[robjmitchell]

Noise is additional random data of limited amplitude that sits on top of the image data. The current practice of blending/averaging noise both destroys detail by spreading data between pixels while loosing dynamic range as the black floor is raised and a lower white level is clipped. However there is no reason you cannot recognise the additional noise data and subtract it back to the original curve at the pixel level, ideally prior to amplificaton. The KP seems to be doing this.

[phoebus]

@reed yes sdm/plm means the soon 50mm and 85mm

I thought PLM was a linear 'motor' and SDM a ring motor???

The DFA* 50 and DFA* 85 are purported to have SDM ring motors....

The only Pentax lens currently specified as a PLM is the HD DA 55-300mm F4.5-6.3 ED PLM WR RE

If you follow that link, you can see that it is actually a linear actuator driven by a rotary stepper motor. It seems that these are only suitable for light-weight focussing groups.

I don't know, but strongly suspect that the DFA * primes do not have light-weight focussing groups. And the name of the 50mm certainly includes the term SDM.

[swanlefitte]

I thought the new 50 is kaf4 so i am thinking it is the fast focus just with a larger motor. That may not be correct.

[Mistral75]

I thought the new 50 is kaf4 so i am thinking it is the fast focus just with a larger motor. That may not be correct.

KAF4 means that the aperture is electromagnetically controlled and commanded instead of mechanically. This has nothing to do with the type of autofocus motor.

[swanlefitte]

Not the motor but perhaps the algorithms.

[normhead]

I couldn't have any it other way!

Do raw file have more vitamins than cooked ones?

[reh321]

I thought PLM was a linear 'motor' and SDM a ring motor???

The DFA* 50 and DFA* 85 are purported to have SDM ring motors....

The only Pentax lens currently specified as a PLM is the HD DA 55-300mm F4.5-6.3 ED PLM WR RE

If you follow that link, you can see that it is actually a linear actuator driven by a rotary stepper motor. It seems that these are only suitable for light-weight focussing groups.

I don't know, but strongly suspect that the DFA * primes do not have light-weight focussing groups. And the name of the 50mm certainly includes the term SDM.

I would expect 50mm prime and 85mm prime to each have a smaller focusing group than the 55-300mm. I would expect the lower mass of a light-weight focusing group to have a natural advantage in quick focusing. I wonder if lenses could be redesigned to make the focusing group lighter and the focusing throw shorter.

[RockvilleBob]

Do raw file have more vitamins than cooked ones?

Only when cooked in the Old World Schwenker style

[swanlefitte]

My thinking is plm was a proof of concept for a new system. Why else another 55-300 unless it was a good candidate to test on? I feel it fell short and that was a dead end or it proved fruitfill but showed new problems to be overcome, like weight. I think the new kaf4 stm is the solution that was needed.

---------- Post added 02-17-18 at 10:01 PM ----------

While complerely changing your future lens-camera interface why would you invest in a new completely reworked camera that may not work with this future interface?

[Reed]

PLM drive will not work with lenses like 50* or 85* because they have a large and heavy focusing groups. They will be SDM with new ring type motors, probably very similar to that used in new SP series primes by Tamron. In zoom lens with variable aperture like 55-300WR RE PLM there is only one single lens moved with very small throw, PLM can give great results. I hope they will incorporate that into DA limiteds and consumer grade primes, line refresh or new models- still it would require them to redesign all fo them. Those lenses are focusingh with whole optical assembly, not single element movement.

[Rondec]

Noise is additional random data of limited amplitude that sits on top of the image data. The current practice of blending/averaging noise both destroys detail by spreading data between pixels while loosing dynamic range as the black floor is raised and a lower white level is clipped. However there is no reason you cannot recognise the additional noise data and subtract it back to the original curve at the pixel level, ideally prior to amplificaton. The KP seems to be doing this.

I have no idea what is actually going on. All I can say for sure is that compared to my (fairly meager) noise reduction skills, using Noiseware or DFine 2, this is technique seems to give excellent results with good retention of detail.

