Blind megapixel test: Medium format vs. G7, can you spot the difference?

[V8Beast]

With all the stink about megapixels lately, I thought I'd put together a quaint little test. Some of these are product shots I took for editorial clients with my 10 megapixel G7 point-and-shoot. The rest aren't my shots, but they're of similar engine parts taken with a 39 megapixels medium format Phase One P45. I read a post today where someone said they'd be happy with the 5DIII if it had at least 25 megapixels instead of 22, so apparently every last MP counts, but does it really?

And please don't cheat by looking at the exif data. I tried to erase it the best I can, but there might be some giveaways to the camera's identities nonetheless. At any rate, I've numbered each image 1-7, so try to guess which images was taken with the G7, and which one was taken by the Phase One

If you look closely enough, you can spot a difference between the G7 and Phase One images. However, I'd argue that a lot of that has to do with differences in technique. My point-and-shoot shots were taken using nothing more than natural light and some white poster board whereas the Phase One images were taken in a studio with fancier equipment. I don't know how fancy, because they aren't my shots and I wasn't there, but considering that they were commission by a rather large mega-budget corporation, it's safe to say they had the means to spend a lot more money than my editorial clients. I'll reveal who the unidentified corporation is later, but don't want to yet as it might give a clue as to which shots came from which camera.

1

2

3

4

5

6

7

Before anyone cries foul, is there any way I can upload the high-res files somewhere for download, if anyone cares to pixel peep? I can upload them onto my website's lightbox, but I have to invite people via e-mail.

 

[V8Beast]

Almost forgot to mention that the Phase One images are press images that the "mystery corporation" offers for download to editorial outlets. As such, they have been sized accordingly by the corporation, not me, for easy download. The point is that I've been downloading these files for years, oblivious to the fact that they were shot with Phase One backs, which is definitely overkill for editorial use. I just happened to check the exif data one day, and was shocked to see that the images were taken in medium format.

 

[Wrathwilde]

I'm thinking 1, 2, 5 and 7 are MF. Didn't do any pixel peeping, or downloading, just looked at the pictures as they appeared on screen.

How far off am I? You don't need to tell me exactly which I got right or wrong (leave that until other people have had a chance to play), just tell me how many of the ones I picked were MF.

 

[Mt Spokane Photography]

For online use, 3mp is adequate, and 5mp excellent. The main advantage of having a high mp original is that you have headroom for adjustments and cropping before downsizing for web.

And the large photosites in a mf body do make for clean images.

I still have my 3.3mp nikon Coolpix 990, and it still takes excellent images, as long as you don't try to blow them up to a very large size.

I took this 3.3mp image of my dogs in 2001, and it is still fine even though it does not show the fine detail that my 5D MK II does.

 

[pdirestajr]

I guess 2,5,7 by just looking on my iPhone. I just think the gradients to the clipped highlights on the metal look smooth in those.

 

[CowGummy]

I too would say 2,5 & 7 are MF - Although no.1 is a very good shot - just something about those shadows that lead me to believe they weren't shot in a controlled studio environment. For what it's worth I'm viewing these as they appear here on a 10" netbook screen.

 

[DavidRiesenberg]

Good post. 2,5,7 is my guess. I did recognize the items in 2 and 5 so that made it easier to guess.
But at a casual glance, from an image quality point of view, it is neigh impossible to tell the cameras apart in these examples.

 

[StevenBrianSamuels]

Coke, Pepsi and RC Cola.

 

[Daniel Flather]

You took photos 1, 4, and 6


*runs*

 

[Daniel Flather]

Maybe you took #3 too, but mostly it's the lighting giving it away.

Good thread.

 

[V8Beast]

I did recognize the items in 2 and 5 so that made it easier to guess.

How dare you be so well informed

Images 2, 5, and 7 are from the Phase One. They are stunningly good images and beautifully lit, although I still find it odd that GM used such insanely expensive gear. I suppose that's the luxury you have when you're a multi-billion dollar corporation backed by taxpayer dollars The largest images like these are ever run in print are 4.5x3, and usually more like 3x2. As such, they are scaled to 10x8 at 300 dpi on GM's media site, so they're essentially throwing all the pixels away.

