Acceleration Chip (K1mkII, KP and K70) - What's the Difference

reh321 · 04-05-2019, 05:42 PM - **1 Like**

Originally posted by rawr

The GR III curve is also an interesting one.

A strong 'accelerator' boost kicks in even earlier than the KP. Ricoh seem to like this tech a lot.

So do I.

Last Fall, when photographing animals out my window with my K-30, I was constantly making compromises in order to keep the ISO at or below 800.

Now, with my KP, I just set the shutter speed and aperture to appropriate values using TAv mode, and if ISO 4400 is needed to make it work - I know the KP can handle that.

Gimbal · 04-06-2019, 02:48 AM

"Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade.

But then even the K5 had noise reduction kicking in at 1600, and back then it was considered cheating as it improved some measurements.

**Rondec** · 04-06-2019, 02:58 AM

Originally posted by Gimbal

"Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade.

But then even the K5 had noise reduction kicking in at 1600, and back then it was considered cheating as it improved some measurements.

The whole thing is overblown. It probably does affect shadow detail a little bit, but not stuff you would see without quite a bit of post processing. And I do think it makes higher iso shots in reach of people who don't have high level noise reduction software and the skills to use it.

But the K-1 and K3 don't have it, so those are options for people who want "purity of RAW images."

Gimbal · 04-06-2019, 06:04 AM

Originally posted by Rondec

The whole thing is overblown. It probably does affect shadow detail a little bit, but not stuff you would see without quite a bit of post processing. And I do think it makes higher iso shots in reach of people who don't have high level noise reduction software and the skills to use it.

But the K-1 and K3 don't have it, so those are options for people who want "purity of RAW images."

Yes, but you also have Sony with the star-eater function.
It's an easy way to improve performance, or at least it looks like the performance have increased.
Why not make it an option instead and everybody could be happy?

**gaweidert** · 04-06-2019, 06:25 AM

Originally posted by Gimbal

"Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade.

But then even the K5 had noise reduction kicking in at 1600, and back then it was considered cheating as it improved some measurements.

The chip itself accelerates math processing by taking some of the load off of the CPU. What Ricoh chooses to do with it is up to them. Whether we like what Ricoh has done with it is up to us. Another function that one reviewer actually did measure was a slight improvement in autofocus speed. Not much, but it was measurable. While Pentax autofocus speed is not up to Nikon etc speeds, it is, in reality, not that far off either, so every little bit helps.

Personally I like my K1-II just a little bit more than my K1. I do have a bit of regret for not sending my K1 in for the upgrade. My K1-II does not suffer from shutter shock and having that range of shutter speed available is a plus in my book.

reh321 · 04-06-2019, 07:02 AM

Originally posted by Gimbal

"Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade.

It is much more than noise reduction. Improvement in Dynamic Range and Color Depth are significant.

The comparison below comes from Pentax KP First Shots: Going head-to-head against the Nikon D500 at ISO 819,200!
Pentax KP is on the left - Nikon D500 on the right. My own experiments on KP have also shown amazing performance at ISO 25600 or higher.

reh321 · 04-06-2019, 07:04 AM

Originally posted by Gimbal

Yes, but you also have Sony with the star-eater function.
It's an easy way to improve performance, or at least it looks like the performance have increased.
Why not make it an option instead and everybody could be happy?

The Ricoh representative at CP+ said that making it optional "would be hard" - which is usually the polite Japanese way of saying "We can't figure out how to do that right now".

reh321 · 04-06-2019, 07:07 AM

Originally posted by Rondec

The whole thing is overblown. It probably does affect shadow detail a little bit, but not stuff you would see without quite a bit of post processing. And I do think it makes higher iso shots in reach of people who don't have high level noise reduction software and the skills to use it.

I could do it, but after a lifetime sitting at a computer, I would rather spend my discretionary time in the field and let automation do what it can,

reh321 · 04-06-2019, 07:09 AM

Originally posted by gaweidert

The chip itself accelerates math processing by taking some of the load off of the CPU. What Ricoh chooses to do with it is up to them.

It appears to be more than just that. Apparently it has specific tasks built into it and it performs those specific tasks on sensor output before it gets to the processor.

**Rondec** · 04-06-2019, 07:12 AM

Originally posted by Gimbal

Yes, but you also have Sony with the star-eater function.
It's an easy way to improve performance, or at least it looks like the performance have increased.
Why not make it an option instead and everybody could be happy?

