[biz-engineer]

I'm not stating the above to argue with you or split hairs

Pixel shift and monochrome sensor aren't at all the same. With pixel shift, we lose 2 pixels horizontally and 2 pixels vertically due to shifting of the frames around, with monochrome we don't lose a single pixel.

[BigMackCam]

Pixel shift and monochrome sensor aren't at all the same. With pixel shift, we lose 2 pixels horizontally and 2 pixels vertically due to shifting of the frames around, with monochrome we don't lose a single pixel.

???

You don't lose anything with Pixel Shift. Without it, the colour data from a 2x2 cluster of pixels is incomplete and has to be demosaiced and interpolated to generate a very close approximation - calculated by the demosaicing algorithm, as best it can - of pixel-level colour data. With Pixel Shift, full colour data is recorded for each individual pixel - no demosaicing and no interpolation required... so a 26MP K-3III Pixel Shift image has 26MP of accurate colour data at the pixel level.

With the dedicated monochrome camera, we get luminance data at a pixel level without having to move the sensor, as there's no Bayer filter sat on top of it. A 26MP monochrome-version K-3III image has 26MP of accurate luminance data at the pixel level without the need for Pixel Shift...

[gatorguy]

???

You don't lose anything with Pixel Shift. Without it, the colour data from a 2x2 cluster of pixels is incomplete and has to be demosaiced and interpolated to generate a very close approximation of what the algorithm calculates - as best it can - as the pixel-level colour data. With it, full colour data is recorded for each individual pixel - no demosaicing and no interpolation required... so a 26MP K-3III Pixel Shift image has 26MP of accurate colour data at the pixel level.

With the dedicated monochrome camera, we get luminance data at a pixel level without having to move the sensor, as there's no Bayer filter sat on top of it. A 26MP monochrome-version K-3III image has 26MP of accurate luminance data at the pixel level without the need for Pixel Shift...

Would that be a reason there's so little noise in a high ISO monochrome compared to the same color image converted to mono?

[BigMackCam]

Would that be a reason there's so little noise in a high ISO monochrome compared to the same color image converted to mono?

The main reason (perhaps the only reason?) for that, I believe, is that with the dedicated monochrome camera there's no color filter array on top of the sensor. In a colour camera, that filter array robs the sensor of considerable light, so greater amplification is needed to achieve the same effective sensitivity (ISO level)... which, of course, results in more noise.

---

Last edited by BigMackCam; 04-10-2022 at 04:16 AM.

[Digitalis]

The astro K70 is wierd IMO.

Hello, you must be new here. Weird is the the shtick Pentax has been running with for decades now.

---------- Post added 2022-04-10 at 10:44 AM ----------

My guess is that they are more like shooting with a camera with a Foveon sensor and I'm not sure if Foveon's give better black and white output or not.

In good lighting Foveon equipped cameras are sublime for B&W but in bad lighting, yeesh. The weaker blue channel and the software sharpened red channel look distinctly artificial on their own.

[Joetitch]

In good lighting Foveon equipped cameras are sublime for B&W but in bad lighting, yeesh. The weaker blue channel and the software sharpened red channel look distinctly artificial on their own.

Does this mean a scene with high DR is not going to work well?

[Digitalis]

Does this mean a scene with high DR is not going to work well?

From what I have seen over the years, High Dynamic ranges are not friendly to Foveon sensors.

[Joetitch]

From what I have seen over the years, High Dynamic ranges are not friendly to Foveon sensors.

One more thing to scrap from my GAS list😊

[Digitalis]

I have spoken to a few colleagues over the years whom have worked with Foveon cameras and while they certainly have a niche advantage* that is most prominent when the stars align. For everyday photography a Pentax K-1 paired with an FA31 Limited or D-FA50 f/1.4 are going to produce images that are every bit as compelling and have a performance envelope vastly exceeding any recent Foveon camera.

* Though the proprietary software for processing Foveon RAW files is a real pain to work with.

[Joetitch]

...... an FA31 Limited or D-FA50 f/1.4 ....

Nooooo the list is supposed to get smaller.

[Rondec]

I have spoken to a few colleagues over the years whom have worked with Foveon cameras and while they certainly have a niche advantage* that is most prominent when the stars align. For everyday photography a Pentax K-1 paired with an FA31 Limited or D-FA50 f/1.4 are going to produce images that are every bit as compelling and have a performance envelope vastly exceeding any recent Foveon camera.

* Though the proprietary software for processing Foveon RAW files is a real pain to work with.

I think Sigma's software most auto-magically turn all of the dials up to 11 because most of the images I have seen from Foveon sensor cameras seems to have sharpening and saturation set to really high levels. I'm guessing these can be adjusted down somewhat, but I haven't used the software or the cameras to know.

[Digitalis]

I think Sigma's software most auto-magically turn all of the dials up to 11 because most of the images I have seen from Foveon sensor cameras seems to have sharpening and saturation set to really high levels

So the default settings are set to Ken Rockwell? that is reassuring.

[Smolk]

Yes, the Foveon sensors (in the DP2 and 3 Merrill) are neither good in low light nor with a high DR, but when they shine they shine if not radiate (metaphoric breakdown warning).
The sharpness seemed to be due to the sensor, not the software I'd guess. Especially with texture. Portraits had to be blurred somewhat in PP.
Noise was not a forte. And the software was a pain indeed. The back screen could be hard to see in bright daylight while there was no EVF.

[Zivelot]

Pity they have not published any Q regarding it, but here is one informal vote for a monochrome K-1. If news hits the public it is coming, I'll start saving.

[Ronald Oakes]

Here we go down the Dead Rabbit Hole.....mehhh.
Pentax really needs to pull something decent out of its hat !
None of this weirdness is doing anything at all for me.....disgusted !