Jump to content

Recommended Posts

Advertisement (gone after registration)

Can anyone explain me this question. A lot of people say that the Leica S with their bigger pixels on the size has almost better image quality than other MF Cameras with more Megapixel resolution.

Is this right? Are more megapixel on an MF sensor not lead to better image quality?

Link to post
Share on other sites

There are tradeoffs. Obviously, one giant pixel covering the sensor will make poor images, and ten billion pixels registering pure thermal noise likewise. There is a range where the tradeoffs of resolution and individual pixel quality balance. That range moves over time. Once within that range, it's difficult to tease out exactly what is responsible for each system's look and appeal. It's not just the sensor, but how the engineers treat the signal, how the software designers deduce color and detail that isn't actually present on the chip, and a great many other things.

In my own tests, resolution of the Leica S didn't start limiting detail in prints until they were at least 5' wide. If you need 7' prints from a single capture that will be viewed from less than 2 feet, then you will want more pixels.

Edited by mgrayson3
  • Like 4
  • Thanks 2
Link to post
Share on other sites

While there are advantages to large sized pixels, I think they are mostly in relation to the same generation of sensors. In other words romanticizing about the 12mp in the D3 is not really productive these days. A D850 will beat the pants off the D3 at high ISO even though it is 40+mp in comparison to 12. Sensor design has come a long way in the last ten or 11 years. I think the Leica S at some times compares well to 50mp sensors, but the difference between 37 and 50mp is really not that large in actual size. Make a mock up in photoshop, you will see the size difference is not that huge. Comparing the S to 80 or 100mp is another story. In that case, if the S pictures compare favorably it has far more to do with Leica's lenses and the tuning of their color than anything to do with their pixels.

 

My experience is a little more critical than Matt's, but very close. The S can make very good prints at 5 feet (100x150cm), but I do see a bit of loss in resolution already at that size. Below 100x60 and it is more or less invisible. If you crop, or consistently use other aspect ratios (1 to 1, 4 to 5 etc), then you have to account for that as well. The S cropped to 4x5 is a 33mp camera, while a Fuji or other MFD camera are mostly 4 to 3, so they will retain more of their resolution and be more like 45mp.

Link to post
Share on other sites

2 minutes ago, Stuart Richardson said:

... if the S pictures compare favorably it has far more to do with Leica's lenses and the tuning of their color than anything to do with their pixels.

Plus the 16-bit S-files, I imagine. 

Link to post
Share on other sites

How do you rationalize print results without knowing what the printer hardware does to your image? Or are you extrapolating from a monitor view, which is entirely different?

Link to post
Share on other sites

Advertisement (gone after registration)

6 hours ago, mgrayson3 said:

f you need 7' prints from a single capture that will be viewed from less than 2 feet, then you will want more pixels.

You might also want a perceptual reboot. Under what circumstances would a < 24" viewing distance be rational for an 84" print?

Edited by pico
  • Like 1
Link to post
Share on other sites

To be clear, I was comparing prints from an S(007) and a camera with about twice the MP. I saw no difference between them until 5'. Both prints were uprezzed and sharpened in PS, and I have since found better tools. Any experiment is just one data point. This is the comparison - printing this crop at 8.5x11 is similar to a 5' wide print. At the pixel level, the difference is clear. On screen, where more detail is visible, it shows up. But the prints looked astonishingly similar. The building is 2 miles away. S(007) on the left, as you can probably tell.

Sorry - I posted the 80" comparison earlier. I have now changed it to the 50".

Edited by mgrayson3
Link to post
Share on other sites

1 hour ago, pico said:

You might also want a perceptual reboot. Under what circumstances would a < 24" viewing distance be rational for an 84" print?

I think that was rather the point, Pico - it isn't.

I don't believe, if you're printing the full frame, at normal (rational) viewing distances, there is an apparent difference between 24MP, and 50MP., or for that matter 100MP.  If you go looking for it, sure, but does it actually improve the actual image?  If you're looking at the image?  I think not - but then, I like to think that when people look at my images, they're looking at the whole image, not peeping at the grain.  Depends on the image, I guess.

Cropping is a different game.

Link to post
Share on other sites

People look at images in different ways. Usually they start from afar, and if the print is big, they walk around a bit and then come close and look at some point that interests them. I do not know anyone who looks at a big print and says, "oh, this one is 100x150 cm, so I have to stay 2 meters from the wall in order to properly see the whole image at once." Whether an image benefits from more resolution and whether it NEEDS more resolution are different questions. Photos from the S rarely need more resolution, though they may benefit from it. 

  • Like 1
Link to post
Share on other sites

vor 27 Minuten schrieb Stuart Richardson:

People look at images in different ways. Usually they start from afar, and if the print is big, they walk around a bit and then come close and look at some point that interests them. I do not know anyone who looks at a big print and says, "oh, this one is 100x150 cm, so I have to stay 2 meters from the wall in order to properly see the whole image at once." Whether an image benefits from more resolution and whether it NEEDS more resolution are different questions. Photos from the S rarely need more resolution, though they may benefit from it. 

My reason for buying the S was this kind of distinctive look and rendering. For many I bought the S Magazine and I saw hundreds of great photos with them and also on the Online S Magazine you can look at different fashion and portrait series. They hey a certain signature which really appreciate. 

