Jump to content

Medium format quality? Test results: M10P vs. M10M vs Sony a7riv vs Phase One IQ4 with optimum lenses


onasj
 Share

Recommended Posts

The forum won't let me upload the comparison matrix... apparently I'm at the total attachment limit but I don't see a way to clear my older attachments 😕

Here is a link to the comparison matrix:

https://www.dropbox.com/s/5a3g3hvk7za19wb/crop-14204 Comparison Matrix.jpg?dl=0

Edited by onasj
Link to post
Share on other sites

Thanks for the test, David.  I believe that most people who compare the M10M to medium format are referring to medium format film, not the Phase One, which few people can practically obtain.  It makes sense that a much larger sensor (2.4x) with nearly 2.5 to 6 times the number of pixels, which themselves are much larger, would outperform the full frame sensors.

Glad to know that the M10P continues to produce respectable images even compared to the other two cameras.  There are so many ways to measure "better."

It is interesting to watch people at gallery exhibitions.  You know the photographers right away because they get right up close to large prints to see how sharp the image is.  Everyone else stands back and marvels at the images.  The first is a function of the equipment, the second of the photographer.

I am fortunate enough to have obtained a signed print of H C-B's, Derrierre la Gare St. Lazarre many years ago, before the price went up to the cost of a Phase One.  It's my favorite photograph of all time and Time magazine called it the most significant photograph of the 20th Century.  The Phase One would blow it away in so many respects, but would it make a better photograph?  Depends on the purpose of the picture I guess.  For wall-sized landscapes - Phase One hands down.  For other work......

Again, thanks for the tests.  These are very tedious to do and I think most people appreciate the effort you took to do them and then to make them available.  Several hours worth of work!

Link to post
Share on other sites

Quote

 ...I believe that most people who compare the M10M to medium format are referring to medium format film, not the Phase One, which few people can practically obtain.  It makes sense that a much larger sensor (2.4x) with nearly 2.5 to 6 times the number of pixels, which themselves are much larger, would outperform the full frame sensors...

 

Let's not lose sight of the big picture while dissecting quarks:  The M10M can make really nice photographs. 

That's a good enough reason to buy one IMHO.

Link to post
Share on other sites

Thanks for the very interesting comparison.  From what I read, I couldn't tell which IQ4 back you used.  100mp Trichromatic or 150mp?  Also it would be great if you could borrow a 150mp Achromatic (monochrome) back.  I attended a photo workshop in Santa Fe, NM, this Fall where every student got to any of the three Phase One backs to their hearts content on an XT or an XF.  After the workshop everyone could instantly pick out which prints at 13 x 19 in came from the Phase One backs.  Other cameras used included the hi-res Canon DSLR.  I don't recall a Nikon D850 being used.  A couple of Sony's were used, but I don't know which models.  It wasn't the sharpest that allowed the prints to be recognized, but it was the tonal quality.  I'm looking in the piggy bank for enough money to buy an M10M to pair with my M9.

Link to post
Share on other sites

I used the IQ4 151 MP color back.  The only “trichromatic” (Phase One’s term for a special narrow-wavelength CFA) to my knowledge was the IQ3-100 MP.  The IQ4 151 MP back comes in regular color and in monochrome (achromatic, to use Phase One terminology).  I agree it would be great to compare to the 151 achromatic but sadly I don’t have access to one (currently).

Link to post
Share on other sites

5 hours ago, onasj said:

Leica’s new M10 Monochrom (M10M) has been touted by multiple reviewers as offering image quality rivaling that of medium-format film.  And Sony explicitly claims that its a7riv camera rivals medium-format cameras.  In this test, I compare side-by-side the Leica M10P (24 MP, color) with the M10M (41 MP, monochrome), the Sony a7riv (61 MP, color), and the “full-frame medium format” (54 x 40 mm sensor) Phase One IQ4 (151 MP, color), currently the largest and highest-resolution imaging sensor available to most photographers.  (Note that the 102-MP Fuji GFX-100 has a cropped (44 x 33 mm) version of the Sony-manufactured sensor In the IQ4.)  For each camera, I used the very best available lens at the best-quality aperture and at base ISO.

