Do the Q2 and Q2 Monochrom have the same processor?

[peter3334]

Advertisement (gone after registration)

Do the Q2 and Q2 Monochrom have the same processor? Do the processors use the same algorithms? When third party companies modify a camera for Monochrom only such as the Nikon D850 and the Nikon D850 Monochrom version, does the Third party also modify the 'processor' algorithm to cater for the absence of the Color Filter array?

[jaapv]

The Q monochrom  outputs a grayscale file on DNG conversion (which can, of course, can easily be changed to RGB). OTOH, provided Leica did not change this for the Q, if you insert an SD card with a  Q 2 colour image on it, the Q2 Monochrom will render it in colour.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[rukez]

When third party company modifies camera to monochrome they have only physical access to Bayer filter (they actually simply remove it) but they can’t modify internal camera software that is generating jpeg files (they still will be generated with approximate data from nearby pixels instead of direct conversion from signal amplitude to grey level) or adds meta data to raw files (including Bayer pattern that is backed-in)

When camera is released as black and white by manufacturer, certainly you will have correct meta data in raw and optimal in camera jpeg creation algorithm (that does not use interpolation of nearby pixels) and out of box support by pc software that supports the exact model (and also understands it’s b/w)

[rukez]

On 12/21/2021 at 3:51 AM, jaapv said:

The Q monochrom  outputs a grayscale file on DNG conversion (which can, of course, can easily be changed to RGB). OTOH, provided Leica did not change this for the Q, if you insert an SD card with a  Q 2 colour image on it, the Q2 Monochrom will render it in colour.

I’ve never used q-m but I am sure you can’t change it’s dng to rgb because main physical difference between q and q-m is Bayer filter absence, and without Bayer filter you can’t split visible light into spectrum parts, needed for generating color image. Instead, you can only register level of light, reaching each pixel (but not wave length). 
Q-m can show color pics from generic Q because 99% of software is unified inside them, so it’s reasonable that engineers use same base modules (actually, I am sure that generic Q has right b/w conversion engine from Q-M inside also, that’s just cheaper and more bugs-proof to develop and test their firmware together rather then add-test-modify-test separately) - only two things needs to be changed for b/w conversion software-wise:

- you need simpler JPEG converter (no need to get average from nearby pixels to combine full-spectrum signal and then write it’s alpha level, you already have alpha level registered)

- you need to re-calibrate exposure meter because you don’t loose light inside Bayer filter

and yep, you need to change meta in raw header to let pc software know you are monochrome 

[jaapv]

All other Monochroms have no problem there. A DNG file is not RGB or an other colour format, that is only assigned on raw conversion. Photoshop converts automatically to grayscale by default - until one changes the setting to RGB in ACR after which it will always be RGB, which is the way that I convert. (just try it)
Which is logical, as the colour format is only set on raw conversion, not in the camera. The result is an RGB file with three identical channels.
Why would one do that? For instance to apply toning, and many plugins and tools in PS don't work in Greyscale, but do so in RGB. What you do to the file after it has been written has no influence on the exposure - that is set on exposure and correctly calibrated by the manufacturer.

 

[rukez]

Yep, raw file color mode is not affected until raw is “developed” - but processing software must understand whether and what filter pattern was used, for supported “true b/w” cameras it’s easy because they use proper data in raw header, for converted camera you will have to select proper deBayer method manually 

In PS yes, rgb color space makes sense for tools compatibility, but it should be applied after raw file is “developed” in proper b/w mode (and I’m sure PS does so for cameras, that it knows to be b/w) and you will be still limited in tools selection, because all tools that require true color mixer (like color filter) wouldn’t work as expected (because there was no color separation in original file)

If you develop b/w raw directly in rgb you will loose a lot of data because you will have to get “complete” (color+alpha) signal averaging neighboring pixels, that makes no sense for Bayer-less sensors (except for some noise reduction but it is probably least effective way because you average all the data but not only pikes)