Negative things about the X2D II so far

[Jared Willson]

Hi,

as much as I am excited by the new X2D II camera and I'd like to upgrade, I think it is good to also mention any shortcomings the new model is coming with.

After watching all video reviews here is my list:

• With older lenses AF-C is not supported, though AF is generally better
• 45p incompatible (one of my favorite lens...)
• Joystick moves the focus point only vertically or horizontally (like in the previous version) but not diagonally. What a bummer
• AF-C good but not great, apparently because of the shutter lag it does not cope with rather fast moving objects
• Battery duration (some 10%) worse
• Slightly worse ISO performance, apparently at high ISO are the files noisier (Carl Taylor's review)

Please correct if wrong, or add yours minuses to the list.

Thanks.

One thing mentioned was sort of a gut punch for me is the cost ($5000+) of the new 35-100 lens to go with the X2DII to have full advantage of continuous AF.

A question is when shutter-lag is mentioned, is this going to affect the acquired focus when subject is moving?

-especially the ability to aquire focus in dark environments, it is going to be a success for Hasselblade. [ if money is no object . . .] The simplicity of the UI is a plus. Hats off to Hasselblade,

Shutter lag has never been a strength for the X series cameras--they just aren't meant for action. I did check mine just now. I no longer have the X2D mk i to compare it to, so I can only give you an idea of where the X2D ii lies.

Here is how the shutter works on the X2D ii, at least with current generation XCD lenses. The camera starts with the shutter open--effectively an electronic first curtain shutter. When you trip the shutter, the camera first stops down the iris to the desired focal ratio. Once that is complete, it closes the shutter. Once that is complete it waits for a little while, presumably as the read process is completed. Once the read is complete, it re-opens the shutter to end the blackout period and start providing fresh data to the viewfinder.

I tried four scenarios:

• lens wide open so the diaphragm would not affect results, 16-bit capture raw
• Lens stopped down to minimum aperture, in case waiting for auto stop down increases lag measurably, 16-bit capture raw
• Lens wide open, 14-bit raw
• Lens stopped down, 14-bit raw

The process was rudimentary, but reasonably consistent. I was getting results within 10% from one run to the next. All I did was set the camera to the fastest shutter speed, manual focus, RAW mode, and using the "voice memo" app on my phone, I recorded the sounds as pressed on the shutter release , and then looked at an audio recording of shutter actuation. I could see clear peaks where the button was pressed, where the shutter (and lens diaphragm when the lens was stopped down) actuated, and when the shutter re-opened. The iPhone voice memo app tells timestamps to the nearest 1/100s. I chose to measure from the middle of the "click" when I actuated the shutter to the end of the "click" when the shutter had clearly finished closing.

Scenario 1:

• 0.16s from the "click" of my finger on the shutter release to the shutter completely closing
• 0.24s from the shutter closing to the point where it starts to re-open, presumably the sensor readout time

Scenario 2:

• 0.23s from the "click" of my finger to the sound of both the diaphragm and shutter completing--quite a bit longer
• 0.24s from the shutter closing to the point where iris and shutter start to re-open

Scenario 3:

• 0.16s from the "click" of my finger on the shutter release to the shutter completely closed
• 0.05s from the shutter completing its close to starting to re-open

Scenario 4:

• 0.24s from the "click" of my finger to the sound of both the diaphragm and shutter completing
• 0.05s from the shutter completing its close to starting to re-open

This was with the XCD 55mm lens. Obviously, timings might be a bit different with different lenses. Hasselblad uses different shutters in different lenses, and some may be quicker than others. I ran the experiment multiple times to ensure consistency in my measurements and to ensure the camera didn't vary too much from frame to frame. I was consistently within about 10% of the above listed measurements. I didn't bother to enter all results into a spreadsheet or compute standard deviation since I was only trying to get a rough understanding of how the camera behaved, not to provide exact measurements. The numbers above should only be considered rough guidelines, but they do give an idea of how the camera behaves. To summarize:

There is significant shutter lag with the X2D, between roughly 0.16s wide open and roughly 0.25s if the lens is stopped down--extra time is needed for the aperture diaphragm to close. The readout time for the X2Dii is still in the 250ms range when shooting in 16 bit mode. One can dramatically reduce readout time, EVF blackout, and rolling shutter when using electronic shutter by using 14 bit mode rather than 16. It's possible that there will be a hit in dynamic range in 14 bit mode.

