Does Pentax applies NR to its RAW files?

[ManuH]

Read this:

DxOMark-review-for-Pentax-cameras

"Noise filtering processing is present on the Pentax K10D (ISO 800 and ISO 1600), the Pentax K20D (ISO 3200 and ISO 6400), and the Pentax K200D (ISO 800 and ISO 3200)."

(Note, for the K200D it's a typo, the K200D does not have a 3200 ISO).

"However, while most other cameras have an interface that permits turning off internal noise filtering and thus providing unprocessed RAW data files, Pentax cameras still apply some noise reduction even when the noise reduction option is supposedly switched off."

That is extremely disapointing if true.

[deejjjaaaa]

Originally Posted by ManuH

Read this:

DxOMark-review-for-Pentax-cameras

"Noise filtering processing is present on the Pentax K10D (ISO 800 and ISO 1600), the Pentax K20D (ISO 3200 and ISO 6400), and the Pentax K200D (ISO 800 and ISO 3200)."

(Note, for the K200D it's a typo, the K200D does not have a 3200 ISO).

"However, while most other cameras have an interface that permits turning off internal noise filtering and thus providing unprocessed RAW data files, Pentax cameras still apply some noise reduction even when the noise reduction option is supposedly switched off."

That is extremely disapointing if true.

at this point you can start googling for gordonbgood AND/OR oleg_v

have a nice reading.

[Marc Sabatella]

As someone who has followed many of their posts over the years, I can provide what I *think* is an accurate summary:

- all models can do "dark frame substraction" (aka "slow shutter speed NR") at slow shutter speeds

- the K20D is the only Pentax camera that in fact cannot turn that type of NR *off*

- all recent models provide "high ISO NR" settings in the menu that affect JPEG only

- the K20D appears to perform "some" (albeit not necessarily a lot) of NR even on RAW files at high ISO, even if you've turned the NR option off (like I said, it's only relevant for JPEG). The proof of this is way over my head mathematically, but Oleg and Gordon know what they are doing and seem 100% convinced their mathematical analysis of the RAW data itself demonstrates beyond any shadow of a doubt that this *is* going on.

- as far as I can tell, no evidence of this sort of slight high ISO NR being applied to RAW files has been found in other Pentax cameras, although I'm not sure they decided for sure one way another regarding the K-m/K2000

The actual amount of high ISO NR applied by the K20D to its RAW files does not seem to be of the sort that should be "extremely disappointing". It seems to have no discernible effect on detail. If you're imagining some sort of heavy-handed Neat Image style of NR, that's *not* what they are talking about. Just mathematical evidence that the numbers have been processed slightly.

BTW, after all this extensive analysis of noise levels at different ISO settings (and comparing different types of noise levels too), Gordon ended up buying a K200D for himself as the best bang for the buck, although I haven't seen many posts lately to get a sense if the camera is living up to his own expectations.

[ManuH]

Originally Posted by Marc Sabatella

- as far as I can tell, no evidence of this sort of slight high ISO NR being applied to RAW files has been found in other Pentax cameras, although I'm not sure they decided for sure one way another regarding the K-m/K2000

The actual amount of high ISO NR applied by the K20D to its RAW files does not seem to be of the sort that should be "extremely disappointing".

I know about Gordon, I read (not understand...) his posts always with great interest. I know what he said about K20D RAW NR, and yes this is not the kind of heavy-handed NR that Sony used in their A700. But nonetheless it's disapointing (ok maybe not extremely) that you can't turn it off at all. That may explain the strange noise pattern at ISO 3200 and above.

According to DXO, the K200D also uses RAW NR at 800 and above. Maybe not much but enough that its score was equal to the K20D. That was strange as all other tests show the K20D performing better at high ISO.

Now the Nikon D90 figures are also suspect considering this the same sensor as used in the D300 and A700. However, nobody (Gordon did also analysed it) has yet been able to prove it. But the pictures are also strangely somewhat soft...

[Marc Sabatella]

Originally Posted by ManuH

That may explain the strange noise pattern at ISO 3200 and above.

It might, or it might be that it would be even stranger without the NR applied. Whose to say?

Actually, I'm just taking you at your word about strange noise patterns. I don't have a K20D, but have seen some images where noise looks "normal", and others where there is a kind of cross-hatched pattern. Kind of like the infamous VPN many people saw with the K10D. *Occasionally* I'll see something like that when doing extreme push processing on the K200D, but mostly not. So I'm assuming that to some degree, this sort of thing *is* "normal".

According to DXO, the K200D also uses RAW NR at 800 and above.

OK; I can't prove or disprove that. I can only say that others who specifically looked for it didn't see it. Not that it matters to me one way or the other.

Maybe not much but enough that its score was equal to the K20D. That was strange as all other tests show the K20D performing better at high ISO.

I'm convinced that high ISO performance depends on so many variables that you can get a comparison to come out however you want depending on how you test.

[monochrome]

Four questions

1. Why does this matter?
2. Who is Gordon B Good?
3. Who pays for DxO labs?
4. Does anyone find the timing interesting, in light of the momentum building for the K-7?