 

Test conditions:

Leica M10P + Leica APO 50 Summicron at f/5.6, ISO 100

Leica M10M + Leica APO 50 Summicron at f/5.6, ISO 160

Sony a7riv + Sony 85/1.4 GM at f/5.6, ISO 100

Phase One IQ4 + Rodenstock HR90 at f/9, ISO 100 (yes, the base ISO of the IQ4 is 100, not 50).

 

Shooting the IQ4 at f/9 gives a comparable depth of focus as shooting the other cameras at f/5.6.  I know from my other tests, and from discussions with Phase One dealers, that at f/9 diffraction is not limiting in practice for the IQ4 sensor (though it can be limiting at apertures smaller than f/9).

 

Everything was shot on a tripod with 2-3-second release delay to avoid shake, at subject distances to give similar subject size and position.  Note that the aspect ratio of the IQ4 sensor is 4:3, rather than 3:2 like the other sensors, so I used frame width to match framing.  Since I was shooting the IQ4 on a technical camera, all four cameras were mirrorless, with no risk of mirror slap vibrations spoiling sharpness.

 

Raw files were exposure-matched and converted to monochrome using Capture One 20.  Then each converted max-quality JPEG was auto-bicubic-scaled in Photoshop to either 7864 pixels wide (the width of native M10M images) or to 14204 pixels wide (the width of native IQ4 images).

 

You can download all the raw files and full-resolution max-quality JPEG files here:

https://www.dropbox.com/sh/scvrfg8h41naux4/AADw8Cfm37db2BWkGTbie3M2a?dl=0

 

A tiny 100% crop from near the center of the 14204 pixel-wide images was composited into a matrix for comparison, attached to this post.  Click on the image to see the comparison full-size.  I choose to use the scaled-up 14204-pixel-wide images since one goal of this test is to determine if any of these imaging systems can approach medium format quality, not to test if they can approach down-sized medium format quality.  You can build analogous matrices from the smaller images as well using the files in the above link if you would like. 

 

Summary of findings:

1) As expected, all of these four imaging systems with optimum glass, ISO, and aperture produce excellent images at their native sizes.

 

2) The Bayer CFA-less M10M does indeed punch above its megapixel weight.  With respect to capturing subject details, I found the M10M (41 MP) and Sony a7riv (61 MP) to be virtually identical.  In fact, to my eye the M10M+50 APO was able to capture slightly more details than the a7riv+85 GM—compare the lint on the upper grey band of the vessel in the attached matrix.

 

3) The Phase One IQ4 eats all the competitors for lunch if you compare image quality at 100%.  But a 14,204-pixel-wide image printed at 200 dpi is 6 feet wide (!).  At any practical viewing distance (assuming you don’t need to crop heavily), I would say that the IQ4, M10M, and a7riv are actually quite similar, even when scaling up the latter two images to the 14,204 pixel width of the IQ4 image.  Which is pretty cool.

 

4) The M10P, as you might expect for a 24-MP camera thrown into a resolution and sharpness gunfight, lags well behind the others, but even so, once I view the M10P image at about 50% magnification it begins to look quite similar to the other images.

 

5) In terms of bang for the buck, Sony offers outstanding value, as usual.  The Sony a7riv + 85 GM lens costs ~$5,200.  Either Leica body + the 50 APO costs ~$17,000.  And the Phase One IQ4 + Rodenstock HR90 costs around $55,000.  And the Sony is the only one of the four tested systems that offers autofocus.

 

So can the Leica M10M or the Sony a7riv offer image quality rivaling that of state-of-the-art digital medium format systems?  Yes, at practical viewing sizes and distances, when using optimal glass.  But if pixel peeping or (very) large prints are your target application, then medium format, for now, has no peer.

 

I like Monochrome Leica, I get what film MF is.

But this is just totally wrong:

Quote

Leica’s new M10 Monochrom (M10M) has been touted by multiple reviewers as offering image quality rivaling that of medium-format film.

Film is film, digital is digital. But online reviewers are not photogs to realise it.  IMO.

Link to post
Share on other sites

Many thanks from doing such a well-explained test. It’s much appreciated.

I came to the same conclusion as your own observations. The Phase One back really is incredible, and its benefit becomes more and more obvious the larger the print size.

For such a controlled subject, though, I kept wondering what we’d conclude if the test included a Panasonic S1R specifically in its “high resolution mode”?

I find “high resolution mode” gives a noticeable resolution and noise improvement over the normal mode. This multishot mode is also expected to be made available for the Leica SL2 in due course. With an SL Summicron prime attached, I expect it could make for a very interesting comparison at some stage.

