6 camera colour comparison

Started 1 month ago | Discussions thread
ForumParentFirstPrevious
Flat view
andrewD2 Veteran Member • Posts: 9,358
6 camera colour comparison
2

Magnar had a nice idea of comparing colour swatches of 6 cameras each with a jpg and raw swatch set, to see if there was a blind preference:
https://www.dpreview.com/forums/post/64109454
He hasn't said what all the cameras are, though some were revealed.

In that thread someone posted a Canon R vs A7III video comparsion.
The colours are very different but so were the WB and exposures. You can draw NO conclusions about the colours of these brands from that video.

Which leads to the first point of this thread,
Until you have matching WB and exposure and contrast, do not attempt to look for colour differences, it is utterly pointless.

i) A matching WB will NOT usually be found by using exactly the same values in a raw

ii) convertor or using the same e.g. cloudy preset in the camera.
Even the same camera will require different white balance depending on the lens used.

iii) Auto white balance can be fooled on ALL cameras, don't believe me, then see here: https://www.dpreview.com/forums/post/56477627
I've done that experiment with a lot of cameras when running a meetup.com group

iv) the same camera will not agree on AWB if the scene changes even if the light is the same.

v) exposure should be obvious, a bright yellow is a muddy brown with more exposure

vi ) contrast affects saturation
It takes some effort to compare even this imperfect comparison took some quite time.

So we need to normalise the swatches to compare.
I first inset the raw swatches of each camera onto the jpgs swatches of the same camera.
I then tried to set the black and white points to be the same and the greys roughly the same.

That is enough to see if the adobe color profile raw conversion is similar to the camera jpg (I have no idea what dpreview uses as setting for this, hopefully the camera defaults).

I wanted to compare between cameras so after that I then tried to normalise the neutral patches between all 6 cameras. Its rough but they are a LOT closer than they were.

I then adjusted a colour checker(CC) to how the colours would look with the same white and black points (probably should have done the reverse, but the difference will still show) These I added on the top left of each swatch.

I'm not going to call colours "good or bad" but if you want a camera that has an adobe color profile similar to the jpg then I think probably camera 5 then 3.

I started to calculate the dE values but it was taking so much time I think I'd be better writing a script than doing it manually.
I think in general the raw conversion is closer to the color checker than the jpgs for all of them. So in this illuminant and using raw all of them should be able to get reasonably accurate colours.

Camera 1 seems to play mostly with blues, cyans and purples 
Cameras 2 and 4 lighten reds in the jpgs relative to raw / CC
Camera 6 has some pretty big differences from raw to jpg.
I'm sure you can see other differences.  
Seems to me that the jpgs for 3 and 5 are created for accuracy, 2 and 4 for lighter skin relative to other colours. Not sure what 1 and 6 are trying to do.

Camera 1 top, down to 6 bottom, compared with the colour patches arranged in one long row:

It should be possible using a LUT or the DNG profile creator to make the colours of say camera 5 look like the jpgs of 2 or 4 (I'm not sure why you'd want to match 1 or 6 but I'd have to see the effect on a wide sample of images).

I'm very curious to know what camera 5 and if it is the same brand as 3. 
And thanks to Magnus for giving me something to do for a few hours on lockdown. 
Images are sRGB. 
Cheers,
Andrew

ForumParentFirstPrevious
Flat view
Post (hide subjects) Posted by
ForumParentFirstPrevious
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow