What is the defference between "Half High" and "Half Standard"?

[Amano Takahisa]

I want to know the difference of processing about debayer quality. We can choose some items when we use "Debayer Quality".
"Debayer Quality" is a setting in "RENDER" button of RED ALERT.
After we push "RENDER", another window appears and we can see "phase1" in the window.
We can choose dabayer quality.We use 4K HD shooting images and change to half resolution images by the debayer.
I tested the "Half High" and "Half Standard", so we already knew the differences of them. "Half High" is better than "Half Standard"
But I don't understand the difference of debayer process.
If someone know the difference of the debayer process, please explain to me.
I use RedAlert! ver20.1.8.

 

[Deanan]

Half Standard is a fast debayer that tends to be perceptually sharper but is more aliasy.
Half High is a nicer/smoother debayer that retains more resolution and is not aliasy.
Full debayer with a scale to half would similar to Half High but with a bit more resolution after the scale.

Both Half High and Full can benefit from a post sharpening at the end of the chain to bring out the perceptual sharpness from the inherent resolution. Think of it as not baking in the sharpness (different from resolution) so you can add the sharpening at the end of the chain and dependent on the output format.

 

[Amano Takahisa]

Half Standard is a fast debayer that tends to be perceptually sharper but is more aliasy.
Half High is a nicer/smoother debayer that retains more resolution and is not aliasy.
Full debayer with a scale to half would similar to Half High but with a bit more resolution after the scale.

Both Half High and Full can benefit from a post sharpening at the end of the chain to bring out the perceptual sharpness from the inherent resolution. Think of it as not baking in the sharpness (different from resolution) so you can add the sharpening at the end of the chain and dependent on the output format.

Thank you for quick reply. I understand your explanation.
Then I am wondering about the way of the pixel calculation.
Normally, debayer calculation uses 4 pixels around one pixel.
i think it is the reason of the quality.
I notice the quality diffence is caused by debayer calcuration.
What canlucation methods are used for "Half High" "Half Standard"?
If you can, please explain to me.

Sincerely

 

[Amano Takahisa]

Deanan
I hope you answer my second question again.
Sincerely

Taka

 

[Deanan]

Deanan
I hope you answer my second question again.
Sincerely

Taka

It's a tough very complicated question to answer. Different debayers can use different numbers of pixels to evaluate for reconstruction. All the way from the simplest direct neighbors up to fairly large kernels. Half fast/standard is at the simple end and half high is a bit more complex. Any more detail, and Graeme would have to start writing some very long whitepapers

 

[Uli Plank]

Very polite way of avoiding to say: "you are getting too close to our secret sauce" …

 

[Amano Takahisa]

Thank you, Deanan. I appreciate your kind reply so much.
I understood I asked complicated question.
Because it may be basis of calculating technology of debayer.
There was no meaning and I was not nosy.
Please don't get me wrong.
There was the reason that I had question.
A friend told me "The Half/High function uses some surrounding pixels.
And the Half/Standard doesn't use surrounding pixels. Half/Standard shows
origin bayerpattern data which aren't calculated by surrounding pixels.
So, Half/High is better color than Half/Standard.".
I am wondering this assumption.
So, I wanted to know about it and asked to you.
If you had duscomfort feeling from my question, my apologies.

>Uli
I hope this reply correct your mistaken impression...

Let's make nice handiwork together all over the world in this year again.

 

[Deanan]

Thank you, Deanan. I appreciate your kind reply so much.
I understood I asked complicated question.
Because it may be basis of calculating technology of debayer.
There was no meaning and I was not nosy.
Please don't get me wrong.
There was the reason that I had question.
A friend told me "The Half/High function uses some surrounding pixels.
And the Half/Standard doesn't use surrounding pixels. Half/Standard shows
origin bayerpattern data which aren't calculated by surrounding pixels.
So, Half/High is better color than Half/Standard.".
I am wondering this assumption.
So, I wanted to know about it and asked to you.
If you had duscomfort feeling from my question, my apologies.

>Uli
I hope this reply correct your mistaken impression...

Let's make nice handiwork together all over the world in this year again.

No worries, some questions just take more time to answer than there is sometimes available.

Both use the neighboring pixels but the amount and complexity of the calculation of weighting the contributions it the primary difference.