[falconeye]

According to Nikon, CIPA has released a new standard for specification of shake reduction. It is effective 2013 July 1. All Japaneses vendors adhere to the standard, i.e., must quote the corresponding numbers in spec sheets. Quote from Nikon:

From the 1st July, 2013 CIPA introduce a standard for Image Stabilization preformance for Digital Cameras. Nikon comply to this standard together with other manufacturers. The CIPA standard allows users to compare product VR performance in a way that is easy to understand.

Here is a link to the standard: CIPA DC-011-2012 Measurement and Description Method for Image Stabilization Performance of Digital Cameras: Home

Here is a link to the new numbers as specified by Nikon: https://nikoneurope-en.custhelp.com/app/answers/detail/a_id/57751
(Rather than "4 stops", it now reads anywhere between 2 stops and 4.5 stops)

Question: Where is this data from Pentax?

Addendum: I understand that Pentax recently started to advertize more modest efficiencies ("up to 3 stops" rather rhan the usual 4). However, this is not allowed by the new CIPA standard. The SR efficiency must be communicated in a more precise manner and has been specified as e.g.:

Description Example 2
Image stabilization performance:
2.0 stops (focal length f = 55 mm, f = 85 mm on a 35 mm equivalent, AB lens is used)
2.5 stops (focal length f = 200 mm, f = 300 mm on a 35 mm equivalent, CD lens is used)

Other examples given are similiar. So, let's inquire the exact specs from Pentax, they are missing!

---

Last edited by falconeye; 07-01-2013 at 06:30 PM.

[carrrlangas]

great info. Interesting that all Nikon´s DX lenses are below 3.5stops vs 4 or 4.5 for some of the FX lenses

[kenafein]

I would like to see the specs on the famous Olympus 5 axis.

[maxfield_photo]

I'm not sure 'stops' is the only measure of a shake reduction system, the magnitude of the vibrations, and the focal length should also be taken into account. For me they should give the rating in cups of coffee.

[Winder]

My personal, subjective experience is that the Olympus IBIS (E-3) seemed to be better than my K-7 and my K-5. Pentax IBIS seems better in the K-5 than the K-7, but it is still lacking. The A900 was probably on par with the K-5. I have not used any newer Sony or Olympus IBIS cameras.

[falconeye]

I'm not sure 'stops' is the only measure of a shake reduction system, the magnitude of the vibrations, and the focal length should also be taken into account. For me they should give the rating in cups of coffee.

You can read the CIPA standard I linked above, it is pretty well thought out.
And as I said, for body-based SR, the lens and focal length used must be specified when specifying the 'stops'. Or a range for a multitude of lenses and focals. The magnitude of vibrations is part of the human tremble model embedded into the standard and its vibration machine. Fore more details, please read http://www.cipa.jp/image-stabilization/documents_e/VMVA_for_DC-X011_E.pdf where two 32s angular shake test patterns at 500Hz sampling frequency are given. The pattern is load-dependent where 400g and 600g are thresholds for light and heavy.

I have seen a few independent tests of SR efficiency, did one myself (cf. 2.5.3 in http://www.falklumo.com/lumolabs/articles/sharpness/index.html finding an average 1.7 stops for K-7 s). I'd say, Pentax scores between 1 and 3 stops in most cases. But unlike optical VR with e.g. Nikon, it isn't normally necessary to turn off VR for best image quality from the lens. I have seen a great test for Olympus E-M5, exceeding 4 stops. And I have seen a catastrophic test for Pentax Q, basically saying that it's SR is doing nothing.

[Daedbird]

This standard is for optical system image stabilization, and not sensor-based, or electronic based stabilization. The standard for measuring says so.

[falconeye]

This standard is for optical system image stabilization, and not sensor-based, or electronic based stabilization. The standard for measuring says so.

No.

Sensor shift is an optical rather electronic (read digital) method. It shifts the lens optical axis wrt the sensor.

Consequently, the standard explicitely lists 'camera body with image stabilization mechanism' and 'interchangeable lens with image stabilization mechanism' (section 5.2).

A bit more care before commenting, please.

[crewl1]

In practice it seems the lens stabilization of my OS Bigma gives better results than the in-body SR of my K-5.
The lens OS drains the battery much faster, though

[Class A]

Here is a link to the standard: CIPA DC-011-2012 Measurement and Description Method for Image Stabilization Performance of Digital Cameras: Home

I find it strange that the certified vibratory apparatuses do not support roll "shake". They are "2-axes" devices only.

