[pschlute]

On the other hand, Pentax allows up to a couple of degree XY plane "rotate" for the purpose of leveling. This could also be an addition via the software. There is no way to really know at the present time.

I don't think this is an addition.

The Pentax K1 has always had up to 24 adjustments in left/right/up/down plus up to 8 steps of rotate (leveling). Have a look at page 71 of the K1 manual. it is almost identical to what is being described in the linked article here.

[JPT]

In my experience, if you're using Google Translate and something looks strange, it probably is. Google often translates the Japanese word for 10,000 as "million", which is where the wrong information about the metering sensor is coming from.

The FAQ only went up today in Japanese. I don't know how long the English translation will take, but when it's ready, it should appear on this page.
FAQ & Support | RICOH IMAGING

[angerdan]

The translation of the whole sentence is wrong, so the relevant part only (8.6万画素) says "86,000 pixels".

[Quicksand]

... I'm taking it somewhat half heartily to mean "tilt" - given the 5-axis image stabilization system, does have a small amount of out of XY plane movement into the XYZ plane with the sensor. Thus, if the 5-axis image stabilization allows a slight amount of tilt, changing the main circuit board with updated software which may allow the user to specify a small amount of out of plane - "tilt" would be new.
...

The 5-axis system that Olympus implemented allows yaw and pitch - so providing "tilt" is not out of the question. It's all a matter of what Pentax actually allows in the mkII, and how they are currently describing their intent.

• A close-up of Olympus’ 5-axis image stabilization system

It would be nice if they did provide some margin of "tilt".

Nah, a "5 axis" SR system still has only three axes of movement at the imaging sensor level -- up-down shift, left-right shift, and in-plane rotation. The other two axes are employed in the accelerometer and processing algorithm only -- up-down shift vs. up-down tilt (pitch), left-right shift vs. left-right tilt (yaw), and rotation (roll).

Having these five axes at the accelerometer/processor stage improves SR performance because required sensor shift for a shift-like movement of the camera is dependent on the distance to the subject, whereas required sensor shift for a tilt movement of the camera is not. Using both together will work better for macro and close-up shots in particular, where blur is particularly influenced by shifting movements.

Having the imaging sensor move out of its plane in the Z axis won't improve SR performance at all. It will just mess with focus.

As you look at that Olympus video linked in your post, carefully observe the reflections from the sensor. The reflections don't move at all, so the sensor isn't tilting.

[Canid]

I always thought they should include tilt in their composition adjustment feature.... I think I even suggested it when they distributed that big survey a while back.

That is tremendously exciting, if the translation is correct. (I want it for macro)

[interested_observer]

I don't think this is an addition.

The Pentax K1 has always had up to 24 adjustments in left/right/up/down plus up to 8 steps of rotate (leveling). Have a look at page 71 of the K1 manual. it is almost identical to what is being described in the linked article here.

Yes, I have used it before and it works well. If indeed they are adding "tilt" - that would be a new addition.

[Class A]

I don't think you can use flash with Pixel shift.

I've done it once as an experiment and it worked (surprisingly).

As you say, there is the issue of recycle rate, but you can either use low flash power, multiple speedlights (each at lowered power, but overall adding up to useful levels), or a dark environment and long shutter speeds.

In my experiment, I took the latter approach and fired the speedlight manually whenever the camera proceeded to taking the next image. As a matter of fact, I'm pretty sure that there will always be a limit on the shutter speed as I don't think the camera issues a flash signal for each PS image taken, so the shutter speed will have to be slow enough for you to reliably get the timing right. EDIT: It may be possible to use four speedlights, each equipped with a manually set delay so that it fires at the right time in the sequence. The Cactus RF60X and/or the Cactus V6/V6II support such delays.

I expected issues with slight brightness variations due to the speedlight not being able to perfectly output the same amount of light each time, but most of the test images looked fine.

[texandrews]

I've done it once as an experiment and it worked (surprisingly).

As you say, there is the issue of recycle rate, but you can either use low flash power, multiple speedlights (each at lowered power, but overall adding up to useful levels), or a dark environment and long shutter speeds.

In my experiment, I took the latter approach and fired the speedlight manually whenever the camera proceeded to taking the next image. As a matter of fact, I'm pretty sure that there will always be a limit on the shutter speed as I don't think the camera issues a flash signal for each PS image taken, so the shutter speed will have to be slow enough for you to reliably get the timing right. EDIT: It may be possible to use four speedlights, each equipped with a manually set delay so that it fires at the right time in the sequence. The Cactus RF60X and/or the Cactus V6/V6II support such delays.

I expected issues with slight brightness variations due to the speedlight not being able to perfectly output the same amount of light each time, but most of the test images looked fine.

Super helpful! Great news for me.

[jbinpg]

The reason I asked is because I thought that the Real Resolution mode did not use electronic shutter which would mean it could send a flash trigger signal at each of the four exposures.

[monochrome]

The reason I asked is because I thought that the Real Resolution mode did not use electronic shutter which would mean it could send a flash trigger signal at each of the four exposures.

Wouldn’t your flash have to recharge in 0.25 seconds?

[jbinpg]

Yes, but if you are shooting up close ar 1/128 power, that is doable.

[Rondec]

Wouldn’t your flash have to recharge in 0.25 seconds?

People have used flash with pixel shift already. The big thing is to have a long enough exposure to allow for the flash to recharge in the interim. Pinhole Camera is the one who has experimented most with it. Pixel Shifted Images - Page 22 - PentaxForums.com

I would think the same would be true for this other version of pixel shift.

[Class A]

Wouldn’t your flash have to recharge in 0.25 seconds?

You can throw equipment at this problem.

Either

1. reduce the power on the speedlight so that it reaches a sufficiently quick recycling rate and compensate by using multiple speedlights (for an 1/8 power setting, you'd need eight speedlights to get back to full power (with full power perhaps not being required), or
2. use "group cycling" with a Cactus V6(II) and four speedlights. After one shot, the group cycling feature will move to the next speedlight in line, essentially allowing for an effective zero recycling rate at any power level for a series of four shots (as there are four groups).

[GMounk]

Wouldn’t your flash have to recharge in 0.25 seconds?

I am seriously thinking about using LED lights. The subject shouldn't be moving when using pixel shift anyways. I am sure at some point in time LED lights will sync with cameras if they don't already.

[Class A]

I am sure at some point in time LED lights will sync with cameras if they don't already.

Some of them do already. They output more power momentarily, very similar to a flash.

The issue with LED lights are power and colour accuracy. Most lights are rather weak in comparison to speedlights and even when they have good CRI figures that won't guarantee good colour reproduction. Standard CRI figures are not very meaningful and one needs to look closer to check whether the light will be of useful quality for one's standards.