Link to post
Share on other sites

11 hours ago, mgrayson3 said:

To be clear, I was comparing prints from an S(007) and a camera with about twice the MP. I saw no difference between them until 5'. Both prints were uprezzed and sharpened in PS, and I have since found better tools. Any experiment is just one data point. This is the comparison - printing this crop at 8.5x11 is similar to a 5' wide print. At the pixel level, the difference is clear. On screen, where more detail is visible, it shows up. But the prints looked astonishingly similar. The building is 2 miles away. S(007) on the left, as you can probably tell.

\Sorry - I posted the 80" comparison earlier. I have now changed it to the 50".

Sigh. I meant, of course, 60". Here are the originals for comparison

 These are test pics only. If I see an arresting image, I may look at interesting details, but looking for sharpness is like looking at how smooth the marble on a statue is. Of course, as the photographer, I wince at my own images printed too large, but I have never seen more or less interest in a print due to resolution.

Link to post
Share on other sites

Always worth remembering that doubling the megapixels only increases resolution of details 1.4x.

I.E. an S2 37.5 Mpixel image is 7500 pixels wide. Double that to 75 Mpixels, and the image width only increases to 10600 pixels wide. It requires a 150 Mpixel sensor (15000 x 10000 pixels) to double the linear resolution of fine lines and point-details.

Trying to actually double one's resolution gets to be an expensive hobby.

It's true that one must use comparable sensor tech when comparing small or large pixel sizes. Today's D850 will beat a 12-year-old 13Mp D700. However, today's 12.2 Mpixel Sony A7S II will eat the hind legs off today's D850 in noise performance, at, say, ISO 51200. Or even 1600.

That leaves for another thread whether 150 Mpixels of boring - umm - manure is somehow more interesting than 12.2 Mpixels of boring manure. Sebastiao Salgados's 60"-wide prints from Tri-X work just fine. Because of the quality of the seeing, not the image quality.

 

  • Like 8
Link to post
Share on other sites

That was another reason I mentioned aspect ratio. I have increasingly changed my compositions to 4 to 5 rather than 3 to 2, and in doing so I am throwing a way a lot of resolution with the S. Having a sensor that matches the way you shoot is also important to maximize resolution.

 

By the way Adan, I am not so sure the A7S II will beat the D850 so handily as you think. I have the A7S, and while it is remarkable for what it can do, so is the A7Rii at 42mp. Size it down to the same 12mp, and it is very similar. I had a gut feeling it was like this, so I went and compared images I shot at ISO 12800 on both cameras, and they were very similar in grain level when sized equally. At 51,200 my A7S looks terrible, and has red sensor blooming from the bottom and edge of the sensor. The A7Rii actually looks better, though I only ever shot a few pictures at this ISO.

Anyway, just something I noticed. I think getting a camera you like to use is the most important part of the equation. All main line cameras are so good these days that most differences are academic for people who do not make very large prints. For those who do, good lenses at optimal apertures and optimizing the use of the sensor area is very important, along with choosing a company whose color science you like.

  • Like 3
Link to post
Share on other sites

7 hours ago, Stuart Richardson said:

Having a sensor that matches the way you shoot is also important to maximize resolution.

Well said Stuart, the kind of landscape I’m in for me demands wide screen prints and therefore I like the 2:3 aspect ratio.

  • Like 1
Link to post
Share on other sites

Am 25.5.2019 um 18:25 schrieb Jan1985:

A lot of people say that the Leica S with their bigger pixels on the size has almost better image quality than other medium-format cameras with more megapixel resolution. Is this right?

No, that's wrong.

If a Leica S picture happens to be better (higher technical image quality) than a picture from some other medium-format camera then it's due to the superior lens quality.

.

Am 25.5.2019 um 18:25 schrieb Jan1985:

Do more megapixel on an medium-format sensor not lead to better image quality?

Yes, they do.

Link to post
Share on other sites

2 hours ago, 01af said:

Yes, they do.

Up to a point, after which it will deteriorate. Technology is pushing the optimum higher and higher, but there is a point that it will run into diminishing returns.

Link to post
Share on other sites

On 5/27/2019 at 5:24 AM, Stuart Richardson said:

By the way Adan, I am not so sure the A7S II will beat the D850 so handily as you think.

One doesn't have to think one thing or the other.

One can just go to dpreview.com for the review of either camera, and call up a comparison sample picture from the other camera at the same ISO, and then one knows for sure.

Pretty easy with this D850 IQ sample page to call up the 7SII sample in one of the other "comparison panes" and see the difference at the same ISOs:

https://www.dpreview.com/reviews/nikon-d850/8

D850 noise at 12800 is about the same as the A7S II at 51200. A 4x advantage.

Link to post
Share on other sites

I think you missed my point Adan. The A7SII does indeed look better if you look at 1 to 1, but that is because it is a much lower resolution image. If you equalize the sizes, the story is a lot different. I downloaded Jpegs from that site and downsized the Nikon images to the same dimensions as the Sony. In that case, both of the Nikon images look better to my eyes, although the 51200 version has some magenta tint coming in. It has better detail, less moire and similar grain, however. I am attaching a screenshot from photoshop. The images are unadjusted. I simply downsized using regular bicubic interpolation. 

From left to right, Nikon at 51200, Sony A7II at 51200, Nikon at 12800. 

https://www.dropbox.com/s/98osmkshoilwlu3/Screenshot 2019-05-31 at 08.30.20.png?dl=0

Link to post
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
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...