M9 v M8 sharpness

[ulrikft]

Advertisement (gone after registration)

A 100% crop is simply looking at one part of the file itself without resizing. It is the file at its native size - a JPEG version (hopefully very lightly compressed) of what was made by the camera. Of course many of us would love to compare prints but that doesn't work for a group of photographers who are spread all over the world. We simply are not all in one physical place and we would need to be in order to compare prints.

 

And even if we were comparing prints, there are a lot of variables we would need to keep in mind - which printer? which paper? what ink? what size? what ppi setting? A print is at least one generation further removed from what comes out of the camera.

 

The print is the digital file's final destination (for many of us) but it is not the file itself or even a crop of it. A crop of a file seen at native size *is* part of the file itself and experienced photographers can actually tell a lot from looking at 100% crops. Much is often made of the (roughly) 72 ppi resolution of monitors. But it is important to understand that all this does is to spread a given number of pixels across a broader area - there's no reduction of information (simply from viewing at that res.) but instead a different distribution of it.

 

PPI is of course pixels per inch. (DPI refers to dots per inch and is a print measure - there of course are no ink dots on a monitor). A print made from a file sized to 240 ppi, for example, is simply spreading the same visual information (in the printing source file) across a smaller area.

 

In many respects, a print is more forgiving than a view (full or cropped) of the actual digital file on screen. Noise is less obvious in the print and some differences in resolution are lost as the ink absorbs into the paper (spreading a bit as it does that).

 

The idea that looking at a file at 100% size on screen is not useful is, in my view, mistaken. Such a view simply shows us a piece of the actual file itself - the source for all kinds of potential prints, etc. I think the important thing is to understand the 100% crop in context - it will often reveal differences that are hidden in a print, especially if one is using a very high quality monitor. And the print, which we see through light reflection, is very different from the monitor image, which we see through light transmission. But with practice one learns how to relate the latter to the former.

 

So unless we're all going to start traveling to a given location to look at prints together, I think the most useful thing is that we consider 100% crops in context. And the more one prints, the better he or she gets at understanding how the look of the digital file itself will relate to the look of a given print.

 

As far as resizing files to match their pixel dimensions, etc. - that can be useful if we want to compare two files at the size at which we would send them to the printer. So it makes sense sometimes, for example, when comparing the noise in a 21 MP file to that in a 12 MP file. And, for that reason, myself and other have been doing this for certain reviews, presentation of information in threads, etc.

 

For better or worse, the monitor (hopefully a very good one) is the medium we (as a group of photographers who communicate over the web) have in common as a way of sharing information in pictures. We can write about our impressions of prints but the monitor is where we "see" (to the extent possible) what another is talking about. That being the case, I think the task is to find out how best to use that medium of the monitor *and* how to relate that information to the specific kinds of prints we (individually) are making.

 

Ansel Adams (who was a great printer - one's feeling about his work itself being beside the point) often wrote that the negative is the score and the print is the performance. For digital work we might think of the RAW file as the negative, the JPEG or TIFF as an intermediary, and the print as the performance. We can't actually look at RAW files so the best we can do is to look at the intermediary. We can't normally fit a large digital file on our screens all once so we often look at it in sections.

 

Can a skilled musician tell a lot about a composition from the score? I would say yes. Can experienced photographers and printers tell a lot about a picture from seeing the actual intermediary file (as a JPEG, TIFF, etc.) - I would again say yes.

 

Cheers,

 

Sean

 

I still disagree, and I notice that you actually don't adress the points I and quite a few others have raised about the problems with 100% crops without resizing and using this both for resolution comparisons and iso comparisons across higher and lower megapixel sensors.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[delander †]

I think that the fact that both the M8 and M9 have the same pixel density makes this a useful test. That is not say other tests aren't useful.

 

The situation is different when cameras have the same sensor size but different pixel densities.

 

For me I'm happy, I would have been more concerned if in my test the M9 crop had been less sharp than the M8. As it is I see them both as the same.

 

Of course the M9 will allow me to produce the same photograph with 18 instead of 10Mp. I've been very pleased with the 10Mp of the M8 so I'm not expecting any resolution problems with the 18Mp of the M9. I trust Leica on that one.

 

Jeff

[EsaKivivuori]

Quote: (....) "The distance to the car was about 1.6 metres.# (....)

 

This means that you didn`t move tripod ? If that is the case and you used same lens and same aperture, there should not be any significant difference because the pixel pich (6,8 x 6,8 um) is the same.

 

If I`m correcty, M8 uses 8 bit DNG-fils and from M9 you can get uncompressed 16 bit DNG files.

 

Although Leica informs in M8/M8.2 technical specifications "DNG-file information: 6-bit color resolution, 10.2 Mbyte file size per picture" (sound odd).

