I was to say that only 1600 is better even from 1250 and 2000, so 320 640 and 1600 for extremes!

youtube compression forever!!!

what can I say, it's a nice instrument lol. As some others agree, iso 320 640 are the "best", ok, do not expect huge differences but there are.

:)

@starios Good question. I use mov on the GH3 because of (1) the higher bit rate (2) easier workflow (3) habit. * 90% of the editing I do at work comes in as ProRes .mov's so that's the habit part. But I love having one single file. Personally I hate using AVCHD for the simple reasons of work flow. Private>AVCHD>packets> stream etc. etc. is a pain in the a** when culling footage, deleting unneeded clips and grabbing one or two clips instead of the entire shoot. As far as the "higher bit rate" is better argument: remember a while back when I shot that wedding and accidentally set the camera to shoot mp4 vs mov? When I look at that footage now I can't emediately tell (visually) what's shot at 28mbs MP4 or 50Mbs MOV! So yeah, why not save the space if the quality is the same? Phycological I guess.

On another note I shot a multi cam interview last week. I'm again so frustrated trying to get the GH3 "pink plastic skin" to match the hacked GH2's footage. This issue alone has become 40% of the work alone...on the entire job. Not happy. >:^(

GH4 going to be a higher, cleaner version of pink plastic? I've invested lots of money and still believe in the m4/3 system. But this weekend is the first time I've begun to intertain the thought of a different brand or system. Sorry to the group for getting off topic.

@maddog15 About changing brand or system, I used in a theater a 5d unhacked, no way, I 1000% prefer gh3, only bmpcc is a reason to change for the DR

@wgtwo yep, you're right

@wgtwo what do you mean by high iso, it is clear even in youtube's sample in the final 200% comparison that till 1600 avchd has less noise, anyway, higher than 1600 is useless. 50mbps must have more detail but I can't see where it is, just like the 60fps vs 30.

I just realized that iDynamic actually works pretty well, but unfortunately its usefulness is very limited. It only kicks in when it sees a large light source in the frame, and as soon as you move away from the light source the shadow boost effect turns off. This may be good for home video where its just set to Auto, but for controlled work its unusable.

Does anyone know if there is a way to force it on, and to keep it on at all times? It would be amazingly useful if so. For now, its just limited to static, or low movement shots, with a big swatch of bright sky. Perhaps Vitaly could look into this when he gets to the GH3 hack? ;)

Is it really any better, though, than just boosting the shadows and mids in post? Does the fact that its boosted them internally before the compression stage make any difference to quality?

I'm using the Olympus 17mm and it only works in high contrast scenes.

@maddog15 ...

First, most of the better footage of the GH4 that I've seen shows marvelously detailed NORMAL skin. Considering it's got both a new sensor and a completely different processing "engine" one would expect something different, and "normal" is the best you can hope for. Thankfully!

That bit of nastiness the Panny engineers did to the GH3 to try and guesstimate what we the users would want in skin tones ... and build that into the processing algorithms ... was an abortion that keeps on giving. Clearly, the cam is designed to recognize "skin" color, then modify it ... to supposedly "our" ideal. Instead, yea, we've had to learn to get around it. It IS doable.

Yak's LUT's are one good starting point. Another is slightly tipping the curve of the mid/upper-mids steeper to raise contrast in the areas of Zones V and VI coupled with a slight drop in red luminance/sat and very tiny up-tick at times in green &/or blue. The contrast gets skin detail back (the cam's clearly designed to moosh over skin details for us ... ) and the color changes give a more balanced color range in skin.

In Speedgrade or most other grading apps (and some NLE's) you can pull a secondary to narrow in on the skin and do that pretty much only to "skin" areas. Or you can apply it "globally" in a primary layer or node and when done in small amounts does ok on the rest of the footage.

@maadog15

Hear your pain. My clients aren't typically that critical of color in the work I do. So I've gotten by though I'd a LOT rather had the skin & color I wanted without messing. Still, as nearly everything I actually "worry" about is a studio-style interview footage, I can pretty quickly apply the standard fixes and not have to worry about messing some other part of the image up.

I've got a couple personal projects I want to get to as possible, and they would be a LOT more affected by color-mangling to correct the wonderful gift of the Panny engineers. Worse comes to worst, I might have a GH4 by then ... ;-)

@rNeil Stop spreading your falsehoods. GH3 can deliver beautiful skin tones. It just requires careful lighting and grading. Here are screen grabs of ungraded footage shot over the weekend. If you can't get a natural look from these files, I assure you, it's not the fault of the camera. Please, no more unverified lies about engineers making pink detail-less skin conspiracy. It's utter, complete nonsense! https://copy.com/EjcvwpX6xfCM