I think that much of the differences between the G7 and Phase One images are due to superior lighting techniques used in the MF images. I'd venture to say if the corporate hot shot photog shot used my G7 and I used the Phase One, the GM images might still look better.

 

[V8Beast]

OK, let's step things up a notch. Here are some product images I recently took with a 5DC, 1D2, a couple of small softboxes, and a pair of 550EXs. They're mixed in with some more GM media photos shot in medium format. Interestingly, despite having just 8 megapixels, my 1D2 didn't suddenly self-destruct out of embarrassment in light of the higher resolution bodies on the market ;D It still took exceptional images.

8

9

10

11

12

 

[Viggo]

I guess this conclude= Light is everything, if you had your shots in a studio, they would be harder to separate. On the other hand, if you printed them rather large I say the difference will be more apparent. But on the third hand : who prints THAT large, I mean, larger than 100% cut at 22mp. I guess the people using Phase One, so good and clever thinking Canon

If the difference between the 18 mp 1d X and the 5d 3 is nothing, why didn't they keep the 5d3 at 18 also? I'm DEF not a mp-guy, I'm just wondering. Is it to make the noise levels separate of the two?

 

[V8Beast]

I guess this conclude= Light is everything, if you had your shots in a studio, they would be harder to separate. On the other hand, if you printed them rather large I say the difference will be more apparent. But on the third hand : who prints THAT large, I mean, larger than 100% cut at 22mp.

I concur. It's all about the lighting. I don't mean to bash megapixels. I'll take as many of them as I can get. If you make HUGE prints or need to crop like a mofo, by all means get a mega resolution camera. It just so happens that if you don't make huge prints, a lower mp body combined with good lighting can go a long way.

 

[meli]

I did recognize the items in 2 and 5 so that made it easier to guess.

How dare you be so well informed

Images 2, 5, and 7 are from the Phase One. They are stunningly good images and beautifully lit, although I still find it odd that GM used such insanely expensive gear. I suppose that's the luxury you have when you're a multi-billion dollar corporation backed by taxpayer dollars The largest images like these are ever run in print are 4.5x3, and usually more like 3x2. As such, they are scaled to 10x8 at 300 dpi on GM's media site, so they're essentially throwing all the pixels away.

I think that much of the differences between the G7 and Phase One images are due to superior lighting techniques used in the MF images. I'd venture to say if the corporate hot shot photog shot used my G7 and I used the Phase One, the GM images might still look better.

If you havent mentioned that they're actual photos, 2 5 & 7 do look like engineering renders, rendered by an intern at that! That superior lightning technique you're referring to, is actually the default look of any GI render engine

 

[caruser]

If the difference between the 18 mp 1d X and the 5d 3 is nothing, why didn't they keep the 5d3 at 18 also? I'm DEF not a mp-guy, I'm just wondering. Is it to make the noise levels separate of the two?

Wasn't there something about the 5D3's horizontal resolution being 1920 * 3 which helps clean video?

So you could ask the other way around, why doesn't the 1DX have the same 22 MPs ;-)

 

[Viggo]

If the difference between the 18 mp 1d X and the 5d 3 is nothing, why didn't they keep the 5d3 at 18 also? I'm DEF not a mp-guy, I'm just wondering. Is it to make the noise levels separate of the two?

Wasn't there something about the 5D3's horizontal resolution being 1920 * 3 which helps clean video?

So you could ask the other way around, why doesn't the 1DX have the same 22 MPs ;-)

Good point. My guess is that a few mp's have a lot to do with noise, they can't double the res and still have two stops cleaner anymore, so every mp steals away from the size of the pixels, making even a small increase of mp shows more noise.

I'll think I'll be quite happy with 18. I was very happy with 16 on the mk4. Printed 60x40 cm recently, one of each of my kids, one shot at 10 mp 1d mkIII and the other 1d mkIV at 16, they looked like they were shot at the same res.