SD1 Merrill to sd Quattro - What have we lost?

TN Args wrote:

xpatUSA wrote:

gw24 wrote:

TN Args wrote:

D Cox wrote:

I think the disadvantage of the Quattro sensor (low colour resolution)

I feel like taking this 'fact' to task, but it is late and I am tired. Show me a number that measures this 'low' colour resolution.

or one divided by four.

You are referring to inputs (the machine), whereas I am referring to outputs (the image). What is the quantitative effect on the image, i.e. what parameter do you measure on the image to detect the magnitude of the effect of the different input quantities?

Again: my argument was a spoof. It's (for most) about the final Image, either on a Screen or printed (I do lots of printing). That's what matters, and you can't put a number on that.

Sigma uses the top layer to measure color just like prior sensors - in addition to measuring brightness.

Step 1 of conversion is to get four sets of values, albeit with the lower layer values the same for each set. Nonetheless, each set is different by virtue of the top layer differences. That means the resolution, in the normal sense of the word, is that of the top layer.

Whether or not the "colors" so obtained are accurate is a completely different issue which arises from Step 2 onward.

Of, in our world, "resolution" means many different things. For myself "color resolution" is the bit depth of an RGB image, i.e. "number of colors". As to D Cox's "hue and saturation resolution", perhaps that should be explained; personally, I have no idea what that might be.

-- hide signature --

Ted

Thank you Ted. Saved me the trouble and said it better anyway.

@gw24, @D Cox, the Quattro produces 20M Red values, 20M Green, 20M Blue, none of them in quads, whereas the Merrill produces 15/15/15. So in terms of spatial colour resolve, Quattro wins.

Wrong - but if you believe so, hat's fine with me.

What about colour accuracy? Is that the question? Like Ted says above, the normal definition of colour resolution is bit depth (and hence number of possible tones) of each photodiode, summed for each layer (or each filter in Bayer). Merrill tech has a problem with signal to noise ratio for its middle and bottom layers, because the pixels are small and photons are being absorbed by higher layers, so when you sum the three bit depths for Merrill vs Quattro, Quattro wins. Sigma could make a very strong technical argument that Quattro produces more accurate colours.

Like Lyon said in 'that' post 2 years ago, the 'thing' (he called it chroma resolution) that is less in a Quattro sensor in terms of data values and hence inputs to the process of producing an image, is effectively imperceptible in its effect on the outputted image. Which is the same as saying that it has zero quality, or no qualitative impact on image quality.

All right: I will leave you to your quest of quantifing what can't be quantified.

-- hide signature --

Arg