Edited by Jon Warwick
Link to post
Share on other sites

vor 15 Stunden schrieb onasj:

Leica’s new M10 Monochrom has been touted by multiple reviewers as offering image quality rivaling that of medium-format film.

That's a ridiculous statement because regular digital 35-mm-format cameras, with their Bayer filters, are exceeding the overall image quality of medium-format film for more than a decade now. Some are even approaching the quality of 4×5" sheet film.

.

vor 15 Stunden schrieb onasj:

But if pixel peeping or (very) large prints are your target application, then medium format, for now, has no peer.

Now that's comparing digital 35-mm format to digital medium format. And so, it's hardly surprising. Bigger is better; it always has been and always will be (given the same underlying technology). The only surprising thing is how small the bigger format's advantage is and how many loops you have to jump to make it perceptible.

Link to post
Share on other sites

I think different reviewers of the M10M in making comparisons with medium format were referring to different aspects— in tonality, vs. MF film, and in detail capture, vs. MF digital.  In any case, I was impressed that the M10M could capture as much (if not more) detail than the a7riv, and as much detail at all but the largest viewing sizes as the IQ4.

Perhaps the real take-home here is the M10M, and the rumored M10R color 41-47-MP cousin, now allow Leica shooters to have it all (except autofocus and IBIS, but the need for the latter is much less when you can shoot very nice images at ISO 25,000)—Leica lenses, Leica design and form factor, Leica functional simplicity, and now state-of-the-art sensors.

I’m eager to see how Bill Claff’s tests turn out for the M10M.  I suspect the low-light performance may set a record, or come close to it (currently held by the IQ4 in his measurements).

Edited by onasj
Link to post
Share on other sites

14 hours ago, onasj said:

I used the IQ4 151 MP color back.  The only “trichromatic” (Phase One’s term for a special narrow-wavelength CFA) to my knowledge was the IQ3-100 MP.  The IQ4 151 MP back comes in regular color and in monochrome (achromatic, to use Phase One terminology).  I agree it would be great to compare to the 151 achromatic but sadly I don’t have access to one (currently).

This is a great thread.  Just for completeness, there is a 100MP IQ4 Trichromatic back.  I did not try this back; I only took a few shots with the 150MP color and Achromatic backs.

https://www.phaseone.com/en/Photography/XF-Camera-System/Camera-Configurations/XF-IQ4-100MP-Trichromatic-Camera-System

Link to post
Share on other sites

Sorry—I forgot that Phase One uses IQ4 to refer to the processor and software running the back, not to the sensor chip technology.  Confusing and somewhat misleading.  Only the 151 MP IQ4 backs have the BSI Sony IMX411 sensor, which provides the extraordinary dynamic range and high-ISO performance (for 3.76 micron pixel size) of the resulting back.  While you can buy 100 MP IQ4 systems, they have the older non-BSI chip with (slightly) poorer dynamic range and substantially worse high-iso performance. Plus the IQ3 back firmware and OS is much less buggy than all the IQ4 firmware releases to dates.  They should really number their IQ systems based on the sensor, not the processor and software...

Edited by onasj
Link to post
Share on other sites

On 1/25/2020 at 1:56 PM, onasj said:

Leica’s new M10 Monochrom (M10M) has been touted by multiple reviewers as offering image quality rivaling that of medium-format film.  And Sony explicitly claims that its a7riv camera rivals medium-format cameras.  In this test, I compare side-by-side the Leica M10P (24 MP, color) with the M10M (41 MP, monochrome), the Sony a7riv (61 MP, color), and the “full-frame medium format” (54 x 40 mm sensor) Phase One IQ4 (151 MP, color), currently the largest and highest-resolution imaging sensor available to most photographers.  (Note that the 102-MP Fuji GFX-100 has a cropped (44 x 33 mm) version of the Sony-manufactured sensor In the IQ4.)  For each camera, I used the very best available lens at the best-quality aperture and at base ISO.

 

Test conditions:

Leica M10P + Leica APO 50 Summicron at f/5.6, ISO 100

Leica M10M + Leica APO 50 Summicron at f/5.6, ISO 160

Sony a7riv + Sony 85/1.4 GM at f/5.6, ISO 100

Phase One IQ4 + Rodenstock HR90 at f/9, ISO 100 (yes, the base ISO of the IQ4 is 100, not 50).