[thibs]

My thinking is plm was a proof of concept for a new system. Why else another 55-300 unless it was a good candidate to test on? I feel it fell short and that was a dead end or it proved fruitfill but showed new problems to be overcome, like weight. I think the new kaf4 stm is the solution that was needed.

---------- Post added 02-17-18 at 10:01 PM ----------

While complerely changing your future lens-camera interface why would you invest in a new completely reworked camera that may not work with this future interface?

Yep, PLM is new but has nothing to do with ring motors used in DFA* lenses coming. This has, I believe, been stated clearly.
The delay to the DFA*50 introduction is due to this new (to Pentax) AF motor. Acain, stated by RI.
If it was PLM, there'd be no need of delay related to AF motors.

[beholder3]

Pentax "PLM" can be thought of as Canon "STM". These are stepper motors designed for precision and some speed, but not for maximum torque to move heavy glass elements.

Look at the Pentax PLM image:
HD PENTAX-DA 55-300mmF4.5-6.3ED PLM WR RE / Telephoto Lenses / K-mount Lenses / Lenses / Products | RICOH IMAGING

And you'll see a certain similarity to the second animation from Canon here:
Lead-screw type STM
Canon Focusing Technology - Canon Europe

STM lenses shoot great photos and even better videos. They’re named after an element of technology called a Stepping Motor, which lets the lens focus smoothly and quietly – two great characteristics to have when capturing video.
Whilst some motors used in lenses make very audible and distracting mechanical noises when they focus, STM lenses focus quietly, allowing you to record more of what you naturally hear in an environment. Canon's STM technology provides a range of extremely quiet lenses that are still fast enough to capture photos in most scenarios.
Lead-screw type STM
The physically larger range of STM lenses features the ‘lead-screw type’ STM – which when compared to the ‘gear type’ offers even greater levels of silence and speed.

[xandos]

Noise is additional random data of limited amplitude that sits on top of the image data. The current practice of blending/averaging noise both destroys detail by spreading data between pixels while loosing dynamic range as the black floor is raised and a lower white level is clipped. However there is no reason you cannot recognise the additional noise data and subtract it back to the original curve at the pixel level, ideally prior to amplificaton. The KP seems to be doing this.

There is though. When reading out the signal directly from the sensor, the electric signals do not come carrying tags 'I am noise' and 'I am signal'. While some of the noise might be easily recognizable as noise, a lot of it will not be. Many of the less sophisticated types of noise reduction use thing like convolutions with kernels (aka moving averages aka low-pass filtering) or techniques such as combinations of dilating/eroding, but these are techniques that do not necessarily retain all of the image details, because they attack the signal and the noise (using the fact that noise often has no or very weak correlations across the image, but so does anything with high local contrast). There are cases in which noise can be identified because it is static across mutliple pictures. Dark frame subtraction works for that. But any electronics (and any optical system) has to deal with many different noise types, amongst which there are types (such as shot noise) that cannot reliably be separated from the original image signal. The more sensitive our cameras get, the it is to have good SNR (signal-to-noise ratio). Pentax is usually very good at addressing this, which is one of my main reasons to like pentax (the other being small lenses and the trickery they like to perform with the shake reduction system: astrotracer, AA-simulator, pixel shift).

TLDR: It is never possible to get rid of all the noise in a signal without losing some of the signal, although for some special cases it is possible to get rid of some special types of noise.

PS: a good non-destructive way to reduce noise is by taking multiple images and averaging them. The noise should go down with the square root of the number of pictures taken (a factor of 2 for averaging 4 images, a factor of 4 with 16 images, etc). That is of course difficult (ie, impossible) to do on moving subjects.

[AyeYo]

There is though. When reading out the signal directly from the sensor, the electric signals do not come carrying tags 'I am noise' and 'I am signal'.

Bingo.

Lots of comments in here have me wondering how many people actually understand how the equipment works.