[texandrews]

Just to throw this out there as "food for thought" to extend the metaphor closer to the breaking (baking? boiling?) point....

The sensor "collects" photons, and they become electrical signals per the design of the sensor. Then there is a processor to....process (uh, cook...) those electrical impulses. So, isn't the accelerator unit just another part of the processing? While in principal a raw that is raw is an ideal, in practice I would have no problem with an additional step that helped filter noise that is noise (and not detail...). Therefore for the moment, and based on the samples I've seen so far, I have no problem with this accelerator cooking, or maybe "pre-heating" ("air dry"? "sifting"?....lots of food metaphors, and the sun's not yet over the yard arm, so nothing to drink yet!) for my shooting.

I could see where this might give astrophotographers pause, but I am not one.

[monochrome]

Just to throw this out there as "food for thought" to extend the metaphor closer to the breaking (baking? boiling?) point....

The sensor "collects" photons, and they become electrical signals per the design of the sensor. Then there is a processor to....process (uh, cook...) those electrical impulses. So, isn't the accelerator unit just another part of the processing? While in principal a raw that is raw is an ideal, in practice I would have no problem with an additional step that helped filter noise that is noise (and not detail...). Therefore for the moment, and based on the samples I've seen so far, I have no problem with this accelerator cooking, or maybe "pre-heating" ("air dry"? "sifting"?....lots of food metaphors, and the sun's not yet over the yard arm, so nothing to drink yet!) for my shooting.

I could see where this might give astrophotographers pause, but I am not one.

Good reason to have both cameras then. Hhhmmmmm.

[credos4u]

Just to throw this out there as "food for thought" to extend the metaphor closer to the breaking (baking? boiling?) point....

The sensor "collects" photons, and they become electrical signals per the design of the sensor. Then there is a processor to....process (uh, cook...) those electrical impulses. So, isn't the accelerator unit just another part of the processing? While in principal a raw that is raw is an ideal, in practice I would have no problem with an additional step that helped filter noise that is noise (and not detail...). Therefore for the moment, and based on the samples I've seen so far, I have no problem with this accelerator cooking, or maybe "pre-heating" ("air dry"? "sifting"?....lots of food metaphors, and the sun's not yet over the yard arm, so nothing to drink yet!) for my shooting.

I could see where this might give astrophotographers pause, but I am not one.

From what i know, I guess that the mod of this thread could be right...

I once experimented around a bit with PHDK and the firmware,
and by standard ... there is something done to the image that is called "dark frame subtraction" (if you wanna know in detail, google is your friend)...
(one can switch this on and off in FW by gettin into debug mode and writing certain commands in a file that once was directed to be named like MODSET.XXX(X is for certain number) now you have to name it somewhat else, i guess...)
this so called "dark frame subtraction"(something that sony cameras, compared to the rest of their performance, perform rather poor at... thats the reason why they are called star-eaters...) reads out the state of the sensor and so it can subtract its standard noise... - i guess this is something that PENTAX engineers can do really good and they seem to have elobareted/polished that feature to its most possible effect- so their you have an explanation why the same sensor does so much better in a PENTAX camera. - (no fairy dust, sry.)
All that formerly was done by the old chipset, without an accalerator unit(i guess pre-raw). Now it has an accelerator processor especially designed to do this.... what would should really free up power from the rest of the cameras chipset...

What in the end is nice to have BUT! > is NO excuse for PENTAX-Ricoh Imaging to NOT upgrade the chipset as a whole(all of its electronic architecture) from ground up(despite minor upgrades) for about nearly exact 9 years by now.
(and generally "UP" the power of its processing units and memory...)
but rest assured, there is no need to talk about that now.. that would be like beatin a dead horse, because with that "K-1II Accelerator-move", they drew their last Ace out of their sleeves...

now they HAVE TO buy into new electronics to speed up the show, if they really want to sell any more (new) cameras.
talking for myself: I would NOT buy another K-1 Mark III after that upgrade with another nifty trick or some more additional wheels.(or a fancy led that tells me air humidity has risen, if i would be so dumb to not buy a WR lens)...

I wouldnt even buy it, if it could hover around in open space...
because you can rest assured, if they calculate that a "hovering feature" would be 15% cheaper than upgrading the electronics,
... ricoh would go and implement a hovering feature instead of upgrading the basic/core features(giving you bigger buffer, higher burst rate and an up-to-date AF-module...) of a modern camera...
That is not because the innovators at that company dont want to or wouldnt feel that they could sell a whole lot of cameras more if they would do the serious upgrades it would take to get to the top, but they are somewhat very conservative in their economical thinkin i guess.