 

BR

 

Esa

[delander †]

You are right I did not move the tripod the only thing changed was the camera body.

 

Jeff

[sean_reid]

I still disagree, and I notice that you actually don't adress the points I and quite a few others have raised about the problems with 100% crops without resizing and using this both for resolution comparisons and iso comparisons across higher and lower megapixel sensors.

 

Actually, I often do resize files to match pixel dimensions between cameras as part of my ISO noise testing. I imagine you haven't seen those reviews but I've been arguing in favor of that relevance for several years now.

 

In fact, the M9's primary noise advantage over the M8 comes when the M9's files are downsized to match the pixel dimensions of the M8 files (for printing) - as seen and discussed in the M9 review. That comparison has been discussed here on the forum at some length already. I've done this also in previous reviews and explained, in some detail, why it is relevant.

 

There's often merit to looking at both: actual size to actual size and, also, matched pixel dimensions.

[ulrikft]

Actually, I often do resize files to match pixel dimensions between cameras as part of my ISO noise testing. I imagine you haven't seen those reviews but I've been arguing in favor of that relevance for several years now.

 

In fact, the M9's primary noise advantage over the M8 comes when the M9's files are downsized to match the pixel dimensions of the M8 files (for printing) - as seen and discussed in the M9 review. That comparison has been discussed here on the forum at some length already. I've done this also in previous reviews and explained, in some detail, why it is relevant.

 

There's often merit to looking at both: actual size to actual size and, also, matched pixel dimensions.

 

I have actually seen those reviews, as I subscribe to your site. I still think that actual size to actual size has little value in most cases, beacause people won't consequently frame differently with one or the other camera. And many people draw the "wrong" conclusions like "the m9 is less sharp than the m8" for instance. When the opposite is clearly true, when one sees the entire sensor against the entire sensor. But i digress..

[sean_reid]

Advertisement (gone after registration)

I have actually seen those reviews, as I subscribe to your site. I still think that actual size to actual size has little value in most cases, beacause people won't consequently frame differently with one or the other camera. And many people draw the "wrong" conclusions like "the m9 is less sharp than the m8" for instance. When the opposite is clearly true, when one sees the entire sensor against the entire sensor. But i digress..

 

Looking at crops from actual size files, I see no reason to think that the M9 shows less resolution than the M8 (unless some kind of confounding variable comes into the results). Confounding variables are potentially a big problem in testing.

 

Once one starts re-sampling files (up or down) they lose some of their resolution because software is either making guesses, averaging data or some similar process. So, for looking at pure resolution, the files at actual size give the best information. And by actual size, I mean RAW conversions that do not resize the file up or down. So while I may resize for part of the noise tests (in addition to comparisons at actual size) I would not choose to do that for resolution comparisons.

 

But the larger point is that no matter how we do comparisons, people who see them must think about them in context - ie: What does this mean? What does this suggest? etc.

 

Gotta run.

 

Cheers,

[Hartmonstr]

Cameras aren't sharp or unsharp - only lenses are. If you photograph an object or scene using the same lens (properly focused) on both an M8 and an M9 from the same position, you can overlay the two files in Photoshop and the M8 file will be an identical subset (cropped) version of the M9 file.

 

Let's say, on the other hand, you photograph the same scene with the same properly focused lens, but this time moving the M9 back the distance needed to get the same coverage as the M8 image. The sharpness difference between these two files will only become apparent when the M8 file begins to exceed a certain magnification limit, whether on a print or a monitor. In other words, at some point the M8 file will begin to break up, losing apparent sharpness. The M9 file will, too, but at a larger magnification. This point of magnification is higher than most photographers will ever practically utilize; I have beautiful, extremely sharp 40-inch prints made from my M8.

 

So an M9 file can be enlarged bigger (or cropped more) than an M8 file. But up until the size where the M8 image starts to break up, in theory there will be no sharpness difference between images made from the two cameras. Of course, that 'theory' thing means that there are other possible variables that could cause differences between images from the two cameras, i.e., file processing, IR filter variations and, most likely, human error. But the main differences for real-world users come down to enlargement and the lens choices between a cropped and full frame sensor.

[lct]

Cameras aren't sharp or unsharp - only lenses are...

Not my opinion i must say. Suffice it to compare the same lenses on more or less sharp bodies like Nikon D100 and D70 or D40 for instance. All sensors are not the same and AA filters play a role obviously.

[delander †]

 

… Let's say, on the other hand, you photograph the same scene with the same properly focused lens, but this time moving the M9 back the distance needed to get the same coverage as the M8 image …

 

 

You have to move the M9 forward not back, thus making use of 18Mp instead of 10.

 

Jeff