So, it's definitely not a sports camera. The lag is significant enough that, coupled with EVF lag, one would easily miss "decisive moment" type shots, and fast moving subjects may no longer even be in focus once the shutter actually trips. Don't plan on using the X2Dii for fast moving action. The reduced readout time in 14 bit mode was a pleasant surprise. I was expecting some improvement, but no a reduction from 250ms to 50ms. I did confirm that rolling shutter is much less when shooting in 14 bit mode, so I'm pretty confident that the time between the shutter closing and starting to re-open is a rough indicator of the readout time.

I'm not sure any of this will change how I use the camera, but it might. For now, I will still use 16 bit mode for landscape shots where shutter blackout is completely irrelevant. Since I rarely use adapted lenses, the rolling shutter improvements with 14 bit mode are not relevant to me. But, for things like family pics? Travel photography on a city street? Things where a bit more responsiveness is needed? Definitely 14 bit mode. And if I were using adapted lenses? Even on a tripod I think I would choose 14 bit mode since moving leaves and branches can cause rolling shutter artifacts on fairly static subjects--a quarter second readout is a long time.

Now to evaluate dynamic range in both 14 bit and 16 bit mode... So I can see whether there is any reason at all to choose 16 bits, even for landscapes...

I found a mistake in my measurements... The sound I was interpreting as the shutter completing its close was actually the sound of my finger releasing the shutter button--that makes a click as well. So, I was over-estimating the shutter lag. Revised measurements...

• 0.06s +/- 0.01s for shutter lag in both 14 bit and 16 bit mode with the aperture wide open--the time from the finger "click" to the end of the sound of the shutter closing, assuming the 1/2000s shutter duration is negligible.
• Read time is about 330ms (from the shutter completely closing to starting to re-open) in 16 bit mode, and about 170ms in 14 bit mode, so not as big an improvement as I reported in earlier.

Since I messed up the first set of measurements, I'd encourage anyone else to do the same test and see if you get similar results to revised numbers. I have no idea whether these numbers are any different from the X2D mk i.

Assuming I didn't mess up a second time, I prefer this approach to Jim's method of photographing a timer, simply because it eliminates anticipation and reaction times affecting the results. This is a direct measurement of the sound of the camera, and it's easy enough to replicate.

 

[Jared Willson]

Yup, I saw Jim's measurements. My one concern is that photographing the clock incorporates reaction times into the "lag". Since that isn't reproducible, I prefer to just measure the clicks.

I am aware of the measurements that showed there was no difference in dynamic range between 14 bit and 16 bit in the Mk i version of the camera, but Hasselblad claims a touch more dynamic range in this version, so it's worth checking. It's easy enough to do with a combination of flat frames, bias frames, and dark frames. Some astronomy software I have--PixInsight--can easily check to see if there is a difference. the catch is avoiding the automatic profile correction, so it will take a little bit of work with the raw files.

 

[as1mov]

A couple more negatives to add to your list...

• As SrMi pointed out to me in another thread, for some reason the new matrix metering is required when HDR is enabled, and it is not available at all when HDR is disabled
• HDR appears not to be available in 'M' exposure mode. Can't think why that would be. Maybe because the camera can't be sure that highlights are protected? Obviously, raw files still contain the full dynamic range of the camera, so one can still process the images in HDR, but they won't show the extended brightness range on the back of the camera
• I'd still love to be able to see images in monochrome on the EVF. This isn't new to the H2D ii, but it's a flaw I would love to see addressed
• I still don't find the focus indicator feature--the partial circle that moves a marker like a clock to indicate how far you are from focus--to be all that accurate. It's too bad because I really like the idea of it, but in real use it is often off a bit. Just zooming in judging by eye seems to be much more consistent.
• Wish I could map the two dials to different functions--not just the "click" but the purpose when turning the dials.
• The AE-L and AFD buttons are not customizable. I virtually never use the AE-L, for example, so would love the ability to map it to something else. We can do that with the M and ISO buttons, so why not the two on the back?
• There is still a lag between when I bring the camera to my eye and when it turns on the EVF

All in all, fairly minor niggles

Thanks, I agree. A firmware update could address most of the issues you mentioned.

I understand you bought the camera? Can you please check the following for me?

• switch to electronic shutter. And take a few shots while slightly shaking the camera. Do you see the same kind of rolling shutter effect as in the previous version? Any better?

I measured the readout time using Jim's approach (link). The readout time is the same as in all 100MP sensor cameras (to be confirmed by Jim).

• mount any adapted , non Hasselblad lens on the camera. Does the joystick works as expected? I am asking this because in the old version I could move the focus point only after zooming in 100% in the center. In other words, with the previous camera it is not possible to position the focus point before zooming in to check the focus

Thanks a lotT

Thanks.

