make it 6 MP and you still get a usable poster size image. I snipped a piece (100% scale) out of the middle as a sample.

Good cameras don't make good pictures, good photographers do.

 

[maxi]

Good cameras don't make good pictures, good photographers do.

I couldn´t have said it better myself. (In fact, I actually tried but ended with a thread full of flaming... har har )

 

[Macanadian]

but what would that prove?

he would need to get the same scene, and different style cameras and show us 100% crops of the images.

At, say, 28mm, the image center should be the same for all, though the larger MP counts will show more of the image, and we can see the difference that you get from say, a 1mp crop versus a 10mp crop.

that's a test. this is...like i said... showing off snapshots.

Shots of the same scene would be a better test. I recall one shot I took... I used a 5mp camera (point and shoot) and my Pentax Z1P, loaded with 50 ASA slide film. Of a mountain scene with clouds in the hovering over the mountain. I was going nuts looking at the same photo (seconds apart). The digital shot had bleachly clouds (no detail) in comparison to the film shot.

I'll try to find the two shots and upload them.

 

[0098386]

Good idea and it's a shame a lot of people took this thread the wrong way.

Though I won't be swopping down to a lower MP camera any time soon (and if I know I won't be making prints of my pics I will shoot in my cameras lowest res for speed/space benefits... with a 4gb microdrive ) as I tend to print out my work on A3 for my mates. They're usually things we do together like a rafting adventure, stunt kiting or spelunking or whatever.

 

[freebooter]

Interesting original post. It does demonstrate that for many purposes (the most common, not pro or gear-headed pixel-peeping purposes) MP size is but a relatively minor factor in determining image quality.
Interesting, too, to contemplate the vitriol that spewed onto the thread in later comments.

 

[Teddy's]

I totally agree with the sentence "Good cameras don't make good photographs, good photographers do."
Also, good cameras will do the job easier.
When it comes to selling, the first rule I learnt was: ask the client what do you need? what do you do? what do you want to do? It is just a process of discovering needs and how to satisfy them. If she still says "3MP is too little" well, give the client what they want and they will be satisfied.
"I need a 3+ MP camera"
Benefits: It will be bigger better (reality 50%, perception 50%)

...and yes, I would like to see some tests of the same scene taken with different cameras (...and less fight between MR members)