Diglloyd tests the P645Z

[BernardLanguillier]

Yes, something like that is what one would hope for. Obviously, large mirrors, and large curtain shutters (especially the faster traveling ones), will need to abruptly move more mass and thus create more opportunity to cause vibrations, duh.

The question then becomes, how well engineered are the counter measures, and do they result in a workable solution? One also has to consider that comparisons between different sensor dimensions need to be normalized, because larger sensors require less magnification for the same targeted output size.

If the pixels are of equal quality, why is sensor size relevant?

Cheers,
Bernard

[Bart_van_der_Wolf]

If the pixels are of equal quality, why is sensor size relevant?

Hi Bernhard,

Two ways of looking at it.

To magnify from a sensor size to an output size, say a physically larger sensor e.g. 32.8mm versus 24mm, will require 26.8% less magnification to achieve the same output size. If the sensor arrays have similar sensel pitch, the physically larger sensor will have 36.7% more sensels and therefore higher resolution at the output size.

Alternatively, the sensel pitch of the physically larger sensor array could also be larger, which would reduce the resolution advantage proportionally with increasing pitch size. If the pitch size would be such that both sensor arrays have the same number of sensels (vertically), then the larger sensels of the physically larger sensor would still achieve a higher MTF, because of larger image magnification (assuming optics with the same angle of view) of the same features that are projected on the sensor, and we know that MTF reduces towards smaller details, higher spatial frequencies.

So that's why the LPPH are needed to normalize comparisons of differently sized sensors.

Cheers,
Bart

[ErikKaffehr]

Hi,

As it seems now, we know we have an issue with vibration on the A7r. Lloyd found a way to mitigate it, using a camera flash as additional weight.
It is a bit odd that Lloyd gets blame for discovering the problem. Some testers have not seen it, I guess they are not observant enough, or just overenthusiastic about their system.

Getting back to the 645D, we know that it may be a problem. The issue was seemingly noted on the 645D, but it seems that the shutter was reworked. The problem is probably observable on very high resolution detail, when shooting from a tripod. Lloyd had a USAF type test target in one of his shot which indicated the problem quite clearly on the Sony A7r.

Jim Kasson also measured the Nikon D800 (or D800E) and it also had a measurable vibration, but almost an order of magnitude less than the A7r.

It is really a part of the problem with high resolution, there are a lot of ways to loose it.

Best regards
Erik

Hi Bernhard,

Two ways of looking at it.

To magnify from a sensor size to an output size, say a physically larger sensor e.g. 32.8mm versus 24mm, will require 26.8% less magnification to achieve the same output size. If the sensor arrays have similar sensel pitch, the physically larger sensor will have 36.7% more sensels and therefore higher resolution at the output size.

Alternatively, the sensel pitch of the physically larger sensor array could also be larger, which would reduce the resolution advantage proportionally with increasing pitch size. If the pitch size would be such that both sensor arrays have the same number of sensels (vertically), then the larger sensels of the physically larger sensor would still achieve a higher MTF, because of larger image magnification (assuming optics with the same angle of view) of the same features that are projected on the sensor, and we know that MTF reduces towards smaller details, higher spatial frequencies.

So that's why the LPPH are needed to normalize comparisons of differently sized sensors.

Cheers,
Bart

[Paul2660]

Erik

I noticed Nikon moved to an electronic first curtain on the D810. They may found the same issue you noted. In my usage with the D800 I really cant say I have noticed the issue either when hand held or on tripod.

Paul

[eronald]

Actually, the thing might resonate with a strong Q factor at a frequency near 1/25s or 1/12s and the Q factor means that the double shock causes the visible degradation there, with the system being fairly well baflled with respect to other frequencies.
Update - this is trash of course if that camera does not have a double shutter shock (electronic first curtain)

Edmund

Hi,

A good question. The difference is the behaviour of the first curtain. On a mirrorless camera the shutter needs to be closed pre exposure and that causes vibrations. But, it seems that the opening of the shutter causes most vibration. With a short exposure it will not affect resolution that much as motion is probably slow. Conservation of momentum essentially says that if shutter is 1g and moves at 10 m/s and the camera is say 500 g it will move at 1/50 m/s. So say with 10 ms exposure we will have 1/5000 mm motion not noticable. But with 1/10 s we will have 2 microns of motion, and that will be visible. With longer times the motion will be absorbed (or arrested?) by the mass of the tripod and sharpness improve again.

No great physics… but I guess you see my point.

These images, using a thermonuclear D50 illuminant at 93e6 miles distance ( http://www.space.com/17081-how-far-is-earth-from-the-sun.html) , show the loss off sharpness: http://blog.kasson.com/?m=20140110

Best regards
Erik

[gdanmitchell]

If the pixels are of equal quality, why is sensor size relevant?

Sensor size relates to system resolution, so lenses with equal lp/mm resolution will produce higher system resolution on the system with the larger sensor.