I don't think the way it is created it can be optional, although I imagine that Pentax could design a "full frame" accelerator chip and have it start at a little higher iso. But after reading everything I can about it, it just doesn't seem like something you can turn on or off without quite a bit of redesign.

joergens.mi · 04-06-2019, 09:30 AM

In my opinion it is a signal processor, which is in the loop after the sensor before the Main-Processor, there may be some inputs to the signal processor to get information on ISO Settings and things like that - maybe that this are the signals that are send to the sensor, therefore it may be hard to find a way to set it to do nothing.

Hi Reh321, i don't think, that you have a chance to break in. I assume that only little Information is send from the prime IV to the acc-chip and for the rest the acc-chip is participating on the signals send to the sensor. And the sensor signals are passing thru the acc-chip before seen by the Prime IV.

Gimbal · 04-06-2019, 11:22 AM

Originally posted by Rondec

I don't think the way it is created it can be optional, although I imagine that Pentax could design a "full frame" accelerator chip and have it start at a little higher iso. But after reading everything I can about it, it just doesn't seem like something you can turn on or off without quite a bit of redesign.

We can only speculate, but as I understand it (which might be way wrong) it kicks in on images above iso640, which means it indeed can pass the image through without "working" on it.
The fact that it now can't be turned off is probably a design decision they made to hide the "true" performance and nothing else.

reh321 · 04-06-2019, 12:12 PM

Originally posted by Gimbal

We can only speculate, but as I understand it (which might be way wrong) it kicks in on images above iso640, which means it indeed can pass the image through without "working" on it.

We know that it reduces noise, increases Dynamic Range, and enhances color - we don't know whether it does anything else; we don't know whether it truly is idling in 'pass through' mode below ISO 640 and/or whether these different functionalities are turned on / ramped up one-at-a-time up to ISO 640. We also don't know whether 'not working on data' is a result of a switch the processor throws or whether its programming causes different behaviors at different ISO levels.

Originally posted by Gimbal

The fact that it now can't be turned off is probably a design decision they made to hide the "true" performance and nothing else.

'Design Decision' doesn't mean that they could turn it off; I'm not sure what you mean by 'true' performance. Proper design could mean that it works best when data is always flowing through it, that in some kind of 'off' mode it would degrade the image.

Rico · 04-06-2019, 01:13 PM

Originally posted by Gimbal

We can only speculate, but as I understand it (which might be way wrong) it kicks in on images above iso640, which means it indeed can pass the image through without "working" on it.
The fact that it now can't be turned off is probably a design decision they made to hide the "true" performance and nothing else.

The Accelerator Unit is always on. The benefits of the of the Accelerator Unit begin to be really visible starting above ISO 640 but that does not mean it is turned on and off above and below those values.

Ricoh developed the Accelerator Unit which is a factor for them in not complying with demands for an on/off switch due to pressure from sites like DPR. Especially when there is no need to turn off the Accelerator Unit.

peterh337 · 04-07-2019, 02:18 PM

"It could possibly be a Gate Array or some other specialized hardware - or a Digital Signal Processor.
Back when I worked on digital radios, the "modems" were implemented in DSPs, because they could do that kind of processing so much faster and efficiently."

IF the sensor outputs digital data then this chip cannot do anything which the camera CPU could not do in its software.

It can probably do it faster...

04-06-2019, 07:02 AM	#36
reh321 Pentaxian Join Date: Nov 2014 Location: South Bend, IN, USA Photos: Gallery \| Albums Posts: 23,177	Originally posted by Gimbal "Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade. It is much more than noise reduction. Improvement in Dynamic Range and Color Depth are significant. The comparison below comes from Pentax KP First Shots: Going head-to-head against the Nikon D500 at ISO 819,200! Pentax KP is on the left - Nikon D500 on the right. My own experiments on KP have also shown amazing performance at ISO 25600 or higher. Last edited by reh321; 04-06-2019 at 07:10 AM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
KP vs K-70 - to KP or not to KP, that is the question	OldChE	Pentax DSLR Discussion	28	11-02-2019 05:29 AM
Currently have a Kx, looking to move up to a KP, K70, or K3-ii. Recommendations?	mojoe_24	Pentax KP	10	11-07-2017 10:05 PM
LR update, May 2017 + GPU Acceleration issue.	BruceBanner	Digital Processing, Software, and Printing	2	07-29-2017 07:35 PM
Noise Reduction Acceleration Unit	reh321	Pentax Q	5	06-15-2016 03:52 AM
Sensor difference due to optics in front of chip	rechmbrs	Pentax K-5 & K-5 II	5	12-04-2010 05:56 PM

