Pentax K-3 III Sensor Measurements at PhotonsToPhotos

Started 4 months ago | Discussions
bclaff Forum Pro • Posts: 11,800
Pentax K-3 III Sensor Measurements at PhotonsToPhotos
28

Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.

These results are not comparable to any other cameras.

Note that the NR is so strong that the camera measures well above what a perfect sensor could accomplish.

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

Pentax K-3 Mark III
If you believe there are incorrect tags, please send us this post using our feedback form.
Tom Lusk Senior Member • Posts: 1,854
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
1

Why would they do that?

It's just asking for trouble.

OP bclaff Forum Pro • Posts: 11,800
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
4

Tom Lusk wrote:

Why would they do that?

It's just asking for trouble.

My complaint would be that it is not under user control; off should mean off.
Other than that. If people are happy with their pictures; great.

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

Tom Lusk Senior Member • Posts: 1,854
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
1

bclaff wrote:

Tom Lusk wrote:

Why would they do that?

It's just asking for trouble.

My complaint would be that it is not under user control; off should mean off.
Other than that. If people are happy with their pictures; great.

I agree, but some reviewers will have a field day with this information.

MightyMike Forum Pro • Posts: 41,079
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
1

bclaff wrote:

Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.

These results are not comparable to any other cameras.

Note that the NR is so strong that the camera measures well above what a perfect sensor could accomplish.

But does it still retain detail regardless of very strong NR?

-- hide signature --

Mike from Canada
"I am not a great photographer! God is a great creator! All I do is capture His creation with the tools He has provided me."
'I like to think so far outside the box that it would require a telephoto lens just to see the box!' ~ 'My Quote :)'
http://www.michaelfastphotography.com/galleries/VP-BDI_3a.jpg
http://www.airliners.net/user/SpeEedy/profile/photos?sortBy=photoFavoriteCount

MarBa Contributing Member • Posts: 807
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
2

bclaff wrote:

Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.

These results are not comparable to any other cameras.

Note that the NR is so strong that the camera measures well above what a perfect sensor could accomplish.

Well ... I would think that there is no way this will not result in a loss of detail. I don't get it why there can't be an option to turn this OFF. I would for example like to know how this NR reduction in RAW compares with NR reduction using PureRAW.

 MarBa's gear list:MarBa's gear list
Pentax K-1 Pentax K-3 Mark III Pentax smc FA 31mm F1.8 AL Limited Pentax smc FA 43mm F1.9 Limited Pentax smc FA 77mm 1.8 Limited +3 more
OP bclaff Forum Pro • Posts: 11,800
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
1

MightyMike wrote:

bclaff wrote:

Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.

These results are not comparable to any other cameras.

Note that the NR is so strong that the camera measures well above what a perfect sensor could accomplish.

But does it still retain detail regardless of very strong NR?

We've debated this before with the KP (this NR is slightly stronger and starts at a lower ISO setting).
I don't measure detail but mathematically it's not possible to do NR without smearing the data somewhat. That said, the effect might be limited to the deep shadows.

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

flektogon
flektogon Veteran Member • Posts: 5,138
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos

Good job Bill, though it will make many people unhappy . I do not mean your job, but the fact that Pentax took such an approach to make "fantastically" clean pictures .

Now the most important information would be the impact on the fine details capturing. It may not be so significant. This "forced" NR actually acts as an AA filter, and if the camera is still able to capture close to 13 "mega details", then everything should be fine.

-- hide signature --

Regards,
Peter

Roland Karlsson Forum Pro • Posts: 29,531
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos

bclaff wrote:

MightyMike wrote:

bclaff wrote:

Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.

These results are not comparable to any other cameras.

Note that the NR is so strong that the camera measures well above what a perfect sensor could accomplish.

But does it still retain detail regardless of very strong NR?

We've debated this before with the KP (this NR is slightly stronger and starts at a lower ISO setting).
I don't measure detail but mathematically it's not possible to do NR without smearing the data somewhat. That said, the effect might be limited to the deep shadows.

You can do NLM noise reduction. Then you can go beyond the theoretical noise reduction without smearing. NLM means non local means. You sample the entire image after similar details and then do mean value. The disadvantage is that the resulting image not really is correct.

NLM is an expensive operation and do need an accelerator.

-- hide signature --

/Roland
Kalpanika X3F tools:
https://github.com/kalpanika/x3f

 Roland Karlsson's gear list:Roland Karlsson's gear list
Sigma DP3 Merrill Sigma dp2 Quattro Sony RX100 III Pentax K-3 Pentax K-1 +14 more
iso rivolta Contributing Member • Posts: 846
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
1

bclaff wrote:

I don't measure detail but mathematically it's not possible to do NR without smearing the data somewhat. That said, the effect might be limited to the deep shadows.

