[biz-engineer]

Out of interest, have you used any raw processor that was able to give you the same colours and tone curve as PDCU?

Yes, Silkypix gives me the exact Pentax colors, but not Rawtherapee.

[BigMackCam]

Yes, Silkypix gives me the exact Pentax colors, but not Rawtherapee.

I guess that makes sense in that PDCU is based on Silkypix. But I'm assuming you haven't had the exact Pentax colours (by which we mean the same colours that the in-camera JPEG engine produces) in any other software?

[biz-engineer]

I guess that makes sense in that PDCU is based on Silkypix. But I'm assuming you haven't had the exact Pentax colours (by which we mean the same colours that the in-camera JPEG engine produces) in any other software?

I tried Afinity, DXO, LR, Rawtherapee, Silkypix, PDCU, compared with OOC JPEGs. DXO, LR, Silkypix and PDCU delivered the same color base to start with, with rawtheparee I was never able to get the same, close but not the same as my Pentax JPEG reference, and I was not happy with the magenta cast of rawtherapee (if I remember correctly, I made this comparison about a year ago).

[BigMackCam]

with rawtheparee I was never able to get the same, close but not the same as my Pentax JPEG reference, and I was not happy with the magenta cast of rawtherapee (if I remember correctly, I made this comparison about a year ago).

It might be worth trying the latest version of RawTherapee with the latest dcraw profiles. The magenta cast is most often a problem with the dcraw profile, rather than a problem with the application.

Just a thought... I know this results in a less efficient workflow, but if you're only doing microcontrast adjustments on a small number of individual images, you could do your main raw conversion in PDCU, export as a full size TIFF, load that TIFF into RawTherapee (or something else) and adjust microcontrast at that point, before exporting to final JPEG. A little long-winded, but it'll give you the result you're looking for.

[biz-engineer]

Just a thought... I know this results in a less efficient workflow, but if you're only doing microcontrast adjustments on a small number of individual images, you could do your main raw conversion in PDCU, export as a full size TIFF, load that TIFF into RawTherapee (or something else) and adjust microcontrast at that point, before exporting to final JPEG. A little long-winded, but it'll give you the result you're looking for.

Correct. That a good idea of a working solution. Thanks for the insight.

[ThorSanchez]

Don't be embarassed. I have lightbulb moments all the time in every aspect of photography and post-processing... often with something I've been doing for months or even years

So, assuming the dcraw profile for your camera exists and has been accurately produced, auto match tone curve will look at the embedded JPEG image within the raw file and adjust the tone curve to achieve a surprisingly accurate recreation. It's pretty awesome, actually, and something I wish Darktable had...

A little embarrassed. Not suddenly appearing in my underwear at work embarrassed.

I've never thought much about the OOC JPEGs, typcially I'm not processing like biz-engineer and actively trying to match the camera's JPG engine. I rarely shoot RAW+, just occasionally when I'm traveleing and might want to share some pictures but won't have access to processing software. So it doesn't/didn't really occur to me to use that as a benchmark. Sometimes I'll use a PP3 like Squirrel Mafia's that used Auto matched tone curves, but sometimes I don't like the look or just want my own tone curve based on the exposure.

[BigMackCam]

Not suddenly appearing in my underwear at work embarrassed.

I hate it when that happens!

I've never thought much about the OOC JPEGs, typcially I'm not processing like biz-engineer and actively trying to match the camera's JPG engine. I rarely shoot RAW+, just occasionally when I'm traveleing and might want to share some pictures but won't have access to processing software. So it doesn't/didn't really occur to me to use that as a benchmark. Sometimes I'll use a PP3 like Squirrel Mafia's that used Auto matched tone curves, but sometimes I don't like the look or just want my own tone curve based on the exposure.

I went through a few different phases with colour and tone matching. Initially, I wanted to make the colours and tone curve close to what the camera would produce in JPEG. Then, for a while, I looked to normalise the output from all of my different cameras to colour chart accuracy, at least as a starting point for further processing. And there's good sense for doing that where product photography is concerned. Now, I'm less concerned about that. As long as the colours and tone curve look OK by default when I import photos into my raw converter, I'm happy to work with that.

