GH2 GH2 - Testing Smooth vs Cinema

cowpunk52

cowpunk52

Veteran
I know this topic has been discussed to death, but I still have a lot of questions over which may be better for me. I'm DP'ing an indie feature with hacked GH2's in January, and at the moment I'm doing tests to determine whether Cinema or Smooth would be the best choice for me. I've already decided to go -2 on everything, so all tests are done with those custom settings as well as the Quantum v2 hack. Here are two frame grabs I took just now. My purpose was to see if Cinema held any more highlight detail and/or if Smooth actually held more shadow detail.

I used identical settings on the two samples:
Canon FD 28mm f/2 lens
ISO 320
Shutter 1/50
f/5.6

I transcoded using 5DtoRGB Batch to ProRes 444, 709 decoding matrix, and full range luminance.

You will see blank bars on the top and right side of the Cinema frame-grab because I wanted to match framing between the shots as close as possible in FCP 7 in order to click back and forth and more closely examine differences in parts of the frame.

With the Cinema clip, I pushed the mid-range from 100 to 111 using FCP 7's 3 Wheel Color Correction tool in order to match the mid-range of the Smooth clip. I found that while the Cinema clip appeared slightly darker before boosting the mid, there was no less shadow detail than what was in the Smooth clip. I did not adjust any other color settings. My recommendation is to open both these stills, with the imgur full size link, in separate tabs and click between them for comparison.

Here is the Cinema framegrab (with mid-range boost):
xLDIe.jpg
full-size imgur link: http://i.imgur.com/xLDIe.jpg

Here is the Smooth framegrab (no adjustment):
tNrsn.jpg
full-size imgur link: http://i.imgur.com/tNrsn.jpg

One of the more interesting things I've noticed, besides the obvious highlight and roll-off differences, is that the Smooth frame-grab appears to be ever so slightly softer than the Cinema frame-grab. I'm interested to hear other observations.

Here is the Cinema frame-grab without mid-range boost:
TpHEW.jpg
full-size imgur link: http://i.imgur.com/TpHEW.jpg
 
Last edited:
Count me in for SMooth, especially with 5DtoRGB and Rec709 (1.8) in play.

The Cinema mode clips way too fast for my tastes.

Whatever Smooth is lacking, as well, can be made up for with NDs and better glass.
 
I thought that you were supposed to set it to none, but I was told by someone who had worked with Canon and 5DtoRGB converted footage to DaVinci workflow to flag it 1.8.

It's one of those cases where I just said "okay" and continued to do it. Now that it's brought up, I'm gonna go and do some research.
 
@kholi - yay for tests, lemme know if there is a difference, the skin tones test at high key that I mentioned in the other thread, I'll find a way to post it, edit: link (sorry for derail cowpunk)
 
Last edited:
First of all, I have to mention that it looks like you sneaked into my house to shoot those photos. That's exactly what it looks like beneath my computer desk! But I'll also say I prefer Smooth and your 2 photos represent the exact same things I've noticed. In CINEMA, the hottest part of your frame (the wood floor at the bottom) looks yellow. I'll bet you that your floor looks in reality much more like the photo shot in SMOOTH. The CINEMA clip looks yellow because the color channels aren't clipping uniformly. Same problem with the AF100. This is why still prefer SMOOTH.
 
Yes the smooth shot seems much better, you can see the clearer part on the white wall is starting to clip while on the smooth one it is well exposed, the wooden floor is also clipping in the cinema settings. So clearly on your example the smooth is much better.
 
Hmmm... I just thought of something.


Does it make sense to use Pro-Res4444 in FCP? Should we be testing in After Effects or something similar (Color or DaVinci) that deals with 16-bit or higher to make sure what we're seeing is right?

IN the end, grading should occur with DPX files (in my opinion), and you wouldn't be doing that in FCP.

I don't doubt, however, that smooth is the right film mode to use still, I wonder if there's even more to eck out of the camera--as if it wasn't being stretched super thin now, if you go through the right channels.

Another plus for Film Mode hack, and after it's been brought to my attention, iDynamic hackin' too.
 
Kholi said:
I thought that you were supposed to set it to none, but I was told by someone who had worked with Canon and 5DtoRGB converted footage to DaVinci workflow to flag it 1.8.
Click to expand...
Using the "1.8" setting is the best representation of the picture as it came out of the camera. It doesn't post-process the image at all and Final Cut Pro, etc. renders it correctly.

I've done away with the gamma options altogether starting with 1.5.6. If you want to apply a gamma correction during transcoding, you can use a shader.
 
Thomas Worth said:
Using the "1.8" setting is the best representation of the picture as it came out of the camera. It doesn't post-process the image at all and Final Cut Pro, etc. renders it correctly.

I've done away with the gamma options altogether starting with 1.5.6. If you want to apply a gamma correction during transcoding, you can use a shader.
Click to expand...

That's exactly what I concluded last night, after looking at some old files with "none "selected.. If you don't add the flag, then you're going to go right back to the Quicktime gamma flag and you do not want that. Smart move on removing the option altogether. So it defaults to 1.8? One less option I need to check!
 
I used Cinema on my last vid and I liked it. But I underexposed slightly, as I almost always do. Using an external monitor really helps identify possible clipping.
 
Dave, you might've touched on something that's been baffling me.

I've noticed that it's actually not very good to underexpose my image, but it may be because I'm using Smooth. It's almost like I have to over-expose slightly to get the image where I want it to be for grading.

However, maybe Cinema is more for underexposing? Too many tests to run.. BLERG.
 
You must log in or register to reply here.
Back
Top