Hi Bill,

Then would it be possible to also analyse grey frames instead of only dark frames? If one could generate almost perfect uniform grey areas, I think it would be interesting to do an analysis of NR as function of exposure level.

 iso rivolta's gear list:iso rivolta's gear list
Pentax K-3 II Pentax KP +3 more
agukha Regular Member • Posts: 238
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
1

bclaff wrote:

Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.

These results are not comparable to any other cameras.

Note that the NR is so strong that the camera measures well above what a perfect sensor could accomplish.

Thank you !

When I'll have time for that, I'll try to take an outdoor image to compare the details, with and without pixelshift, between K3 ii, KP, K3 iii and 645Z. (I don't know someone near me having a K1, we are still limited to 10km here by the way)

And then see if this theorical loss in details is viewable for human eyes ^^

OP bclaff Forum Pro • Posts: 11,800
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
4

Roland Karlsson wrote:

bclaff wrote:

MightyMike wrote:

bclaff wrote:

Except between ISO 100 and ISO 160 there is very strong Noise Reduction (NR) baked into the raw data even with NR set to off.

These results are not comparable to any other cameras.

Note that the NR is so strong that the camera measures well above what a perfect sensor could accomplish.

But does it still retain detail regardless of very strong NR?

We've debated this before with the KP (this NR is slightly stronger and starts at a lower ISO setting).
I don't measure detail but mathematically it's not possible to do NR without smearing the data somewhat. That said, the effect might be limited to the deep shadows.

You can do NLM noise reduction. Then you can go beyond the theoretical noise reduction without smearing. NLM means non local means. You sample the entire image after similar details and then do mean value. The disadvantage is that the resulting image not really is correct.

NLM is an expensive operation and do need an accelerator.

FWIW, 2D Fourier Transforms indicate some sort of nearest neighbor type of computation. That doesn't sound like NLM.

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

OP bclaff Forum Pro • Posts: 11,800
Re: Pentax K-3 III Sensor Measurements at PhotonsToPhotos
2

iso rivolta wrote:

bclaff wrote:

I don't measure detail but mathematically it's not possible to do NR without smearing the data somewhat. That said, the effect might be limited to the deep shadows.

Hi Bill,

Then would it be possible to also analyse grey frames instead of only dark frames? If one could generate almost perfect uniform grey areas, I think it would be interesting to do an analysis of NR as function of exposure level.

I will look into this.

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

PortrayingLife
PortrayingLife Forum Member • Posts: 87
Or...

Tom Lusk wrote:

Why would they do that?

It's just asking for trouble.

It's taking the steps necessary to guarantee success.

Cheers... M

-- hide signature --

Michael
"Be the kind of person chance finds easy to help.” Anon

 PortrayingLife's gear list:PortrayingLife's gear list
Pentax K100D Pentax K110D Pentax Q7 Pentax K-3 II Pentax K-1 II +12 more
left eye
left eye Senior Member • Posts: 2,427
Pentax STOP this amateurishness!
10

So it looks like Pentax is not listening to it's photographers, it's listening to people who don't like to see grain (noise) on their screens. When printed, the presence of grain/noise actually helps in making large prints. Even the real world when you look closer and closer contains 'grain', cells, sands, dust, etc, which we don't make much sense of, but it helps.

This is exceptionally annoying ...to have detail smudged away, upon which I shall then need to add grain!

And why this ludicrous need to present ultra-high ISO gain values? - esp when this camera has class-leading IBIS.

I'm out.

I dearly hope the K-1 III doesn't go this route (- I realise it already partly has with the K-1 II, but I was hoping Pentax would pull back with this K-3 III not go OTT).

Anyway many thanks Bill, for your wok in seeing behind the curtain!

 left eye's gear list:left eye's gear list
Olympus E-300 Fujifilm GFX 50S Panasonic GH5 Pentax smc DA 15mm F4 ED AL Limited Pentax smc DA 35mm F2.8 Macro Limited +8 more
Holger Bargen Veteran Member • Posts: 4,593
A look at shaddow improvement
2

Dear Bill,

Once again, great jpb! I find your measurements very helpful and especially for Pentax it shows the improvements in an almost daramtic way.

I just prepared another graph from your side - the shaddow improvement, comparing K1, K1ii and K3iii.

I use these cameras as K1 shows the old foundation without NR improvements, K1ii was the first camera that implemented these new technics and K3iii is the latest step in this development:

https://www.photonstophotos.net/Charts/PDR_Shadow.htm#Pentax%20K-1,Pentax%20K-1%20II,Pentax%20K-3%20III

In the chart showing just the dynamic range vs. ISO we found that K3iii starts at much lower ISO with corrections - the higher ISO part is almost same as K1ii (which is an enormous development as K1ii is FF and K3iii is APS-C.