 

[as1mov]

A couple more negatives to add to your list...

• As SrMi pointed out to me in another thread, for some reason the new matrix metering is required when HDR is enabled, and it is not available at all when HDR is disabled
• HDR appears not to be available in 'M' exposure mode. Can't think why that would be. Maybe because the camera can't be sure that highlights are protected? Obviously, raw files still contain the full dynamic range of the camera, so one can still process the images in HDR, but they won't show the extended brightness range on the back of the camera
• I'd still love to be able to see images in monochrome on the EVF. This isn't new to the H2D ii, but it's a flaw I would love to see addressed
• I still don't find the focus indicator feature--the partial circle that moves a marker like a clock to indicate how far you are from focus--to be all that accurate. It's too bad because I really like the idea of it, but in real use it is often off a bit. Just zooming in judging by eye seems to be much more consistent.
• Wish I could map the two dials to different functions--not just the "click" but the purpose when turning the dials.
• The AE-L and AFD buttons are not customizable. I virtually never use the AE-L, for example, so would love the ability to map it to something else. We can do that with the M and ISO buttons, so why not the two on the back?
• There is still a lag between when I bring the camera to my eye and when it turns on the EVF

All in all, fairly minor niggles

Thanks, I agree. A firmware update could address most of the issues you mentioned.

I understand you bought the camera? Can you please check the following for me?

• switch to electronic shutter. And take a few shots while slightly shaking the camera. Do you see the same kind of rolling shutter effect as in the previous version? Any better?
• mount any adapted , non Hasselblad lens on the camera. Does the joystick works as expected? I am asking this because in the old version I could move the focus point only after zooming in 100% in the center. In other words, with the previous camera it is not possible to position the focus point before zooming in to check the focus

Thanks a lot

Rolling shutter is as bad as ever. Just took a 1/60s exposure of a book shelf while moving the camera horizontally. Books are reasonably sharp--as much as one would expect from a 1/60s when waving the camera around. The shelf is leaning at a 30 degree angle. Definitely no more usable than its predecessor with respect to rolling shutter.

Edit: rolling shutter in 14 bit mode is MUCH less of an issue than in 16 bit mode. I did a rough measurement just based on timing the clicks of the shutter closing (beginning of readout) to shutter re-opening (end of readout), and the time goes form about 250ms to readout in 16 bit mode down to 50ms to readout in 14 bit mode. This isn't quite enough to eliminate rolling shutter with adapted lenses, but it takes things from unusable with moving subjects or panning to actually pretty decent. I did confirm this by waving the camera around as I took a picture. The rolling shutter is much lower in 14 bit mode. No idea if this was also true on the X2D mk i or not--this may not be an improvement--but it definitely would be an attractive option if you use adapted lenses.

I mounted a Leica M lens using the Novoflex M-X adapter. The focus point selector disappeared from view, so no way to use the joystick. However, by double tapping on the part of the screen I was interested in, the viewfinder zoomed in to 100% at the location where I tapped, so I can compose and focus just fine. Image stabilization worked once I put in the focal length of the lens. Focus peaking worked, both zoomed in and when viewing the full field of view. So, if you can live with the rolling shutter, it would be usable with non-XCD lenses. You can focus where you want. You can't use the joystick to do it, though.

- Jared

Thanks for testing this for me. A bit of a bummer. The the only reason to upgrade for me would be the AF-C

Cheers

 

[Philip Eihuyar]

Shutter lag has never been a strength for the X series cameras--they just aren't meant for action. I did check mine just now. I no longer have the X2D mk i to compare it to, so I can only give you an idea of where the X2D ii lies.

Here is how the shutter works on the X2D ii, at least with current generation XCD lenses. The camera starts with the shutter open--effectively an electronic first curtain shutter. When you trip the shutter, the camera first stops down the iris to the desired focal ratio. Once that is complete, it closes the shutter. Once that is complete it waits for a little while, presumably as the read process is completed. Once the read is complete, it re-opens the shutter to end the blackout period and start providing fresh data to the viewfinder.

I tried four scenarios:

• lens wide open so the diaphragm would not affect results, 16-bit capture raw
• Lens stopped down to minimum aperture, in case waiting for auto stop down increases lag measurably, 16-bit capture raw
• Lens wide open, 14-bit raw
• Lens stopped down, 14-bit raw

Now to evaluate dynamic range in both 14 bit and 16 bit mode... So I can see whether there is any reason at all to choose 16 bits, even for landscapes...