04-05-2019, 05:42 PM - 1 Like	#31
reh321 Pentaxian Join Date: Nov 2014 Location: South Bend, IN, USA Photos: Gallery \| Albums Posts: 23,177	Originally posted by rawr The GR III curve is also an interesting one. A strong 'accelerator' boost kicks in even earlier than the KP. Ricoh seem to like this tech a lot. So do I. Last Fall, when photographing animals out my window with my K-30, I was constantly making compromises in order to keep the ISO at or below 800. Now, with my KP, I just set the shutter speed and aperture to appropriate values using TAv mode, and if ISO 4400 is needed to make it work - I know the KP can handle that.

04-06-2019, 02:48 AM	#32
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106	"Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade. But then even the K5 had noise reduction kicking in at 1600, and back then it was considered cheating as it improved some measurements.

04-06-2019, 02:58 AM	#33
Rondec Loyal Site Supporter Join Date: Mar 2009 Location: Gladys, Virginia Photos: Gallery Posts: 27,652	Originally posted by Gimbal "Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade. But then even the K5 had noise reduction kicking in at 1600, and back then it was considered cheating as it improved some measurements. The whole thing is overblown. It probably does affect shadow detail a little bit, but not stuff you would see without quite a bit of post processing. And I do think it makes higher iso shots in reach of people who don't have high level noise reduction software and the skills to use it. But the K-1 and K3 don't have it, so those are options for people who want "purity of RAW images."

04-06-2019, 06:04 AM	#34
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106	Originally posted by Rondec The whole thing is overblown. It probably does affect shadow detail a little bit, but not stuff you would see without quite a bit of post processing. And I do think it makes higher iso shots in reach of people who don't have high level noise reduction software and the skills to use it. But the K-1 and K3 don't have it, so those are options for people who want "purity of RAW images." Yes, but you also have Sony with the star-eater function. It's an easy way to improve performance, or at least it looks like the performance have increased. Why not make it an option instead and everybody could be happy?

04-06-2019, 06:25 AM	#35
gaweidert Loyal Site Supporter Join Date: Mar 2012 Location: Rochester, NY Photos: Gallery Posts: 5,325	Originally posted by Gimbal "Accelerator" is it more than a fancy name for noise reduction? I'm not so sure I like it...and that is why I didn't upgrade. But then even the K5 had noise reduction kicking in at 1600, and back then it was considered cheating as it improved some measurements. The chip itself accelerates math processing by taking some of the load off of the CPU. What Ricoh chooses to do with it is up to them. Whether we like what Ricoh has done with it is up to us. Another function that one reviewer actually did measure was a slight improvement in autofocus speed. Not much, but it was measurable. While Pentax autofocus speed is not up to Nikon etc speeds, it is, in reality, not that far off either, so every little bit helps. Personally I like my K1-II just a little bit more than my K1. I do have a bit of regret for not sending my K1 in for the upgrade. My K1-II does not suffer from shutter shock and having that range of shutter speed available is a plus in my book.

04-06-2019, 07:04 AM	#37
reh321 Pentaxian Join Date: Nov 2014 Location: South Bend, IN, USA Photos: Gallery \| Albums Posts: 23,177	Originally posted by Gimbal Yes, but you also have Sony with the star-eater function. It's an easy way to improve performance, or at least it looks like the performance have increased. Why not make it an option instead and everybody could be happy? The Ricoh representative at CP+ said that making it optional "would be hard" - which is usually the polite Japanese way of saying "We can't figure out how to do that right now".

04-06-2019, 07:07 AM	#38
reh321 Pentaxian Join Date: Nov 2014 Location: South Bend, IN, USA Photos: Gallery \| Albums Posts: 23,177	Originally posted by Rondec The whole thing is overblown. It probably does affect shadow detail a little bit, but not stuff you would see without quite a bit of post processing. And I do think it makes higher iso shots in reach of people who don't have high level noise reduction software and the skills to use it. I could do it, but after a lifetime sitting at a computer, I would rather spend my discretionary time in the field and let automation do what it can,

04-06-2019, 07:09 AM	#39
reh321 Pentaxian Join Date: Nov 2014 Location: South Bend, IN, USA Photos: Gallery \| Albums Posts: 23,177	Originally posted by gaweidert The chip itself accelerates math processing by taking some of the load off of the CPU. What Ricoh chooses to do with it is up to them. It appears to be more than just that. Apparently it has specific tasks built into it and it performs those specific tasks on sensor output before it gets to the processor.