If we look at shaddow improvement, K1ii shows the lifted curve above K1. However in this graph, K3iii shows a much higher lifting of the curve of the dynamic range in shaddow areas.

If we look at the read nosie vs ISO and compare K1, K1ii and K3iii:

https://www.photonstophotos.net/Charts/RN_ADU.htm#Pentax%20K-1_14,Pentax%20K-1%20II_14,Pentax%20K-3%20III_14

... now, let's look at ISO 800 at K1 (I found this a very useful ISO for K1 and try to keep it in my work), we find same nosie level for K1ii at about ISO 2000 and for K3iii at about ISO 4000!

I realize that K3iii rocks! And K1ii would have been a good upgrade for my work, too.

Best regards

Holger

 Holger Bargen's gear list:Holger Bargen's gear list
Pentax K-5 Pentax K-S1 Pentax K-1 Pentax smc DA 55-300mm F4.0-5.8 ED Sigma 70mm F2.8 EX DG Macro +7 more
Fotoni Senior Member • Posts: 1,046
K-1 Mark I, the first and the last...
6

Their only rather modern and proper camera without cheating. Why they keep doing forced noise reduction? Nikon also cheats with their raw video by using line-skipping. Both are deaf even when people are screaming not to do those things again.

So much for believing that Pentax target is great value cameras. K-1 Mark I was a shooting star in that area because they decided to cheat with Mark II and then also raised the price. There have been better options at least two years.

 Fotoni's gear list:Fotoni's gear list
Pentax K-1 Pentax FA* 50mm F1.4 SDM AW Canon PowerShot SD4000 IS
agukha Regular Member • Posts: 238
Re: Pentax STOP this amateurishness!
5

left eye wrote:

So it looks like Pentax is not listening to it's photographers, it's listening to people who don't like to see grain (noise) on their screens. When printed, the presence of grain/noise actually helps in making large prints. Even the real world when you look closer and closer contains 'grain', cells, sands, dust, etc, which we don't make much sense of, but it helps.

This is exceptionally annoying ...to have detail smudged away, upon which I shall then need to add grain!

And why this ludicrous need to present ultra-high ISO gain values? - esp when this camera has class-leading IBIS.

I'm out.

I dearly hope the K-1 III doesn't go this route (- I realise it already partly has with the K-1 II, but I was hoping Pentax would pull back with this K-3 III not go OTT).

Anyway many thanks Bill, for your wok in seeing behind the curtain!

I have not yet seen pictures where a loss of details can be seen between K1 and K1 ii.

And the "pure" RAW is a myth I think, so many implementing and processes choices having an impact on it.
It's like making a difference between "natural" and "chemical", everything is based on chemical processes, and it's not bad just because of it. Or eat cyanide, it's purely natural, can't be bad. ^^

Holger Bargen Veteran Member • Posts: 4,593
Re: Pentax STOP this amateurishness!
3

left eye wrote:

So it looks like Pentax is not listening to it's photographers, it's listening to people who don't like to see grain (noise) on their screens. When printed, the presence of grain/noise actually helps in making large prints. Even the real world when you look closer and closer contains 'grain', cells, sands, dust, etc, which we don't make much sense of, but it helps.

This is exceptionally annoying ...to have detail smudged away, upon which I shall then need to add grain!

Does Pentax harm details in their technology?

When I saw first approaches of getting rid of noise by smearing pixels (first saw it at old Panasonic Lumix FZ50), I did not like it. However, if I look at the Pentax fine detail resolutions of photos taken with K3iii, I don't see that pattern.

If you want to do post-processing of your RAW files, noise is a heavy burden. No matter if professional or amateur - anybody will be glad to get rid of noise.

For this reason I may ask you, why professionals should not like the photos taken with K3iii.

And why this ludicrous need to present ultra-high ISO gain values? - esp when this camera has class-leading IBIS.

I'm out.

I dearly hope the K-1 III doesn't go this route (- I realise it already partly has with the K-1 II, but I was hoping Pentax would pull back with this K-3 III not go OTT).

Anyway many thanks Bill, for your wok in seeing behind the curtain!

 Holger Bargen's gear list:Holger Bargen's gear list
Pentax K-5 Pentax K-S1 Pentax K-1 Pentax smc DA 55-300mm F4.0-5.8 ED Sigma 70mm F2.8 EX DG Macro +7 more
OP bclaff Forum Pro • Posts: 11,800
Re: Pentax STOP this amateurishness!
4

agukha wrote:

..

I have not yet seen pictures where a loss of details can be seen between K1 and K1 ii.

That's good. You realize that the NR kicks in at ISO 640 on the K-1 II; right?

-- hide signature --

Bill ( Your trusted source for independent sensor data at PhotonsToPhotos )

Keyboard shortcuts:
FForum MMy threads