Really i think the upgrade IS worth its 500 bucks...(actually only 200... but add to that the the financial and logistic effort it takes and you have your 500 bucks) but its up to you if you want to honor their upgrade politics.
(I understand everyone that doesnt think its worth it.)

BTW - why is it always so smelly in chinatown?

[richandfleur]

NO excuse for PENTAX-Ricoh Imaging to NOT upgrade the chipset as a whole[/U][/B](all of its electronic architecture) from ground up(despite minor upgrades) for about nearly exact 9 years by now.

Agree, via the firmware hacking thread, we know the chipset changed around 2012 with the move to the K-01/K-30 from the K-5 era processor. From memory this was around the Ricoh takeover time as well.
Video wise this was when mechanical stabilisation was dumped for the rubbish 'Movie SR' approach etc, and saw the introduction of focus peaking.

That core processor style hasn't changed since then, and there's a very good chance that stance accounts for the lack of real tracking AF and eye focus etc.

[reh321]

Do these processes affect RAW files, or is it for JPEG only?

We do not know. I was hoping on regular reporting from K-70 and KP owners who process 'raw', but that hasn't happened. From the pov of design engineers, firmware is a just replacement for no-longer needed resistors, capacitors, etc. As long as they don't change desired characteristics of 'raw', I see no reason for anyone to freak out.

[photoptimist]

Just to throw this out there as "food for thought" to extend the metaphor closer to the breaking (baking? boiling?) point....

The sensor "collects" photons, and they become electrical signals per the design of the sensor. Then there is a processor to....process (uh, cook...) those electrical impulses. So, isn't the accelerator unit just another part of the processing? While in principal a raw that is raw is an ideal, in practice I would have no problem with an additional step that helped filter noise that is noise (and not detail...). Therefore for the moment, and based on the samples I've seen so far, I have no problem with this accelerator cooking, or maybe "pre-heating" ("air dry"? "sifting"?....lots of food metaphors, and the sun's not yet over the yard arm, so nothing to drink yet!) for my shooting.

I could see where this might give astrophotographers pause, but I am not one.

Raw is not ideal. The photon collection process is inherently imperfect due to manufacturing imperfections in the sensor at the pixel level. Each pixel has a different black level, sensitivity, and color profile. And although some of those effects look random in the image, they aren't. Those can be corrected and are guaranteed to improve the quality of the image with absolutely no effect on resolution or artifacts. The output of these corrections would actually be much closer to the ideal image. (And it's a lot easier to correct these in-camera because they require camera-specific calibration files that are created in the factory.)

There's a second set of potential corrections that are a bit more contentious because they rely on analyzing and adjusting statistic properties. There are differences between the statistic properties of the noise and the structural numerical relationships and statistical properties of valid images. Certain patterns of neighboring R, G, and B values may be physically impossible in a noise-free image. Hence if those anomalies are in the data, it evidence of noise that can be subsequently removed. These methods can be contentious because they may have some slight effect on the "true" image. Optimized correctly, the gain in resolution from suppressing noise would more than offset any slight loss of resolution from "baking" the image.

The empirical data seems to show that Pentax RAWs may be baked to perfection rather than over-cooked. Those sample shots at ISO 12800 show improvements in both noise and resolution. The same is true of the DPreview images taken with the KP. So far, no owners of the K-70 or KP have found evidence of over-baking.

We shall see!

[richandfleur]

As long as they don't change desired characteristics of 'raw', I see no reason for anyone to freak out.

Reason I ask is Sony users are going through a situation whereby the noise reduction processes are 'killing stars' from their astro images. I'd like to know if the extra processing this new hardware chip offers will affect RAW images, because of that possible situation.

Not freaking out by any means, but am interested to know.

[texandrews]

Raw is not ideal. The photon collection process is inherently imperfect due to manufacturing imperfections in the sensor at the pixel level. Each pixel has a different black level, sensitivity, and color profile. And although some of those effects look random in the image, they aren't. Those can be corrected and are guaranteed to improve the quality of the image with absolutely no effect on resolution or artifacts. The output of these corrections would actually be much closer to the ideal image. (And it's a lot easier to correct these in-camera because they require camera-specific calibration files that are created in the factory.)

There's a second set of potential corrections that are a bit more contentious because they rely on analyzing and adjusting statistic properties. There are differences between the statistic properties of the noise and the structural numerical relationships and statistical properties of valid images. Certain patterns of neighboring R, G, and B values may be physically impossible in a noise-free image. Hence if those anomalies are in the data, it evidence of noise that can be subsequently removed. These methods can be contentious because they may have some slight effect on the "true" image. Optimized correctly, the gain in resolution from suppressing noise would more than offset any slight loss of resolution from "baking" the image.