I found a mistake in my measurements... The sound I was interpreting as the shutter completing its close was actually the sound of my finger releasing the shutter button--that makes a click as well. So, I was over-estimating the shutter lag. Revised measurements...

• 0.06s +/- 0.01s for shutter lag in both 14 bit and 16 bit mode with the aperture wide open--the time from the finger "click" to the end of the sound of the shutter closing, assuming the 1/2000s shutter duration is negligible.
• Read time is about 330ms (from the shutter completely closing to starting to re-open) in 16 bit mode, and about 170ms in 14 bit mode, so not as big an improvement as I reported in earlier.

Since I messed up the first set of measurements, I'd encourage anyone else to do the same test and see if you get similar results to revised numbers. I have no idea whether these numbers are any different from the X2D mk i.

Assuming I didn't mess up a second time, I prefer this approach to Jim's method of photographing a timer, simply because it eliminates anticipation and reaction times affecting the results. This is a direct measurement of the sound of the camera, and it's easy enough to replicate.

Another way to test for how great the effect of shutter-lag is to compare by measurement the speed of a moving subject. Thanks for all the detailed work on this.

 

[nurf]

• no phone support for Android
• no image transfer to an FTP server

 

[nurf]

The images also appear to be more detailed, the new camera just seems to have the sensor based noise reduction reduced by a bit.

 

[SrMi]

The images also appear to be more detailed, the new camera just seems to have the sensor based noise reduction reduced by a bit.

There is no in-camera noise reduction in X2D, and AFAIK, none on Mark 2 either.

 

[nurf]

Photos are processed by an image processor, noise reduction is part of that process. It's on a hardware level. Different image processors create different outputs from the same sensor.

 

[SrMi]

Photos are processed by an image processor, noise reduction is part of that process. It's on a hardware level. Different image processors create different outputs from the same sensor.

Some cameras apply in-camera NR, some don't, and some do it only at higher ISOs.

P2P measurements can detect whether NR was applied.

 

[nurf]

We got the same sensor with different processors.

One is less sharp and less noisy.

One slightly sharper and with a bit more noise.

Seems pretty clear.

 

[SrMi]

We got the same sensor with different processors.

One is less sharp and less noisy.

One slightly sharper and with a bit more noise.

Seems pretty clear.

What is the source of "less sharp and less noisy?"

 

[nurf]

The samples shown in the youtube video being discussed here.

 

[SrMi]

The samples shown in the youtube video being discussed here.

That is not a reliable source.

I tested it as thoroughly as I know how, and I do not see any difference in sharpness or high ISO noise.

There are a couple of traps that one can fall into when comparing the two models. I hope that I have avoided all of them.

 

[Knorp]

why-o-why ?
..............................................................................................................................
Bart

 

[JAKORO]

why-o-why ?
..............................................................................................................................
Bart

probably left over parts from X2D/existing mould

 

[SrMi]

why-o-why ?
..............................................................................................................................
Bart

I love the strap lugs. Easy and quick to attach and detach, and safe (with locking lugs available). The strap lug adapters are compatible with any standard strap and are easy to untangle, even when the camera is hanging on the strap.

What is the negative?

 

[Schrodingers cat]

is the negative?

if you like them, I would say there is none.

My purely personal preference is for a strap that doesn’t tangle to begin with and a strap lug that I perceive to be more secure whether in actuality it is or not.

just a matter of what you like.

--
Personal travel snapshots at https://www.castle-explorers.com
1. Making good decisions is generally the result of experience.
2. Experience is generally the result of making bad decisions.
3. Never underestimate your capability for doing incredibly stupid s**t.

 

[JAKORO]

why-o-why ?
..............................................................................................................................
Bart

I love the strap lugs. Easy and quick to attach and detach, and safe (with locking lugs available). The strap lug adapters are compatible with any standard strap and are easy to untangle, even when the camera is hanging on the strap.

What is the negative?

First off the lugs only rotate on one axis, like others have said it can create a mess sometimes.

Second, I recently saw a friend's camera, Obviously this is after a long period of use but modern strap/lugs have been designed to prevent this ...



Lastly, i've said it when x2d first came out, why try to fix something when its not broken.(I know i know the lugs are a classic design...)

 

[Schrodingers cat]

That is nasty looking.



--
Personal travel snapshots at https://www.castle-explorers.com
1. Making good decisions is generally the result of experience.
2. Experience is generally the result of making bad decisions.
3. Never underestimate your capability for doing incredibly stupid s**t.