04-06-2019, 07:12 AM	#40
Rondec Loyal Site Supporter Join Date: Mar 2009 Location: Gladys, Virginia Photos: Gallery Posts: 27,652	Originally posted by Gimbal Yes, but you also have Sony with the star-eater function. It's an easy way to improve performance, or at least it looks like the performance have increased. Why not make it an option instead and everybody could be happy? I don't think the way it is created it can be optional, although I imagine that Pentax could design a "full frame" accelerator chip and have it start at a little higher iso. But after reading everything I can about it, it just doesn't seem like something you can turn on or off without quite a bit of redesign.

04-06-2019, 09:30 AM	#41
joergens.mi Veteran Member Join Date: Jan 2012 Posts: 408	In my opinion it is a signal processor, which is in the loop after the sensor before the Main-Processor, there may be some inputs to the signal processor to get information on ISO Settings and things like that - maybe that this are the signals that are send to the sensor, therefore it may be hard to find a way to set it to do nothing. Hi Reh321, i don't think, that you have a chance to break in. I assume that only little Information is send from the prime IV to the acc-chip and for the rest the acc-chip is participating on the signals send to the sensor. And the sensor signals are passing thru the acc-chip before seen by the Prime IV.

04-06-2019, 11:22 AM	#42
Gimbal Pentaxian Join Date: Apr 2007 Location: Sweden Posts: 2,106	Originally posted by Rondec I don't think the way it is created it can be optional, although I imagine that Pentax could design a "full frame" accelerator chip and have it start at a little higher iso. But after reading everything I can about it, it just doesn't seem like something you can turn on or off without quite a bit of redesign. We can only speculate, but as I understand it (which might be way wrong) it kicks in on images above iso640, which means it indeed can pass the image through without "working" on it. The fact that it now can't be turned off is probably a design decision they made to hide the "true" performance and nothing else.

04-06-2019, 12:12 PM	#43
reh321 Pentaxian Join Date: Nov 2014 Location: South Bend, IN, USA Photos: Gallery \| Albums Posts: 23,177	Originally posted by Gimbal We can only speculate, but as I understand it (which might be way wrong) it kicks in on images above iso640, which means it indeed can pass the image through without "working" on it. We know that it reduces noise, increases Dynamic Range, and enhances color - we don't know whether it does anything else; we don't know whether it truly is idling in 'pass through' mode below ISO 640 and/or whether these different functionalities are turned on / ramped up one-at-a-time up to ISO 640. We also don't know whether 'not working on data' is a result of a switch the processor throws or whether its programming causes different behaviors at different ISO levels. Originally posted by Gimbal The fact that it now can't be turned off is probably a design decision they made to hide the "true" performance and nothing else. 'Design Decision' doesn't mean that they could turn it off; I'm not sure what you mean by 'true' performance. Proper design could mean that it works best when data is always flowing through it, that in some kind of 'off' mode it would degrade the image.

04-06-2019, 01:13 PM	#44
Rico Veteran Member Join Date: Mar 2007 Posts: 1,204	Originally posted by Gimbal We can only speculate, but as I understand it (which might be way wrong) it kicks in on images above iso640, which means it indeed can pass the image through without "working" on it. The fact that it now can't be turned off is probably a design decision they made to hide the "true" performance and nothing else. The Accelerator Unit is always on. The benefits of the of the Accelerator Unit begin to be really visible starting above ISO 640 but that does not mean it is turned on and off above and below those values. Ricoh developed the Accelerator Unit which is a factor for them in not complying with demands for an on/off switch due to pressure from sites like DPR. Especially when there is no need to turn off the Accelerator Unit.

04-07-2019, 02:18 PM	#45
peterh337 Veteran Member Join Date: Jan 2009 Location: London Posts: 573	"It could possibly be a Gate Array or some other specialized hardware - or a Digital Signal Processor. Back when I worked on digital radios, the "modems" were implemented in DSPs, because they could do that kind of processing so much faster and efficiently." IF the sensor outputs digital data then this chip cannot do anything which the camera CPU could not do in its software. It can probably do it faster...