[ErikKaffehr]

Hi,

I guess that electronic first curtain is a property of the sensor, not the camera.

Regarding the shutter vibration on the A7r, it is more measurable than visible. You can observe it as a softening of the image at certain exposure times, or drop in MTF. No double contours, or so.

The D800 has something like 1/8 of the vibrations compared to the A7r, as I recall, so it would not cause a visible degradation.

Here is an actual measurement by Mike Colette:

Full size:
http://www.kasson.com/ll/a7R_shutter_annotated.png

Full article here: http://blog.kasson.com/?p=4459

Best regards
Erik

Erik

I noticed Nikon moved to an electronic first curtain on the D810. They may found the same issue you noted. In my usage with the D800 I really cant say I have noticed the issue either when hand held or on tripod.

Paul

[gdanmitchell]

Hoping for discussion of the Pentax 645z, and a bit disappointed to see that this thread has wandered off to other subjects.

[ErikKaffehr]

Hi,

Good point.

Here is how I see it. Lloyd has reviewed cameras for a long time, so he has a good insight in the qualities of different systems. He tests a lot of gear, and he likes some more and some less. The point is that he has a wide experience.

The point has been made that he is not into medium format, which is quite true. He has tested several MF format cameras. He tested the Mamiya DL28 with a Leaf Aptus II-6, the Pentax 645D, Leica S2, Hasselblad H4D50 and now he is testing the Pentax 645Z. Of the cameras he tested I guess the only one he really liked is the Leica S2. The Hasselblad was quite OK, I think, but he felt it was to large. He was also impressed by the DL28, but times are changing and competition from affordable DSLR technology is more intense.

Lloyd generally has issues with AF, he finds it is not reliable enough. I don't know what is reality. Lens rentals had a series of articles on AF on Nikons and Canons, and it is pretty clear that neither system can focus accurately, except latest generation Canon bodies and the latest generation Canon lenses seem to focus very well. Lloyd generally finds that live view based focusing is needed for accurate focus, period.

Regarding his testing he essentially does a lot of comparison shots. The best one is probably a mosaic in front of some university close by. Very good detail at near infinity. He makes aperture series. Lloyd spends a lot of effort on presentation, you can instantly switch between images, which is good for comparison. Most of the images are at actual pixels.

He also has some table top images, that add depth. In such an image only a small part is in accurate focus, but it is possible to study bokeh. He also shoots some outdoor subjects, these are mostly intended for checking rendition. Focus in these cases is not crucial. You simply look for a point that is in focus.

Getting back to the Pentax 645Z. I would say that he likes the camera. It is probably the most advanced and well designed MF DSLR. But, he has some reservations. Mostly the price. He feels that the lenses are good enough, but not really excellent. AFAIK he has to date tested two lenses on the P 645Z, the 90 mm macro and the 25/4. Both lenses are good, but the 25/4 lacks the bite of a really excellent lens. He compared it with the Sony Alpha 7r using a Zeiss 21/2.8 lens. The Zeiss 21/2.8 and A7r combo costs less than the Pentax 645Z lens alone, but in the test shot he made the Sony A7r/Zeiss combo outperformed the Pentax combo at all apertures and across the field. So, he asks why, pay 2-3 times the money for no better results?

He usually test stuff for a couple of months, so we are at the beginning of this test.

Issues he doesn't really go into is tethering, flash, skin tones etc. He usually looks at DR and noise. I am pretty sure that the P645Z excels in DR. Regarding colour rendition I guess it is OK. What I think is that much effort goes into profile development at Phase One and Hasselblad, giving those cameras some subtle benefits in skin tones. I guess that the Pentax can produce 'accurate' color, that is accurate in the measurable sense. DR is probably just fine. It is a Sony sensor and they used to be excellent.

So, to sum it up:

Lloyd likes the camera, but he is skeptical about the value. The camera itself is great, but the lenses are OK but not as good as the latest generation of DSLR lenses from Zeiss and Sigma. Canon and Nikon can make nifty lenses, too. The Canon T&S lenses are quite workable with medium format backs, for instance.

Best regards
Erik

Hoping for discussion of the Pentax 645z, and a bit disappointed to see that this thread has wandered off to other subjects.

[gdanmitchell]

Thanks.

[sim4nee]

I had an opportunity a few days ago to witness my photographer friend  testing both Pentax 645Z and Leica S. Pentax with the tele 150mm and Leicas S with Macro 120mm f2.5, ha gave me couple of raw file to do the comparison and this is the result. Both shot at same EV, WB. The skin tone in Pentax looks flat and couldn't render red very well, not sure if this has anything to do with being CMOS sensor.

[Bart_van_der_Wolf]

I had an opportunity a few days ago to witness my photographer friend  testing both Pentax 645Z and Leica S. Pentax with the tele 150mm and Leicas S with Macro 120mm f2.5, ha gave me couple of raw file to do the comparison and this is the result. Both shot at same EV, WB.

