Of course it does, the lower the bit rate is the more blocks of pixel you'll discover when you push the grade... in the sky, in shadows, nearly everywhere... With pixel count, the more you have pixels in a face for exemple, the more colors and nuances you'll get so more information to play with....
GH2 For Color Grading: Hacked GH2 or Canon DSLRs...?
- Thread starter Breezer
- Start date
Of course it does, the lower the bit rate is the more blocks of pixel you'll discover when you push the grade... in the sky, in shadows, nearly everywhere... With pixel count, the more you have pixels in a face for exemple, the more colors and nuances you'll get so more information to play with....
PaPa
Major Contributor
^^ What he said. It absolutely matters. No question. Don't convince yourself otherwise.
I do not see any parameters to the inquiry. He simply asks if it grades well. I said yes, gave an example. The image can be pushed much further and still hold up very well.
Here are two extremes
nuff said.
http://img210.imageshack.us/img210/8671/extremelight.jpg
http://img824.imageshack.us/img824/2176/extremeheavy.jpg
Last edited:
gonzo_entertainment
Veteran
No doubt, and not slamming you at all, just observing.
If somebody really wants to show what it can do, shoot something in the wrong white balance and try to correct it, or do a day for night, or something where you're really pushing it around.
If somebody really wants to show what it can do, shoot something in the wrong white balance and try to correct it, or do a day for night, or something where you're really pushing it around.
PaPa
Major Contributor
Challenge accepted.
PaPa
Major Contributor
First off, these were shot on 24L at 66 MBps instead of the 24H at 156MBps, which right there makes a bit difference in the fine noise quality. so it gets better than this. But at the same time, this is a pretty extreme example with different white balances all over. The left portion of his face ( his right cheek ) is getting a lot more tungsten than his left cheek, so i chose to balance the left cheek otherwise the entire image would go too blue. I believe this was shot at 800 ISO but may actually have been 1600 ISO. So this really is a comparison of an image in a worse case scenario.
You can see the noise detail on the red column and its still very acceptable considering how heavy a grade this is to begin with. Also look at the halos of light from the christmas lights and how they are affected. Very good IMO and these curves went pretty bendy. So this is very subjective, because to me, if i were shooting anything important ( which i believe the GH2 is not a good run and gun camera, but requires a decent setup with preparation ) i would never get anything this off white balanced ( except here lol ) But to know that even something like this that's so far gone is still fixable is fantastic. My af100 would have macro blocked to hell. Once again, jpg recompressed and low bitrate hack. I believe the comparison is between hacked GH2 or canon DSLR as well. GH2 wins hands down.
http://img811.imageshack.us/img811/2267/yellowbefore.jpg
http://img220.imageshack.us/img220/6470/yellowafter.jpg
You can see the noise detail on the red column and its still very acceptable considering how heavy a grade this is to begin with. Also look at the halos of light from the christmas lights and how they are affected. Very good IMO and these curves went pretty bendy. So this is very subjective, because to me, if i were shooting anything important ( which i believe the GH2 is not a good run and gun camera, but requires a decent setup with preparation ) i would never get anything this off white balanced ( except here lol ) But to know that even something like this that's so far gone is still fixable is fantastic. My af100 would have macro blocked to hell. Once again, jpg recompressed and low bitrate hack. I believe the comparison is between hacked GH2 or canon DSLR as well. GH2 wins hands down.
http://img811.imageshack.us/img811/2267/yellowbefore.jpg
http://img220.imageshack.us/img220/6470/yellowafter.jpg
Last edited:
mentatDUKE
Veteran
Nice examples Pops.
Jester2138
Well-known member
Ding ding ding - with all the Canon's KO'd and drooling on the mat, the lone GH2 raises his arms in victory as the awed spectators begin to cheer madly. Almost immediately the hottest cheerleaders from around the country climb onto the mat and dump water over themselves and the GH2 before they all go back to shower off and bask in the glory of the underdogs clear victory.
This thread makes funny:
Jumps back and forth between grades, aggressive and light... and that was in ETC mode on QV2...
^^^^ Non hacked 1080/60i
They're both 8-Bit 4:2:0. The difference is the hack bitrate, which undoubtedly WILL give you a better base to grade on.
Jumps back and forth between grades, aggressive and light... and that was in ETC mode on QV2...
For safe measure, I went back to the FulL Range and dipped the color contrast a bit more so it wasn't as milky.
And here are the "raw" looks:
Broadcast Range + 1.8 Gamma
Full Range + 1.8 Gamma
There you go.
^^^^ Non hacked 1080/60i
They're both 8-Bit 4:2:0. The difference is the hack bitrate, which undoubtedly WILL give you a better base to grade on.
Dynamic Range in a fixed format (AVCHD/H264) only gives a BASE upon which to grade, and can easily be "faked". It has no technical baring on how an image grades, it allows for more leeway.
Furthermore, USABLE DR is more important than theoretical DR: it doesn't matter if the 5D can net 9 stops in video mode if you can only use 7 because of the noise floor/macroblocking (which is probably a close estimation from my experience), when the GH2 may do 8~8.5 and the usable DR is 7~8 because of the hack's noise floor. It's true, if we had access to picture profiles and could find something SLIGHTLY flatter, we would be able to underexpose another 1/3 stop for highlights, which would give the appearance of "better DR".
I'm only talking hack here, as well. Not unhacked, at this stage a lot of us aren't even thinking in unhacked space anymore; the hack is the camera.
gonzo_entertainment
Veteran
Yep, it does correct out pretty well, wouldn't say it's "superior" to the Canon's, but seems at least as good. Not on the right monitor for full on pixel peeping.
I will say it's superior, because I have used both and own both and have graded both. I've sat in front of a massive screen with footage from both and flopped. No, again, the current crop of Canon DSLRs cannot hang in this arena. And it's basically because of the Hack.
Gary said it in another thread, but honestly, you really just have no idea until you put your hands on it. A month or so before I started toying with the Gh2 + Hack I was pretty convinced that no DSLR could suffice, and most of the examples on the web weren't convincing enough.
Once I picked the camera up that all went away.
All that said : I would trade it out for a 5D with IQ close to this. That was really my only issue, the muddy results.
gonzo_entertainment
Veteran
Off Topic:
It seems the 5D corrects better with bit rate hacks as well, but the last I heard on the magic lantern bit rate hack it seemed it still had stability issues once you take it high enough to really have an impact.
I wish someone around here rented them but nobody does. Couple places with ALL the Canons (5D2, 7D, 60D, t2i, etc...), but nobody rents the panny.
It seems the 5D corrects better with bit rate hacks as well, but the last I heard on the magic lantern bit rate hack it seemed it still had stability issues once you take it high enough to really have an impact.
I wish someone around here rented them but nobody does. Couple places with ALL the Canons (5D2, 7D, 60D, t2i, etc...), but nobody rents the panny.
Right. You can't get it high enough to actually make any difference. Which sucks. And even then, it won't help with the aliasing/detail etc. =T I hope 5DIII addresses all of this stuff.
Makes sense. Maybe there's a user near you with one? This community is massive...
gonzo_entertainment
Veteran
At day job, but I'll try to remember to post some of my more extreme 5D grades tonight. We had a bad white balance shot in my last film and I have some pretty good examples of extreme low light shot in cinestyle then graded,
I had this as well (7D off from properly balanced 60D) that I fixed in Color Finesse. It's merely "okay" in my opinion (the shot was barely used), and I'll post examples as soon as I can make it public.
gonzo_entertainment
Veteran
I had a tard for a 1st AC. He killed me on some focus pulls as well.