Even though yaw and pitch shake will dominate image blur, one could consider the omission of roll shake to be catering to lens-based stabilisation systems.

Of course, Olympus could claim that the impact of translational shake should be measured as well.

[Fogel70]

There is some thing I don't like about the CIPA test. Blur is determined by checking A5 prints with 65-80 cm viewing distance and the stops calculated is using 1/focal length x "crop factor" shutter speed.

Using that small print with that long viewing distance will probably mask a lot of camera shake, and many experienced photographers can probably hand hold camera without image stabilization 1.5 - 2 stops below 1/focal length x "crop factor" shutter speed and not detecting any blur at that test scenario.

IMO A3 prints should have been use at that viewing distance, not A5.

[falconeye]

There is some thing I don't like about the CIPA test. Blur is determined by checking A5 prints with 65-80 cm viewing distance and the stops calculated is using 1/focal length x "crop factor" shutter speed.

Using that small print with that long viewing distance will probably mask a lot of camera shake, and many experienced photographers can probably hand hold camera without image stabilization 1.5 - 2 stops below 1/focal length x "crop factor" shutter speed and not detecting any blur at that test scenario.

IMO A3 prints should have been use at that viewing distance, not A5.

Again, I believe this statement is due to misunderstanding (I for myself only glimpsed through it, but anyway...).
The blur is determined very much as described in my paper about "Understanding Image Sharpness" (cf. my blog or hp). They even "use" my formulas about how to combine static and motion blur. It is determined by digital image analysis and measured in microns. Nothing gets printed. My own determination of SR efficiency was very much like the CIPA procedure, replacing the vibration apparatus by human tremble

Using 1/equivalentFocal makes sense, it is the established rule of thumbs. At 10 mrad/s tremble (again, cf. my paper), this creates about 10/1000 focal/equivalentFocal mm or 10 microns/cropFactor blur. Or about 1/3 the max. depth of field blur. So, the rule of thumb is a compromise between visible in print from a distance and visible in pixel peeping.

A measurement of rotational blur in the standard would have been difficult. It depends on the location in the image and the current standard allows to determine blur at any location in the image.

[Pål Jensen]

This is only a good idea if the standard is representative for real life usage.
Otherwise we may end up with systems designed for performing well in standardized tests as opposed to real life (which might be very different).

[froeschle]

Where is this data from Pentax?

In the Japanese FAQ I recently found:

Q: Image stabilization effect What stage are you?

A: Is 3.5 rounds.
※ CIPA compliant (f = 135mm (35 mm format equivalent f = 207mm), smc PENTAX-DA18-135mmF3.5-5.6ED AL [IF] DC WR is used)

K-3 - FAQ

[Lowell Goudge]

I find it strange that the certified vibratory apparatuses do not support roll "shake". They are "2-axes" devices only.

Even though yaw and pitch shake will dominate image blur, one could consider the omission of roll shake to be catering to lens-based stabilisation systems.

Of course, Olympus could claim that the impact of translational shake should be measured as well.

Probably because the standards are developed in conjunction with manufacturers. This means usually that they test the minimum of all systems capability, therefore eliminating roll shake, because it is optically impossible. That is only for sensor based systems.

Since I have personally achieved 4.5 stops with the K7 and a 500mm lens, (1/40th of a second) I don't bother with what gets reported, what is important to me is I know what my technique can do.

There is some thing I don't like about the CIPA test. Blur is determined by checking A5 prints with 65-80 cm viewing distance and the stops calculated is using 1/focal length x "crop factor" shutter speed.

Using that small print with that long viewing distance will probably mask a lot of camera shake, and many experienced photographers can probably hand hold camera without image stabilization 1.5 - 2 stops below 1/focal length x "crop factor" shutter speed and not detecting any blur at that test scenario.

IMO A3 prints should have been use at that viewing distance, not A5.

actually an A 5 print flies in the face of traditional wisdom, where "acceptably sharp" has always been defined as 1/100 of an inch when enlarged to an 8" x 10 print.

That old standard definition is what applies to depth of field, shutter speed for hand held lenses and motion blurr etc. etc.