 

Shooting the IQ4 at f/9 gives a comparable depth of focus as shooting the other cameras at f/5.6.  I know from my other tests, and from discussions with Phase One dealers, that at f/9 diffraction is not limiting in practice for the IQ4 sensor (though it can be limiting at apertures smaller than f/9).

 

Everything was shot on a tripod with 2-3-second release delay to avoid shake, at subject distances to give similar subject size and position.  Note that the aspect ratio of the IQ4 sensor is 4:3, rather than 3:2 like the other sensors, so I used frame width to match framing.  Since I was shooting the IQ4 on a technical camera, all four cameras were mirrorless, with no risk of mirror slap vibrations spoiling sharpness.

 

Raw files were exposure-matched and converted to monochrome using Capture One 20.  Then each converted max-quality JPEG was auto-bicubic-scaled in Photoshop to either 7864 pixels wide (the width of native M10M images) or to 14204 pixels wide (the width of native IQ4 images).

 

You can download all the raw files and full-resolution max-quality JPEG files here:

https://www.dropbox.com/sh/scvrfg8h41naux4/AADw8Cfm37db2BWkGTbie3M2a?dl=0

 

A tiny 100% crop from near the center of the 14204 pixel-wide images was composited into a matrix for comparison, attached to this post.  Click on the image to see the comparison full-size.  I choose to use the scaled-up 14204-pixel-wide images since one goal of this test is to determine if any of these imaging systems can approach medium format quality, not to test if they can approach down-sized medium format quality.  You can build analogous matrices from the smaller images as well using the files in the above link if you would like. 

 

Summary of findings:

1) As expected, all of these four imaging systems with optimum glass, ISO, and aperture produce excellent images at their native sizes.

 

2) The Bayer CFA-less M10M does indeed punch above its megapixel weight.  With respect to capturing subject details, I found the M10M (41 MP) and Sony a7riv (61 MP) to be virtually identical.  In fact, to my eye the M10M+50 APO was able to capture slightly more details than the a7riv+85 GM—compare the lint on the upper grey band of the vessel in the attached matrix.

 

3) The Phase One IQ4 eats all the competitors for lunch if you compare image quality at 100%.  But a 14,204-pixel-wide image printed at 200 dpi is 6 feet wide (!).  At any practical viewing distance (assuming you don’t need to crop heavily), I would say that the IQ4, M10M, and a7riv are actually quite similar, even when scaling up the latter two images to the 14,204 pixel width of the IQ4 image.  Which is pretty cool.

 

4) The M10P, as you might expect for a 24-MP camera thrown into a resolution and sharpness gunfight, lags well behind the others, but even so, once I view the M10P image at about 50% magnification it begins to look quite similar to the other images.

 

5) In terms of bang for the buck, Sony offers outstanding value, as usual.  The Sony a7riv + 85 GM lens costs ~$5,200.  Either Leica body + the 50 APO costs ~$17,000.  And the Phase One IQ4 + Rodenstock HR90 costs around $55,000.  And the Sony is the only one of the four tested systems that offers autofocus.

 

So can the Leica M10M or the Sony a7riv offer image quality rivaling that of state-of-the-art digital medium format systems?  Yes, at practical viewing sizes and distances, when using optimal glass.  But if pixel peeping or (very) large prints are your target application, then medium format, for now, has no peer.

Thanks for doing this test. is there a reason why you did not include the M246?

Link to post
Share on other sites

  • 3 weeks later...

I decided to compare the M10M to an 007 S.

The sensor of the M is more modern and more pixels. I used a zoom on the S vs an APO50  prime on the M.

I set the ISO at 800 for each and f8 for each, 1/2000 hand held.

The noise on the M is a bit more visible on the M10M but I prefer the look of the M10M.  Is this a test of a zoom vs a prime more than an M10M vs S?

Who knows. It tells me the the M10M is a nice substitute for now.

The three images are M crop, S crop, M full image

Link to post
Share on other sites

vor 5 Stunden schrieb LeicaS2:

I forgot that the M10M without the Bayer color filtration is really effective 82 mp. It perhaps shows.

It isn't

If you leave away the filter you don't increase the number of pixels. You might say that the effective number of  pixels of a sensor with color filters is decreased but not the other way round.

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

×
×
  • Create New...