How to get the best raw exposure, part 2

If one channel of the sRGB image is clipped, you can see color shifts. But are you talking about raw clipping or clipping in the developed image?

In the raw file, or in the output file?

That make me think you were probably talking about raw clipping all along. Right?

But what histogram are you talking about? The in-camera one?

What kinds of artifacts? Done right, there should be no color shifts.

Histograms aren't perfect. And you've put your finger on one of their problems. Sometimes, there is no substitute for knowing your subject and knowing your craft. Sounds like you've been there.

Sounds like you know what suits your clients, and yourself. But there are many other kinds of photography than the kind that you do, and for other kinds of photography, other working methods are appropriate.

Plagiarism is presenting work or ideas from another source as your own,

The facts, never said it was my own. I referenced it . others just assumed it was my writings and couldn't help themselves to personally attach it.
I would, and have, posted images. to prove my points.

Thanks again.

When one color channel is clipped/saturated in the raw file and the others are not, there is necessarily a color shift or a saturation shift if one uses all of the bits. Quantization error can also produce shifts when applying the gamma curve is done by conversion from integer to floating point and back to integer. The "Highlight Tone Priority" camera setting which I would not use tries to avoid those shifts ( cam.start.canon/en/C003/manual/html/UG-03_Shooting-1_0090.html ) by making the lowest available ISO 200 instead of 100 and thus decreasing the quantization error in the highlights in the out of camera JPEG. The only histogram available to me when making the photo is the one in the camera. In Canon DPP, I could distort the gamma curve to extremes by adding many control points so that there are visible artifacts.

Thanks again for writing about this.

Without attribution. Please don't so that again.

The most recent image of yours that I commented on did hot proved the point you said it did. Quite the contrary.

There was no indication that it was anyone else's words.

Lr and ACR are quite good at avoiding these color shifts if only one raw channel is clipped.

I'm assuming you're working in 16-ish bit precision in your raw developer before writing out the JPEG file. So you'll not see significant quantization error. I know of no raw developer that uses floating point to apply a gamma curve, although when I'm writing image manipulation software, I often use floating point for all internal calculations.

I'd suggest using RawDigger on some of your files to see how well you're doing. You'll be better prepared the next time you encounter a simialr subject.

Define

I already did in my posts where I pointed out you altered what I posted in your quotes of my posts.

If you need to refresh your memory look up those posts.

Try using Ps and set the blending mode to Luminosity .

That is one the most stupid arguments I have seen anywhere.

You are effectively admitting the images in your posts without any verifiable attribution are most likely not yours either :-)

On the R5, the invariant range starts at ISO 800. Despite it's 'ISO-invariant', I still don't want to push it more than ~3 stops in post. The exposures I use in such conditions (like concert/stage shooting) would require ISOs ~3200-6400 with evaluative metering. So if I set ISO to 800, it gives a lot of headroom for highlights. That's basically the primary use of the ISO-invariance idea - avoid highlight clipping with no (or little) image quality loss.

My Bad 🙄, i did think about it, and always quote the writer, but it looked like more fun not to in the context and post i was replying to.😁

You have been caught out plagiarising and now you are back pedalling :-)

The plagiarising was deliberate because you said you thought about it.

It's too late now because you can't unscramble an egg :-D

I see. You have a situation where you know the light will not be so bright to cause clipping at ISO 800.
I just realized that the exposure indicator has correct metering feedback even when the exposure simulation is turned off. So that can be used to catch situations where clipping could occur at ISO 800.

BTW, Canon R5 has some read noise weirdness between ISO 800 and ISO 1600 and the most invariant range starts at ISO 1600 (link). It probably does not matter in practice.

People have you just ignore them

No the camera just recorded what was there for that given light

Maybe you should think about what this means, maybe just maybe that is what the camera recorded in the raw file

When you know how your camera records data it can be easily done on the fly

If I ever have issues getting the best out of raw, I'll be sure to let you know.

Danny.