Did that same EV setting result in equal exposure levels of Raw data? IOW, are both cameras equally sensitive, e.g. when set to ISO 100? The White Balance setting upon shooting is not relevant for Raw's, the WB as determined by the Raw converter is.

The skin tone in Pentax looks flat and couldn't render red very well, not sure if this has anything to do with being CMOS sensor.

No, it has nothing to do with CMOS. The only thing that matters are the characteristics of the Bayer CFA filters used. More relevantly, which camera profile was used for the Leica S and for the Pentax 645Z?

It's known that C1 does not always handle DNGs all that well as other native Raw formats, so that may say more about that than about the camera ... Did you also compare in e.g. Lightroom?

Cheers,
Bart

[sim4nee]

Did that same EV setting result in equal exposure levels of Raw data? IOW, are both cameras equally sensitive, e.g. when set to ISO 100? The White Balance setting upon shooting is not relevant for Raw's, the WB as determined by the Raw converter is.

No, it has nothing to do with CMOS. The only thing that matters are the characteristics of the Bayer CFA filters used. More relevantly, which camera profile was used in C1 for the Leica S and for the Pentax 645Z?

It's know that C1 does not always handle DNGs all that well as other native Raw formats, so that may say more about that than about the camera ... Did you also compare in e.g. Lightroom?

Cheers,
Bart

Yes both were set equally, WB were set is LR, temp-5000 and tint-0, and camera/lens profile disabled.

[ErikKaffehr]

Hi,

Lloyd wrote about Lightroom having two profiles for the Pentax 645Z. There is the Adobe Standard profile and an embedded profile coming from Pentax.

They are quite different.

Best regards
Erik

Did that same EV setting result in equal exposure levels of Raw data? IOW, are both cameras equally sensitive, e.g. when set to ISO 100? The White Balance setting upon shooting is not relevant for Raw's, the WB as determined by the Raw converter is.

No, it has nothing to do with CMOS. The only thing that matters are the characteristics of the Bayer CFA filters used. More relevantly, which camera profile was used in C1 for the Leica S and for the Pentax 645Z?

It's known that C1 does not always handle DNGs all that well as other native Raw formats, so that may say more about that than about the camera ... Did you also compare in e.g. Lightroom?

Cheers,
Bart

[eronald]

I had an opportunity a few days ago to witness my photographer friend  testing both Pentax 645Z and Leica S. Pentax with the tele 150mm and Leicas S with Macro 120mm f2.5, ha gave me couple of raw file to do the comparison and this is the result. Both shot at same EV, WB. The skin tone in Pentax looks flat and couldn't render red very well, not sure if this has anything to do with being CMOS sensor.

Hello,

 This is interesting. Could you post the dng files?
 If you have space issues, you can dropbox or send them to yourself with yousendit.com and post the links.

Edmund

[eronald]

Bart,

 how do you *know* that? yes, the CFAs define much of camera color, but texture is a different story. From the out of sensor internal Raws (not camera raws) I've seen from sensors similar to Blackmagic's, a huge amount of filtering sometimes needs to be done to solve pattern noise issues and make a CMOS image usable. Intensive filtering will destroy the texture, you won't get it back.

Edmund

No, it has nothing to do with CMOS. The only thing that matters are the characteristics of the Bayer CFA filters used.

Cheers,
Bart

[Bart_van_der_Wolf]

Lloyd wrote about Lightroom having two profiles for the Pentax 645Z. There is the Adobe Standard profile and an embedded profile coming from Pentax.

They are quite different.

Hi Erik,

Then that's an issue. It might be interesting to try in RawTherapee if that supports the 645Z DNGs or PEFs, e,g, with the IQ250 profile from Capture One. Alternatively one could create a DNG profile from a colorchecker for use in ACR/LR.

Cheers,
Bart

[eronald]

Hi Erik,

Then that's an issue. It might be interesting to try in RawTherapee if that supports the 645Z DNGs or PEFs, e,g, with the IQ250 profile from Capture One. Alternatively one could create a DNG profile from a colorchecker for use in ACR/LR.

Cheers,
Bart

You can also use Adobe's DNG color editor to play with the primaries and create a matrix that looks good on the fly. Portrait use is not about accurate color, it's all about pretty color.

It's clear that in those pics something is very different. But that doesn't mean it's the camera's fault, nor does it mean it's just the profile. For all I know the lighting may have changed.
 

Edmund

[Bart_van_der_Wolf]

You can also use Adobe's DNG color editor to play with the primaries and create a matrix that looks good on the fly. Portrait use is not about accurate color, it's all about pretty color.

I agree.

Cheers,
Bart

[Jim Kasson]

The best one is probably a mosaic in front of some university close by.

Erik, that's Memorial Church at Leland Stanford Junior University. There is no Leland Stanford Senior University.

Jim