Thank you! I think I started to reply with more stuff but this is more than enough info. Thanks again for the help, @thepalalias.

@thepalalias Oops! I forgot that I'd said "...correct?" Sorry, haha. Anyway, thank you so much for helping me so far.

Questions:

1. "The weakest channel is in this pic, and the strongest is in that pic", "lower your less detailed channel". Does your use of weakest here refer to "more detailed" and strongest "less detailed"?

2. I know you just went into a huge explanation of it, so thank you very much. But just to confirm, does balancing RGB channels truly bring back a natural texture to the skin? Initially, I wasn't so much concerned with the color of the skin at that show, but when you said "If I had been you, I would've started off warmer" I figured you know best.

I specifically chose the second set of shots not because of the RGB balance (I'm familiar with doing that in curves but since, as you also pointed out, the lighting was cooler I left it as it was since their music is pretty dark). Balancing RGB aside, I'll just go with what you said about the skin: those second set of pics were easier from the get go due to the contrast on their skin (ie. tattoo), etc.

You said what I was describing didn't match up with how things work, but it comes down to this: I wish I knew how to describe the skin of the guy in the first set of pics vs the second haha (his looks dry - and by dry, i'm not talking about lightness; not soft, not smooth or natural, whereas the other set of pics look like silky/smooth, natural skin), but I think you've figured out what I was getting at. Sure, there's detail in the original set I posted, ie. veins in his hands - but we were able to see that before RGB balance, etc.

So I guess texture is what I've been getting at.

@dreo Regarding the ISO bug, I didn't think you hadn't read what I wrote. I was just responding to the question you asked when you said "correct?" Since in my experience there is rarely a situation that called for any one thing, I just mentioned the options. :)

The thing is this - the words you are using to describe the problem and the solutions you are talking about don't entirely gel with the way they are commonly used. Texture is more commonly used to refer to consistency, sharpness and detail than it is to value, so I'm having to work a little harder to interpret what you mean.

It sounds like you aren't so much talking about "texture" as "overall grading" so I'll take a shot at addressing that.

Update 1: You said the venue was "very well lit" and I wasn't there, but judging from the two sets of pictures, it looks like the lighting in your more recent set of pictures was brighter/harsher than in the previous venue. While there is a lot that can be done in post to make the more recent pictures more pleasing, it's not going to end up looking like the older set of images.

For example, the majority of the guitarist's skin in the new series of shots is highlight. Every member in the previous photos has their skin in a range of values, either from shadows to midtones or from shadows all the way up to highlights. But in all cases, highlights account for the minority of the values across the skin for the old pictures. The largest concentration of highlight is on the guitar arm in one of the old pictures, but even here we generally see a more usable spread of values.

It looks like the more recent venue was lit more for clarity/high energy wheras the old one was lit more for mood.

Anyway, I would suggest using RGB curves to try and get the most out of the values you have at the moment.

Edit 2: Here's a trick you might want to try to get started. I've included the relevant extractions below.

Look at comparisons between the individual channels, both in relation to the other channels on the same image and in relation to the image you want to match.

In both the old and the new images, we see that the blue channel is less detailed in representing the skin than the other channels because well... it is human skin. :)

But at the old show, the blue channel was the weakest channel and at the new show it was the strongest. I've run into this issue at shows sometimes, too - it doesn't affect RAW files, but for both video and JPEG stills, white balance starts to get pretty important in adverse lighting conditions and you can start to lose detail because of it.

In your curves, you want to downplay your less detailed channel (the blue) and then balance the red and green ones. Use the histograms (or your scopes) as a guide for where to allocate more/less of your values for the new show vs. the old one.

Also notice that tattoos that two of the band members have in the older show allow their arms to maintain significant contrast, even if lighting were more adverse so the old one was easier to shoot from the get go.

Anyway, it doesn't look like your biggest problem is macroblocking as such, so it may very well be that switching hacked settings won't solve the problem. But to get a sharper green channel, you might try using CM Night and then making sure that your WB doesn't over-emphasize the blue.

Edit 3: Just attached a Red channel before and after with the associated curve comparison. Notice the additional definition in the shadows.

Edit 4: Attached the channels and the final result.

Obviously the above edits do not result in something that looks as nice as the original and there are some issues. One is the fact that the back of his neck has more definition but also more banding. However, I think this gives a general idea of what you could do with curves. There are of course other approaches available.

@olegkalyan Yes, it is impossible to get the ISO bug with ISO 200, 400, 800, 1600, etc.. The top row ISOs may be less noisy if you take the steps to avoid the ISO bug, but with middle row ISOs you don't have to worry about encountering the ISO bug even if you forget.

@dreo I'm not saying don't use the top row of ISOs, I am only saying that when it comes to avoiding the ISO bug, you either make sure to approach from the middle row by moving upward, or you only use the two bottom rows. As I said, by using ISOs that aren't on the top row, you don't have to remember to deal with anything ISO-wise. It's not magic of course, so you might need to adjust your aperture, ND or shutter speed if you need to decrease/increase by 1/3 stop. It's all question of how big a deal remembering to check your ISO vs adjust your other settings/equipment is for you. :)

What I meant about shadows and highlights was that I had brought the highlights down a fair amount in the PNG files I posted earlier, so I wanted to know if you meant something more dramatic than that.

Interesting, some people, thought on GH2 better using iso 160, 320, 640.. etc, noise is less than 200, 400, etc.. Not proven that..

Is it in fact impossible to get ISO bug with (200, 400, 800, 1600) ?

Thanks for the quick response and the white balance tip, @thepalalias. It's been driving me nuts, haha. Since I did this last year, I really wish I could remember all of the settings, but I can't. If I had to "guess" it was in SH Mode, SS 50 - as there wasn't really wild movement - and either Standard or Smooth. I understand that other settings movie modes work better, but I also know that it depends on what you are shooting. I've read a little bit about that bug, but no, I didn't think it would've been an issue - what steps would I have needed to take to avoid it?

Edit:I guess I'm just worried about the texture of the skin which I clearly messed up while shooting and I'm not sure how to salvage it in the editing process.

Edit 2: I feel like I may have to experiment with the highlights a bit and recover them. Thoughts?

I haven't been using other cameras, but here are some screencaps of other GH2 footage I've shot to better illustrate what I was going for:

Here are a few of those images: