A small demo of 100MP vs. 60MP

[ErikKaffehr]

Hi,

Below is an attachment from Doug's studio shot of the IQ3-100MP compared to the IQ3-60. The 100 MP image was downscaled to 60 MP.  On both images Capture One default settings were used (at least I hope so).

Now, there is some cheating involved. I chose the f/11 shots from each. I know from experience that f/11 almost eliminates aliasing on 4.5 micron pixels, while larger pixels may need f/16.

In my humble opinion, this illustrates that there is a significant advantage to the 100 MP back compared to the 60 MP back if proper rendition is required.

Another point I feel needs to added is that the IQ3-100MP probably has gapless microlenses which effectively enlarge the aperture of the pixels, thus doing more of area sampling and less of point sampling. That may loose some crispness but improves correctness of rendition.

So, I have no doubt, the 100 MP back is better. I would think that the IQ350 performs very similarly, but with a crop factor.

Best regards
Erik

Ps. Thanks a lot to DigitalTransitions for making this test and making the raw files available for download:
https://digitaltransitions.com/massive-still-life-shootout/

[eronald]

Hi,

Below is an attachment from Doug's studio shot of the IQ3-100MP compared to the IQ3-60. The 100 MP image was downscaled to 60 MP.  On both images Capture One default settings were used (at least I hope so).

Now, there is some cheating involved. I chose the f/11 shots from each. I know from experience that f/11 almost eliminates aliasing on 4.5 micron pixels, while larger pixels may need f/16.

In my humble opinion, this illustrates that there is a significant advantage to the 100 MP back compared to the 60 MP back if proper rendition is required.

Another point I feel needs to added is that the IQ3-100MP probably has gapless microlenses which effectively enlarge the aperture of the pixels, thus doing more of area sampling and less of point sampling. That may loose some crispness but improves correctness of rendition.

So, I have no doubt, the 100 MP back is better. I would think that the IQ350 performs very similarly, but with a crop factor.

Best regards
Erik

Ps. Thanks a lot to DigitalTransitions for making this test and making the raw files available for download:
https://digitaltransitions.com/massive-still-life-shootout/

Erik,

 Thanks for pointing out the test - I think that like most here I will draw my own conclusions.

Edmund

[ErikKaffehr]

Hi Edmund,

Whatever conclusions you draw, I hope that you share.

Just to say a very interesting test with a lot of good stuff.

Best regards
Erik

Erik,

 Thanks for pointing out the test - I think that like most here I will draw my own conclusions.

Edmund

[voidshatter]

I would say this is a really great test to show the superiority of the Sony CMOS sensor over the Dalsa CCD.

Thanks again to Doug!

[torger]

With 6um pixels I prefer f/16. Too much aliasing with f/11. The 100MP looks like a good f/11 camera.

The majority seems to be one stop ahead of me though (ignoring aliasing and preferring higher pixel peep contrast) and shoot f/11 on 6um so maybe f/8 will be the new f/11 for the 100MP?

Of course for portraits you shoot wider to get a shorter DoF, and then I've always thought not having that AA-filter is broken design but the more pixels you have the smaller the problem becomes.

[ErikKaffehr]

Hi Anders,

Pretty much my point. I chose f/11 for this demo because I guessed f/11 would be aliasing free on the 100 MP sensor. By the way, I did see significant aliasing on the IQ3-100MP at f/8. And yes, I have seen aliasing at f/8 on my Sony A7rII which has the same pixel size, even with a humble Canon 16-35/4 zoom.

One point which has been demonstrated by Doug's test is that in most cases the sensor simply outresolves the subject. There is very little fine detail in that test. The surface of the ColorChecker Passport is such a detail and it happens to be in near perfect focus.

Best regards
Erik

With 6um pixels I prefer f/16. Too much aliasing with f/11. The 100MP looks like a good f/11 camera.

The majority seems to be one stop ahead of me though (ignoring aliasing and preferring higher pixel peep contrast) and shoot f/11 on 6um so maybe f/8 will be the new f/11 for the 100MP?

Of course for portraits you shoot wider to get a shorter DoF, and then I've always thought not having that AA-filter is broken design but the more pixels you have the smaller the problem becomes.

[Bart_van_der_Wolf]

With 6um pixels I prefer f/16. Too much aliasing with f/11. The 100MP looks like a good f/11 camera.

Hi Anders,

I agree, depending on the lens quality. A 6 micron pitch sensor is only virtually aliasing free (due to diffraction blur, MTF near zero) at f/20 or narrower, with some lens aberrations added maybe stop wider, say f/14 .
A sensor pitch of 4.6 micron will be virtually aliasing free at f/16 or narrower, so with a stop wider for lens aberrations, say f/11 should be good. Anything narrower will lose effective resolution, anything wider will be sharper but with increased aliasing.

The majority seems to be one stop ahead of me though (ignoring aliasing and preferring higher pixel peep contrast) and shoot f/11 on 6um so maybe f/8 will be the new f/11 for the 100MP?

I don't care about what others do, because I'm kind of allergic to aliasing artifacts which will also look worse when resampling.

Of course for portraits you shoot wider to get a shorter DoF, and then I've always thought not having that AA-filter is broken design but the more pixels you have the smaller the problem becomes.

Yes, and there is less DOF, so more will be defocused. Defocus will also kill aliasing very quickly. But aliased hair still looks weird.

Cheers,
Bart

[Bart_van_der_Wolf]

One point which has been demonstrated by Doug's test is that in most cases the sensor simply outresolves the subject. There is very little fine detail in that test. The surface of the ColorChecker Passport is such a detail and it happens to be in near perfect focus.

I'm not sure if there is not another cause for that, lens aberrations and diffraction blur combined with the sensor element apertures, or camera shake. I have not seen any good tests that quantify camera shake/vibrations, and how to minimize it (not just in theory, but in actual practice). All camera/tripod combinations have some, and with fine enough resolution, and long enough focal lengths, some of that effect will creep into these tests.

Cheers,
Bart.

[eronald]

I'm not sure if there is not another cause for that, lens aberrations and diffraction blur combined with the sensor element apertures, or camera shake. I have not seen any good tests that quantify camera shake/vibrations, and how to minimize it (not just in theory, but in actual practice). All camera/tripod combinations have some, and with fine enough resolution, and long enough focal lengths, some of that effect will creep into these tests.

Cheers,
Bart.

the new phase back or cam has a "seismometer". I think it can wait for vibrations to die down to take a pic. It's a neat trick.

Edmund

[Bart_van_der_Wolf]

the new phase back or cam has a "seismometer". I think it can wait for vibrations to die down to take a pic. It's a neat trick.

Hi Edmund,

Indeed, and probably for a reason. However, the release of the shutter (or aperture), and to a lesser extent the closing of it, will also initiate vibrations. Electronic first curtain will alleviate that to a certain extent, but there are several shutter configurations possible if I understand it well.

Cheers,
Bart

[Wayne Fox]

the new phase back or cam has a "seismometer". I think it can wait for vibrations to die down to take a pic. It's a neat trick.

Edmund

It's in the XF camera body.  Personally I"m not sure I trust it, it seems to fire way to fast after I push the button, like nearly instantly.  I've calibrated it but at this point I've gone back to just letting it count down.  I think the feature would be better if you could add a delay, so after the seismometer decides it's still enough it would wait another "x" seconds.

I've watched the readout on the seismometer and it still shows slight movement when it fires.  So there seems to be a "tolerance" in there which means it crosses a threshold that might not be as still as it might be.  As I mentioned, it almost always fires almost instantly after pressing the shutter button .. .maybe 1/2 second delay is the most i"ve seen unless I intentionally move the camera tripod.

cool feature, but I need to test it more.

[bbrantley]

I played with seismo on an xf the other day, and it was a little slower than yours, Wayne.  On my cube with a 120 macro LS, it usually waited one or two seconds after the press.  I didn't actually look to see if that yielded maximum sharpness, though.