The empirical data seems to show that Pentax RAWs may be baked to perfection rather than over-cooked. Those sample shots at ISO 12800 show improvements in both noise and resolution. The same is true of the DPreview images taken with the KP. So far, no owners of the K-70 or KP have found evidence of over-baking.

We shall see!

Great to hear, and thanks for the explanation. I enjoy reading your posts. And I usually learn something, always a good thing in my book. Nevertheless, although I had details wrong/missing, your post re-affirms something I've suspected for a while now.

[MJKoski]

Any KP (or K-70) owner can take multiminute exposure and find that small details looks like they are wrapped in very thin transparent plastic shell for the lack of better word. That ISO12800 comparison was interesting but shows heavy handed noise reduction when looking for crisp details. Sure, at first look it looks clean.

[jbinpg]

Some appropriate edge sharpening in post can likely take care of soft edges.

[derekkite]

Any KP (or K-70) owner can take multiminute exposure and find that small details looks like they are wrapped in very thin transparent plastic shell for the lack of better word. That ISO12800 comparison was interesting but shows heavy handed noise reduction when looking for crisp details. Sure, at first look it looks clean.

compared to what? iso 12800 is pretty bad. To get something presentable out of a raw file of that iso is very difficult, and typically post processing would mean smoothing out the background and enhancing some of the details.

I am far more interested in the lower iso, 3200 - 6400 range. I regularly get quite good croppable shots at 4000 iso, with good detail. But there are thresholds where the quality falls of a cliff, even at lower iso values, depending on light, contrast etc. I don't get very many presentable shots at 6400.

[Drinkkeri]

What Im really interested in is how is the buffer clearing speed affected by the accelerator or if it's affected at all.

[richandfleur]

What Im really interested in is how is the buffer clearing speed affected by the accelerator or if it's affected at all.

Given it's advertised as an product to give additional noise reduction capabilities, I don't think it will have any affect on buffer clearance/file writing speeds at all.

[Snakeisthestuff]

Reason I ask is Sony users are going through a situation whereby the noise reduction processes are 'killing stars' from their astro images. I'd like to know if the extra processing this new hardware chip offers will affect RAW images, because of that possible situation.

Not freaking out by any means, but am interested to know.

Yes it will.
Read the dpreview review of the KP on the high ISO performance.
They improve color and contrast in cost of slightly less details (saw only the comparison on ISO 25.600 live so can not comment much on lower ISOs).

If you are worried about this you should wait until the K-1 II reviews arrive.

[gatorguy]

Raw is not ideal. The photon collection process is inherently imperfect due to manufacturing imperfections in the sensor at the pixel level. Each pixel has a different black level, sensitivity, and color profile. And although some of those effects look random in the image, they aren't. Those can be corrected and are guaranteed to improve the quality of the image with absolutely no effect on resolution or artifacts. The output of these corrections would actually be much closer to the ideal image. (And it's a lot easier to correct these in-camera because they require camera-specific calibration files that are created in the factory.)

There's a second set of potential corrections that are a bit more contentious because they rely on analyzing and adjusting statistic properties. There are differences between the statistic properties of the noise and the structural numerical relationships and statistical properties of valid images. Certain patterns of neighboring R, G, and B values may be physically impossible in a noise-free image. Hence if those anomalies are in the data, it evidence of noise that can be subsequently removed. These methods can be contentious because they may have some slight effect on the "true" image. Optimized correctly, the gain in resolution from suppressing noise would more than offset any slight loss of resolution from "baking" the image.

The empirical data seems to show that Pentax RAWs may be baked to perfection rather than over-cooked. Those sample shots at ISO 12800 show improvements in both noise and resolution. The same is true of the DPreview images taken with the KP. So far, no owners of the K-70 or KP have found evidence of over-baking.

We shall see!

I've not seen any negative impact on my RAW's taken with the K-70 compared to my K-S2's. On the contrary they are cleaner at most ISO's so if there's any baking taking place with the accelerator chip it's being done by a skilled chef IMHO.

---------- Post added 03-12-18 at 10:02 AM ----------

Yes it will.
Read the dpreview review of the KP on the high ISO performance.
They improve color and contrast in cost of slightly less details (saw only the comparison on ISO 25.600 live so can not comment much on lower ISOs).

If you are worried about this you should wait until the K-1 II reviews arrive.

I find just the opposite, much improved detail from the K-70, at least at ISO's 12,800 and under. I don't recall ever shooting at a higher ISO tho.