RustyBug
Offline
• • • • • • •
Upload & Sell: On
|
pmeheut wrote:
AFAIK, bayer matrix filters and spectral sensitivity of the sensor have some influence...
+1
Yes, this can be seen in the DXOMark data on different camera sensors. They present the contribution of the individual R-G-B filter weighting.
https://www.dxomark.com/Cameras/Leica/M11---Measurements
Also, the density of the filters can vary among mfr's / models, which (iiuc) influences things like DR and ISO response (vs. gain, etc.).
IF it were solely a software matter alone, then everyone could just select a given profile ... no matter which camera they were using, and everything would be an exact match. However, that isn't the case. Case in point, can you just click on a a file from a Sony and apply a Hasselblad profile to get HNCS colors? or a Canon to get Fuji, etc. The physical filters dictate the light signal alterations, en route to the sensor. That base information is the starting point for the software calcs for a given profile.
OEM mfr's know where that starting point is for their models. Thus, the profiles they build in the software pick up from there. Non-OEM profiles, have to reverse engineer that information starting point to build theirs. That, or just apply general / universal profiles. This is where Hassy has "raised the bar" with their HNCS, in the calibration to land things as they do.
I tend to think of each camera as being a given film stock, and the standard processing provides for a given look. However, if I send a roll of film to the lab, I can ask them to process it in a non-standard way ... Hey, give it a longer bath to boost the contrast, etc. Leave it in the developer more / less, I may have exposed incorrectly. Then, from there, the printing processing (again, standard vs. custom), etc.
For us in digital realm, our camera's present the film stock. After that, we present as the lab. Standard processing = X, non-Standard processing = the entire alphabet of choices. 
For ease of comparison ... take note of the scale on the far right. Notice how one camera uses a relative equal balance of red vs. blue. While another camera uses an asymmetric balance of red vs. blue. Two different amounts of energy (info) being transmitted to the sensor from the same light source. The range of how diff mfr's / models vary this relationship presents differences that can be detected.
I'd be curious to compare two diff models with vastly different weighting of the blue filter ... relative to how skies generate noise differently. So, it would seem that the choices for the BFA could be weighted toward skin tones R/G or skies B sensitivities by a given model / mfr ... or striving to achieve an overall balance of R/B relative to G for more general use purposes.
In the case of the M11 vs. M8, the R/B ratios are ballpark 1:2, but the density is different relative to G (0.3 / 0.6 vs. 0.4 / 0.8). Meanwhile the SL2-S is more like 1:1 for R/B at 0.5 / 0.5. So, it would seem that the noise levels of skies could differ for an M8 vs. an SL2-S by nearly a stop (0.8 vs. 0.5 is almost 2X), etc. More to it, but you get the gist.
With film, the spectral response was baked into the film stock (Fujichrome vs. Ektachrome vs. Kodachrome). With digital, the spectral response is baked into the BFA. With film, the post-capture of the spectral response went to the lab for development (E-6 vs. K). So, the combination of spectral response + post spectral response capture = final. With digital, the spectral response capture is then post-capture processed by the mfr's Standard profile. To the degree that we choose to employ standard processing vs. custom processing profiles ... infinitely variable (it's what we dreamed about in the pre-digital era).
Making Ektachrome look like Fujichrome or Kodachrome wasn't something readily achieved. Making an M11 look more like an M8, yeah, that's more doable, of sorts. But, starting out with a bias of sensitivity toward R vs. B vs. balanced, just presents the starting point. Where / how you land it ... how you get it there ... your choice.
So, yeah ... it's a bit more than just software (i.e. BFA + software) ... and, diff models do have different colors. It's baked in to the BFA DNA from the start, available to be changed to your hearts desire.
|