Having said that, I do like the auto match tone curve feature in RawTherapee very much. It seems to do a good job of matching to the embedded JPEG...

[house]

So, assuming the dcraw profile for your camera exists and has been accurately produced, auto match tone curve will look at the embedded JPEG image within the raw file and adjust the tone curve to achieve a surprisingly accurate recreation. It's pretty awesome, actually, and something I wish Darktable had...

It should be using the Pentax embedded profile when using DNG not the dcraw profile and the "Camera standard" setting under color management. Hover checkbutton to see a tooltip. Knowing for sure which of the options is active is a bit of a mystery though...

Even with those settings I find that the exposure and white balance is not quite the same. But with those sorted the exported raw colours are very similar to the jpegs on the K-3.

[biz-engineer]

I've never thought much about the OOC JPEGs, typcially I'm not processing like biz-engineer and actively trying to match the camera's JPG engine

If you have no reference, how do you know your colors are right? Judging with own eyes looking at the display?

[ThorSanchez]

If you have no reference, how do you know your colors are right? Judging with own eyes looking at the display?

Yes. If it looks good to me I run with it.

[biz-engineer]

Yes. If it looks good to me I run with it.

I see. The problem for me is that I'm not very good at evaluating colors just by looking, I'm not good enough, I need a baseline to start with.

[ThorSanchez]

I see. The problem for me is that I'm not very good at evaluating colors just by looking, I'm not good enough, I need a baseline to start with.

It helps that there are no real consequences if I don't get the colors "right". Most of the time no one outside of my family and friends see my photos. Even when I share here, it's viewed on monitors that are calibrated who knows how accurately or not at all.

It might be different if I was photographing things to get paid.

[BigMackCam]

If you have no reference, how do you know your colors are right? Judging with own eyes looking at the display?

The thing is, Pentax JPEG colours aren't exactly "right"... There's a sort of signature Pentax JPEG colour look (rather a nice one, admittedly), and the same can be said for each of the other manufacturers too. But in-camera JPEG output is rarely - if ever - a particularly accurate representation of what we saw. It tends to be aimed more towards creating an overall pleasing look. As a result, straight-out-of-camera JPEGs aren't a good choice if using them for professional product photography, as colours will be (often a lot more than) slightly skewed.

If we want the colours to be more accurate, we need (ideally dual-illuminant) profiles that match to recognised colour charts. I went down that route for a while, creating dual-illuminant profiles for each of my cameras, such that photos from every one of them were more-or-less colour accurate when those profiles were used. The approach works... but after a while, I became more enamoured with the profiles offered by the various raw converters. They may not reproduce the same results as the in-camera JPEG engine, and they still might not be completely accurate, but they're as individual as, say, a specific film stock.

If I should ever need accurate colour, though, I'd go straight back to using dual-illuminant profiles.

[normhead]

I can't imagine even caring about accurate colour except in maybe catalogue work where people may be choosing things to go with other things they already own. Hold the jacket up beside the image....is it the same colour? Photograph a colour chart in the corner of your images. Hold the image up beside the colour chart, adjust to get it as close as possible.

[BigMackCam]

I can't imagine even caring about accurate colour except in maybe catalogue work where people may be choosing things to go with other things they already own.

It's definitely important for product photography - catalogues, advertising, online product configuration choices, fine art reproduction etc...

Imagine selecting the colour for your new car then finding it's a completely different shade upon delivery?

Hold the jacket up beside the image....is it the same colour? Photograph a colour chart in the corner of your images. Hold the image up beside the colour chart, adjust to get it as close as possible.

That's a nice idea, but adjusting each colour typically alters other related colours (e.g. adjust blues, and your purples change... tweak the purples, and now your pinks look wrong).

When I set out to emulate the K10D's colour reproduction in Lightroom using K-3 raw files, I started with the primary red, green and blue channels - which got me in the ball park - then went on to adjust each colour individually. It took me many, many attempts and revisions to get it close, and even then it wasn't 100% accurate. In truth, 100% accuracy wasn't my goal and didn't matter in that application, but I mention it as an example of how difficult it is to achieve reasonable colour accuracy when adjustment of each colour affects the others.

For general photography, I don't believe it's important, and that's why I don't use chart-matched colour profiles any more. But there are applications where it can matter...