Sent from my iPad using Tapatalk

[ErikKaffehr]

Hi,

Should be helpful when shooting in the wind…

Best regards
Erik

It's in the XF camera body.  Personally I"m not sure I trust it, it seems to fire way to fast after I push the button, like nearly instantly.  I've calibrated it but at this point I've gone back to just letting it count down.  I think the feature would be better if you could add a delay, so after the seismometer decides it's still enough it would wait another "x" seconds.

I've watched the readout on the seismometer and it still shows slight movement when it fires.  So there seems to be a "tolerance" in there which means it crosses a threshold that might not be as still as it might be.  As I mentioned, it almost always fires almost instantly after pressing the shutter button .. .maybe 1/2 second delay is the most i"ve seen unless I intentionally move the camera tripod.

cool feature, but I need to test it more.

[tjv]

I would say this is a really great test to show the superiority of the Sony CMOS sensor over the Dalsa CCD.

Thanks again to Doug!

What are the processing settings in C1 for these? What is the exposure / shadow push here?

[Chris Livsey]

Hi,

Should be helpful when shooting in the wind…

Best regards
Erik

Under suitable conditions, that is very unsuitable ones, would it delay indefinitely?

[ErikKaffehr]

Hi,

Just some small reflection. If 100 MP is not needed, I would be pretty sure that 50 MP cropped sensor in the IQ-250, Hasselblad 50c, Pentax 645Z would deliver the same image quality per pixel. So, if 50MP is enough and the 1.3X crop factor is not an issue, I would expect that the 50 MP CMOS sensors would perform very similarly to the full size 100 MP sensors cropped to 50 MP.

Best regards
Erik

[Wayne Fox]

Under suitable conditions, that is very unsuitable ones, would it delay indefinitely?

the camera will fire when it reaches the end of the programmed delay countdown regardless.  Typically you would set the camera to an 8 second delay to allow maximum time, so if the seismo doesn’t trip the shutter, it will fire at 8 seconds .... presumably the most stable it will become.

[Wayne Fox]

the camera will fire when it reaches the end of the programmed delay countdown regardless.  Typically you would set the camera to an 8 second delay to allow maximum time, so if the seismo doesn’t trip the shutter, it will fire at 8 seconds .... presumably the most stable it will become.

OK, this is off topic, but needed to follow up on my statements.  My concern for the seismo trip technology came from a shoot where several images were not sharp, and there was no sharp point in the entire image, even though I had distances from a few feet to several hundred yards.  I thought it may have been motion, blaming the seismo for tripping too quicly. Still not sure why they were blurry, but I tested the seismo trip technology today.

Even the motion of gentle  tapping a tripod leg (very solid RRS tripod) was enough to prevent the shutter from tripping.  As soon as I stopped tapping, about a second later the shutter would trip.  After about 10 experiments, it appears this technology works extremely well. I retract my concerns it doesn't work.

[dchew]

...and the 1.3X crop factor is not an issue...

Best regards
Erik

That's the rub with me. It has taken me a while to settle on a set of tech camera lenses I really like: 40hr, sk60xl, 90hrsw, sk150xl (see I'm brand agnostic ).

If I upgrade from the 180 to the 3100, it will not be for sensor image quality. It will be for live view and to keep me in the upgrade circle. The 150/250 would probably be fine but I can't give up the lenses and view angles I've come to love.

Dave

[ErikKaffehr]

Hi,

I would have considered buying one of Hasselblads VF50C backs, was it not for the crop factor.

Now, I feel that the A7rII with a T&S adapter gives me the use of Hassy lenses, and with a Canon 24/3.5 TSE I also have the short end covered.

I don't think I would ever afford the high en backs.

Best regards
Erik

That's the rub with me. It has taken me a while to settle on a set of tech camera lenses I really like: 40hr, sk60xl, 90hrsw, sk150xl (see I'm brand agnostic ).

If I upgrade from the 180 to the 3100, it will not be for sensor image quality. It will be for live view and to keep me in the upgrade circle. The 150/250 would probably be fine but I can't give up the lenses and view angles I've come to love.

Dave